EPA-AA-TSS-PA-85-3
                     Evaporative Emissions
                     from the Fourteen Car
               DOE Gasoline/Methanol Blend Fleet

      (Report from Work Assignment 7, Contract 68-03-3192
              with Southwest  Research  Institute)
       Harry E. Dietzmann,  Southwest Research  Institute
                     Craig  A. Harvey, EPA
                           July 1985
                            NOTICE

Technical  Reports  do   not   necessarily  represent  final  EPA
decisions or positions.   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 with may form the  basis  for  a final EPA
decision, position or  regulatory action.

                    Technical  Support Staff
             Emission Control Technology Division
                    Office of  Mobile Sources
                  Office of Air and Radiation
             U.S. Environmental Protection Agency

-------
                              -2-

I.    INTRODUCTION

     The.purpose  of  this work was to  investigate  the effect on
evaporative-   emissions   of   mileage   accumulation   witn   a
gasoline/methanol/cosolvent   splash    blend    versus   mileage
accumulation  with the  base  gasoline.   This  test  program  was
carried out  under Work Assignment 7  of EPA Contract 68-03-3192
with Southwest  Research  Institute (SwRI) in  conjunction  with a
one  year  demonstration   program  at  SwRI  sponsored  by  the
Department of  Energy .  The  same vehicles  (fourteen  1984 Ford
Escorts)  were   included   in   the  previously  completed   Work
Assignment 4,  which  involved evaporative emission testing with
the gasoline-alcohol  splash blend,  the  control  gasoline,  and a
gasoline-alcohol   blend   matched  for   FEVI*   to   the  control
gasoline.  Those tests  were  run after about 10,000  miles  had
been accumulated  on  each vehicle, and  the  final tests reported
here were  run after  approximately 4,000 more  miles were  put on
each vehicle.   A  copy  of  the  Final  Data Report for Assignment 4
is included as Appendix A for reference.

     Work Assignment  7  involved  evaporative emission testing of
all  fourteen vehicles with their original  charcoal canisters,
and retesting  after  the  canisters had been switched between the
control  and  blend vehicles  (i.e.,  blend  vehicles  tested with
the  control   canisters and control  vehicles  tested with  the
blend   canisters).   This   was  done   to   separate  canister
differences   from vehicle  differences.   Canisters  from  two
control  and   two  blend  vehicles were  also weighed  before  and
after  each  segment  of  the SHED  test  to help in understanding
the  load-purge   cycle.    The  control  fuel  at  the  time  of
evaporative  emission  testing  was a winter  grade (11.5 psi RVP)
unleaded  fuel,  but   two  vehicles  from  each  group  were also
tested  with  the  summer  grade control  fuel  (9.1  psi  RVP) with
and  without   switched   canisters.    This   testing  provided  a
comparison back  to the original evaporative emissions data from
Work Assignment 4.


II.  SUMMARY

     Evaporative  emission  tests were conducted  on a 14-vehicle
fleet  undergoing driveability testing  for  DOE  at SwRI.   Seven
vehicles   (200   series)   accumulated   mileage  on  an  unleaded
control  fuel,  and  seven  vehicles  (100  series)  accumulated
mileage  on  a  gasoline-alcohol  blend  containing  four  percent
methanol, two  percent ethanol,  and two percent t-butyl alcohol
(TBA).   The   results  of  experiments conducted  with  these
vehicles are summarized below:


* FEVI: Front End Volatility Index = RVP +0.13  (% evap @  158°F)

-------
                         -3-

A.    SHED  evaporative  emissions  from the  blend  vehicles
      at standard FTP temperatures were 32 g/test compared
      to  9 g/test  for   the control  vehicles when  tested
      -with  the  winter  grade  control  fuel  (11.5  psi  RVP,
      EM-620-F) .  Due to  the  use of this higher volatility
      fuel, these  results can not  be  compared  directly to
      SHED tests using 9.0 psi RVP Indolene.

B.    SHED  evaporative   emission tests  conducted on  four
      vehicles  (two  of  the 100  series and  two  of the 200
      series)  with  the  summer grade  unleaded control fuel
      (EM-601-F)  produced  evaporative  emissions  of  3-5
      g/test.     Blend   and   control  vehicles   produced
      essentially  equivalent   SHED   emissions   with  the
      summer grade unleaded fuel  (EM-601-F).

C.    When  canisters  were randomly switched between blend
      and   control  vehicles   (i.e.  blend  vehicles  with
      control  canisters  and  control  vehicles  with  blend
      canisters),  the   SHED   evaporative  emissions  were
      essentially   equivalent    at    30-31   g/test   with
      EM-620-F.    Several  vehicles   in   each   group  had
      unusually  high  hot  soak   losses  which  tended  to
      overshadow any changes due  to canister  switching.

D.    The  diurnal contribution to the total SHED  emissions
      ranged from  52  to 72 percent for both vehicle groups
      with  EM-601-F.   With the  winter grade  control fuel
      EM-620-F,  the  diurnal contribution  ranged from 8 to
      36 percent.

E.    Canister   weight    gain   appeared   to    be   more
      vehicle-related than a  function of previous fueling
      history.   Canisters  which  gained  the most  weight
      during the  hot  soak also  had the most breakthrough.
      No  differences  in  blend or  control canisters could
      be   observed  from  canister  weight  gain  or  SHED
      evaporative  emissions during the  canister  weighing
      test  series.

F.    In general,  36  to  42 grams of  hydrocarbons/alcohols
      were  purged  from  the canister  during FTP operation,
      which was 4.3  -  5.5%  of   the  purged weight  of the
      canister.   Canister weights  at  the  end of  the hot
      soak  were  within  11 grams   (i.e., two percent)  of the
      weight at the start of the  diurnal.

-------
                              -4-

II.  DESCRIPTION OF VEHICLES AND FUELS

     A.    Vehicles

     The fourteen  vehicles  tested in this  Work  Assignment were
obtained   immediately    upon   completion   of   a  DOE   fleet
demonstration  program  which  was  under  contract  to  SwRI  and
administered  by the  Fuels  and Lubricants  Technology  Division.
The  fleet  consisted  exclusively  of 1984  Ford  Escorts,  seven
accumulating  mileage  on  the control fuel  (unleaded Amoco)  and
seven accumulating  mileage  on  the blend fuel.   Table 1 provides
a  list  of  the  vehicles  tested during  this  Work  Assignment and
the  nominal   odometer  readings at  the  time the  vehicles were
delivered for testing.

-------
                              -5-
                   TABLE  1.  DOE  TEST VEHICLES
Vehicle
  No.

  101
  102
  103
  104
  105
  106
  107

  201
  202
  203
  204
  205
  206
  207
Mileage
Accumulation
Fuel
Blend
Blend
Blend
Blend
Blend
Blend
Blend
Control
Control
Control
Control
Control
Control
Control
Nominal
Odometer
Reading
15,100
24,200
21,000
20,600
18,100
17,000
15,000
17,700
24,400
21,000
28,700
29,900
18,200
16,200

License
No.
261 EGK
264 EGK
260 EGK
266 EGK
265 EGK
273 EGK
259 EGK
263 EGK
267 EGK
272 EGK
271 EGK
268 EGK
262 EGK
258 EGK


Miles3
Winter
3,100
4,840
4,200
4,120
3,620
3,400
3,000
3,540
4,880
4,200
5,740
5,980
3,640
3,240
Summer
12,000
19,360
16,800
16,480
14,480
13,600
12,000
14,160
19,520
16,800
22,960
23,920
14,560
12,960
    Of  the  mileage accumulation on  the winter grade version  of
    the  blend  and  control  fuels,   it is  estimated  that  the
    majority  of  the  mileage  accumulation was  obtained  during
    the  winter  of 1983-84  (December 1983-February 1984).   Only
    two  or  three  tankfuls  were used at the  end  of the  mileage
    accumulation  in December  1984.

-------
                              -6-

     B.    Fuels

     Mileage  accumulation was  accomplished  using  two  fuels, a
control  f-ual  and  a  blend  fuel.    The   control  fuel  was  an
unleaded Amoco  fuel.   The blend was prepared by adding methanol
(4%  volume) ,  ethanol  (2%  volume)   and  TBA  (2%  volume) .   The
nature of  the DOE demonstration program  did not provide strict
control over  the  RVP  (or  FEVI)  of  the test fuels, and the fuels
varied with  the seasons.   That  is,  summer grade unleaded Amoco
was  used  during  the majority  of the  mileage  accumulation,  but
winter  grade  was  used  during December  1983,  January  1984,
February 1984,  and December 1984.    Table  1 also summarizes the
mileage estimated to  be  accumulated on winter and summer grades
of gasoline.  The pertinent fuel inspection data are summarized
in Table 2.

     Based on obvious differences  between the original control
fuel  (EM-601-F)  and the control fuel  in  use at  the  end of the
DOE  mileage  accumulation,  several  vehicles were tested with the
limited amount  of EM-601-F remaining to provide some comparison
back  to  the  original  data from Work Assignment  4.   The front
end  volatility  index  (FEVI)  of the summer grade fuel  (EM-601-F)
was  12.7, compared to  15.7 for  the winter  grade  (EM-620-F) .

-------
                              -7-
                 TABLE 2. FUEL INSPECTION DATA
                                             Fuel Code
Specification        ASTM

API Gravity 60 F    D-287

Distillation, F     D-86
   IBP
   10%  ,
   50%
   90%
   EP
Recovery, %
Residue, %
Loss, %

% Evaporated 158 F

RVP, psig             D-323

FEVIC
EM-601-F3
59.5
87
119
207
341
409
98.0
1.0
1.0
28.0
9.1d
12.7
EM-620-F0
60.0
81
109
198
331
383
98.0
1.0
1.0
32.0
11.5
15.7
c

d
EM-601-F was a  summer  grade unleaded Amoco gasoline used in
the   original   emission   testing   conducted  under   Work
Assignment 4 of Contract 68-03-3192.

EM-620-F was  a winter grade  unleaded  Amoco  gasoline  that
was being  used when the  final testing of  the DOE vehicles
was initiated.

Front end volatility index, FEVI = RVP +0.13  (% Evap 158°F)

A different  test  program which used  this  fuel measured the
RVP to be 9.8 psi.

-------
                              -8-

III. Test Procedure

     Employees used the  cars  in everyday driving, but used only
the  mileage • accumulation  fuels.  Following  the  completion  of
mileage  accumulation,   the  fuel tanks  of  the   seven  vehicles
operating  on  the  blend  fuel  were  run  to near  empty,  and  a
tank'ful  of the  control  fuel   (EM-620-F)  was  then  run  through
each blend vehicle over  approximately  a week's  time  using the
same  normal  driving  procedure  as  was  used  throughout  the
program  for  mileage accumulation.   This  was  done  to  minimize
any  short  term effects  of  the  blend fuel  so  the testing would
show mainly any long term effects of high volatility blend use.

     After  all  mileage  accumulation  and  the  tank  of  base
gasoline  in  the blend  vehicles was  used,  there  was a variable
length  of  time  ranging  from  about  one to  five  days before the
evaporative emission tests  were run.  All evaporative emissions
tests   were  conducted   according   to  the  Federal  SHED  test
procedure,  except  for   the   use   of  test  fuels  other  than
Indolene.  This means  that  following a fuel tank drain and fill
a  single  LA-4  preconditioning  cycle  was  run on the dynamometer
and  then  within 12-36  hours the tank  was drained and filled to
40%  of  capacity and the diurnal heat  build test was run.  This
was  followed   by   cold  and   hot   start  LA-4  cycles  on  the
dynamometer and then the hot soak part of the SHED test was run.

IV.  RESULTS

     A.    Original Vehicles with Original Canisters

     First,  all   fourteen  vehicles  were  tested  with  their
original canisters  using EM-620-F,.   These results are presented
in  Table  3,  and   illustrate  that  the   blend  vehicles  had  an
average  SHED  emission  rate of  about 32  g/test compared  to  9
g/test  for the  control  vehicles.  Although the test plan had an
adequate  number  of vehicles,  no replicate  tests were conducted
to   investigate   individual  vehicle   variability.    Since  no
repeats were run during this  set of  experiments, and there is a
wide vehicle-to-vehicle and  test-to-test variability  as shown
by  this   and  data  from  the  later  set  of  experiments,  the
conclusions are not as  firm as  they  otherwise might  be.

     The  average  diurnal  evaporative  emissions  of  the blend
vehicles  contributed about  one-third  (10 grams)  to the overall
SHED  emissions  compared to the two-thirds  (22  grams)  from the
hot  soak  segment  of   the  SHED  test.   The  average  diurnal
contribution  to   the   evaporative   emissions  of   the  control
vehicles was 0.8 gram,  whereas  the  hot soak produced an average
of  8 grams.  The  difference  in diurnal emissions  between the
blend and control  vehicles was statistically significant at  a

-------
                              -9-

99.5% confidence  level,  but the  difference  in hot  soak losses
was  not  Significant at  a confidence level  of 95%, due  to the
wide vehicle-to-vehicle  scatter.   The difference  in total SHED
emissions*"Was significant at a 99% level.*

     Initial  results  from  these tests  indicated  that  several
vehicles  (both control  and  blend)  had  noticeably higher  hot
soak  losses  than  several  apparently  similar vehicles.   Some
researchers  feel  that  the hot  soak  losses are more related to
carburetor  losses  rather  than  canister  hydrocarbon  losses.
Blend  vehicles which  had  higher-than-normal  hot  soak  losses
were  102,  104,  105  and  107.   Control  vehicles  which  had
higher-than-normal  hot soak  losses  were 202,  203  and  205.  The
significance  of these  high hot soak losses  is discussed in the
next  section  with  the  discussion  of  the  effect  of  canister
switching.

     B.    Canisters Switched Between Control and Blend Vehicles

     Canisters were switched  randomly  between control and blend
vehicles  to  investigate  the effect  on  evaporative emissions.
Test results  are  summarized  in  Table  3.   For the blend vehicles
equipped with  the control canisters,  the composite average SHED
evaporative emission rate was about  31  g/test with the 11.5 psi
RVP  fuel.   The diurnal segment  averaged  about 4  grams compared
to  27  grams produced during the  hot  soak.   Blend vehicles 102,
103, 104  and 107  produced noticeably higher  hot  soak emissions
than the remaining vehicles.

     When  the  control  vehicles  were   tested with the  blend
canisters with  the  same  11.5  psi RVP fuel,  the SHED evaporative
emissions went  from 9  g/test in  the  stock  configuration to  30
g/test.  Control  vehicles 202,  203,  and 205 produced higher hot
soak emissions  than the   remaining vehicles  in that group.  The
diurnal  emissions  were   about   11  g/test  while  the  hot soak
emissions  were   almost   20   g/test.    To  check  for  canister
differences   as   separate  from  vehicle  differences,  we  can
combine  these data from  the  switched canisters with  the above
data  from  the  original   canister/vehicle  tests.    The  fourteen
tests  with  the   100 series  canisters  can  be  compared  to the
fourteen tests  with the   200 series canisters.   This comparison
shows  the  diurnal  emissions  from  the  blend  canisters  (100
series  average 10.4  grams)   to  be  greater  than  the  gasoline
     The  statistical  test used was  the T  test  for independent
     groups of data.

-------
                              -10-

canisters (200 series,  average  2.3  grams)  at a 99.5% confidence
level.  Due  to  the wide scatter  in hot soak data,  there  is no
significant difference  in  hot soak or  total  emissions  at  a 95%
confidence level.    However,  if the  higher hot  soak  tests are
omitted  from the  data,  the  blend  canisters average  6.5  grams
per hot  soak  compared to 3.3 grams for the gasoline canisters,
and the difference  is  significant at  a 99% level.  In this case
the  total  SHED   emissions  average 15.6  grams  for the  blend
canisters versus  5.6 grams for the gasoline  canisters,  and the
difference is significant at a 99.5% level.

     It is  interesting  to  note  that in tests with the canisters
from the  blend vehicles the hot soak emissions were about twice
the diurnal  emissions.   For those tests with control canisters,
the hot  soak was  7-11  times greater  than the  diurnal emission
rate.   Table  4  summarizes  the  diurnal   contribution  to  the
overall SHED  emissions.   In general,  the diurnal segment of the
SHED accounted for  50-75 percent  of the composite SHED emission
rate  when  the   blend  and  control  vehicles  were  tested  with
EM-601-F.  When  these  same  vehicles were tested with the winter
grade gasoline,  the diurnal segment accounted for an average of
8  to  12 percent  of  the SHED emissions for  the control canisters
and  31  to   36  percent  of  the SHED   emissions  for  the  blend
canisters.

     Based  on the  data presented  in  Table 3,  it appears that
the higher hot soak losses  attributed to vehicles 102, 104, 105
and 107  were also observed on  three  of these vehicles when the
canisters were switched and these vehicles were tested with the
control canisters  (vehicles 102,  104 and 107).  A similar trend
was observed for  the  control vehicles  where 202,  203  and 205
produced  noticeably  higher hot  soak  emissions  with  both the
blend  and  control  canisters.   Since  the  hot  soak  emissions
accounted  for the  majority  of  the  SHED  emissions  when  using
EM-620-F  and since  it appears that  the  vehicle  may influence
hot soak  emissions  more than the  canister, it is very difficult
to determine  the effect of  canister switching.

     Although these data indicate  some increase in evaporative
emissions from the  blend canisters  (100 series) relative to the
gasoline  canisters  (200  series),  when tested with the 11.5 psi
RVP gasoline, this  can not  simply be attributed to the presence
of  alcohol  in the  mileage  accumulation fuel.   The  other  major
difference  between  the mileage  accumulation  fuels  was  their
volatility,  since  the  blend fuel  was a  splash blend  with no
attempt at volatility  matching.   Therefore it is uncertain from
these  data  alone   whether  mileage   accumulation  on  an  alcohol
blend  fuel  with  a  matched  volatility  to  a  base gasoline would
show any difference  in evaporative  emissions.

-------
                     -11-

                   TABLE 3

SUMMARY OF SHED EVAPORATIVE EMISSIONS RESULTS
           (11.5  psi  RVP TEST  FUEL)
                          SHED Evaporative Emissions, g
Vehicle3
101
102
103
104
105
106
107


201
202
203
204
205
206
207


101
102
103
104
105
106
107


201
202
203
204
205
206
207


Fuel
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F


EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F


EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F


EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F


Canisterb
101
102
103
104
105
106
107
Avg.
S.D.
201
202
203
204
205
206
207
Avg .
S.D.
202
201
207
204
206
205
203
Avg.
S.D.
102
101
107
104
106
105
103
Avg.
S.D.
Diurnal
6.28
4.92
14.50
10.22
16.13
3.77
13.20
9.86
4.94
0.50
0.32
0.44
0.33
2.07
0.37
1.31
0.76
0.67
1.69
3.85
- 6.80
2.04
2.17
10.26
0.34
3.88
3.48
10.20
17.23
4.78
11.37
15.26
6.85
10.38
10.87
4.36
Hot soak
7.13
37.44C
7.22
57.82C
12.25C
5.38
24.68C
21.70
19.74
2.67
10.98°
12.35C
2.63
20.02C
2.20
7.90
0739
6.61
1.25
58.44C
52.56C
47.84C
3.77
2.40
23.97C
27.18
25.47
9.37
43.20C
39.90^
5.03
27.53C
3.40
8.27
19.53
17.03
Total
13.41
42.36
21.72
68.04
28.38
9.15
37.88
31.56
20.11
3.17
11.30
12.79
2.96
22.03
2.57
9.21
9.15
7.08
2.94
62.29
59.36
49.90
5.94
12.66
24.31
31.06
25.62
19.57
60.43
44.68
16.40
42.79
10.25
18.65
30.40
18.78

-------
                             -12-
  101
  102
  201
  202
  101
  102
  201
  202
                     Table 3   (continued)

         SUMMARY OF SHED EVAPORATIVE EMISSIONS RESULTS
                    (9.1 psi  RVP TEST FUEL)
EM-601-F
EM-601-F
EM-601-F
EM-601-F
EM-601-F
EM-601-F
EM-601-F
EM-601-F
202
201
Avg.
S.D.

102
101
Avg.
S.D.

101
102
Avg.
S.D.

201
202
Avg.
S.D,
1.68
2.24
1.96
0.40

0.43
4.45
2.44
2.84
1.09
4.32
2.71
2.28
2.77
6.57
4.67
2.68

1.43
5.79
                                                         3,
                                                         3,
                                                  61
                                                  08
                        6.30
                        2.29
                        47TO
                        2.84
                                              0.06
a)  100  Series Vehicles  accumulated  mileage  on alcohol  blend
   fuel  containing  4%  methanol,  2%  ethanol  and  2%  tertiary
   butyl  alcohol  added  to  the  baseline  commercial  unleaded
   Amoco gasoline.   200  Series Vehicles  accumulated  mileage on
   commercial unleaded Amoco gasoline.

b)  The  "canister"  column  indicates  that  tests on  the  11.5 RVP
   fuel  were  first  run  with  the  vehicle's  original canisters
   and  then   canisters were  switched between the  two  sets of
   vehicles.    Then  testing  was  done with  the  9.1  RVP  fuel,
   first with  the  canisters switched and  then with  them back on
   their original vehicles.
c)  Higher hot  soak test  results  which  may  be related  more to
   carburetor factors than canister factors.

-------
                              -13-

     C.    Comparison of Vehicles to Original Tests

     Several SHED  tests were conducted  with vehicles operating
on  EM-60r-"F- to  provide  a  comparison   back   to  the  original
baseline.   Only two  vehicles of  each  group  were  included in
this  study.  These  vehicles  were tested  with their  original
canisters  and  with  the  canisters  switched.  These  results are
presented  in Table  5,  and are  summarized  and compared  to the
original tests  conducted in Work Assignment  4.   The difference
between blend and  control  SHED evaporative  emissions was not as
apparent  as  it  was  during  the  original  tests.   Switching
canisters did not  have  any effect  on  SHED emissions.  It should
be  noted  that   only two  vehicles,  rather  than  the  original
seven, were  used  in  these  tests.   The intent of these tests was
to  establish that  the  vehicles had  not shifted  significantly
since  the   original  tests   in Work  Assignment  4.   The  blend
vehicles  were  down  slightly  and  control  vehicles  were  up
somewhat, but in neither case was a major shift observed.

     D.    Canister Weighing

     Two blend  vehicles (101 and  102)  and  two control vehicles
(201  and  202)  were  selected for  additional  experiments   that
included  canister  weighing  during  different  segments of  the
SHED  test.   The purpose of these  tests  was  to investigate the
relationship between the  hydrocarbons adsorbed  on the charcoal
and  the  hydrocarbons measured in  the SHED.   The hydrocarbons
adsorbed onto  the carbon  were determined by canister weighing.
These  plus  the  hydrocarbons measured  in the SHED test (i.e. the
evaporative  emissions)   would  be   the  total  vapor  generated.
These  results  are presented  in  Table 6  for  vehicles 101,   102,
201  and  202 with their  original canisters.  In reviewing these
data,  it is important to recall  that  two of these vehicles  (202
and  102) were  identified  as  possibly  having higher soak losses
due  to  carburetor  losses  rather than canister  breakthrough
losses.   The  possibility  is  also  apparent  with  these  two
vehicles during  these tests.   For  example,  vehicle 101 produced
2  g/hot  soak  while vehicle  102  had about  20 g/hot  soak.  A
similar  trend was observed with vehicle 201,  which produced  7
g/hot  soak  as compared  to 18 g/hot soak for  vehicle 202.   Two
of  the three hot soak  tests  for  vehicle 201 were  1 g/hot  soak
but  the third test had higher  (about 17 grams) hot soak losses.

     Canister weight gain  appeared to  be  more related  to the
specific   vehicles  than   to   the   fuel   used   for   mileage
accumulation.    For  example,   vehicles  101   and   201   had
essentially  equivalent  canister weight  gain (36  to 38 grams),
although  one  was  a  blend  vehicle  and  one  was  a  control
vehicle.  Vehicles  102  and 202 were identified as vehicles  with
high   hot   soak   losses.   This   was   also   observed  in  these

-------
                              -14-

experiments where  canister weight  gain for these  two vehicles
was 45 to  50  grams,  with the majority  of  the  weight gain being
observed in the hot soak segment of the test.

     Table 7  summarizes the  actual canister  weights  that were
observed during these  experiments.   The canister weights before
diurnal  testing  were  800  +  15  grams  for canisters  from both
blend  and  control vehicles.   The  diurnal canister  weight gain
was  16+6  grams  with  control  and  blend  vehicles  being
essentially  equivalent.   After  the  vehicles  had  completed the
FTP cycle, canisters  were re-weighed and  FTP  losses during the
canister purge were  determined.   These data are also presented
in Table  7 and illustrate that  the  canisters  generally lost 33
to  42  grams  during  the  FTP  purge.   No  difference  between
canisters  from blend  and control was observed  for losses during
the  FTP  purge.    The  after-hot  soak  canister  weights  were
generally within 11 grams of the before-diurnal canister weight.

V.   QUALITY ASSURANCE

     The   areas   requiring  quality  assurance   on  this  work
assignment  are SHED  calibration  and  retention check,  HC FID
calibration,   and   balance   for    canister   weighing.    SHED
calibrations,  SHED  retention checks   and  HC  FID calibrations
were performed using  procedures and equipment specified in the
Federal  Register.   Records of these  calibrations are available
to  confirm  that   the  SHED  and  FID-related  calibrations  were
valid during the period when emission testing  was conducted.

     The  balance  used  for  weighing  canisters  was calibrated
prior  to  each individual test using  a  2  Ib and a 1 Ib weight.
These  data were   included  on the^  canister-weighing  data sheet.
The balance  accuracy  for reference  weights  was  0.04 percent or
better  for  both   the  1  Ib  and  2  Ib  weights,  for  all  tests
requiring canister weighing (Table 8).

     No problems  were  encountered  with operation or calibration
for  the   SHED,   FID,   or   balance,   and  all   data  should  be
considered   valid  from   a  strictly   analytical  standpoint.
However,  it  must  be  recognized  that  vehicle  repeatability in
the  SHED  test,    particularly  with  vehicles  operating  on  a
relatively high  volatility fuel,  is not  as good  as  a typical
new  certification  vehicle.   Since  no   replicate  tests  were
conducted  and   several  vehicles   (blend  and   control)  were
identified as  having  higher than  normal  hot soak emissions, it
will be  difficult to draw  any conclusions as to  the  effect of
canister switching.

-------
                              -15-

VI.  RECOMMENDATIONS

     Investigation of evaporative  emissions  from vehicle fleets
provides   valuable   information    for   typical   evaporative
emissions.  However,  when  using these  fleets  to look for small
changes  due  to  fuel,  canister, and other  items,  it  is quite
conceivable  that  the  variability  of  the  vehicles  and  test
procedures  will  overshadow  any  differences  due  to  canister
switching.   If it  is desirable to pursue  additional  research
along  these  lines,  SwRI has  several suggestions that  might  be
of interest.

     1.    Vehicle  fleets  are  expensive to operate  and  it  is
           not   always   possible   to   "piggy-back"   emissions
           testing  on a  fleet test  that  is  using  the alcohol
           blend  of  interest.   Previous work  assignments  have
           used synthetic  blends to simulate gasoline and blend
           vapors  in  loading  mini-canisters.   Although  these
           studies have  provided  useful  information  on loading
           and  purging  characteristics,  it  is  difficult  to
           relate these  data  to equivalent  vehicle evaporative
           emissions.  The desirable approach would  be to have
           a  bench  scale  system  that  would  load  and  purge
           full-scale  canisters   using  actual  gasoline  and
           blend   vapors.   The   system  could   be  used  on  a
           24-hour  basis  to   simulate  mileage  accumulation  on
           many  canisters  simultaneously.   It  would  also allow
           flexibility  for testing  canister  capacity,  as  well
           as  evaporative  emission  testing  of the  canisters on
           slave vehicles.

           Another  advantage  of  -this  type of   system  is  that
           various  fuel  blend  formulations could  be evaluated
           using  various combinations  of  alcohols, co-solvents
           and gasoline.   This system would allow  investigation
           of  azeotrope  formation  with  alcohol-gasoline blends.

     2.    Using   the   aforementioned   evaporative   emission
           generator,    it   would    be   possible  to   conduct
           experiments    that    would    provide    a    better
           understanding  of  azeotrope  formation  with  various
           alcohol blend combinations.   Vapor  composition could
           be  monitored  for methanol,  co-solvent,  and detailed
           hydrocarbons  to establish  the  relationship between
           vapor  composition   in a  typical  gas  tank  and those
           vapors adsorbed during  normal and hot  vehicle purges.

-------
                             -16-
TABLE  4.  DIURNAL CONTRIBUTION TO  OVERALL SHED EMISSIONS FROM

BLEND AND CONTROL  VEHICLES WITH VARIOUS CANISTER COMBINATIONS
Vehicle
Group
Blend
Control
Blend
Control
Blend
Control
Blend
Control
Blend
Control
Fuel
EM-601-F
EM-601-F
EM-601-F
EM-601-F
EM-601-F
EM-601-F
EM-620-F
EM-620-F
EM-620-F
EM-620-F
Canister
Blend3
Control3
Blend
Control
Control0
Blendb
Blend
Control
Control
s Blend
TTtT\* /"*«.w^**-^j'«l4-
jn.i-*LJ jjui*t.»«j*v»»«-*r
% Diurnal
67
52
72
65
42
68
31
8
12
36
ca_m-7i Q7
           aWor< Assignment. •*, c.m i_v..».*.,-— ^_ ,_	
           bOnly four vehicles/canisters:  101, 102, 201, 202

-------
                              -17-
TABLE 5. COMPARISON OF SHED EVAPORATIVE EMISSIONS FROM BLEND
     AND CONTROL VEHICLES WITH FUEL EM-601-F  (9.1 psi RVP)
                                                            b
                                  SHED Evaporative Emissions
Vehicle
Blend3
Control3
Blend
Control
Blend
Control
Canister
Blend
Control
Blend
Control
Control
Blend
Diurnal
3.70
1.31
3.10
2.07
1.96
2.44
Hot Soak
1.85
1.19
1.20
1.10
2.71
1.17
Total
5.55
2.50
4.30
3.17
4.67
3.61
        aWork Assignment 4, EPA Contract 68-03-3192.
        ^Numbers given are average of tests on two vehicles
            (101 and 102 with the blend or 201 and 202 with
            the control fuel used for mileage accumulation.)

-------
                                        -18-
         TABLE  6. SHED EVAPORATIVE EMISSION RATES AND CANISTER WEIGHT GAIN,
              SUMVIARY FOR VEHICLES,  101, 102, 201 AND 202 WITH EM-620-F
Vehicle
   101
                     SHED Evaporative Emissions,g    Canister Weight Gain,g
   102
   201
   202
Run
1
2
3
Avg
S.D.
1
2
3
Avg
S.D.
1
2
3
Avg
S.D.
1
2
3
Avg
S.D.
Diurnal
8.63
11.84
7.51
9.33
2.25
10.76
12.72
13.18
12.22
1.29
7.38
7.99
6.54
7.30
0.73
10.41
6.38
15.67
10.82
4.66
Hot Soak
2.14
1.27
1.45
1T62
0.46
32.90
11.92
17.10
20.64
10.93
1.26
1.05
17.41
6.57
9.39
6.39
16.94
30.06
17.80
11.86
Total
10.77
13.11
8.96
10.95
2.08
43.66
24.60
30.30
32.85
9.78
8.64
9.04
23.95
13.88
8.73
16.80
>23.32
45.73
28.62
15.17
Diurnal
21.2
20.9
22.9
21.7
1.1
18.4
18.2
17.5
IsTo
0.5
17.3
17.5
21.3
18.7
2.3
15.0
18.8
10.0
14.6
4.4
Hot Soak
16.7
14.2
13.0
14.6
1.9
34.0
29.9
31.3
31.7
2.1
22.9
15.7
18.2
18.9
3.7
28.9
31.6
32.3
3079
1.8
Total
37.9
35.1
35.9
36.3
1.4
52.4
48.1
48.8
49.8
2.3
40.2
33.2
39.5
37.6
3.9
43.9
50.4
42.3
45.5
4.3

-------
             TABLE 7. SUMMARY OF CANISTER WEIGHTS DURING SHED TESTING ON VEHICLES
                              101,  102, 201 AND 202 WITH EM-620-F
Canister Weight,
Before

Diurnal
814.6
816.4
807.2
812.7
4.9
812.8
814.9
815.4
814.4
1.4
787.4
788.2
781.2
785.6
3.8
793.0
790.0
802.9
795.3
6.8
After

Diurnal
835.8
837.3
830.1
834.4
3.8
831.2
833.1
832.9
832.4
1.0
804.7
805.7
802.5
804.3
1.6
808.0
808.8
812.9
809.9
2.6
Diurnal

Gain
21.2
20.9
22.9
21.7
1.1
18.4
18.2
17.5
18.0
0.5
17.3
17.5
21.3
18.7
2.3
15.0
18.8
10.0
14.6
4.4
Before

Hot Soak
793.7
793.6
789,9
792.4
2.2
795.2
794.4
792.4
794.0
1.4
767.9
767.9
766.8
767.5
0.6
770.0
775.4
774.8
773.4
3.0
g
After

Hot Soak
809.0
807.8
802.9
806.6
3.2
829.2
824.3
823.7
825.7
3.0
790.8
783.6
785.0
786.5
3.8
798.9
807.0
807.1
804.3
4.7

Hot Soak

Gain
16.7
14.2
13.0
14.6
1.9
34.0
29.9
31.3
31.7
2.1
22.9
15.7
18.2
18.9
3.7
28.9
31.6
32.3
30.9
1.8

FTP
a
Losses
42.1
43.7
40.2
42.0
1.8
36.0
38.7
40.5
38.4
2.3
36.8
37.8
35.7
36.8
1.1
38.0
33.4
38.1
36.5
2.7


b
Canister
+5.6
+8.6
+4.3
+6.1
2.2
-16.4
-9.4
-8.3
-11.3
4.4
-3.4
+4.6
-3.8
-0.9
4.7
-5.9
-17.0
-4.2
-9.0
7.0
Vehicle Run

   101   1
         2
         3
        Avg
        S.D.

   102   1
         2
         3
        Avg
        S.D.

   201   1
         2
         3_
        Avg
        S.D.

   202   1
         2
         3
        Avg
        S.D.

aFTP losses = (weight of canister after  diurnal)- (weight of canister before SHED hot soak)
bCanister = (weight of canister before diurnal)-(weight of canister after SHED hot soak)
10
i

-------
                                    -20-
      TAPLE 8. SUWIARY OF REFERENCE WEIGHT CHECKS ON BALANCE
                    USED FOR CANISTER WEIGHING
               2 Ib Weight(907.4g)
                      1 Ib Weight(453.6g)
Vehicle
101



102



201



202



Run
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
Actual
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
907.4
Observed
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
907.6
% change
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
Actual
453.6
453.6
453.6
453.6
453.6
453.6
453.6
453T6
453.6
453.6
453.6
453.6
453.6
453.6
453.6
453.6
Observed
453.8
453.8
453.8
453.8
453.8
453.8
453.8
453^8
453.8
453.8
453.8
453.6
453.8
453.8
453.8
453.8
% charge
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
%   change defined as
(Observed-Actual)  x 100%
      Actual

-------