EPA-AA-EOD/TPB-87/2
                               Technical Report
                   Evaluation of Several Methods to Measure
                           Volatility of Motor Fuels
                                   July 1987
                        Marcus  E.  Haubenstricker,  Ph.D.
                                Carl  A.  Scarbro
                                    NOTICE

Technical  reports  do  not   necessarily   represent   final   EPA  decisions  or
positions.   Their  publication   or  distribution  does  not   constitute  any
endorsement  of  equipment  or  instrumentation  that  may  have   been  evaluated.
They are intended to present  technical  analysis of issues using data which are
currently  available.    The  purpose   in  the  release  of such  reports   is  to
facilitate the  exchange of technical information  and to inform  the  public of
technical developments  which  may  form the basis for  improvements in emissions
measurement.

                           Testing Programs  Branch
                        Engineering Operations Division
                           Office of Mobile Sources
                        Environmental  Protection Agency
                              2565 Plymouth Road
                          Ann Arbor,  Michigan  48105

-------
Abstract

    The  U.S.  Environmental  Protection  Agency  is   proposing  regulations  to
reduce  the  amount  of  hydrocarbons  released  to the  atmosphere  due to  the
evaporation  of  automotive  fuels.   The  new  regulations  may  define  upper
volatility  limits  for  these  fuels  based  on  seasonal  climatic  patterns.
Volatility  of  gasoline fuels  is  typically  quantified  by measurement  of Reid
vapor  pressure  (RVP).    Although  an  established   procedure   exists   for  the
assessment  of  this  parameter (ASTM  D  323),   there is  question  as to  the
accuracy  of   the   procedure   when  evaluating   fuels  with  water-interactive
constituents,  e.g.,  alcohols  and  ethers.  ASTM P 176 is  a procedure which has
been  proposed  as  its  replacement.   It  addresses the problem  of water and  is
known as the "dry" version of ASTM D 323.

    This correlation  study was designed to  evaluate the proposed  ASTM  P 176
procedure  for  repeatability  and  reproducibility.   The  results of  this  study
will  help  establish a  basis  for  enforcement tolerances.  Also  evaluated was
the  equivalency  of  results   from  two   commercial   instruments  versus  those
obtained by the traditinal "gauges and bath"  technique.

    Thirty-nine  laboratories   within the United  States  and   Canada  analyzed
duplicate  samples  of four  fuels.   The  work  was  performed during  the  fall  of
1986.

    The  results  of  the  study indicate  a level  of  precision  similar  to that
determined  by  ASTM  in 1985.   Precision may be improved by  calibrations  or
other enhancements  to  the  technique.  It was also determined  that  the results
from  the commercial instruments were comparable  to   those from the gauges and
bath methods.

-------
                                      -2-
Background

    The  U.S.  Environmental Protection  Agency has  placed  a limit  on  the lead
content  of  automotive fuels as  part of a process  to  limit adverse automotive
emissions.   To counteract  the resultant  decrease  in octane,  refineries have
increased amounts of  lighter hydrocarbons  in final  fuel  blends.  EPA continued
its  original  actions  by  placing  limits  on  other  emissions products.   As
exhaust  by-products  decreased,  evaporative  emissions became  more noticeable
for two  reasons:  Evaporative  emissions constituted a  larger fraction of total
hydrocarbon  emissions  as  exhaust  hydrocarbon emissions  decreased, and fuel
blends  became  more  volatile   due   to  the  increases  in   amounts   of  lighter
hydrocarbon fractions.

    In an additional  effort to control evaporative  emissions,  EPA  may develop
upper  limits  for fuel  volatility.   If adopted,  these limits  would likely be
geographically based on seasonal climatic trends.

    Volatility is historically  associated  with the  results of a vapor pressure
assessment  using  the  Reid method.   The  American  Society  for   Testing  and
Materials  (ASTM)  adopted  this  test  in 1930,  and  presently  documents  it  in
their procedure D  323.   A reference to Reid vapor  pressure (RVP)  implies this
procedure,  which  prescribes a closed  cylinder and the  determination  of  the
headspace pressure  of a sample which has  been heated  from 32°F to  100°F.  The
units of measurement are generally expressed in pounds per square inch (psi).

    Because of various  market  economics, short-chained alcohols and ethers are
being  used  as  fuel  additives.  In  some   cases,   these  additives   comprise  a
significant percentage  of a fuel.   Due  primarily  to chemical hydrogen bonding,
these  substances  interact  with water,   resulting in  lower  observed  vapor
pressures.   To  prevent  these  potential   inaccuracies,   the  RVP  should  be
determined under dry (non-aqueous) conditions.

    In ASTM P 176,  a  proposed  general specification for gasoline, ASTM  has
outlined  a  new procedure for  measurement  of  volatility.   This procedure  is
identical to ASTM  D 323 except that safeguards are taken  against  inclusion of
water.   ASTM  conducted   a  correlation  study  in   1985   to  determine  the
statistical acceptance  of the  procedure.   The  study included  17 laboratories,
18  fuels,  and  271  observations.   The   results   of  the  study  produced  no
statistical difference  between the  "wet"  (ASTM D 323) and  "dry" (ASTM  P 176)
methods.

Purpose

    The   purpose    of  this   study   is   to  evaluate    the   repeatability,
reproducibility,  and  accuracy  of several  methods  to  measure RVP  as  they  are
applied  at  a  number  of  laboratories  throughout North  America.   The  results
will be used in the development of regulations of fuel  volatility.

-------
                                      -3-
Program Design

Equipment and Methods:

    In  addition to  the ASTM  P 176  method  using  the traditional  gauges and
bath,  there  was  also  a  need  to  assess  the  capability  of  more  automated
methods.   Two  types  of  instruments are  in general  use —  the German-built
Herzog  semi-automatic  unit distributed by UIC,  Inc., Joliet,  IL;  and a micro
semi-automatic  unit  developed  and  manufactured  by Southwest Research Institute
(SwRl), San Antonio, TX.

Laboratories:

    In  order  to  obtain  reasonable results  for  each  of  the several  test
methods, a number  of  laboratories  would need to participate.  Twenty to thirty
was  an early  estimate.   Laboratories  which participated  in  the  prior  ASTM
survey  were  identified  as  candidates.    Other  laboratories   known  to  EPA
personnel would also be invited to participate.

Fuels:

    Pour fuels  were chosen  for the  survey.  Two of these are  used  at EPA's
Motor  Vehicle  Emissions  Laboratory  (MVEL) as  test  fuels.   One  was  a  high
octane, non-oxygenated,  unleaded  test  gasoline  with  a nominal vapor pressure
of  9.0 psi  and is  used  by  EPA and  the  automotive  industry as  an  emission
certification fuel.   It was  procured from Howell  Hydrocarbons  of San Antonio,
TX.   The  other  test  fuel was  a "commercial"  unleaded test gasoline  procured
from Phillips Chemical  Company  of  Bartlesville,  OK.   Its  RVP was approximately
11.5 psi.

    The  third  fuel  was  cyclopentane   (90$ pure),   procured   through  Curtin
Matheson  (order number  NCX2414-1).   The  theoretical RVP of  this  hydrocarbon
compound is  10.15  psi.  This  particular  fuel  component  was chosen because of
its purity and therefore stable vapor pressure.

    The fourth  fuel was a  commercial grade gasohol (nominally 10$ ethanol) and
was  procured  from a  local service  station.  The  vapor  pressure of  this  fuel
was  over  12 psi.   The  procured sample was stored  in a  closed  drum  at below
60°F until bottled.

    Duplicate samples  of  each  fuel  were  to be  sent  to  each  laboratory.   The
containers were coded but the contents  were  not identified.

Sample Handling:

    Sample handling  was a critical  factor  in  the program.  Every  effort was
made  to  ensure  that samples  of each fuel  were  identified and  that  they  were
packed properly  for  shipment.   Concern for  possible  changes in vapor pressure
during the  process was addressed  by a statistical analysis of results  versus
the time between bottling and analysis.

-------
                                      -4-
Data Processing:

    The  results from  each laboratory  were evaluated  from the  standpoint of
repeatability,   reproducibility,   and  accuracy   using  accepted  statistical
techniques.

Conduct of the'Program

    During June and  July,  a  number of laboratories were contacted.  Almost all
agreed   to  participate.    These  contacts   also  resulted   in  unsolicited
responses.  Ultimately, a total of 49 laboratories agreed to participate.

    Most  of  the  laboratories were  equipped for  the  gauges  and  bath method.
Almost as  many used a Herzog instrument,  and  a few had  the  SwRI unit.  Three
offered  to  perform the analyses  using  a gas chromatograph.   A listing of the
laboratories  and  their  type  of  equipment  is   contained in Attachment  A.
Attachment B is a  copy of the letter used to describe  the  program and  formally
solicit participation.

    Acceptable  containers  for  storage  and  shipment  of  the  samples  were
obtained  from  All-Pak Corporation   of  Pittsburgh.   These  shippers  met  DOT
exemption  E-9168  for  hazardous  materials  and  were  comprised of  a  one-quart
PVC-coated  flint  glass  bottle with  teflon cap,  a  foam-filled  metal  can to
contain  the  sample  bottle,  polypropylene tape to  seal the bottle and can,  a
plastic  bag to  further  seal the can,  a  copy  of  the DOT  exemption  and  a
specially  constructed  cardboard   shipping  box,  and  shipping  tape  to  seal  it.
The teflon-lined caps  were replaced  by  poly-seal  caps to ensure an even better
aeal.

    The  four  fuels  were  bottled  on  October 7,  9,  16,  and  21 and  boxed  for
shipment.  Each  type of  fuel  was bottled in the  course of a  single  day by a
bottom-fill  method  through  1/4"  tubing.   Each  bottle   was filled  to  15%
capacity and securely capped.

    When the samples were  packaged  for  shipment,  a special seal (Attachment C)
was placed  over the  can  before  it  was placed in the plastic bag.   This  was
done to  direct the  samples  to certified  analysts before  being  opened.   Also
placed   in sample  box  #1 was  a   copy  of  the  ASTM  vapor pressure  procedure
(Attachment  D),  a  data  report  sheet  to  be  filled  out  by  the  analyst
(Attachment E), and  a  sample  transfer tube  fabricated for the one-quart sample
bottles.

    A contract  was  awarded to United Parcel Service  for  shipping the samples
to laboratories in the United  States.   Shipments  to  Canadian laboratories were
arranged  through  Consolidated Preightways.   All samples left MVEL by  the  end
of October.  Overall, 520 samples were sent to a total of 49 laboratories.

    Results began  to return  during the  first week of  November,  and by the  end
of  November  approximately  25 laboratories had  responded.   Ultimately,   39
laboratories submitted  results which accounted  for 80$  of the  samples  which
were sent out.

-------
                                      -5-
Results

    Attachments  P,  G, and  H are listings  of the results  from  each reporting
laboratory   for   the  P  176,  Herzog,   and   SwRI  techniques,   respectively.
Anonymity of the laboratories has been maintained by use of the code numbers.

    Statistical  analyses  of the  data  for  repeatability  and  reproducibility
began  in mid-December.   The methods  employed were  identical  to  the  methods
used  by ASTM  on the  original  survey  in 1985,  and  are  found  in  "Manual  of
Determining  Precision   Data for  ASTM  Methods  on  Petroleum  Products  and
Lubricants," (RR D-2-1007).  According to ASTM:

       "Repeatability is  a  quantitative expression of the random error
       associated  with   a   single   operator   in  a  given  laboratory
       obtaining repetitive  results with same apparatus under constant
       operating conditions  on  identical test material.  It is defined
       as   the   difference  between  two  such   results   at   the  95$
       confidence level.

       Reproducibility  is  a  quantitative expression of the  random
       error associated  with operators  in  different  laboratories each
       obtaining  a  single   result  on  an identical  test  sample  when
       applying  the   same method.   It  is  defined  as the  difference
       between  two  such single  and  independent results  at the  95$
       confidence level."

    After evaluating  the  data,  it was observed that  some  laboratories reported
consistently different results  than others.   Assuming that calibration or some
other  systematic difference  was  the  cause,  it  was  decided  to  normalize the
data based  on  the  accepted RVP of  the  cyclopentane.   The average value of the
raw data  for cyclopentane for  each laboratory was therefore  divided  by 10.15
psi.   The   ratio  was then  applied  to  the  raw  data  for  each  fuel  for  each
laboratory  and  the   results recalibrated.    This normalization  resulted  in
generally improved values, for reproducibility.  The summary  of  all  results  is
contained   in   Attachment   I.    The   statistics  for   the   calculations  of
repeatability  and   reproducibility  for   each   test  method  are   found  in
Attachment J.

    Concerns about the effect of  storage time on vapor pressure were addressed
by  evaluation  of vapor  pressure  vs.  time as  shown on  Attachment K.   For this
purpose, the total vapor pressure of all eight samples was plotted against the
number of days between the date  of bottling and analysis date.

Discussion

    The  statistical   results generated   in  this  EPA  survey are  similar  in
magnitude  to  the  results   of  the  1985 ASTM  RVP survey.   Normalization  to
cyclopentane improved the calculated  reproducibility  values for  all three test
methods.

-------
                                      -6-
    In  1985,  ASTM  deleted  data above  and beyond  the  criteria  used  in their
RR D-2-1007.   This  is   considered   acceptable   if  the  data  is  judged  as
contributing to the majority  of the  variance for a sample or a lab.  Deletions
did not extend beyond ASTM RR D-2-1007 directives in the present EPA aurvey.

Conclusions

    The  statistical  calculations  of the  correlation  data  indicate  similar
precision  as  determined  by  ASTM in  their 1985  study  for the  same  three dry-
methods  (bath  and  gauges,  Herzog,  and  SwRl).   All  of the  studies  indicate
worse precision than the  ASTM published  values for ASTM D 323 (repeatability =
0.25 psi,  reproducibility =  0.55  psi).   The precision  does not  change  due to
the magnitude of the sample measurement, and the cyclopentane
displays the same  variance  as the other  fuels.   The variability appears to be
due to the methods of RVP analysis.

    Based  on the  correlation coefficient  and  the slope of  the  best  fit line,
there appears to be no relationship between storage time and vapor pressure.

Recommendations

    Improvement  due to  normalizing  the  data indicates  that  differences  in
technique   used   by  the   individual  labs   may  have  contributed   to  the
reproducibility  of  the  results.    Quality  assurance  techniques,   such  as
calibration using pure components, may  enhance accuracy and reproducibility of
all the test methods and should be considered.
1818c

-------
                                    ATTACHMENT  A

                EPA Vapor Pressure Correlation Testing Laboratories

    Laboratory                        Number     Manual   P 176             Gas
                                      Samples    'Tank    Herzog   SvRI     Chrom.

Andre Jasmin                            16
Ultramar Canada Inc.
CPP 2055
165 Rt. Des lies
St. Romuald, Quebec  G6W 5M4
(418) 837-3641

E. Lyn Turner                            8
AM Laboratories
6207 Tri-port Court
Greensboro, NC  27410
(919) 854-0747

Richard Yurek                            8
E.W. Saybolt & Co., Inc.
400 Swenson Drive
Kenilworth, NJ  07033
(201) 245-3100

O.L. Salter                              8
Charles Martin Analytical Laboratory
P.O. Box 1558
Pasendena, TX  77701
(713) 477-1542

Roger Penstermaker                       8
Phillips Petroleum Co.
1186 Adams Bldg.
Bartlesville,  OK  74004
(918) 661-3508

Harold Honaker                           8
Ashland Petroleum Co.
Automotive/Product Application Lab
P.O. Box 391
Ashland  KY  41114
(606)  329-5495

T. Hok Gouw                             24
Chevron Research Co.
P.O. Box 1627
576 Standard Avenue
Richmond,   CA  94802
(415) 620-2417

-------
                                ATTACHMENT A (cont.)

                 EPA  Vapor Pressure  Correlation Testing Laboratories

     Laboratory                        Number     Manual   P 176              Gas
                                       Samples      Tank    Herzog   Swrl      Chrom.
 Leo Zafonte                               8           1
 GARB
 Haagen-Smit  Laboratory
 9528 Telstar Avenue
 El  Monte,  CA  91731
 (818)  575-7054

 Baron  Munchausen                         8           1
 Litton Core  Laboratory
 P.O.  Box 34282
 8210 Mosley  Road
 Houston, TX   77234
 (713)  943-9776

 Bob Snavley                             16           1                          1
 Koch Refining Co.
 P.O.  Box 2608
 Corpus Christi,  TX   78403
 (512)  289-8511

 Larry  White                               8           1
 Diamond Shamrock Refining/Manufacturing Co.
 P.O. Box 490
 Three  Rivers,  TX  78071
 (512)  786-2536

 David  Brown                             16           11
 Phillips 66  Company
 P.O. Box 866
 Sweeney, TX   77480
 (409)  491-2265

 Jack Gilliland                          16           1                          1
 Chevron USA
 Quality Control Laboratory
 P.O. Boc 1272
Richmond,  CA  94802
 (415)  620-2100

 C.  Iverson                              16           11
 Phillips 66 Company
Refinery - NGL Laboratory
P.O. Box 271  SPUR 119
 Borger,  TX  79007
 (806)  273-2831

-------
                                ATTACHMENT A (cont.)
                EPA Vapor Pressure Correlation Testing Laboratories
     Laboratory
 Dave  Pillon
 Shell Canada Limited
 Sarnia Manufacturing Centre
 Refinery Lab
 Corunna, Ontario  NON 1GO
 (519)  481-1301

 Ian Taylor
 Shell  Canada Limited
 Oakville Research Centre
 3415  Lakeshore Road, West
 P.O.  Box 2100
 Oakville, Ontario  L6J 5C7
 (416)  827-1141

 Terry  Boyle
 Iringin Oil
 P.O.  Box 1260
 St. Johns
 Mew Brunswick, Canada  E2L 4H6
 (506)  633-3000

 Calvin LaBauve, Jr.
 Murphy Oil USA, Inc.
 Meraux Refinery
 P.O. Box 100
 Meraux, LA  70075
 (504)  271-4141

 William Marshall
 NIPER
 P.O. Box 2128
 Bartlesville, OK  74005
 (918) 337-4345

Rick Schomaker
 Oldsmobile Division
 920 Townsend Street 40-2
 Lansing,  MI  48921
 (517) 377-5019
Number
Samples

   8
Manual
 Tank
P 176
Herzog
                                                                   SwRI
Gas
Chrom.
   8
  16
   8
   8

-------
                                ATTACHMENT  A  (cont.)
                EPA Vapor Pressure Correlation Testing Laboratories
    Laboratory
Rich Macintosh
Petro-Canada
Trafalgar Refinery
Oakville, Ontario  L6J 5B5
(416) 825-1703

Jane Jackson
Alberta Research Council
P.O. Box 8330, Station P
250 Karl Clark Road
Edmonton, Alberta  T6H 5X2
(403) 450-5100

Tony Novak
ARCO Petroleum Products Co.
1801 E. Sepulveda
Carson, CA  90745
(213) 584-8201

ftoy Sabourin
National Research Council of Canada
Montreal Road, Bldg M-9
Ottawa, Ontario  K1A OR6
(613) 993-2186.

Jeff Mann
AC Spark Plug Division, CMC
1601 N. Averill Avenue
Flint,   MI  48556
(313) 257-8246

Dr. Leo Duffy
AMOCO Corporation
Amoco Research Center
P.O. Box 400, Bldg 600
Naperville,  IL  60566
(312) 420-5225
Number
Samples

  16
Manual
 Tank
P 176
Herzbg
                                                                   SvRI
Gas
Chrom.
  16
 24
   8
  16

-------
                                ATTACHMENT  A  (cont.)

                EPA Vapor Pressure Correlation Testing Laboratories

    Laboratory                        Number     Manual   P 176             Gas
                                      Samples     Tank    Herzog   SwRI     Chrom.
Michael Davis                            8                   1
Department of the Army
Corpus Christi Army Depot
Mail Stop 27
SDSCC-QLC
Corpus Christi, TX  78419
(512) 939-3555

Bob Pateraon                             8                   1
Corpus Christi Petrochemical Co
P.O. Box 10940
1501 McKenzie Road
Corpus Christi, TX  78410
(512) 241-6450

Charett Navarrete                        8          1
Curtis & Tomplins Lits.
401 Canal Street
Wilmington, CA  90748
(213) 549-6727

Sue Frederick                            8                   1
EXXON Company, USA
P.O. Box 9000
E. 22nd Street
Bayonne, NJ  07002
(201) 925-6976

Edwin Calvin                            16          1        1                 0
EXXON Research and Engineering Co.
1900 E. Linden Avenue
Bldg. S105
Linden, NJ  07036
(201) 474-2516

Ted Eckman                              8           1
General Motor Corporation
Milford Proving Grounds
Bldg 31
Milford, MI  48042
(313) 685-6032

Syd Feldesteen                           8          I
Chrysler Corp; Fuels & Lube Lab
12800 Oakland Avenue
Bldg. 138-2W, Code 5830-u
CIMS 418-17-07
Highland Park, MI  48203

-------
                                ATTACHMENT A (cont.)

                EPA Vapor Pressure Correlation Testing Laboratories

     Laboratory                        Number     Manual   P 176
                                      Samples     Tank    Herzog   SwRI

Bill Okamoto                             8          1
Ford Motor Co.
P.O. Box 2053
SRL Rm 3198
Dearborn,  MI  48121
(313) 594-1016

R.L. Akers                               8          1
Marathon Petroleum Co.
1300 S. Fort Street
Detroit, MI  48217

Paul Michaluk                            8          1
Petro-Canada R&D
2489 N. Sheridan Way
Sheridan Park
Ontario, Canada
L5K 1A8
(416) 896-6805

R.E. Campell                            16         1                  1
Port Arthur Product Control Lab
Texaco USA
P.O. Box 712
Port Arthur, TX  77641
(409) 982-5711

Clark Ellison                            8          1
Coastal Refining & Marketing, Inc.
P.O. Box 521
1300 Cantwell Lane
Corpus Christi,  TX  78403
(512) 887-4252

Campos Roschetzky                        8          1
Champlin Petroleum Co.
P.O. Box 9176
1801 Nueces Bay Blvd
Corpus Christi,  TX  78408
(512) 887-3362

Carl Scarbro                            16          11
USEPA
2565 Plymouth Rd
Ann Arbor,  MI  48105
(313) 668-4209

-------
                                ATTACHMENT A  (cont.)

                EPA Vapor Pressure Correlation Testing Laboratories

    Laboratory                        Number     Manual   P 176             Gas
                                      Samples     Tank    Herzog   SwRI     Chrom.

Frank Gallagher                         16
U.I.C.
P.O. Box 865
Joliet, IL  60434

Marvin Jackson                           8
General Motor  Corporation
Research Lab
3400 Mound Road
Warren,  MI  48090
(313) 947-1778

Dale Burton                              8
Automotive Testing Laoboratories, Inc.
Route 33, Bldg 40 AT TRC
P.O. Box 289
East Liberty, OH  43319
(513 666-4351

Doug Probert                             8
Shell Canada Limited
Scotford Refinery
P.O. Bag 23
Fort Saskatchewan, Alberta  T8L 3T2

Bill Finley                             16
Texaco USA
P.O. Box 37
Convent,  LA  70723
(504) 562-7681

J. Weyman Carson                         8
Texaco Chemical Co.
P.O. Box 968
Port Arthur, TX  77641
(409) 982-5711

Dale Fuller                              8
Valero Refining Co.
5900 Up River Rd.
P.O. Box 9370
Corpus Christi,  TX 78469
(512) 289-3254
I846c

-------
                                 Attachment B

                        ENGINEERING OPERATIONS DIVISION
      Evaluation of Several Methods to Measure Volatility of Motor Fuels
                            Basic Letter of Inquiry

                                    (date)
(Name of Laboratory)

Dear (Name)
    As a  part  of its effort to reduce  levels  of ozone,  the EPA is considering
ways to restrict the amount  of hydrocarbons released into the atmosphere.  One
approach  would  be  to  set limits  on the  volatility of gasoline.   A possible
mechanism for enforcement would be through the measurement of vapor pressure.

    The  purpose  of this  letter  is  to  confirm  our  telephone  conversation
regarding the  involvement of  your organization  in a program  to  evaluate the
accuracy,  reproducibility,   and  precision of  ASTM's  proposed procedure  for
measurement  of  vapor  pressure.   This  procedure is  contained  in ASTM  P 176,
"Proposed Specification  for Automotive  Spark-Ignition Engine  Fuel".  Details
can be found on page 1059 of  the  1986  Annual  Book  of ASTM  Standards.   We are
also interested  in evaluating any  comparable  methods such  as  those using the
Herzog  unit  from  UIC,   the  automated  instrument  from  Southwest  Research
Institute, or other methods,  e.g.,  gas chromatography. (Merge*)

    Each  of  the  twenty or more laboratories in  the program will receive a set
of eight  blind  samples  for  each measurement  technique.   Included  in  the set
are  several  examples  of gasoline,  a  pure  component, and  an  alcohol  blend.
They will be shipped  in one-quart  bottles  using  DOT-approved  containers  and
should arrive  within  two weeks.   Instructions, a data  sheet,  and a  sample
transfer connection will accompany the shipment.

    I  will   collect  and  analyze   the  data.   A   copy of  our  report will  be
forwarded to you upon  completion  of the program.   In addition  to the  results
of your   analyses,  we   are   interested  in  your   opinions  concerning sampling
procedures  aa   they  apply  to  dispensers  found   at  typical  gas  stations  and
"ideal" containers for  vapor pressure samples.

    The timing  of this  project  is somewhat critical to  the schedule  for the
overall program.  As a result, your  results must be here  by October 20 so that
we can analyze  the data  and  make  our  recommendations  on  test  methodology,
outpoints for enforcement  actions  and directions for further evaluations.   If
you will  have trouble meeting this  deadline, please  contact me.
    The sentence merged  here  confirmed the type(s) of  equipment  to  be used by
    the particular laboratory.

-------
                                      -2-
    We appreciate your involvement  in  the  program and  welcome your suggestions
and  comments.   The   success   of  this  program  will   help  develop  better
communication on technical  issues between EPA and the  private sector.  If you
have any questions, please feel free to call me.
                                  Sincerely,
                         Marcus Haubenstricker, Ph.D.
                         Fuels and Chemistry Services
                            Testing Programs Branch
                     EPA Motor Vehicle  Emission Laboratory
                              2565 Plymouth Road
                           Ann Arbor, Michigan 48105
156lc

-------
                             Attachment C

                   ENGINEERING OPERATIONS DIVISION
  Evaluation of Several Methods to Measure Volatility of Motor Fuels
             Seal for Certified Analysis and Sample  Label
                                 Seal
CAUTION:  This container contains an EPA vapor pressure correlation sample
          and is not to be opened except by a qualified analyst
          experienced in gasoline vapor pressure measurement.   If leakage
          is evident please note the sample number and call Carl Scarbro
          or Dr. Haubenstricker at (313) 668-4209 or (313)  668-4378 for a
          replacement.   The sample bottle must be chilled to 32-34°P
          before opening.   Please note all precautions in ASTM P 176
          before opening the container and performing the analysis.
                             Sample Label
                       Sample No:  12345678

                       Method/Equipment:
                       (  )   Water  Bath and Gauges
                       (  )   Herzog Instrument
                       (  )   SWRI Instrument
                       (  )   Gas Chromatograph

-------
    A2.4. J ripfi Kltrr <* ttulh.
    A2.4.4 Rrfrafliimrtcr.' temperature compcnulcxl.'
  hand held. 0 lu 10 * Oiii (I.3330- 1.3479).

  A 1.5 Rragrufi   j
    A2..VI  I'tailv t) Kraft-nit— Hragriil-grndc clirmi-
  ral* will he used iff ill lesls. Unless iilhrrwi.tr indicated.
  il il intended  thai all reagents shall conform  lo  ihc
  specifications of the Committee on Analytical Kcngrnls
  of (he American Chemical Society, where luch jnccifi-
 nitons art available.'* ()lher grades may be used, pro-
  vided il it firs! ascertained lhal Ihe reagent is ofsuffi-
 cienl  purity lo prrmil  its use wilhoul lessening  the
 accuracy of Ihc determination.
   A2.5.2  Purity uf  Holer—Unless  otherwise  indi-
 cated, reference lo water shall be understood to mean
 distilled water of water of equivalent purity. See Spec-
 ification I) 119). Type IV.
   A2.5.3  .-firriMir  (Danger—I-ilremely  flammable.
 See Anne* A I.))
   A2.S.4  *  llcplinr (Wirnliig—flammable. Harmful
 if inhaled. See Anncn AIM conforming  lo  material
  lived  in 7CM Methods I) 3KI. I)hi I. D 2699.  and
  1)2700. or (Klroleum spin!  MI/KO (U'.rnlng-ll.ini-
  nuhle. Vajior harmful  See Annci  AI 6)  conforming
 lo II* specification or equivalent. Commercial grade
' quality il adequate foi these solvents.

 A l.t  l*rtp*rallo
. vi lhal II
 lomeler
 until opt

 .» r,

 room ten
                           Attachment  D

            ENGINEERING  OPERATIONS  DIVSION
   EvaluaU-on  of  Several  Methods  to  Measure
                Volatility  of  Motor  Fuels
P  176  Procedure  for  Measurement  of  Vapor  Pressure
                                             nrier
                                             0 V
                                             llrii.
                                             table
dilioni  on idrntiral  trsl  material would, in the long
run, and in Ihr normal and correct o|>er«lion of Ihcirsl
method, eircrd Ihe following values only in one cast
in twenty:                                   I
 and pipel filler and Moppet Ihe cylinder
   A2.7. 3 Invert the cylinder, holding Ihe shipper wiih
 a finger, and allow the waler 10 flow into Ihe sample.
 Keeping Ihe cylinder in a lioii/onl.il altitude, shade it
 vigorously for exactly 2 mm. two lo three sliokes |*r
 second, using  12)  lo 250 mm (5 lo 10 in.) strokes.
   A2.7.4 Immediately place Ihe cylinder in a vertical
 position on a tiliuiion-frec suifncc and allow the run
 lenls lo settle undisturbed for 5 mm.
   A2 7.5 Open the rover plate of the refriictnmeier
 and using a soft clolh, or soil tissue paper moiMcneil
 with waler. wipe Ihe prism and  covei plale  Oiy hnih
 surfacn with a soil, dry clolh or tissue. Then close Un-
 cover plale.
   A2 7.6 Air dry I lie pi|x:i and install the pipel Win
 Krmove Ihe *lop|ici from llie giadiialrd olimlci. lilt
 Ihe  cylinder. ,ind  insrn  Ihe pi|X-| Milh  Hie Up in Ihr
 water phave. Squee/e Ihe pi|>el  fillef and disrhaigc -in
 through Ihe pi|>rl  lo puige any I'uel lhal  may  havr
 enlcreil the  pipel lip. Draw up mlo Ihc pipel ahoul one-
 half of Ihc  waler phase in Ihe bottom of the cylinder
 1 ry to avoid drawing in any emulsion.

   Noil A2.2 — If the water is totally arrvoihed into Ihe
 furl phase,  this indicates lhal the sample contains  i
 high coordination  (greater  than  25  volume %)  uf
 elhanol or heavier alcohol.

   A2.7.7 Remove Ihe pipel from the giadualed cylin-
 der. dry Ihr outside of Ihc pi|X'l wiih lissur lo remove
 any fuel on Ihe suiface. and then disihaicc a few diop*
 of waler phase  to purge any fuel in Ihe lip. Till Ihe
 lefracloineler with Ihe eyepiece up. Wulioul lilting the
 rover plate, place  the pipel Up at Ihe top of the co»er
 plale between Ihe hinges ami diM haigc enough of Ihe
 waler phase lo rover al leasi half of the prism.
   A2.7.H lininc-di.ik-ly poinl Ihe instrument toward a
 window or  oilier source of h^lil  s» the light enters Ihe
 piism fmni above Look lliroiiKh Ihe eyepiece and take
 Ihe reading  al the poinl where the dividing line between
 light and dark crosses the scale.
   Null A2 1— lake Ihe reading within III s of dis-
 chatging Ihc walei phase onto Ihe refrat lomeler pn*ni.
 Ihc  reading  may  change with lime as constituent*
 evaporate or Ihe sample otlici-wise changes in compo-
 sition.

   A2.7.9 Record (he scale leading.

   1 the AO In.luMii.il Muid leMei M.»lfl HH«I A
  I rnificraiurc Ciiiiirvnuiril 1 1 j nd Krlijchinieirr ha> net" I
 tuiuMr Al*ti an oiuivjlt-nl mjy he iivtl.
   ""KiMjrnl ChrinH ah. Amriii-jn I lirniiial Si«-«l> S|«f'»
 i-Jlinnl.- Am. < 'hrniii »l S.IT . WaUunflon. IX 1 1" IUM""""|
 cm Ihe lr*lm| u( reu|>.riiii nol hMeil hy Ihr Amrnun I **"'""'
 S<»ieir. *ee  -Krjfrni Chciuu-alt ami Sumlinh." hi <«""*
 R.'S.n. I). Vjn N,,MunJ( o . liu . New V«nl. NV.iml -|l»n«"
 Suie* l*)uiinai-«iiicia ~
                                             •rcial
                                             «e in
                                             rfore
concluding lhal an alcohol is  present, a minimum
leading of 0.5*  llrii mini be observed.  However, it is
still possible lhal a sample having a rrfrarlomclrr read-
ing of leu than 0.5* llrii may contain  a low level of
alcohol.

A 2.9  Precision  tnd  DUs
   A?*) I  Ihe prrrision of this  proposrtl trsl  method
as deirrminrd hy stalisliral e»aminalion of mlerlalxv
ralory results is  as follows:
   A2.9 !(/)  Hr/irainliilily—Tht diffeiencr  hriwrrn
successive lest rrsiills. obtained  hy  Ihr  same  operator
with the same Apparatus under conslanl  operating con-
                                                              Range
                                                       0-1.0 vol % ukoliol
                               Rr|erilion of
Ihe lest method, eiceed  the following values only in
one case in Iwenly:
          Range
   0-1.0 vol % alcohol
                                                                                                                           Rrproduribihly

                                                                                                                        0.251 (Rrfmrlomrtcr
                                                                                                                              Reading)*
   A).
                     Ir.SI MKTIIOI)  KOK VAPOR
                                       HIM.  (DRV MKIIIOD)
   A2.1.2  nia\ — There being no criteria for measuring
bias in these leM product combinations, no statement
ol bias ran he made

             Of  SPARK-IGNITION  KNGINK
A 1.1  S*-«|>*
   A.I. 1.1  I his pro|Nisrd lest melhod rovers Ihe deter-
mination nf ihe absolute vapor pressure (Nolr A J.I) of
gasolines and gasohne-otygenalr blends.
   Noll A.I.I — Dccause theeilcrnal almospherir pies-
sure is counteracted hy Ihe atmospheric pressure ini-
tially present  in Ihc air chamber, Ihe "vapor pressure"
is  an absolute pressure al I GOT (37 8"C) in pounds-
force per square  inch  or (kilopascals (kPa - kN/rn').
1 his vapor pressure differs from Ihe true vapor pressure
of the sample due  lo some small sample vapori/jlion
and Ihe presence of air in Ihe confined space.
   A.VI.2  Ihe values  staled in  inch-pound units are
siandaid.

A J.I  .Summary of Method
   A.V2.I The fuel chamber of Ihe vapor pressure ap-
paratus is filled wiih Ihc chilled sample and connected
lo ihc air chamber al I DOT (37.8'C). The apparatus is
immersed in a hath al  100*1- and is shaken periodically
uniil  a constant  pressure is observed on Ihe gage at-
tached lo Ihe apparatus. 1'hr gage reading, suitably
coricclcd. is reported as the vapor pressure.

AJ..1  Significance and IJ*c
   A.V3 I Test Method l> 323 cannot he used to deter-
mine the va|>or pressure ofgasolinr-onygenaie blends
which rontain wairr-eilrarlablc oiygenales because the
fuel sample comes into contact with waler. This pro-
nosed lesl method is a modification ofTcsl Method
I)  J2J where contact with waler haj been eliminated.

A.1.4  Apparajm
   AJ.4.1 Ihc tonsliuclion of the required apparatus
is drs. lined in Anne« AI of Test Melhod U J2J.
 A.V5 HeigenM
   A.I 5. 1  I'linly <>(  Rrnxrnli—lJtt  rea|;enl  glide
chemicals in all icsis.  Unless otherwise indicated,  it is
inlrndrd lhal all reagents conform to the specifications
of Ihe Commillce on Analytical Reagents of Ihe Amer-
ican Chemical Society where  such specifications are
available.' Olhrr grades may be used, provided il is first
ascertained that the reagent is of sufficiently high purity
lo permit its use without lessening Ihe accuracy of the
determination.
   A3. 5. 2  Atelonr  (Danger — tutremely flammable.
See Anno Al.l)
   A3. V 3  Naphtha  (Danger— r»liemely flammable.
See Anne. A I 2).

A3.6 Handling of Sumplei
   A3. 6. 1 The  enlrcme sensitivity of vapor  pressure
measurements lo losses through evaporation  and Ihc
resulting changes in composition is such as  lo require
Ihe utmost precaution and Ihe most meticulous cart in
Ihe handling of samples.. The provisions of this section
apply lo all samples for vapor pressure deirrminjiions.
   A3.6.2  .Sample in accordance wiih Practice  D 4057.
e«cerH lhal  waler displacement (1 1. 3.1. 8 of  Practice
D4057) must nol be uved
   A.lft.3 Sample i'lwiiami'r .V/;r--The  si/r  of  Ihe
sample container from which Ihe vapor pressure sample
is taken is I  ql (I L). Il will  be 70 to 80 % filled with
ihe sample.
   AJ.6.4 /Viviiur/iMK
  A3 6.4(0 Deierminr vapor pressure as Ihr fust ivsi
run  on a sample. Do nol withdraw more  than  one
sample from the sample container for this lest.
  A. 1.6.4(7) Protect samples from eicessive beat prior
lo lesling.
  A 1 6 KJ)  Do nol mi samples in kjky conuinen.
Uivard them and obtain new samples.

-------
                                          0-2 PropoMl P 176
   A3.A.4«) Discard samples lhal have separated inio
(wo phases and obuin new samples (see Note A3.4).
   A J.6.3 Samplr Handling Temperature—In all cases,
cool the sample container and contend  lu J2 lo 34'l;
(0 lo  I*C) before the container is opened. To ensure
sufficient lime luf reach this temperature, directly  mea-
uire trie icmperiture of a similar liquid in a like con-
tainer placed in the cooling bath at Ihe  same lime as
the sample.

AJ.7  rrrp«nlk» for Tnl
   A3.7.1 Vetificalton of Sample Container  Filling—
With  the sample al a temperature of 32  lo 34'F  (0 lo
I'C),  lake Ihe container from  the cooling bath,  wipe
dry with an absorbent material, unseal il. and examine
its ullage. The sample content, ai determined by use of
a suitable gage,  must be equal lo 70 lo 80 % of Hie
container capacity.
   A).7.l(/> Discard Ihe sample' if its volume is less
than 70 % of Ihe container capacity.
   AJ.7 !(.?) If the container is more than 80 % full,
pour out enough sample lo bring the container contents
within Ihe 70 lo 80 % range. Under  no  circumstance
may any sample poured out be relumed lo Ihe  con-
tainer.
   A.) 7.2 Ate Saturation nf Sample  in Sample  Con-
lamn
   A) 7 1(1) With Ihe umple again at a temperature
of .12  lo 34T (U lu I'C) take the conlainei  from the
cooling  bath, wipe it dry wiih  an absorbent  material.
unseal il momentarily,  taking care  10 prevent  water
entry, meal il. and shake il vigorously. Return il lo the
bath for a minimum of 2 min.
   A).7.2(J) Repeal AJ.7.2 (I) twice mort. Return the
sample to Ihe bath and keep il there until  Ihe beginning
of the procedure (A3.8).
   AJ.7.3 Preparation of Fuel Chamber—Observe the
apparatus preparation procedure of A.3.8.J. then  store
the stoppered  fuel chamber and Ihe  sample transfer
connection in a refrigerator or ice-water bath for a
sufficient lime lo allow the chamber and Ihe connection
lo reach a temperature of 32 lo J4'F (0  lo I'C).  If an
ice-water bath is used, keep Ihe chamber upright and
not immersed over Ihe lop of the coupling threads. The
transfer connection is inserted into a plastic bag lo keep
il completely dry during cooling.
   A3.7.4 Preparation of Air Chamber—Observe Ihe
apparatus preparation procedure of  A3.8.5. Connect
Ihe gage lo (he air chamber and close the lower opening
securely with a dry No. 6Vi rubber stopper. Make sure
the stopper is inserted far enough lo securely dose the
vent hole in Ihe air chamber connection. Immerse Ihe
air chamber lu al least I  in. (2) mm) above its lop in
Ihe water bath maintained al 100 ± 0 2'F( J7.8 ± 0 I'C)
for not less than 20 min. Do not remove the air chamber
from  Ihe Water bath until the fuel chamber  has  been
filled with the sample as described in A3.8.1.

AJJ  ProrHwt
   A3.8.1 Sample Transfer—With everything in  read-
iness, remove Ihe chilled sample container  from Ihe
bath, dry il with absorbent material, uncap il. and dry
and insert Ihe chilled transfer apparatus (see Fig. A3.1).
Quickly  place the chilled fuel chamber, in an inverted
position, over  Ihe sample delivery luhe of the transfer
 apparatus. Invcil the mine system rapiilly MI lli.ii inr
 fuel chamber is upright, with the end of the delivery
 luhe lourhing Hie bottom of Ihe fuel chamber. Till inc
 fuel chamber to overflowing. Withdraw I he deliveiy
 lube from Ihe fuel chamber while allowing the sample
 lo continue flowing up  lo the moment of complete
 withdrawal.
   A3.8.I (I) Caution—Make  provision for suitable
 collection and disposal of Ihe overflowing fuel to avoid
 Tire hazard.
   A3.8.2 A\scnihly nf Apparatus—Immediately re-
 move Ihc air chamber from the water bath and imme-
 diately dry Ihc etlerior of Ihc chamber with absorbent
 material giving particular carr  to Ihe connection be-
 tween the air chamber and the fuel chamber. Remove
 Ihe stopper after drying  and  immediately couple the
 two chambers. Not  more than IU s shall be consumed
 in coupling Ihe two chambers.

   Not I- A.V2—When Ihe air chamber is removed fiiini
 Ihe water l>alh. is dried,  and  the stopper is  removrtl.
connect il lo Ihe fuel chamber without undue move-
 ments through Ihc air. which could promote  enchange
of room temperature air with Ihc IUOT (37.8V) air in
Ihe chamber.

   A3.8.3 lntrtnliictit>n ttf Apparatus inltt llnth—I urn
 Ihe assembled vapor pressure apparatus upside  down
 lo allow the sample in Ihe fuel chamber In run into ihe
air chamber. With Ihe apparatus still inveilrtl. shake il
 vigorously eight limes in a  direction parallel lo  the
length of (lie a|>paralus. With (he gage end up. immerse
 Ihe assembled apturalus in Ihe bath, maintained al 11*1
 ± 0.2T (37.8 t O.rC). in an inclined  position so Ihil
Ihe connection of Ihe fuel and  air chambers is below
Ihe waler level and may be carcfull, eiamincd for leaks.
If no leaks are observed, further immerse Ihe apparatus
lo at least  I in. (2) mm) above the  lop of Ihe air
chamber. Observe Ihe apparatus for leakage throughout
Ihe test. Discard Ihe test at any lime a leak is  delected.

   NOTI A3.3—Liquid leaks are more difficult lo detect
than vapor leaks, and because the coupling between ihe
chambers is normally in  Ihe liquid section of Ihe ap-
paratus, give Ihe coupling particular attention.
   NOTK A3.4—After the apparatus has been immened
in Ihe hath,  check  ihc remaining umple for phase
separation. If the sample is contained in a glass con-
tainer, this observation can be  made prior lo sample
transfer (A3.8.1). If Ihe sample is contained in a non-
Iransparcm container, shake Ihe sample vigorously for
5 s arid then immediately pour a portion of Ihe remain-
ing sample into a clear glass container Immediately
after shaking this sample again for 3  s. observe the
sample for phase separation. If this, sample is  not clear
and bright, and free of a second phase, discard Ihe lest
and Ihe sample.

   A.V8.4 Arra.il/rf/itivi/ of Vapor Piruurr—After Ihf
assembled vapor pressure apparatus has been immersed
in Ihe hath  for al least 3 min, lap  Ihe pressure gage
lightly and observe Ihc reading. Withdraw the apparatus
 from Ihe hath and repeal A3.8.3. Al  intervals of not
Ins than 2 min. perform A3.8.3 until a total of not leu
 than five shakings and gage readings have been made:
continue thereafter,  if necessary, until Ihe  last I*"1
consecutive gage readings are constant, indicating equi-
                                                                                                   D-2 Propoaal P 176
lihrium attainment. These operations normally require
211 lo 3(1 min  RCM\ the final gage pressure In Ihe nearest
0(15 |>si (0 2) kl'a) lor gages with intermediate gradu-
ations of O.I |ni (II.) kfa)ur less and lo Ihe nearest O.I
psi lor gages with graduations of 0 2 lo 0.3 psi (1.0 lo
2.3 kl'a). and  record Ihe value as Ihe "uncorrecled vapor
pressure" of the sample.  Without undue  delay temove
the pressure gjge and. without attempting lo remove
any liquid which may be (rapped in the gage, check  its
reading against  lhal of Ihe manometer while both are
Subjected  lo  a common steady pressure which is  no
nmrc than 0.2 psi (1.0 k Pa) different from Ihe recorded
"uncorrecled vapor pressure." If a difference is observed
between the gage and manometer readings. Ihc differ-
ence shall he added lo or subtracted from Ihe "uncor-
rct'ird  vapor  pressure" recorded  for Ihe  sample being
tested, and Ihe resulting value shall be recorded as Ihe
vapor pressure of Ihe sample.
   Nun A3.3—Cooling the assembly  prior lo discon-
necting Ihc gage will facilitate disassembly and reduce
ihe amount of hydrocarbon vapors released into Ihe
loom.
   Nun A.I d—I'l-ri/icuiinn urn'i,ige
by directing a small jcl of air mm us Hourdnn tube for
al least  5  min.  Kinse both chambers and the sample
transfer connection several limes with petroleum naph-
Iha. then  several limes with acetone, then  blow dry
using dried air. Stopper Ihe fuel chamber and place it
in Ihe refrigerator or ice-water bath for Ihe neit test.
   NOIE A3.7—If the purging of Ihe air chamber is
done in a bath, be sure lo avoid small and.unnoliceable
films of floating sample by keeping the bottom and lop
openings of Ihe chamber closed as they pass through
the water surface.

A3.* Precautions
   A 3.9 I  Grots errors can be obtained in vapor pres-
sure measurements if Ihe prescribed procedure is  not
followed carefully.  The following list emphasi/cs  Ihe
importance of strict adherence lo Ihe precautions given
in Ihe procedure.
   A.3.9.!(/)  Chfrkmg the Fressure (Sate—Check  all
gages against a manometer after each lest in onler lo
ensure high precision of results (A) I 4) Read all gaR"
while ihe gages are in a vertical  position and after
lapping them lightly.
   A3.9.1(7)  Shake Ihe c6nlainer vigorously lo ensure
equilibrium of Ihe sample with Ihe air in Ihe conlainei
(A3.72)
   A39.l(.()  Cherking for Isak.i—Check Ihe 1 prior a
lus before and during each test for  both  liquid  ami
vapor kaks (Al.1.6 of Test Method D 323 and Note
3).
   A3.9.1(1)  Sampling—Because initial sampling ami-
Ihe handling of samples will greatly affect Ihe final
results, employ  the utmost precaution  and Ihe  mosi
meticulous care lo avoid losses  through evaporation
and even slight changes in composition (A3 6.3  ami
A3.8.1). In no case shall any part of Ihe apparatus nxll
be used as Ihe sample container previous  to actually
conducting the lest.
   A3.9.1(5)  Purging  the  Apparatus—Thoroughly
purge the pressure gage, Ihe fuel chamber  and Ihe air
chamber lo be sure they are free of residual  sample
(Ihis is most  conveniently done  at the  end of Ihe
previous lest. See A3.8.5.) Il is important to remove all
waler from Ihe  apparatus before cooling Ihe gasoline
chambers and healing Ihe air chamber. In high-humid-
ity conditions be alert for and  avoid condensation on
Ihe transfer connection and interior walls of Ihe ipp*-
ralus
   A.3.9. l(rt)  Coupling the Apparatus—Carefully ob-
serve Ihe requirements of A3.8.2.
   A3.9.1(7)  Shaking the Apparalut—Shake the appa-
ratus "vigorously"  as  directed  in  A3.8.3 in order 1.1
ensure equilibrium.

A.t.lO  Rrpml
   A3. III.I  Ki'i»irii"f Keiulit—Report lo the  nearest
0.115 psi (0 25 kPa) or O.I psi (0.5 kPa) the  gate result
observed in A3.8.4. after correcting for any difference
between the gage and manomeler. as the "vapor pres-
sure" in pounds-force per square inch (or kilopascals)    I
without reference lo temperature.                      '
                                                1060

-------
                                           O-2PropOMlP 178
                                                                                                                                                         D-2 Propotal P 176
  AJ.II Prrcbioa a«4 BUi

    A J.I 1.1 fr,rifiim—The precision of lhi$ proposed
  ten  method  (us  foi  been determined, bul is under
              Chi fled Sample
            fansfer Connection
                                                    Sludy.
                                                       A3.11.2 Dim—There being no criteria 'for measur.
                                                    ing bias in these test-product combinations, no state
                                                    menl of bias can be made.
                                                         Chilled Gotolint Chombv
FIG  AJ I
                 la)
                  Contanrr
                fo Transfer
              of Sam;*
            Sl.,urk4 Skcir
                                 (b)
                               ™ Ck»un>
                                  I* Smlr
      (c)
Gtsolnr Chamber
»Vtd Ovw l«jjd
 Orlivrry 1ub«
IVwilion of Syclrm  for
   5am(4f Transfer
                                      olkW •( Tnuihnlm S*«>k> to Cuollw CkMbrf ham Oa«*-1 >•« rMlilont
    A4. PROPOSED TEST METHOD FOR VAPOR  PRESSURE  OF  SPARK-IGNITION ENGINE
                           FUEL USING THE AUTOMATIC APPARATUS
 A4.I  Scop*
   A4.I.I This proposed lesl method cuvci> * deier-
 minalion of vapor pressure of gasoline and gasoline-
 oiygenalc Mends.
   A4.I.2 The values suied in inch-pounds uniu are
 the standard.

 A4.2  S»«urr of Method
   A4.2.1 The cold sample cup of the automatic vapor
 pressure  instrument is Tilled with the chilled, air-salu-
 rated fuel sample and connected to the instrument. The
 instrument operation is juried and the vapor pressure
 is automatically determined.

 A4J  Affuntm
   A4.3.1 Vapor Pressure Instrument''
   A4.3.2 Mercury Manometer, as prescribed in Test
 Method 0323. AI 6.
   A4.3.3 Cooler, appropriate apparatus consisting of
a cooled air or water bath, and preferably a refrigerator
lo cool and maintain samples, sample cups, and sample
transfer connection* at 32 lo 34'F (0 lo I'C) Do not
 use solid  carbon dioiide (dry ice) for this purpose.
   A4.3.4 Sample Transfer Connection, as illustrated
in Fig. I  of Test Method O 323.

A4.4  RnfeMi aarf Materials
   A4.4.I  C'yrlopealane.V.Vl psi( 68.3k Pa) vapor pres-
sure. (Dugef—Eilremcly flammable and harmful if
inhaled. See Anne> A 1.2).
   A4.4.2 Mrthyi C>*l,ipeniane.  4.50 psi (31.0  kPa)
 vapor pressure  (Dmngrr—Euremely flammable and
 harmful if inhaled. See Annei A 1.2).
                                                      A4.4.3 n-Penlane,  15.57 psi (107..1 kPa) vapor pres-
                                                    sure. (Danger—fiilremely  flammable and harmful if
                                                    inhaled See Anno AI 2).

                                                    A4.S Safely Prmullons
                                                      A4.S.I Recogni/e  the potential for fire and r t* ohuinfll
                                                   from Southwell Rewarch Inililulc An et|ui»aknl may
he the TIIM test lun on a umple: do mil u\c more than
our \-iniplc Iroin the siimple container fin this lesl.
   A4 h.4  (^) I'mlccl samples'from eneessive heal prior
to testing.
   Ain for Trtl
   A4 7 I  I'm/iiuiiiin i,j Sami'li' (imlainft filling—
With the sample at a lem|ieraliire 32 lo 34'F (0 lo I'C)
lake  the container from the cooler, wipe  dry  with an
jtnorhcnl null-rial, unseal it. and eiamine  ils ullage.
I he umple content, as determined by  use of a suitable
g:i£C.  shall be  equal lo  71) lo HO 71 nf the  container
ranauly.
   A4.7 !(/) Divjnl the umplc if its volume is less
than 70 "/„ of the container capacity.
   A'l 7 !(.') If the container  is more  than 80% full,
(Kiiu out enough uniple In bring I he innuincr ronlrnls
within the  70 to  Kl) % range. Under no circumn
inverted position over the sample delivery lube of the
transfer apparatus Invert the  entire system rapidly sn
that Ihe sample cup  is upright, with the end of Ihe
delivery lube touching the bottom of Ihe cup. (The end '
of Ihe  delivery lube  is cut  al an angle lo prevent
interference with  the delivery flow.) I'lll the  cup in
overflowing and withdraw  the delivery lube from the
cup.
   Nun A4.2—Caution—Make provisions for suitable
restraint and dis|iosal of any spilled gasoline lo avoid <
fire ha/ard.
   A4.9.4 .ItirfriWr <•/ Jp/*ir
-------
                                         0-2 Proposal P 176
                                                    1
                                         0-2 Propo««l P 176
   NOTE A4..1—Aflcr initialing the lot. check  the re-
 maining sample Tor phase separation. If the sample is
 contained in a glass coniaincr. this observation  can he
 made prior to sample transfer (A9.9.J). If the sample is
 contained in a  •nnlransparenl container, shake the
 sample vigorous^ for 5 s and then immediately pour a
 portion of the rrVnaining sample inlo a clear-glass con-
 tainer  Immediately after shaking this sample again fur
 5 s. observe the  sample for  phase separation.  If this
 sample is nol  clear and bright, and free  or «  second
 phase, discard the lest and the sample.
   A4.9.6 Al the end of the instrument operation, read
 and record the displayed DPM reading.
   A4.9.7 Perform the Leak Test described in Section
 V. Part E. of Ihe instrument manual." Discard any
 result obtained where leakage is indicated.

 A.IO Calrublio* of Result*
   A4.IO.I  Use   Ihe  OPM  readings  obtained  in
 A4.S.3 (/) for Ihe pure compounds versus their known
 vapor pressures lo construct a graph relating DPM and
 true vapor pressure.
   A4.I0.2 Using Ihe graph, determine fhc true vapor
 pressure corresponding lo the DH>1 reading of each of
 the unknown samples.

 A4.ll  Report

   A4.I I.I Report the DPM readings obtained for rath
 of Ihe pure- compounds.
   A4.11.2 Rrpoit Ihe DPM readings obtained for each
 of Ihe unknown samples.
   A4.I 1.3 Report Ihe indicated I rue vapor prcvsure of
each unknown as determined in A4.11.2.
                                    *

A4.I2  Precision and Bias
   A4.I2.I Precision—The precision of this proposed
test method has nol  yel been determined.
   A4.12.2 Bias—There being no criteria for measur-
ing bias in these test-product  combinations, no stale-
menl of bias can be made.
    AS.  PROPOSED  TEST METHOD FOR WATKR TOLERANCE (PI I ASK SEPARATION) OK
                                 SPARK-IGNITION  ENGINE  FUEL
A5.I  Scop*
   A).1.1  This proposed Icsl method  determines the
ability of gasoline-oxygenate blends lo retain water in
solution or in a stable suspension at the lowest temper-
ature to which they are likely lo be exposed in use.
   A3.1.2  The values staled in SI units are Ihe standard.

A5.J  Summary of Method
   A5.2.1  The sample  of fuel  is cooled  slowly to its
expected use temperature and is observed for phase
separation. The apparatus of Test Method D 2500 or a
dry ice/isopropyl alcohol bath  may be used. The pro-
cedure of Test Method  D 2500. in which  the sample is
cooled rapidly lo Ihe lest temperature by immersion in
a balh that is maintained at a considerably lower tem-
perature, must not he used due lo the large temperature
gradient employed, and because phase separation in
gasoline-oxygenate blends has a relatively long bul un-
predictable induction period.

A5J  Significance and  Use
   A5.3.1 Some oiygenale-containing  fuels, gasoline-
alcohol blends in particular, have a very limited ability
lo retain water in solution or in stable suspension, and
if Ihe amount of water in the blend exceeds this limit.
Ihe fuel  will  separate inlo a  lower  oxygenale-rich
aqueous phase and an upper oxygenale-lean hydrocar-
bon phase. The most important factor governing Ihe
ability of a specific fuel to retain water without such
separation is ils temperature. This  test method is in-
tended to determine Ihe  maximum  temperature al
which Ihe  fuel will separate. The I Oth percenlile 6-h
minimum temperatures or  10'C (50"F).  whichever is
lower, for Ihe lime of year and geographic area of the
United Slates in which  Ihe fuel may be used are tabu-
lated in  Table 4 of this proposed specification. These
temperatures  represent  Ihe  maximum temperatures
above which the fuel blend must nol separate inlo two
distinct phases
   A5.3.2  Nole lh.il in this lest actual separation of the
sample inlo two distinct phases separated by a single
common boundary (which  may consist of a layer of
emulsion) is Ihe criterion for failure. Formation of a
haze without such separation inlo distinct phases is  not
cause for rejection.

AS.4  Apparatus
   AS.4.1  Sample Conlainfr—This may be as specified
in 4.1, of Test Method 1)2500, bul any glass coniaincr
of about  100-mL capacity capable of accommodating
a thermometer may be used.
   A5.4.2  TherniHineten mccling Ihe requirements of
Specification 1.  I, wilh a range appropriate lo the lesl
require mcnl.
   AS.4.3  ('ink. lo  fit sample container borcil cenli:illy
for Ihc lest thermometer.
   AS.4.4  Citiiling  ttmli may he of similar dimensions
lo those specified  in Tesl  Mcfhod  O2500. 5.7 and
provided wilh n jacket, disk, and gasket as specified by
5.4 through 5.6  of Test Method D2500, filled wilh an
equal-volume mixture of water and "permanent" anti-
freeze, and provided wilh refrigeration coils capable of
reducing ils temperature lo -4()"C (-40T). Alterna-
tively, any balh  of adequate  site may he used.

AS.S  Procedure               '
   A5.5.1  Rinse oul Ihe sample coniaincr wilh some of
Ihe fuel lo be tested. Drain.
   A5.5.2  Pour about 40 mL of Ihc fuel inlo Ihe sample
container. The  precise amount  is nol critical, hul il
must  be enough lo submerge Ihe thermometer bulb
adequately, without being so much as lo require  an
excessive amount of cooling lime.
   A5.5.3  Seal the sample  container.  Locale  a ther-
mometer of Ihe appropriate temperature range wilh  the
bulb approximately al Ihe center of Ihe fuel sample.
  A5.5.4 Cool ihc sample by intermittent immersion
jn or circulation of Ihc coolant. The fuel is either stirred
conlinously or vigorously shaken. Starting al a lemner-
jiure about  I6"C (30°F) above  Ihe lesl temperature,
cool Ihe sample al a maximum  rate of 2*C  (4*F) per
ininiilc lo ihc temperature given in Table 4 for Ihe
water tolerance required. If phase separation is observed
prior lo reaching Ihe lesl temperature. |>er AS.3.2, Ihe
"cooling" It.i/c ixuni temperature is recorded. Then Ihe
sample is  allowed lo warm while being shaken fre-
quently or stirred and a "warming" haze point temper-
ature is recorded.  The "cooling" and "warming" tem-
peratures arc averaged  lo  determine Ihe actual  hare
point.  Ihc "cooling" and "warming" haze point tem-
peratures may be repealed for improved accuracy.

A5.6  Report
   A5.A. I  Report "pass" if no separation occurs al Ihe
specified temperature for Ihe water tolerance class re-
quired, otherwise note Ihe temperature for phase scpa-
ration (see A5.3.2) and report "fail."

AS.7  Precision and  Ilia*
   AS.7.1  I'rffisinn—The precision for this proposed
lesl melhod has nol been determined.
   AS.7.2 Bias—There being no criteria for measuring
bias in these  lesl-produci combinations, no statement
of bias can be made.
                                           APPENDIXES

                                    (Nonmandalory Information)

 XI. PROPOSED TEST METHOD KOR VAPOR-LIQUID RATIO OK  SPARK-IGNITION  ENGINE
                     KIEL (MODIFIED TEST  METHOD  D 1533 PROCEDURE)
XI.I Scope
   XI I.I  This proposed lesl  melhod covers a proce-
dure lor mcDSuimg Ihc volume of vapor formed al
atmospheric pressure from a given volume* of gasoline
or gasoline-oxygenate blend. The ratio ol these volumes
is expressed as the vj|xir liquid (V/L) ratio of Ihe fuel
al Ihe temperature of Ihe lest.
   XI.1.2   Ihe  values staled in inch-pounds units are
Ihc standard.

XI.2 Summary of Method
   XI.2.1  A measured volume of liquid fuel al 32 lo
4UT (0 lo 4'C) is introduced through a rubber septum
inlo » mercury-filled burcl. The charged huret is placed
in a Icmperaiure-conlrolled waler balh. The volume of
vapor in equilibrium wilh liquid fuel is measured al Ihe
desired  temperature or temperatures and Ihe specified
pressure, usually  760  mm Hg. The  vapor-liquid ratio
(>'/;.) is then calculated.
   XI.2.2  If it is desired to know Ihe temperature cor-
responding lo a given I'//.. Ihe vapor-liquid ratio is
determined al  several temperatures and  Ihe selected
pressure. The results  are plotted and Ihe  temperature
read al  Ihc given  Vfl..

X1.3 Significance end Use
   XI.3.1  The tendency of a  fuel lo vaporize in auto-
 mobile fuel systems  is indicated by Ihe  VfL ratio of
 that fuel at conditions approximating  those in critical
 pans of Ihe fuel systems.
   X 1.3.2  Tesl Melhod D 2533 is nol applicable  lo
 fuels containing  alcohol, ethers, or other compounds
 soluble in glycerine. This proposed test melhod substi-
 tutes mercury for glycerine as Ihe confining fluid.
   X 1.3.3  A relationship between vapor-liquid ralio of
 gasoline-oxygenate blends and vehicle performance has
 nol been dcfer,mined.

 XI.4 Definition
   XI.4.1  vapor-liquid ralio of a fuel, al  any specific
 temperature and pressure, is Ihe ratio, al thai temper-
 ature and pressure, of Ihe volume of vapor in equilib-
 rium wilh liquid lo Ihe volume of sample charged, as a
 liquid, al 32'F(0'C).
   Noir XI.I — fhis ralio differs from  the  ahsolulr
 vapor-liquid ralio because corrections are nol m.ide for:
 (/) liquid sample  expansion wilh increasing tempera-
 ture. (!) decrease  in liquid sample volume by vapori-
 zation, (0 dissolved air in Ihe  liquid sample, and (/I
 deviation from Ihe perfect gas law.

 XI.5 Appinius
   X1.5.1  y/L Uiiri-i," constructed of borosilicalc glau
 in accordance wilh dimensions shown in Fig. X 1. 1.1 he
 short bottom arm is closed wilh a rubber serum  Nil lie
 stopper, a U.S. Army Medical Corps type.
   X1.5.2  rrrsMirr Cimlriil Equipment:
   XI.5.2(7) A  250-mL leveling bulb containing mer-
 cury, attached In Ihe \'/L hurel by rubber tubing at
 shown  in Fig. XI.2. The lop of Ihc leveling bulb must
 be filled with a  drying lube containing mercury  vapor
 absnibcnl" packed between  balls of glass wool. This
 drying lube is used lo minimize Ihe escape of mercury
 vapor.
    XI.5.2(2)  Means for  measuring Ihe  difference in
 liquid  level  between the I'/l.  burel and Ihe leveling
 bulb. A calhelomeler or similar optical leveling device
 is suitable. A millimetre scale can be used to provide a
 rough estimate.
    X I.5.2(J) Baromi-lrr. accurate lo 0.5 mm llg.
   XI.5.2(4) A  mercury manomeler with I mm divi-
 sions,  required  only for measurements  al pressure
 appreciably above or below Ihe  prevailing almosphciii
 pressure. The manomeler shall be connected wilh rub
    " The V/l. bum No. 247I8-M2 and rilihrinon tiuppci Nn
 247HI-CS. manufanuml by Continental Olan Blowing. Co .
 26J6 S. Hill Si..  Un Anitln. CA 90007. have been foun.1
 laliilanory for lhi> purpoic.
    " Rnhorto available from J. T. P-ikrr Co.. Phillipst-ufj. NI
 008AS has hern found saliiTaciorjr (CM (hit purrhrw.
                                                1064
                                                                                                                                                              1065

-------
                                 Attachment E

                        ENGINEERING  OPERATIONS DIVISION
      Evaluation of Several Methods to Measure Volatility of Motor Fuels
                     Vapor Pressure  Correlation Data Sheet
Laboratory Number: '                    Laboratory
Date Received:      	            Address
Date of Analyses:  	
                                 Instructions:

    Each  of  the  enclosed sample  containers  have  been  filled  to  70 to  80$
capacity as  required  by ASTM P  176.   After you have cooled the samples  to 32
to 34°F, continue sample  preparations  according to  sections  A  3.7.2 or A 4.7.2
(depending on which equipment is used).

    The table  below is for  the  results  from each of eight  samples.   The test
operator should fill  out  the sample's vapor pressure  (in  psig)  in the labeled
rows for each method/equipment used.

    If you have  any comments or suggestions for assessment of  gasoline  vapor
pressure,  please list them on the back of this form.
                             Results of Vapor Pressure Analyses (psig)
                       	 Sample Number 	
Method/ Equipment         1      2      3    ' '4	5	6    ' '7      8
P 176/gauges & bath:
Herzog Instrument:
SwRI Instrument:
Gas Chromatograph
Note:  The results above are to be expressed in psig.


Analyst:		  Telephone;     	  Date:
    When  finished  with testing  please  telephone the above data  to  Dr.  Marcus
Haubenstricker at (313) 668-4378 or (313) 668-4209 and mail this form to:

                                United States Environmental Protection Agency
                                Engineering Operations Division
                                2565 Plymouth Road
                                Ann Arbor, Michigan 48105
                                Attention: Dr. Marcus Haubenstricker

-------
                              Attachment  F

                      ENGINEERING OPERATIONS DIVISION

Evaluation  of  Several  Methods  to  Measure  Volatility of  Motor  Fuels
                  Reported Manual Tank and Gauges Results
Replicate :
Lab. -
No.
1
2
3
4
5
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
42
43
44
45
46
47
48
49
Results

1
9.06




8.4
8.73
9.3
9.00
9.00


8.6


8.60

8.90
10.00

9.13

8.40

8.60

8.1





8.5
8.0

8.8
8.99
10.0
9.0
10.1

8.5
8.8
8.3
9.74
8.72
8.73


2
11.82




11.1
11.40
11.9
11.75
11.50


11.0


11.00

11.60
12.00

11.55

11.20

11.35

11.0





11.7
10.8

11.7
11.79
12.7
11.8
11.2

11.1
11.7
11.5
11.71
11.51
11.51
-Sample Identification Numl
3
12.40




11.3
11.40
12.4
11.90
11.60


10.7


10.70

11.65
11.50

11.50

11.10

11.65

11.3



9.70

11.65
10.8

11.6
12.43
12.8
10.6
10.9

11.2
11.5
9.7
12.03
11.35
11.25
4
10.37




10.4
9.85
10.6
10.55
10.20


9.1


9.10

9.95
9.50

9.98

9.80

10.00

9.4





10.1
9.4

10.1
9.38
11.3
10.3
10.4

9.4
10.1
10
10.1
9.73
9.93
5
8.63




8.3
8.85
8.8
9.05
8.80


9.2


9.20

8.90
8.00

8.90

8.40

8.25

8.2





8.8
8.0

8.9
8.3
10.1
8.9
8.2

8.7
8.8
8.7
8.99
8.43
8.7


6
11.89



•
12.3
12.40
12.7
12.70
12.30


12.3


12.30

12.30
12.00

12.38

12.40

11.80

11.4



9.75

12.3
7.9

12.3
12.18
13.0
12.0
12.5

12.3
12.4
12.5
12.96
11.57
12.77


7
10.08




10
10.00
10.7
10.30
9.80


9.6


9.60

9.85
9.10

9.93

9.30

9.80

9.5





9.2
9.4

9.9
10.02
11.1
10.0
10.2

9.6
10.1
9.1
10.72
10.01
10.06


8
11.17




11.9
12.40
13.0
12.25
12.20


11.6


11.60

12.25
12.00

12.28

12.60

12.25

11.1



9.20

12.4
11.6

11.8
12.97
13.5
12.2
12.3

12.1
12.3
12.3
12.78
12.41
12.62

-------
                              Attachment   G

                      ENGINEERING OPERATIONS DIVISION

Evaluation  of  Several  Methods  to  Measure  Volatility  of  Motor  Fuels
                          Reported Herzog Results
Replicate Results

No.
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
[25]
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
42
43
44
45
46
47
48
49

1
8.64

8.92
8.7

6.95
8.49



8.60
9.40

8.55
9.41
8.40

8.80
9.60
8.46


8.4
4.05


8.20
8.98
8.80

9.4













8.99
8.44


2
11.54

11.64
11.15

9.85
10.66



11.25


11.15
12.08
10.80

11.45
11.60
11.11


11.5
10.00


11.20
11.58
11.50

12.0













11.75
10.97

-Sample Identification Numl
3
11.49

11.64
11.8

9.85
11.16



11.15
12.10

11.25
11.9
11.00

11.60
11.90
10.98


11.4
7.35


11.20
11.55
12.00















11.76
9.68

4
9.99

10.02
9.7

8.00
9.4



9.15
10.70

9.80
10.45
9.40

9.90
10.10
9.46


9.4
6.60


9.20
10.02
10.00

8.5













10.15
9.68

5
8.73

8.92
8.6

8.40
8.49



8.40
10.10

8.45
9.3
8.40

8.95
8.60
8.51


8.8
5.45


8.20
8.80
8.90

9.0














8.29



6
12.44

12.66
12.1

11.70
11.53



12.15
13.00

12.05
12.87
12.50

12.20
12.80
11.98



9.20


11.40
12.07
12.70















12.86
11.30



7
10.02

10.00
9.55

8.10
9.51



9.40
10.70

9.70
10.62
9.40

10.00
9.90
9.46


9.9
5.55


9.70
9.87
10.10

9.3













10.12
9.71



8
12.04

12.62
12.05

11.60
12.22



12.10
12.90

12.00
12.88
11.80

12.20
12.50
11.99



7.15


11.80
12.32
12.80















12.61
12.26


-------
                               Attachment   H

                       ENGINEERING OPERATIONS DIVISION

Evaluation  of  Several  Methods  to  Measure  Volatility of  Motor  Fuels
                           Reported SwRI Results
Replicate Results
T oK _ _
LidD . — 	 — 	
No. 1
1
2
3
4
5
6
7 8.9
9
10 8.76
11
12
13
14
15
16 8.46
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
42
43
44 8.3
45
46
47
48
49


' 2






11.7

11.54





11.06


























11.3





                            Sample Identification Number-
                             11.7     10.0      8.9     13.2     10.1     13.2

                            11.55     9.70     8.98    12.59     9.74    12.58
                            11.10     9.39     8.44    12.11     9.52    12.13
                             11.3      9.4      8.4     12.3      9.8     12.4

-------
                                                               Attachment  I
                                                    ENGINEERING OPERATIONS DIVISION
                           Evaluation of Several Method*  to  Measure the Volatility of  Motor Fuel
                                            Comparison of Repeatability and Reproducibility
   Measurement
    Technique
   Treatment of Data
Gauges £  Bath    "All"  data  (Note  1)
in accordance with
ASTM  P 176       Outliers  removed  (Note 2)

                   After  add'l processing
                                                           This  Study-
Afo. of  Repeat-   Reproducibility
 Labs   ability   Actual  Normalized

   28    0.75 psi  2.48 psi   2.52 psi

   26      0.68      1.33       1.02
                                                                    ASTM's  1986  Study   	ASTM D  323	
              Repeat-    Repro-
              ability  ducibility

              0.82 psi   1.96 psi

                0.77       1.42

                0.71       1.3
Repeat-    Repro-
ability  ducibility
                                                                                            0.25 psi   0.55 psi
Herzog  (dry)
"All" data (Note 3)            21      0.76      1.59      1.31

Outliers removed (Note 4)      21      0.48      1.57      1.01
                                           0.55

                                           0.43
                           1.44

                           1.16
SwRI   Instrument All  data
                    (there were no outliers)
                                                        0.28
                                               1.32
                              0.42
                                                                       0.31
                                                     1.32
                   The data  from this study and ASTM's 1986 effort were processed in accordance with Section RR D-2-1007
                   of "Manual  of Determining Precision Data for ASTM Methods on Petroleum Products and Lubricants".
                   Note 1:
                   After substitution for
                   single reported  results
                   (lab 32,  sample  2) and
                   substitution for missing
                   pairs (lab 32, pairs 145
                   and 447)
                           Note 2:
                              Lab
                              No.
                              20
                              32
                              38
                              35
                              41
                              46
       Sample (s)
        excluded
          1 45
          all
          all
            6
          1 &5
          243
Note 3:
After substitution for
single reported results
(lab 13, sample 2; lab 32,
sample 2; lab 47,  sample 5)
and substitution for missing
pairs  (lab 24,  pair 648;
lab 32, pair 6s8)
Note 4:
Lab
No.
7
20
32
48

Sample (s)
excluded
1 US
1 &5
447
243
                                                                                                                     updated: 9/2/87 JTW

-------
                                           Attachment  J

                                    ENGINEERING OPERATIONS DIVISION

             Evaluation of  Several  Method*  to Measure Volatility of  Motor  Fuel*
                                          List  of Statistics
Method
Level of Data Reduction
Number of Sample Pairs
Number of Labs
Degrees of Freedom Replicates
Degrees of Freedom by Laboratories
Degrees of Freedom of Laboratories
Missing Single Results
Missing or Substituted Replicate Pairs
Mean Squares of Laboratories
Mean Squares by Laboratories
Mean Squares Replicates
Correction Term
Sums of Squares Samples
Sums of Squares Laboratories (missing data)
Sums of Squares Laboratories (no missing data)
Sums of Squares Sample Pairs
Sums of Squares by Laboratories
Sums of Squares Sample Replicates
Random Error
Estimate of Variance between Laboratories
Estimate of Variance by Laboratories
Variance of Repeatability
Trial ratio f
Mean Square Coefficients
to
Kl
K2
Estimate of Variance in Laboratories for Reproducibi 1 ity
Estimate of Variance between Laboratories for Reproducibility
Estimate of Variance in Laboratories for Reproducibility
Variance for Reproducibility
Degrees of Freedom for Reproducibility
Precision
Repeatability
Reproducibility
P 176
Before Cut
4
28
107
78
27
3
2
0.0712
1.0829
2.6466
24562.4111
414.5952
71.4578
NA
570.5218
84.4688
7.6232
0.0356
0.1955
0.5058
0.0712
0.0658

0.4971
0.3738
0.1285
0.0354
0.4048
0.3402
0.7804
95.1757

0.75
2.48
P 176
After Cut
4
26
100
75
25
0
4
0.0579
0.0453
1.4311
23328.9228
380.3452
35.7780
NA
419.5190
3.3957
5.7922
0.0290
0.1732
0.0063
0.0579
1.2793

0.4979
0.3839
0.1184
0.0288
0.0174
0.1695
0.2157
40.0680

0.68
1.33
P 176
Normalized
4
26
99
74
25
1
4
0.0585
0.0279
0.7599
23239.5704
380.3897
18.9977
m
401.4499
2.0624
5.7922
0.0293
0.0915
0.0153
0.0585
2.0992

0.0498
0.3839
0.1184
0.0291
0.0107
0.0900
0.1298
50.4569

0.68
1.02
Herzog
Before Cut
4
20
74
53
19
4
2
0.0726
0.0910
1.8822
17597.8992
292.0917
35.7613
1R
332.6744
4.8215
5.3689
0.0363
0.2239
0.0092
0.0726
0.7975

0.4914
0.3853
0.1240
0.0357
0.0351
0.2333
0.3041
31.8117

0.76
1.59
Herzog
After Cut
4
20
76
55
19
2
2
0.0290
0.1994
1.6358
17614.2634
302.5768
31.0800
NA
344.6254
10.9686
2.2048
0.0145
0.1795
0.0852
0.0290
0.1455

0.4913
0.3824
0.1272
0.0143
0.0763
0.2080
0.2985
37.3541

0.48
1.57
Herzog
Normalized
4
20
76
55
19
2
2
0.0290
0.0740
0.6636
18211.0198
316.7506
12.6088
NA
333.4314
4.0720
2.2048
0.0145
0.0737
0.0225
0.0290
0.3918

0.4913
3824.0000
0.1272
0.0143
0.0283
0.0844
0.1270
41.1084

0.48
1.01
SwRI
Before Cut
4
4
16
9
3
0
0
0.0081
0.0402
0.7493
3581.7532
73.0824
NA
2.2478
75.6915
0.3614
0.1298
0.0041
0.0886
0.0160
0.0081
0.2019

0.5000
0.3750
0.1250
0.0041
0.0151
0.0937
0.1128
4.3107

0.28
1.32
SwRI
Normalized
4
4
16
9
3
0
0
0.0081
0.0283
0.0345
3833.3714
76.1060
NA
0.1035
76.4646
0.2551
0.1298
0.0041
0.0008
0.0101
0.0081
0.2861

0.5000
0.3750
0.1250
0.0041
0.0106
0.0043
0.0190
18.2432

0.28
0.42
NA,  not applicable

-------
       Attachment K: Total  Vapor Pressure vs. Days  Since Sampled
M
PJ

M
O
ft
(0
4J
O
E-i
         92 1
         90 -
         88 -I
         86
         84 -
         82 -
         80 -
         78
                                       y = 84.420   R = 0.02
                        10
                                     20
                                                  30
                                                               40
         Days Since Sampled (Note: one outlier  was not included)

-------