EPA-AA-TSS-PA-86-02

                   Technical  Report
  Vehicle Driveability with Gasoline/Alcohol Blends
                          By

                    Jonathan Adler
                          and
                   Craig  A. Harvey
                      May,  1987
                        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 which 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

-------
       VEHICLE DRIVEABILITY WITH GASOLINE/ALCOHOL BLENDS
1.0  INTRODUCTION

     Gasoline/alcohol  blends  such  as  gasohol  (90%  unleaded
gasoline/10%   ethanol)   and  ARCO's   Oxinol   blend   (unleaded
gasoline with  a  maximum of  4.75% methanol  and 4.75%  tertiary
butyl  alcohol   [TEA])   have  been  successfully  marketed  for  a
number of years  now.   Also,  EPA has approved a  waiver  from the
requirements in  section 211 of the Clean Air  Act for  a  blend
prepared by DuPont.  This  blend includes 5% methanol  with  2.5%
cosolvent alcohols  such as  ethyl, propyl,  or  butyl  alcohols.
It has not been commercially marketed  yet  because  of  industry's
apparent concern  about the  feasibility  of  economically meeting
the   volatility    restrictions   associated  with   the   waiver
approval.   These  volatility restrictions  are  designed  to  help
assure  that the  evaporative emissions  of  vehicles  using  the
blend  do not  increase compared  to  those  of the  same vehicles
using the gasoline it would displace.

     Many oxygenated blends have  been  evaluated in a  variety of
test   programs   conducted  by  oil   companies,   other   fuel
developers,'  vehicle   manufacturers,   EPA,    DOE,   and   .other
interested  parties.   While  there are  many  issues  involved  in
the decisions  about  the desirability  of such  fuels  (economics,
alcohol   supply,   octane  enhancement,   vehicle  driveability,
emissions,  fuel  volatility levels, etc.),  this paper addresses
only  the issue  of  vehicle  driveability  with   gasoline/ethanol
and  gasoline/methanol/cosolvent  blends.   This  paper  does  not
.examine  driveability with  gasoline/MTBE blends  because there is
less data available.   There  is  current research in the  area  of
gasoline/MTBE  blends  which has  been  sparked by recent interest
in the fuels,  and more data may be forthcoming.

     Various  programs   have  been  conducted evaluating  vehicle
driveability   with   these   blends  and  comparing  it  to  that
obtained  with  gasoline.    Some  of   these  programs  include
subjective  evaluations of  the  performance  of  fleets of in-use
vehicles.   These  evaluations  provide  information  about  the
performance of the fuels  under many  driving situations.  Other
programs  have  been run with limited  numbers of vehicles using
more  objective driveability tests with trained observers over
controlled  conditions.  These  tests   do  not cover  all  of  the
driving  conditions  encountered  during in-use driving.   This
report summarizes  the  results  of  several  testing  programs.  No
attempt  is  made  to reevaluate  the conclusions  of  the  individual
studies.

-------
                              -2-
1.1  Oxygen Content

     Probably   the   most   important   factor   in   determining
driveability of oxygenated  fuels  is the  oxygen  content of  the
fuel.  This  is  usually expressed in weight  percent of  oxygen.
Neat methanol  has an  oxygen  content  of  50%  by  weight,  while
that of ethanol is 35%.   The maximum oxygen content  of  methanol
blends   is   limited  to  3.5%  and  3.7%  by  weight  under  the  EPA
waiver   decisions   for   Oxinol   and   the   DuPont   waiver  blend,
respectively.  Gasohol contains  3.7%  oxygen by weight,  because
the  ethanol content  is  specified  at  10%.   Thus,  the  volume
percent  of  methanol  is  limited   to   7.0%   and   ethanol   10%.
However, in the case of Oxinol  and  the DuPont waiver blend,  the
methanol content  is  limited to 4.75%  and 5%  respectively,  due
to  the  addition  of  cosolvent  alcohols.   Cosolvent   alcohols
reduce   the  possibility  of  phase  separation  in  the blend,  and
they help prevent problems  related  to materials  compatibility.
They are  required for  gasoline/methanol  blends  (in at  least  a
1:1 ratio for Oxinol and a  1:2 cosolvent/methanol  ratio for  the
DuPont   blend)   for   these   and  other  reasons.   Including  the
oxygen  content  of  the cosolvent alcohols,  the  oxygen content of
the  gasoline/alcohol blend  will  frequently  be  close to  if  not
equal to the 3.5%  or 3.7% by weight  limit.

     Oxygen content  is an important driveability  factor because
the  presence of oxygen  in  the  fuel makes the  air/fuel mixture
leaner   than if the  fuel consisted  only  of  hydrocarbons.   With
oxygenated  fuels  there is more  oxygen  available to burn a given
quantity  of carbon  and  hydrogen.   Also,  the oxygen  displaces
some of  the carbon  and hydrogen  in  a given  volume  of  the fuel.
Since  fuel  metering  systems  are  generally  volumetric,   less
hydrogen  and  carbon  are delivered  to the  cylinders   during  a
given  cycle,   and  the mixture  is  enleaned.   Since an  engine
usually operates  best  within  a  certain narrow range of air/fuel
mixtures, anything that  puts  the  air/fuel mixture  outside that
range  (especially on  the  lean side)  will  tend  to  degrade  the
driveability.  Therefore, the greater the oxygen  content of  the
fuel,  the   leaner  the  mixture will  be and  the  more  likely it
becomes that the  mixture will  be  outside  of the  optimal range.
As   the  mixture  moves  away   from  this  optimal  range,  the
driveability worsens.  This  phenomenon is  especially   important
for  older cars  without  closed  loop  control systems which adjust
the  air/fuel  ratios  to  within  the  desired  range.  Many  newer
cars with closed  loop  control  systems  adjust air/fuel   ratio for
optimal  operation of  the  engine and  the catalytic converter.
Driveability  changes  due   to  use  of  gasoline/alcohol  blends
should  be less significant  with these vehicles.

-------
                              -3-
1.2  Volatility

     Another fuel  property  that can  influence driveability  is
volatility.  Volatility  affects  the  tendency of  the fuel  to
evaporate  enough  to  start  a   cold  engine,   its  ability  to
vaporize  fully  and  be  distributed uniformly  to   all  cylinders
during  and  after  warm-up,  and  the  possibility  during  hot
operation  of  producing vapor  in  the  fuel metering  system so
that vapor  lock  occurs.   Standard  tests have  been  developed  to
measure  properties  which are related  to these fuel  qualities.
Reid Vapor Pressure (RVP)  is a measure of the  vapor pressure  at
100°F;   the  ASTM   D-86   distillation  procedure   provides   a
temperature versus  percent  evaporated  curve;   and vapor/liquid
ratio  is  measured  as  a  function  of  fuel  temperature.   In
addition,  the   heat  of  vaporization  of  a  fuel  provides  an
additional  indication  as   to   effects  that are   heat  transfer
dependent as opposed to equilibrium dependent.

     A  motor  fuel  with  optimized volatility   should  evaporate
just  after  it  is  metered  into  the  intake stream.    If  it
evaporates  too early,  then  it may  cause vapor  lock or sluggish
performance.   If  it  evaporates too  late,  then  it  may  cause
uneven  distribution among  the  cylinders  and   other  conditions
which  would   adversely   affect   driveability.    Also,    highly
volatile  fuel  can excessively  load  the carbon canister,  which
can  cause hot-starting problems  when  the  canister  is  purged.
Since different  vehicles  handle fuel  in different  ways,  there
is  no single  measure of  volatility which  correlates directly
with the way fuel behaves in fuel systems.

     Since  the addition of  alcohol to  gasoline normally changes
the  fuel's  volatility characteristics,  it  can  have an effect on
.driveability.    If   blended  with   ordinary   gasoline  (splash
blended),  both  ethanol  and  methanol   will  raise  the RVP and
increase  the percent  evaporated in the  140 -  200°F range.  The
effect  on RVP  is more pronounced  with  methanol blends.    If the
RVP  of  the  base gasoline  is   adjusted  to  compensate  for this
increase   by   removing  or  leaving  out   light   ends   (those
hydrocarbons which  evaporate   at  lower temperatures,  such  as
butanes  and possibly pentanes), then  the  blend RVP can be held
constant or increase  only  slightly.   The addition of  cosolvents
can  also  reduce  the  effect of  ethanol  and  methanol  on RVP.
However,  the higher  temperature volatility characteristics will
still be  different from those  of  typical  gasoline because most
of   the   alcohol   evaporates   within  a   relatively    narrow
temperature  range.    Special  formulation  of   the  base gasoline
can negate this effect, but  that is  not considered necessary or
economically feasible  for  most  purposes.  The   type and quantity
of  alcohol  in  the  blend  dictates  the  effort   and  expense that
would be necessary to adjust the volatility in  a given way.

-------
                              -4-
1.3  Heat of Vaporization

     The heat of  vaporization is the quantity of heat  required
to  vaporize  a  given quantity of  fuel.   This  heat  is  absorbed
from the  surrounding air  and engine  components  thus having  a
cooling  effect   on  them.    Alcohols  have  a   higher  heat   of
vaporization than gasoline and,  therefore,  require  more heat  to
vaporize as shown by the figures  below.

     Heat of Vaporization:    Gasoline   940 BTU/gallon
                             Ethanol    2600 BTU/gallon
                             Methanol  3320 BTU/gallon

Any  possible  problems  related  to  this difference  would  most
likely  show up  as  difficulties   in  starting  a  cold engine  or
keeping it  running  once  it  is started.   The standard ASTM  fuel
volatility  tests  mentioned  above   do  not  address  this  fuel
property since they simply provide as much heat as  necessary  to
achieve  the  required  temperatures  for   fuel  distillation  or
vapor formation.

1. 4  Oc'tane

     Another  driveability-related   fuel    property   that   is
affected  by alcohols  is  octane.   Alcohols  raise  octane  when
added  to  gasoline,  so  the  degree of  engine  knock  experienced
with  gasoline/alcohol blends  will   tend  to be  less than  that
with a  base gasoline of  lower  octane.   However,   some  concern
has  been  expressed1  that  the difference  between  research  and
motor  octanes  (known  as  the sensitivity  of   a  fuel)  may  be
larger  for  alcohols  than  for  hydrocarbon  compounds typically
present  in gasoline.   Gasoline   is  frequently  blended  for  an
acceptable  average  of  research  and motor octanes  (the "pump"
octane, which equals  [research +  motor  octane]/2),  so the motor
octane  of  the  gasoline/alcohol blend may  be lower  than that  of
a hydrocarbon only  fuel  of  the same pump  octane  rating.   A low
motor  octane  can result in  engine  knock  especially  under  high
speed or load conditions.  Thus,  use of gasoline/alcohol blends
could   result   in  some  engine  knock  not  experienced  with
gasoline.  Knocking under conditions of high speed  and  load can
lead to catastrophic failure of the  engine2.   However,  other
components  commonly  used   in gasoline,   such  as  toluene  and
xylene,  have sensitivities  as  high as  alcohols3.  Some  data
are.    available     indicating    that    the    sensitivity    of
gasoline/alcohol  blends  may  be somewhat  more  than   that  of
gasoline   itself"   and  other   data   indicate3    that   this
situation may not occur.

-------
                              -5-
1. 5  Intake System Deposits

     Deposits in  the  intake system can  affect  driveability,  as
well  as  fuel  economy  and  emissions.   Their   formation   is
affected by  fuel  composition, engine  condition,  and vehicle  use
pattern.  Deposits can  accumulate  anywhere from  the  carburetor
to  the  intake  valves,   and they  can restrict  air  flow,  clog
vacuum  hoses,  change swirl  patterns, and otherwise  interfere
with devices in the intake system.

     Some  gasolines   include  detergent  additives  in order  to
minimize the  accumulation  of  these  deposits  near  the beginning
of  the  intake  system.   These  additives  may not inhibit deposits
forming  farther   downstream  near  the EGR  duct  or  the  intake
valves.   Other  additives,  such as  carrier  oils  may serve  to
carry  the  deposit-forming  materials  into the cylinder,  where
they may be  burned or exhausted.  Of interest in this report is
how  the use  of  blended  fuels can  affect   the   formation  of
deposits in the intake system.

1.6  Other Parameters

     There  are  many  other  factors which  can  affect the  outcome
of  driveability  tests,  such   as   base  gasoline  composition,
additives to  prevent  corrosion  and phase separation  (other than
cosolvent  alcohols),  and  specific  points on  the  distillation
curve  of fuels.   These parameters  are  useful  in  interpreting
the results of individual  test  programs.   However,  it is beyond
the  scope  of  this report  to  examine  them  in detail  for each
study.

2.0  DRIVEABILITY TESTING PROGRAMS

     Many  organizations  have   conducted  driveability  testing
programs.   This  section will summarize  some  of these programs.
The smallest  of  the  programs  evaluated the performance  of five
vehicles,  while  the  largest  included   several  hundred.   The
programs  generally did  not evaluate driveability  in the same
way.   Some  of  the smaller programs  included  controlled tests
that were developed  by  the Coordinating  Research Council, Inc.
These   tests  involve  drivers   trained   in  the  evaluation  of
driveability   problems,    as    well    as   specified   driving
procedures.   The  larger  programs  were  conducted with  in-use
fleets,  so  they  were  generally  not   able   to  conduct  the
controlled   driveability   tests.    The   evaluations  that  were
produced by these programs are  the  result of longer term  use
under  normal operating  conditions,  often over  the  course  of
more than a year.

-------
                              -6-
2.1  Fleet Testing Programs

     Southwest Research Institute conducted  a  variety of  fleet
tests over five and  a  half  years as  part  of  the  U.S.  Department
of  Energy's  Project  for Reliability Fleet Testing  of Alcohol/
Gasoline  Blends5.   Most of  the individual  fleets  participated
in  this program for  approximately one  year.   These fleet  tests
were  conducted  in  several  geographic   locations  using   four
different  gasoline/alcohol   blends.    The fuels  were   gasohol
(apparently  a splash  blend with  10% by volume  of  anhydrous
ethanol),   a  volatility  controlled   ethanol   blend   (10%   by
volume),  Oxinol,  and a 4.2% methanol blend  which  included 2.1%
ethanol and  2.1%  tertiary  butyl  alcohol as  cosolvents.    The
tests  were  conducted  under  normal  operating  conditions  with
established   fleet   operators.    A   total  of   552   vehicles
accumulated 6,643,936 miles over  the course  of  the program.   Of
these,  218  vehicles were  operated  on   unleaded   gasoline  as
experimental  controls.   The following  fleets  with a total  of
264 vehicles  participated  in the driveability  portion  of  this
project:  Contra   Costa  County,   CA;  Tennessee  Valley Authority
(two  separate  fleets);  the  State of  New   Jersey;  the State  of
Minnesota;   U.S.  Border Patrol  in El  Paso, TX;  and  Southwest
Research  Institute.   Vehicles   in  Contra  Costa   County   were
tested with  two blended  fuels  (gasohol and a gasoline/methanol/
TEA blend); the other fleets were tested  with only  one blend.

     The  report   includes   analyses   of   the   differences   in
driveability  between the cars  that  were   operated  with  blended
fuels  and the control  vehicles  in  each   fleet,  as  well  as  a
composite  analysis  of  the  effects of gasohol  on   three of  the
fleets.   The  following  problem  areas  were  examined;  starting,
stalls  during   warmup,  stalls   in  traffic,  rough   idling,
hesitation, loss  of  power,  pinging,  and dieseling.   Performance
problems  are  rated in terms of  the number of  driver  complaints
divided  by  the  number  of  total  reports.   In  five  of  these
areas,  the  blended fuels  performed  significantly worse  than
regular unleaded  gasoline.   Table 1  (reproduced from page  6-5
of  the  report)  shows the results of  significance  tests  of the
measured  differences  in   performance  between  the   test  and
control vehicles  within each fleet.   Table  2  (reproduced from
page  7-6  of  the  report)  summarizes  for three of the  fleets  the
differences  in  performance  between   the vehicles  which  used
gasohol and the vehicles which used gasoline.

-------
                         Table 1.   Summary of  Driveability Tests Using Blends

                          Department of  Energy/Southwest  Research Institute

                            Table  reproduced  from page  6-5 of Reference 5.
FLEET
CONTRA COSTA
COUNTY
TVA 1 & i
FUEL TYPE
Gasoliol
Gasoliol
NEW JERSEY Gasolt*!
CA. ENERGY COMM.
Sacramento M94.S Vs EIOII
Los Angeles Melllanol (M94.5)
MINNESOTA
UORDER PATROL
CONTRA COSTA
COUNTY
SWRI
ET-2
ET-2
Oxinol 50
MeOII/EIOII/TDA
CRANKING

-------
                              -8-
Table 2.   Comparison of Driveability with Gasoline  and  Gasohol
           Department  of   Energy/Southwest  Research  Institute
           (Table reproduced from page 7-6 of  Reference 5.)
        COMPOSITE ANALYSIS - TENNESSEE VALLEY AUTHORITY,
           STATE  OF NEW JERSEY AND CONTRA COSTA COUNTY
PERFORMANCE PROBLEMS

CRANKING

STALLS WHILE STARTING

STALLS IN TRAFFIC

ROUGH IDLING

HESITATION

LOSS OF POWER

PINGING

DIESELING
                         FREQUENCY* OF OCCURRENCE
GASOLINE
.0165
.0151
.0039
0.065
.0106
.0099
.0091
.0047
GASOHOL
.0443
.0373
.0209
.0156
.0406
.0254
.0065
.0017
SIGNIFICANCE**
<0
<0
<0
<0
<0
<0
<0
<0
.001
.001
.001
.001
.001
.001
.025
.001
       Frequency = Total  complaints/total  reports,  where gasohol
       total reports = 12,846 and gasohol total report = 11,743

       Value indicates significance level of test.

-------
                               -9-
     Gasohol  performed significantly  better  than  gasoline  in
the  areas  of   rough   idling,   pinging,   and   dieseling,   and
significantly worse  than gasoline  in  the other  problem  areas.
Three  fleets  were  used  for   the  analysis  listed  in  Table  2
(Contra Costa  County,  New Jersey, and  the TVA fleets)  totaling
108  vehicles  from  model  years   1974-1981.   Table  1  includes
these  fleets,   as  well  as  the  Minnesota fleet,  which had  60
1980-82  vehicles,  the  El  Paso   fleet,   which   had  52  1980-83
vehicles, the Contra Costa  fleet operating on Oxinol,  which had
30  1975-78  vehicles,   and  the  Southwest  Research  Institute
fleet, which  had 14  1984 Ford Escorts.   There  were  no reported
incidences  of   vapor   lock   or   problems  with  cold  weather
operation   which  were   attributable   to   the  use  of   the
gasoline/alcohol  blended  fuels.    These  reports  indicated  that
the  differences  in   driveability  with   the  gasoline/alcohol
blends   were    statistically   significant    and    definitely
perceptible.  But the  reports  did not necessarily indicate that
the differences were large.

     The Tennessee Valley  Authority (TVA),  in  cooperation  with
ARCO,  conducted a separate  fleet  test  of  140  vehicles  during
1983  and  part of 19846.   This  program tested  a base  fuel  and
three  blended fuels.   The  blends  included  varying  amounts  of
methanol and  tertiary  butyl  alcohol to achieve  oxygen contents
of  3.5%, 4%,  and 5%  by weight (4.75%,  6.0%,  and 8.2% by volume
of   methano-1) .    The  volatility   of   these  blends   was   not
controlled.   The fleet  was  divided  into four  groups,  one for
each  fuel.    One of  the  sub-fleets  used  gasoline   during  the
entire  test  period.   The  other  fleets  used  blended  fuels
roughly  half  of the time;  gasoline  was  used  the rest  of  the
time  to allow   evaluation  of  the  drivers'   responses  with the
same  vehicles.   Drivers recorded  performance information  on   a
.daily  basis  during  the   test.   Driveability  was observed  by
recording  the number  and  severity  of  occurrences  of specific
driveability  problems.   These were hard  starting,  rough idle,
stalls   during   idle,   stalls   during   driving,   hesitation,
backfire,   dieseling,   and  knock.   The  data  were  analyzed
separately  for  the  four  seasons.   The  analyses of  these  data
showed  no   significant  differences  between  the test  and  base
phase  for  any of the fuels tested,  except  for  the  3.5%  and 5%
oxygen blends  during the  spring  phase of the program.  However,
in the case of the 3.5% oxygen blend,  the report concludes that
the  observed  differences are  the  result  of  factors  other  than
oxygenate  content.   A  few of  the vehicles  showed  exceptional
susceptibility  to  problems when fueled with the 5%  oxygen fuel
in the spring phase.   The report concludes  that  the  3.5% and 4
oxygen   fuels   provide  driveability   equivalent  to  that  of
hydrocarbon-only  fuels,  and  that  the  5% oxygen  fuel provides
driveability  equal  to  that  of  hydrocarbon-only fuels  for  most
but not all vehicles.
•6

-------
                              -10-
     Alberta  Gas  Chemicals,   Ltd.,   the  Ontario  Ministry  of
Transportation  and  Communications,   and  Suncor  Sunoco  Group
conducted  tests  beginning  in  1982  on  a  fleet  of  34  in-use
vehicles   in   Canada7.   The   fuels  used   in  this   program
contained  7%  methanol  and 3%   isobutanol.   These  fuels  were
blended to  meet  the seasonal volatility  specifications  used by
the supplier.   The driveability  testing portion  of  this  program
was conducted  in four phases, using test  criteria  adapted from
various  CRC reports.   The tests evaluated  cold  start,  cold
drive,  warm start,  and  warm drive  performance.  The  blended
fuel  gave  worse  driveability, but the  report states  that this
difference  is  not  statistically  significant, and that  the test
drivers rated the fuel performance as acceptable.

      In  support  of  its  waiver   application  for  Oxinol,  ARCO
provided a  wide  range of driveability test data  for  Oxinol and
other  alcohol  blends8.  One  program  involved fleet  testing  of
150 employee-owned  vehicles with fifty  operating on each  of  a
base  gasoline,   Oxinol,   or  gasohol.   Results   indicated  that
driveability  of   an Oxinol  blend  was  equivalent  to   that  of
gasohol,  and  cold-engine  performance was the   only  operating
mode  where these  blends  showed more operational problems than
the  base  gasoline.  A similar  test  program was  conducted  to
evaluate cold  weather  operation  using  the same  blends  as well
as a  blend of  16%  TEA.  In this  program  all  four fuels yielded
equivalent  driveability.   A  smaller  scale  testing  program9
examined vehicles in two fleets  which had  been  using  Oxinol for
at  least  a year.   One  of  these  fleets  consisted  of  eleven
vehicles that were  owned and  operated by  ARCO  employees, while
the other  consisted of five  matched pairs of vehicles  (half of
which  used gasoline).   The program examined the  condition of
the intake  systems  at  the end of the period  of  Oxinol  use, and
found   that  the  deposits   in   the  Oxinol   fueled   vehicles'
carburetors  were  typical  of  those  which  would  be  found  in
gasoline fueled vehicles of the  same mileage.

      American  Methyl  Corporation submitted  some  driveability
test  data  along with  its  waiver application  for  Methyl-1010.
This   fuel  consists   of  methanol,   cosolvent  alcohols   and  a
proprietary additive  blended  with  unleaded  gasoline such that
the  final   oxygen  content  is  no  more   than  5.0%  by  weight.
Fifteen 1974 -  1981  model year  vehicles were evaluated by their
owners  over a  six  month period  of  operation on  the  test fuel.
In general, it ,was  found that older vehicles had less tolerance
for  oxygenates,  and  closed  loop emission  control  systems  on
newer   vehicles  were   found   to  alleviate  most  driveability
concerns.

-------
                              -11-
     Ashland  Oil  Company  conducted  two  short  duration  fleet
tests  (approximately  90  days  each)   of  1%  and  2%  methanol
blends11.   Eighteen  vehicles   were   evaluated  with   the   1%
blend,  and  four vehicles were  evaluated with the 2% blend.   For
the  1%  blend  each  of  the  drivers of  the  test  vehicles  were
interviewed regarding their experience  with  the fuel.   Comments
ranged from neutral  to  positive.   The comment most heard was  a
decrease  in  pinging  when  the  cars   began  the  test.   Other
comments noted  improved  fuel  economy or  improved  acceleration.
One  vehicle  with  electronic   fuel  injection  experienced  hot
start  problems early  in  the   test.    No  other  problems  were
noted.   For the 2%  blend one  vehicle experienced no significant
change  in  startability,  but   there  was  a  slight  intermittent
change  in  hot startability.   A  decrease  in  part-throttle knock
intensity was  the only  significant  difference  noted with  this
vehicle  using  the test  fuel.   Another  vehicle  experienced more
significant hot start problems,  which may have been  related to
the use of fuel injection.

     The  Bank  of  America  has  conducted  fleet  tests with  a
variety   of   blends12.    One   program   involved   testing   67
unmodified  vehicles  fueled with blends  of  up to  8%  methanol.
Driveability  was  measured  on  the  basis  of  starting,  warm-up,
idling,  cruise,  power (acceleration),  and  shutdown performance
with  drivers   who  did  not  know which  fuel  they  were  using.
Vehicles  without   closed   loop   control   systems  experienced
decreasing  driveability  with   increasing  oxygen  content,  but
vehicles   with  closed   loop    systems   maintained   comparable
driveability to gasoline even with 8% methanol fuel.

     The  TVA  has  also  participated  in  a  program  of  more
controlled  driveability testing,  using  the Driveability  Test
Procedure   developed   by  the   Coordinating   Research   Council
(CRC) ' 3 .   The  tests  were  conducted  on  sixteen  1983  and   1984
model  year  vehicles  which  were  leased from  local car  rental
agencies.  These  cars had  odometer mileages in the range of 500
to  15,000.   The  program tested  five   fuels;  a base fuel   from
which  the  blends  were developed,  a splash  blend  of 10% Oxinol,
a blend  of  10% Oxinol with the  RVP  adjusted to within 0.5 psi
of  that of  the base gasoline,  a  blend  of  10%  Oxinol  with  a
reformate  added to  match  the  distillation  curve  of   the   base
gasoline  as closely  as  practical,  and a  splash  blend  of  10%
ethanol.  The  report notes that  the  last  of  these  reflected the
characteristics of  many commercially  available  gasohol blends.
The  report  concludes that  the  different  fuels did  not affect
cold-start  driveability of the vehicles  either  as a  group or
when they  are  stratified  by  fuel system  or  emission control
system.

-------
                              -12-
     The  Coordinating  Research  Council   conducted   extensive
tests  on  fourteen  cars  from  the  1980 model  year1"'15.   The
driveability  portion  of   these  tests  consisted  of   the  CRC
Intermediate Temperature Cold  Start  and Driveability Test,  and
the  CRC  Vapor Lock  Test.   These  tests were  conducted  in  two
phases;  phase  one  included  ethanol  blends,   and  phase  two
included methanol  and  methanol/cosolvent blends.   The  vehicles
tested in  this  program represented  a  mix  of emission  control
and  fuel systems  available  in  1980;  two had fuel injection,  the
rest  had  carburetors;   seven  had closed  loop emission  control
systems calibrated  to  1980  California emission  standards,  and
seven had open loop systems calibrated  to  1980 Federal  emission
standards.   Only  ten of the vehicles were  tested  in  phase two,
due  to funding constraints.  The  following fuels  were  included
in phase one  of  the  program;  a base fuel,  a splash blend of  10%
ethanol, a  blend of 10%  ethanol  adjusted  to match the  RVP  of
the  base fuel, and  a blend of 10% ethanol adjusted to match  the
RVP  and  percent   evaporated  at  158°F  of  the base fuel.   The
blends  had  significantly  worse  driveability  than  the  base
fuel.  In  its discussion of  the  driveability data,  the report
states that  the  driveability  is  probably  adversely affected by
the  leaning effect of the blends, and that  increased volatility
seems to improve  driveability.   The  blends also  showed a higher
tendency to  vapor  lock than  the  base  fuel.   The  tendency  to
vapor  lock   increased  as   the  percent  evaporated  at  158°F
increased.   In the  second  Phase of  the program,   six  fuels were
tested; a base fuel  (not  the same as,  but  very  similar to that
in the first  phase),  a  blend of 3.8%  methanol,  a  blend of 3.3%
methanol  and 1.1%  isobutanol,  a blend  of  10.0% methanol,  a
blend  of  8.8% methanol and  2.9% isobutanol,  and  a   blend  of
14.0%  methanol   and  2.0%  isobutanol.  As  with  the phase  one
fuels,  driveability with  the  blends  was  significantly  worse
than  that   with   the  base  fuel.   The  report  states   that  the
driveability  degradation  seems  to  be  related   to  the  oxygen
content of  the fuel,  and that the presence of cosolvent had no
effect.

     Mobil  Research  and  Development  Corporation conducted  a
number of driveability  test  programs comparing Oxinol blends to
gasolines16.   The  first  of   these   involved   six  1983  cars
tested with  gasoline,   a matched  distillation Oxinol blend,  and
a  lower  volatility  gasoline.   Tests  were  conducted  under
temperature  controlled  conditions,   at 60, 45,  and  25°F.   The
cold  start  and driveability  demerits for the Oxinol  blend were
50  -  100%  greater  than  the  matched  gasoline  and   roughly
equivalent to the  lower volatility  gasoline.   A second  program
using  fifteen 1981-1983  cars  and a  consumer-type driveability
test  determined   the  cold  start  driveability   at  temperatures
ranging from  0  - 60°F.   A  12.5 psi  RVP  gasoline was  best with
6% unacceptable trips,   followed by a  14.8  RVP Oxinol blend with

-------
                              -13-
9%,  and a  9.8  RVP gasoline with 10% unacceptable trips.   It  is
interesting to note  that  the blend averaged  only  half as  many
stalls on  start-up  as the  12.5  RVP gasoline.  A similar  study
used ten 1984  cars  at 60  -  95°F with both  cold  and hot  start
evaluations.    The  Oxinol  blend  fuel  had  RVP and  distillation
properties  between  those  of the  two  test  gasolines (11.3  and
11.8  psi  RVP). The  cold  start  results  showed  the percent  of
unacceptable  trips  with  the  blend to  fall  between  the  two
gasolines.   Under  hot start conditions,  however, the blend had
a much greater percentage  of stalls on  start-up than either  of
the   gasolines,   and   therefore,   a  greater   percentage   of
unacceptable driveability (14% versus  0-4%).

     Esso Petroleum Canada conducted a program in the  summer  of
1984  to  compare  a  base gasoline  with  an  Oxinol   blend17.   In
this  program  two   similar   30   car   fleets   represented   the
1977-1984 model year Canadian car  population.  The program was
followed  in  winter/spring  1985  with   another  program  which
included two 25 car  fleets to represent  1980-1985  Canadian cars
fueled with a  winter grade  gasoline, Oxinol  blend or  10%  MTBE
blend.  The  summer  Oxinol  blend  resulted  in roughly  twice  as
many  complaints as  the  corresponding  gasoline,  including  hard
starting,  rough idling,  and  sluggishness.   For  the winter fuels
the  Oxinol  blend  did  not  have  any  significant  effect  on
start-up or  idling,  but there were roughly  three  times as many
complaints of  sluggishness  relative to  the gasoline.   The MTBE
had no significant effect on driveability.

     From September  1983 to September 1984,  Texaco conducted a
one year test  using  200 1976-1983  model  year domestic  vehicles
with  gasoline  and  two  blends  of  6% methanol  and  2%  TEA; one
with RVP equal  to the gasoline and one with  RVP of 1.0-1.5 psi
higher.  The  volatility  of  the base  gasoline  used  to prepare
the blends was  adjusted to provide good warming-up performance
when blended with  the alcohols.   The  pump octane ratings of the
blends  were   slightly   higher   than   that   of  the  gasoline.
Vehicles operated  1/3 of  the  time on each  of  the  three fuels.
The program  was  designed  to  allow analysis  of  the  effects  of
ambient   temperature,    warming-up   driveability,   warmed-up
driveability,  and knock.   The  warming-up  driveability of all
three  fuels  was  roughly equivalent;  the  warmed-up driveability
of  the blends  was  lower.   The  report   states  that  midrange
volatility might  be a  factor which caused  this  degradation.  No
cases of vapor  lock occurred,  even with  the higher  RVP blend.
The knock performance  of  the alcohol  blends was slightly better
than  that  of  the  gasoline,  reflecting   their  slightly  higher
octane ratings.

-------
                              -14-
3.0  SMALLER TESTING PROGRAMS (10 or  Fewer  Vehicles)

     General    Motors    Research     Laboratories     conducted
driveability  tests   using   methanol   and   methanol/cosolvent
blends19.    Six  1983  and 1984  GM  cars  chosen  to  represent  a
mix of fuel  systems  were  tested  on  five  different fuel  blends;
a  base fuel,   a  blend  with 9.5%  Oxinol,  a blend  with  8.2%
methanol  and  2.7%  TEA,  a blend  with 3% methanol,  and a  blend
with  7%  methanol.    Two  versions  of each  of these blends  were
used, one  to fit ASTM class  C specifications, and one  for  class
D.    The  fuels were  tested  with  a modified  version  of the  CRC
hot-weather driveability and vapor  lock  procedure.   Statistical
analysis  of  the driveability  data   showed  that  the  volatility
class of the  fuel was  the most  significant  fuel-related  factor
in  determining  the  driveability of  these  cars,  and that  the
alcohol content was  not as important.  It was noted  though that
an  isolated  problem with  fuel  foaming  occurred  on  one of  the
carbureted  vehicles  with   a   blended  fuel  containing   8.2%
methanol  and  2.7%  TEA.   Fuel foaming  is apparently  an  unusual
phenomenon  involving  formation   of  a  gasoline  foam  in  the
carburetor  bowl which  results  in   excessively  rich  operation
(less  than  11.7:1  air:fuel  ratio).   It  is not  evaluated  in
conventional  driveability  tests.    Formation  of  the  gasoline
foam is caused in part by excess fuel  volatility.   Intuitively,
most  fuel  injected  vehicles should  be less  susceptible  to the
problems  of  foaming  and  vapor  lock,  because  the  high pressures
in  the metering system prevent the  fuel  from vaporizing  before
it  is  added  to  the  intake stream.   Low pressure  systems,  which
are  less common  than the high pressure systems,  would have less
of  this resistance.

     General  Motors  also conducted   tests  on five  1980 - 1981
model  year  GM  cars,  using  three  different  fuels;  a  base
(Indolene), a splash blend of 10% alcohol,  and  another with 18%
alcohol20.    The alcohol  used was   a  2:1  mix  of methanol  and
butanol.   The study showed that driveability with  the 10%  blend
was much worse  than that with the base fuel.  Driveability with
the  18% blend was not evaluated.   The report cites  the  leaning
effect of  the alcohol,   and  the  cooling of  the  intake manifold
by  the alcohol as reasons for the diminished driveability.

     A presentation  given by Sun  Tech in 1983 shows  a graph of
total  average  demerits  versus  fuel  oxygen content  for  eight
1980 vehicles using  the  CRC cold start  and  driveaway procedure
with   hydrocarbon-only,   ethanol   blend,   and   Oxinol   blend
fuels21.    The curve has  a  significant,  nearly  linear positive
slope   indicating    roughly   equivalent    deterioration   of
driveability  for  ethanol and  methanol  blends  of equal  oxygen
content.

-------
                              -15-
     As  part  of  the  Tennessee  Valley  Authority  program  to
evaluate  gasoline/alcohol  blends,  seven  1983-1984  model  year
vehicles  were  evaluated  for  driveability using  a  cold  start
followed   by    3.6    miles    of   driving    through    various
maneuvers22.    Five of  the  test  vehicles  were carbureted  and
two  had throttle  body fuel  injection,  and  all  vehicles  were
tested  by  a  single trained  driver.   The  fuels  included a base
gasoline,  a  10% methanol  blend,  a  5%  methanol/3.2%  ethanol
blend,  and a 5%  methanol/6.6% ethanol blend.   Each vehicle/fuel
combination was  tested in  duplicate.   The tests  indicated that
the  driveability  of   methanol  blends  was independent  of  the
cosolvent  used.   The  vehicle-to-vehicle  variation was  greater
than  any  variation  associated  with  different oxygen  levels.
While some vehicles performed poorly  with  methanol  blends, they
also performed relatively poorly with the base gasoline.

     Texaco  conducted a  50,000  mile  testing  program with  six
vehicles  from  the  1982  model year2'1.   Three  of  these used  a
blend  of  6%/2%  methanol/TBA  blend,  while  the  others  used
gasoline only.   Both  fuels  contained a detergent  additive;  the
blended   fuel    contained   50%   more.   The   only   substantial
difference  found  on   examination  of  the   two groups  was  the
cleanliness  of  the intake valves.  The  cars  which  used blended
fuel had heavy deposits on the intake valves,  while  those which
used  gasoline  had much  cleaner  valves.   The quantity  of  the
deposits  va'ried  from  cylinder to  cylinder within  each  of  the
cars  that  used  the blend.   The report makes no determination as
to  whether  the  greater  concentration  of   detergent  in  the
blended fuel affected  the formation of deposits.

     VW  has  reported  that  greater  detergent  levels  are  needed
to ensure  that  deposits  do not build up in the intake manifold
.and   carburetor2''.     Also,   greater   alcohol   levels  require
greater  amounts   of  detergent.   The  report   does  not  discuss
deposits on  the  intake valves,  and it only discusses  results of
testing with gasoline/methanol blends.

     Ashland Petroleum Company examined gasohol  performance in
dynamometer  and  in-use tests25.   Three  pairs of cars and some
light  trucks were used.   One  pair  of  vehicles  was  run  on
dynamometers  for  a  50000  mile   test which  examined  lubricant
performance,   emissions,  driveability,  and   other  performance
criteria.    One   vehicle  was   fueled   with   regular  unleaded
gasoline,  and  the other with a  10%  ethanol   blend.   This test
found  little  or  no difference  between the two  vehicles  in the
fuel-related  areas   of   comparison,   including  fuel  economy,
driveability,  and  deposits  in  the   intake  and   combustion
chamber.  The other  vehicles used in this  testing  program were
tested  in-use  under  less  controlled  conditions.   Driveability
was  evaluated  in  formal  reports by the  drivers.   The  report
concludes, based on  the  results of  all  of   these  tests, that
driveability is  not affected by the addition of ethanol.

-------
                              -16-
4.0  Conclusions

     As  a  general   observation,   the   driveability  of   many
vehicles  on  low  level  alcohol  blends  appears  to  be  roughly
equivalent to that of gasoline.  However, there  are  indications
of   problems   with   some   vehicles   and   certain   operating
conditions.   These problems seem to occur less on  vehicles  with
fuel  injection  and closed-loop  fuel  control systems which  are
more  prevalent  in recent model  years.   These problems  seem  to
occur  somewhat   more  on  vehicles   with  carburetors  especially
those  without  feed-back  to  control  air:fuel   ratio.    While
driveability  problems of  the  type discussed  below are  noted
more with blends  than with gasoline,  it  is  not  clear that  the
problems would  be serious and/or  objectionable, just  that  they
would  be  perceptible   to  some   degree.    Nevertheless,   the
possibility  of   some  increased  driveability  complaints  due  to
increased use of gasoline/alcohol   blends cannot be  ruled  out,
although such complaints may not be numerous.

4.1  Oxygen Content

     The  closer  a vehicle is  to its lean  operating limit with
gasoline,  the  less  fuel oxygen  it will  be  able  to  tolerate
before  experiencing  lean mixture  driveability problems  such as
hesitation and s'ome loss of power  as  noted  in the  Department of
Energy project.

      It  appears  that the  richer a  vehicle  operates on gasoline
 (whether  due   to  carburetor   adjustment   or  high  altitude
operation,  although   no  driveability  tests  have been run  with
gasoline/alcohol  blends  at high altitude)  the more  fuel oxygen
it   should   be   able  to  tolerate  before   experiencing  any
deterioration  of  driveability.   (Any   vehicles which  operate
excessively  rich on  gasoline may  even  experience  improvements
in driveability with  oxygenated  fuels.)

     As  mentioned above,  closed loop fuel  metering  systems can
alleviate most  if not  all  of the  normal  driveability problems
that   can  be   associated   with  the  mixture  enleanment  of
oxygenated fuels, up  to  the adjustment limits of the  system.

     Without  closed   loop   systems,   increasing   fuel   oxygen
content   can  be   expected    to   increase    the   incidence  of
driveability problems.

4 . 2  Excessive Fuel Volatility

     For  blends  the  higher volatility  in the 140  - 200°F  range
can  result  in  increased driveability problems due to increased
cases  of vapor   lock,  fuel  foaming, and hot-starting problems.

-------
                              -17-
Also,  a  loss of power  in traffic is  often  noted before  vapor
lock  occurs.   Such  problems  would be  more  likely to occur  in
high  altitude  areas.   Blends  showed  an  increased tendency  to
vapor  lock  in  the CRC  vapor  lock procedure  than  gasoline,  but
vapor  lock  was not  specifically  found to be  a  problem in  the
studies  of   in-use  vehicles.    Some  of the  in-use studies  did
find increased hot-starting problems.

     These problems can be mitigated  somewhat  by keeping  blends
in compliance  with ASTM volatility recommendations.   Additional
control   of   volatility,   such   as   maintaining  the   percent
evaporated  at   some  midrange  point   in the  distillation  curve
equal  to  that  of  typical gasolines or  additional  reductions  in
RVP,  might  help more.  Adjusting  the entire distillation  curve
to  be  equal to  that of  typical  gasolines  would  help quite  a
bit.   It is  not  known,  however,  how  such  an  adjustment  would
interact  with  the high  heat  of  vaporization (of alcohols)  or
other  parameters  to  affect driveability under other  conditions,
such  as  cold starting.   The high cost  of  an adjustment  such  as
this,  as  well  as  other factors, would  make  it unlikely to occur
in actual practice.

4.3  Heat of Vaporization  (cold starting)

     Ethanol and  methanol require approximately  three  times  as
much   heat'  to  vaporize   as   does   gasoline.    Under  certain
conditions this can make  it a  little  harder  to  start  and keep a
cold  engine running  with blends  that contain  these alcohols.
However,  a  high heat of vaporization  might  reduce somewhat the
likelihood of  vapor  lock  by cooling the fuel  system.

4.4  Octane  (knock, dieseling)

     Since methanol  and ethanol increase the pump octane rating
of  blended   fuels,  there  is  a  tendency   for  blends   to  reduce
incidences of  knock  and dieseling (run-on)  relative  to the base
gasoline.  Research  octane may benefit more  from alcohols than
motor  octane,  so  under certain operating  conditions  where motor
octane is important  (e.g., high  load)  there  may be no effective
octane  benefit from  the alcohol.  However,  the data available
on   this  point  do   not  conclusively  indicate  whether  the
increased sensitivity of  the  blended fuels  over gasoline would
pose a problem.

4.5  Intake  System Deposits

     Increased  amounts of  deposits,   especially  on  the  intake
valves,  have  been  reported   in  some studies26, while  others
have  indicated that  there is  no  such  increase.   It   is possible
that  the differences may  be  due  to  variations  in the quantity
of detergents  used  in the fuels.  Evidence  is  not available to
conclusively resolve  this  issue.

-------
                              -18-
4.6  Other Parameters

     Fuel  additives  are  mentioned in  many of  the  reports  as
being crucial to  the successful operation  of  a  vehicle with  a
gasoline/alcohol  blend.   These  additives  are   formulated  to
inhibit  corrosion,  prevent  the  gasoline  and   alcohol   from
separating, and  help clean fuel system parts.  Their  effect  on
driveability is less direct than the other  fuel  properties,  and
their value  becomes  more  apparent when  examining the.issue  of
materials compatibility.

-------
                              -19-

                           REFERENCES


1.    United  Nations  Economic  and  Social  Council,   "Use   of
Replacement   Fuels    and   Fuel    Components   in    Europe",
TRANS/SC1/WP29/R.328,  May 2,  1984.

2.    "Alcohol Blend Operational Problems of Concern  to  Chrysler
Corporation",  part  of  a  presentation  by  Gordon   Allardyce,
Chrysler  Corporation  to  EPA  Workshop   on   the   Effects   of
Gasoline/Alcohol  Blends  on  Vehicles  and  Vehicle Emissions  at
High Altitudes,  Aug 20, 1986.

3.    James  M.   Dejovine,   Edward  G.  Guetens,   Jr.,   George  J.
Yogis,  Brian C.  Davis,  Walter  H.   Douthit,  Paul E.  Hagstrom,
"Use  of  Oxinol  and  Other  Alcohol   Blending   Components   in
Gasoline",  National  Petroleum Refiner's Association,  November,
1982.

4.    N. D. Brinkman, N.E.  Gallopoulos,  "Exhaust  Emissions,  Fuel
Economy,    and    Driveability    of    Vehicles   Fueled    with
Alcohol-Gasoline Blends", SAE paper 750120, 1975

5.    John  D.  Tosh,  Anna F.  Stulgas,  Janet P.  Buckingham,  John
A.   Russell/  and John P. Cuellar, Jr.,  "Project  for  Reliability
Fleet  Testing  of  Alcohol/Gasoline  Blends",   Contract  report
prepared  for U.S. Department of Energy, April  1985.

6.    John  R. Morgan,  "Recent TVA  Experience with Methanol Blend
Fleet  Testing",  Alcohol  Week   Conference,   Washington  D.C.,
November,   1984 .

7.    K.   Taylor,   "Methanol/Gasoline   Blend   Vehicle   Fleet
Demonstration - A  Joint  Project  of Alberta Gas  Chemicals Ltd.,
Ontario  Ministry  of  Transportation  and  Communications,   and
Suncor  Sunoco   Group",  presented  by  Brian Davis   at  the  6th
International Symposium on Alcohol Fuels Technology,  May 1984.

8.    ARCO  fuel waiver application for Oxinol,  April 27,  1981.

9.    David  J.  Miller, David A.   Drake,  and James M.  Dejovine,
"Material   Compatibility  and  Durability  of   Vehicles  with
Methanol/Gasoline   Grade   Tertiary   Butyl   Alcohol   Gasoline
Blends", SAE paper 841383, 1984.

10.  American   Methyl   Corporation   waiver   application   for
Methyl-10, June 3, 1983.

-------
                              -20-
11.   Comments  submitted  by Ashland  Oil Company concerning  the
first DuPont waiver application,  December 3,  1982.

12.   Comments submitted by Bank of America concerning  the  first
DuPont waiver application, November 18,  1982.

13.   John  R.  Morgan  and  Uwe   Zitzow,   "Effects  of  Alcohol
Blending   Procedures   on  1983-1984   Vehicle   Driveability",
presented  at  The U.S.  Blend Methanol Experience -  and Outlook,
May 17-18, 1984.

14.   "Performance Evaluation of Alcohol-Gasoline Blends  in 1980
Model Automobiles", Coordinating  Research Council  Report  #527,
July 1982.

15.   "Performance Evaluation of Alcohol-Gasoline Blends  in 1980
Model   Automobiles   Phase  II   -   Methanol-Gasoline   Blends",
Coordinating   Research   Council   Report   prepared   for   the
Department of Energy,  January,  1984.

16.   Letter  from  R.H.   Perry,  Mobil  Research and  Development
Corporation  to  A.  M.   Biervio,  Chrysler   Corporation,   with
attachments  on  fuel  properties  and driveability,  November  2,
1984.

17.   Esso  Petroleum Canada, summary of  oxygenate  driveability
test  program at  its Sarnia Research Centre  conducted  in summer
1984 and winter/spring 1985.

18.   Texaco,  Inc.,  "Consumer  Reaction  Program",   200  employee
owned cars, conducted September 1983 - September 1984.

19.   Philip   A.   Yaccarino,   "Hot  Weather   Driveability  and
Vapor-Lock Performance with Alcohol-Gasoline  Blends",  SAE Paper
852117, 1985

20.   Robert  L.  Furey   and  Jack  B.  King,  "Emissions,  Fuel
Economy,  and Driveability Effects of Methanol/Butanol/Gasoline
Fuel Blends", SAE Paper  821188, 1982.

21.   B.  Davis,  "Methanol/Gasoline  Blends  Cosolvent  Needs  and
Content  Control",   Sun   Tech,   Inc.,   slides   of  a  presentation
given in Arlington, VA,  October 17-18,  1983.

22.   J.  Morgan,  "Driveability  Characteristics of  Blends  Using
Methanol  With Ethanol Cosolvent",  Tennessee  Valley Authority,
1984 .

23.   Texaco,  Inc.,  "Evaluation  of 6/2 MeOH/TBA Gasoline-Alcohol
Blend  Performance",  Road  Simulator  Test  No.  30,  Presented  to
Chrysler.

-------
                              -21-
24.  Holger  Menrad,   Bernd   Nierhauve,  "Engine   and   Vehicle
Concepts for Methanol-Gasoline Blends",  SAE paper 831686,  1983.

25.  Estel M. Hobbs, Victor L. Kersey,  "Test  Vehicle Experience
with  Ethanol  Extended  Fuels -  Driveability  and  Corrosion",
Ashland  Petroleum  Company,   presented  at  American  Petroleum
Institute's 51st Midyear Refining Meeting, 1986.

26.  Letter  from  Charles  L.  Gray,  Jr.,   Director,  Emission
Control  Technology   Division,   EPA,   to  Wilhelm   Hall,   BMW
Corporation,  discussing   BMW's   examination   of  intake  valve
deposits in in-use vehicles,  March 27, 1986.

-------