[Federal Register Volume 66, Number 52 (Friday, March 16, 2001)]
[Notices]
[Pages 15236-15241]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 01-6725]


=======================================================================
-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY

[FRL-6953-3]


Notice of Availability for Public Comments--Infineum 
Corporation's VEKTRON 6913 Gasoline Fuel Additive Test 
Program

AGENCY: Environmental Protection Agency (EPA).

ACTION: Notice of availability for public comment.

-----------------------------------------------------------------------

SUMMARY: Today EPA is requesting comment on the results of a gasoline 
additive emissions test program submitted to EPA for technical review 
by Infineum Corporation. Infineum conducted a test program to determine 
the effect of the gasoline fuel additive VEKTRON 6913 in 
reducing tailpipe oxides of nitrogen ( NOX) emissions from 
on highway motor vehicles. This notice solicits specific comments on 
Infineum's test program, emission test results, and statistical 
analysis of the NOX emission impacts. Infineum conducted a 
series of tests on a sample population of current automobiles and light 
duty trucks that use gasoline that contains the VEKTRON 6913 
additive compared to vehicles operating without this specific additive. 
EPA invites comment to inform its decision-making concerning the 
evaluation of the emissions reduction testing program and the resulting 
conclusions made by Infineum associated with use of this additive.

DATES: The EPA is establishing a 45-day comment period, ending April 
30, 2001.

ADDRESSES: Comments should be submitted (in duplicate, if possible) to: 
Air and Radiation Docket and Information Center (6101), Attention: 
Docket No. A-00--XX, U.S. Environmental Protection Agency, 401 M Street 
SW, Room M-1500, Washington, DC 20460, telephone (202) 260-7548, 
between 8 a.m. and 5:30 p.m., Monday through Friday, excluding legal 
holidays. A reasonable fee may be charged for copying. Comments and 
data may also be submitted electronically by following the instructions 
under SUPPLEMENTARY INFORMATION of this document. Any confidential 
business information (CBI) should be submitted through e-mail.

FOR FURTHER INFORMATION CONTACT: For specific questions and comments on 
this guidance, contact Mr. Michael Ball, U.S. EPA, OAR/OTAQ/TRPD/TMIG, 
2000 Traverwood, Ann Arbor, MI 48105, telephone (734) 214-4897 
``[email protected]''

SUPPLEMENTARY INFORMATION: Electronic Availability--A World Wide Web 
(WWW) site has been developed so that you can obtain a copy of this 
announcement and supporting information for review and comment. The 
Uniform Resource Location (URL) for the home page of the web site is 
http://www.epa.gov/ttn/oarpg. You can find the protocol and supporting 
information under the heading titled ``What's New.'' If you need 
additional assistance with these web sites, call the TTN Helpline at 
(919) 541-5384. If you lack access to the World Wide Web, you may 
request a copy of the protocol and supporting information from the 
individual listed above under FOR FURTHER INFORMATION CONTACT.
    The EPA has established a docket for materials relevant to this 
notice (which will include the test program and supporting information, 
plus any public comments) under EPA air docket number A-2001-05. A 
public version of this record, including printed, paper versions of 
electronic comments--but excluding any information claimed as 
confidential business information (CBI)--is available for inspection 
from 8 a.m. to 5:30 p.m., Monday through Friday, excluding legal 
holidays. The official record is located at the address in ADDRESSES at 
the beginning of this document. Electronic comments can be sent 
directly to EPA at: [email protected]. Electronic comments must be 
submitted as an ASCII file avoiding the use of special characters and 
data will also accepted on disks in WorkPerfect in 5.1 file format or 
ASCII file format. Electronic comments on this document may be filed 
online at many Federal Depository Libraries.

Table of Contents

I. Overview
II. Information of Infineum's Test Program
III. Issues Which EPA Requests Comment On
    1. Test Fleet
    2. Vehicle Acquisition and Randomness of Vehicle Assignment
    3. Base Gasoline Composition
    4. Representativeness of Reference Additive
    5. Mileage-Dependency of Vektron 6913's Impacts
    6. Statistical Procedures Used to Identify Outliers
    7. Exclusion of Statistical Outliers
    8. Carryover Effect in the Analysis
    9. Carryover Effect and the Exclusion of All Run 2 Data
    10. Appropriateness of the Analysis on the Individual Test 
Results and the Averages
    11. Emissions Effects of the Vektron Additive Using 
the FTP Emissions and the US06 and HFET Emissions
    12. Combining Data Reflecting Continuous and Alternating Use of 
Vektron
    13. Appropriateness of a Single Point Estimate to be Developed 
from the FTP Emission Results or a Weight of Evidence Approach
    14. Calculating the Impact on HC and CO Emissions
IV. Conclusion

I. Overview

    Infineum USA L.P. manufactures a variety of gasoline and diesel 
fuel additives. In 1996, Infineum (then Shell Chemical) approached EPA 
regarding whether the Agency would accept the use of its 
Vektron 3000 series gasoline additive as an air quality 
control measure in a State's Implementation Plan (SIP) for purposes of 
emissions credit generation and trading. Based on a new fleet test 
program, Infineum has prepared and submitted to the EPA a

[[Page 15237]]

statistical analysis on the use of Vektron 6913, a close 
relative of the 3000 series, which contains a conventional detergent 
plus a NOX reduction component called Vektron 
1200. Infineum's analysis indicates a statistically significant 
reduction in oxides of nitrogen (NOX) emissions (estimated 
at 10 percent) from automobiles and light-duty trucks with no adverse 
impacts on emissions of hydrocarbons (HC) or carbon monoxide (CO), or 
on fuel economy. Infineum would like to promote the Vektron 
6913 additive as a method to generate emissions reductions (i.e., 
tradeable credits) which in turn could be resold to stationary sources 
in need of supplemental emissions reductions. Infineum has stated that 
it is not seeking a national mandate for its gasoline additive in 
reducing NOX emissions.
    Vektron 6913 is currently certified under EPA's gasoline 
deposit control additive program and registered under EPA's fuel and 
fuel additives additive health effects program.\1\ Infineum states that 
Vektron 6913 acts to reduce NOX emissions when a 
vehicle is operated over time using commercially available gasoline 
containing the additive.
---------------------------------------------------------------------------

    \1\ See 40 CFR part 80 regarding the requirements of EPA's 
gasoline deposit control additive and 40 CFR part 79 for the health-
effects requirements.
---------------------------------------------------------------------------

    The following discussion outlines the background of EPA's 
evaluation of the potential impacts of Vektron 6913 on 
emissions and summarizes the new data provided by Infineum. The purpose 
of this notice is to request comment to aid EPA in evaluating 
Infineum's fleet test and analysis regarding the impact of 
Vektron 6913 on the emissions of the in-use vehicle fleet. 
The comments we receive should provide useful information in 
determining the extent to which the use of Vektron 6913 in 
gasoline provides an emissions benefit. This is important in evaluating 
the use of Vektron 6913, under an emissions trading program 
or other state-sponsored market incentive initiative, in a state's SIP 
submission to the Agency.

History of EPA's Evaluation of Vektron 6913

    Infineum first approached EPA regarding the potential emissions 
benefits of its Vektron 3000 series gasoline additive in 
1996. At that time, EPA concluded that the available information was 
insufficient to quantify the overall emissions impacts of the gasoline 
additive.\2\ Infineum's initial request to EPA was primarily based on 
data that Infineum used to support its emissions trading protocol for 
Vektron 3364A (a similar package to Vektron 6913) 
under Ontario's Pilot Emissions Reduction Trading (PERT) Project in 
1997. Under the PERT project, the use of Vektron 3364A in 
gasoline supplied from several gasoline terminals generates 
NOX emissions credits which are being resold and used by 
stationary emissions sources in Canada (such as power plants) to meet 
emissions control requirements. The PERT project is a demonstration 
project that caps the amount of emissions credits which can be 
generated under a given protocol.\3\ PERT reviews Infineum's emission 
credit protocol for the use of Vektron 3364A at the 
conclusion of each ozone season.
---------------------------------------------------------------------------

    \2\ Letter from Charles N. Freed, Director, Fuels and Energy 
Division, U.S. EPA to Peter Chant, Consultant to Shell Chemical 
Company, August 12, 1996, II-C-01.
    \3\ The documents associated with Infineum's application for 
Emission Reduction Credits (RTC) for the year 2000 PERT project 
period are contained docket item II-G-02. This includes the 2000 
version of Infineum's emission credits protocol, and a summary of 
the emission credits generated in 2000 under the PERT program.
---------------------------------------------------------------------------

    To facilitate the evaluation of Infineum's emission credits 
protocol for Vektron in the U.S., Infineum submitted its 
protocol to the State of New Hampshire's Department of Environmental 
Services (NHDES) requesting acceptance for use in the State. The NHDES 
arranged for a public hearing on the protocol. That public hearing was 
held in Concord, NH, on October 1, 1997.\4\ The emission credits 
protocol proposed by Infineum in the context of this hearing was based 
on essentially the same data that was used to support Infineum's 
application under the PERT project in 1997.\5\ Comments on the public 
hearing were received from various parties, including EPA.
---------------------------------------------------------------------------

    \4\ Proposed Discrete Emissions Reductions (DER) submitted to 
the New Hampshire, Department of Environmental Services, Air 
Resources Division by Shell Chemical Company for its 
Vektron Gasoline Additives, July 1997, docket item II-G-
01, attachment c.
    \5\ The following documents associated with the public hearing 
in New Hampshire are contained in docket item II-G-01: Infineum's 
proposed emissions credits protocol, Comments on the public hearing, 
Shell's response to comments, and New Hampshire's conditional 
approval of Infineum's (then Shell Chemical's) ``Protocol for the 
Reduction of NOX through the use of VEKTRON 
3000 additive.
---------------------------------------------------------------------------

    At this time, EPA again concluded that the available data was 
insufficient to quantify the emissions impacts of Vektron 
3000 series. Based on comments from the public hearing, the state of 
New Hampshire issued a conditional approval for Infineum's protocol 
which granted a small, and according to Infineum, non-economically 
viable emissions reductions credit and which detailed the issues that 
would need to be addressed by Infineum before larger and more 
economically viable (to Infineum) emission credits could be generated 
from the use of Vektron 3000 series additives in the State of 
New Hampshire.\6\
---------------------------------------------------------------------------

    \6\ Conditional approval by the state of New Hampshire of 
Shell's proposed ``Protocol for the Reduction of NOX 
through the use of VEKTRON 3000 additive, December 12, 
1997, docket item II-G-01, attachment n.
---------------------------------------------------------------------------

    Following the public hearing in New Hampshire, in December of 1997, 
Infineum engaged EPA in discussions, seeking expert advice, regarding 
what additional testing and technical documentation Infineum might 
provide to facilitate our evaluation of the impacts of 
Vektron 6913 on vehicle emissions.\7\ Acting in consideration 
of technical input from EPA and representatives of U.S. automobile 
manufacturers Infineum designed and executed the emissions test program 
that is the subject of today's notice.
---------------------------------------------------------------------------

    \7\ Letter to Randall Evans, Infineum, and Peter Chant, 
Consultant to Infineum, from Deborah Wood, Acting Director, Fuels 
and Energy Division, February 10, 1999, docket item II-C-02.
---------------------------------------------------------------------------

II. Information on Infineum's Test Program

    The test program that Infineum conducted to evaluate the emissions 
and fuel economy impact of Vektron 6913 on a sample of 
vehicles, the test data from this program, and the statistical analysis 
of these data have been documented by Infineum. Infineum also provided 
information regarding issues that were not specifically addressed in 
its test program which have been included in the docket for this 
notice.\8\
---------------------------------------------------------------------------

    \8\ Fleet Test Evaluation of Fuel Additive Performance on 
Emissions, Final Report from Southwest Research Institute (SwRI) to 
Infineum USA LP, July 2000, docket item I-B-01.
    Test Data from Infineum's Test Program to Evaluate that Emission 
Impacts from Its Vektron 6913 Additive, Infineum USA LP, 
docket item I-B-02.
    Infineum Emissions Reduction Gasoline Additive Technology, 
Infineum USA LP, docket item I-B-03.
    Statistical Design and Analysis of Vektron 6913 
Emissions Fleet Trial, Infineum USA LP, docket item I-B-04.
---------------------------------------------------------------------------

    Infineum postulates that Vektron 6913 acts by some method 
other than a reduction of combustion chamber deposits (CCD) levels 
(mass). This impact may be due in part to a change in the properties of 
CCD over time as a vehicle is operated on gasoline that contains 
Vektron 6913.

III. Issues Which EPA Requests Comment On

    Since the mechanism by which Vektron 6913 may impact 
emissions is not well understood, there are issues regarding how to 
adequately account for potential interactions of variations in

[[Page 15238]]

vehicle technology, fuel composition, and vehicle operating cycle in 
the design of a test program to evaluate the impacts of 
Vektron 6913. These issues are summarized below. The majority 
of these issues were discussed in a letter from EPA to Infineum in 
February of 1999, during the October 1997 hearing in New Hampshire 
(referenced in section II) and during several Infineum and EPA 
technical meetings beginning in April 1999. We are also requesting 
comments on Infineum's statistical design of their test program and the 
analysis of the data from this program. We request comments on the 
issues associated with Infineum's evaluation of the impacts of its 
Vektron 6913 additive on the emissions performance of the 
vehicle fleet. The following sections broadly identify the basic 
subject areas included in EPA's evaluation of Infineum's request. EPA 
specifically invites comment on the issues identified below as well as 
any other related issues which commenters believe will inform EPA's 
decision making process. Comments on the issues outlined below will be 
most useful if they include a detailed rationale and technical 
discussion. Comments on any other relevant issues not raised in this 
notice are welcomed.

1. Test Fleet

    The test vehicles selected will impact the representativeness of 
the data in predicting the impacts of Vektron 6913 on the 
emissions of the current and future in-use fleet. Infineum states that 
the test fleet was composed of a broad spectrum of in-use vehicles 
selected based on sales volume which they believe adequately represents 
the impact of Vektron 6913 in the range of vehicle 
technologies present in the in-use vehicle fleet. We request comment on 
this issue. Specifically, we request comment on whether there are 
vehicles in the current or anticipated future fleet (e.g. tier 2 
vehicles), not represented in Infineum's test program that would 
respond in a substantially different manner to the use of 
Vektron 6913.

2. Vehicle Acquisition and Randomness of Vehicle Assignment

    In designing a test program it is important to acquire vehicles 
that would be representative of the current in-use fleet with regard to 
such factors as mileage accumulation, maintenance, and similar factors 
which could effect emissions performance. Additionally, in constructing 
a testing regime, random or other means of unbiased assignment is 
important to reduce the introduction of test result bias. The 
contractor that conducted the vehicle testing for Infineum (Southwest 
Research Institute) either purchased or leased all of the 28 vehicles 
used in the test program on Infineum's behalf. The test program 
criteria for vehicle acquisition included a minimum odometer mileage 
accumulation of 15,000 miles and a maximum of 75,000 miles. Infineum 
states that vehicles were assigned to a fueling regime randomly.\9\ 
Infineum orally related to EPA that as test vehicles became available, 
they were assigned the next available number in the relevant vehicle 
group. For example, the first vehicle available in the GP vehicle group 
was designated as GP-1, the second vehicle that became available was 
designated as GP-2, and so on. Vehicles in a group were assigned to a 
fueling regime, fueled according to their vehicle number (i.e., 1, 2, 
3, 4--see table 1 below). We request comment on the randomness of the 
vehicle selection and assignment to a fueling regime under Infineum's 
test program. We also request comment on the potential impact on 
program results due to the test vehicle fleet selected.
---------------------------------------------------------------------------

    \9\ Infineum Emissions Reduction Gasoline Additive Technology, 
Infineum USA LP, docket item I-B-03.

                                 Table 1.--Infineum Test Program Fueling Regime
                               [Fueling regime during mileage accumulation phases]
----------------------------------------------------------------------------------------------------------------
                                Pre-test phase  (1,000
            Vehicle                     miles)           Run 1  (8,000 miles)         Run 2  (8,000 miles)
----------------------------------------------------------------------------------------------------------------
Fuel:
  1...........................  Reference Fuel........  Reference Fuel........  Test Fuel.
  2...........................  Reference Fuel........  Test Fuel.............  Reference Fuel.
  3...........................  Reference Fuel........  Reference Fuel........  Alternating Fuels.
  4...........................  Reference Fuel........  Alternating Fuels.....  Reference Fuel.
----------------------------------------------------------------------------------------------------------------

3. Base Gasoline Composition

    The base gasoline composition is an important consideration in 
projecting the emissions impacts of the fuel additive on NOX 
when added to the range of typically available in-use fuels. The base 
gasoline used in the Infineum test program according to Infineum 
approximates a non-oxygenated California reformulated gasoline (e.g. 
low deposit forming potential). Infineum states that they are in 
agreement with industry experts, that gasoline composition can affect 
emissions. Infineum states that the base gasoline used in their test 
program would yield a conservatively small estimate of the impact of 
Vektron 6913 on emissions. We request comment on this issue. 
Specifically, we request comment on whether the impact of 
Vektron 6913 on emissions comparable to its impact related to 
in-use fuels would be substantially different in gasolines that have an 
aromatics content or T-90 distillation point or other differences from 
the test fuel used in Infineum's test program. In considering this 
issue, we request that commenters refer to the technical literature 
provided in Infineum's support document.\10\
---------------------------------------------------------------------------

    \10\ Section 7, Infineum Emissions Reduction Gasoline Additive 
Technology, Infineum USA LP., docket item I-B-03.
---------------------------------------------------------------------------

4. Representativeness of Reference Additive

    The reference additive used in the test program determines the 
baseline against which the emissions impacts of Vektron are 
evaluated. Thus, it is important that the reference additive be 
representative of current in-use deposit control additives as these 
might effect emissions performance. Infineum states that the test fuel 
(that contained Vektron 6913) and the reference fuel (that 
contained the reference additive) contained the same type and amount of 
detergent-active ingredients.\11\ The

[[Page 15239]]

reference fuel contained the reference deposit control additive, 
referred to as Infineum F7721, at a concentration of 154 pounds per 
thousand barrels (PTB) of gasoline. The test fuel contained 
Vektron 6913 at 234 PTB.\12\ We request comment on whether 
the reference additive as used by Infineum in this test program is 
representative (in composition and concentration) to other deposit 
control additives on the market with respect to its efficacy in 
controlling intake valve deposits and its impact on combustion chamber 
deposits (CCD). We are also interested in the extent to which the 
emission performance of vehicles operated on fuel using Infineum's 
reference additive accurately estimates the emission performance 
anticipated if the vehicles were operated on fuel using other typical 
additives. We specifically request comment on whether the test results 
presented by Infineum are predictive of Vektron 6913's 
impacts relative to the range of in-use additives on the market.
---------------------------------------------------------------------------

    \11\ Both the reference additive and Vektron 6913 have 
been certified by EPA under our gasoline deposit control program as 
suitable for use in gasoline to satisfy the requirement under 
section 211(l) of the Clean Air Act that all gasoline in the U.S. 
must contain additives capable of limiting the formation of deposits 
in engines and fuel supply systems. EPA's deposit control additive 
program (40 CFR part 80) requires that to be certified for use, an 
additive's ability to control fuel injector and intake valve 
deposits must be demonstrated using EPA-specified procedures.
    \12\ Infineum states that the reference additive is a 
conventional deposit control additive composed of pibamine detergent 
components, a synthetic carrier oil, and aromatic solvents. Infineum 
states that the Vektron 6913 additive contains the same 
detergent components and aromatic solvents with the synthetic 
carrier oil substituted by the Vektron 1200 
polyoxyalkylate component.
---------------------------------------------------------------------------

5. Mileage-Dependency of Vektron 6913's Impacts

    The issue of how quickly the emission impacts of Vektron 
6913 take effect is important to projecting its potential in-use 
emission benefits, because Infineum states that the impact of 
Vektron 6913 on emissions and fuel economy manifests itself 
only after a vehicle is operated using Vektron over time. 
Infineum intends to market its Vektron 6913 as an additive 
for fuel used only during the ozone season (for example, spring to late 
summer). Thus, the amount of mileage that must be accumulated using 
Vektron before it has its full impact on emissions is an 
important input in the calculation of the emissions benefits from the 
use of Vektron over the period it is used. Infineum states 
that after a vehicle is operated for 1,000 miles using 
Vektron 6913, the full effect of Vektron 6913 on 
emissions is realized.\13\ Infineum's test program evaluated 
Vektron 6913's impacts after the test vehicles were operated 
using Vektron 6913 for 8,000 miles. Infineum did not provide 
any publicly available data in support of their conclusion that the 
full impacts of Vektron 6913 are realized after 1,000 
miles.\14\ The issue of how quickly Vektron 6913 acts to 
impact vehicle emissions is significant in determining a set quantity 
of NOX emissions reduced over a given amount of time during 
which Vektron 6913 is used. We request comment on this issue.
---------------------------------------------------------------------------

    \13\ Section 6, Infineum Emissions Reduction Gasoline Additive 
Technology, Infineum USA LP., docket item I-B-03.
    \14\ See section 7, Infineum Emissions Reduction Gasoline 
Additive Technology, Infineum USA LP., docket item I-B-03.
---------------------------------------------------------------------------

6. Statistical Procedures Used to Identify Outliers

    The procedure used to identify outliers is important, as the 
exclusion of data from the analysis can significantly affect the 
emission change attributed to use of Vektron 6913. Infineum 
used a generalized linear model (SAS procedure PROC GLM) and 
Studentized-deleted residuals to the baseline and Run 1 data to 
identify two outliers: a Pontiac Grand Prix (GP1) and a Ford F-150 
(FF4). Vehicle FF4 was identified early before vehicle GP1 was run and 
gave a Studentized-deleted residual greater than 4.0 and was dropped 
from the analysis. When vehicle GP1 became available and was tested in 
the statistical model, its Studentized-deleted residual was 3.78 and 
consequently Infineum dropped this vehicle from the final analysis.
    A different statistical model was used to identify outliers (PROC 
GLM) than to determine emission effects (PROC MIXED). One outlier (FF4) 
was deleted from the database using a portion of the database and then 
further tests for outliers (e.g., GP1) were performed without the 
original outlier in the database. The presence of outliers was assessed 
only on the baseline and Run 1 data (i.e., excluding Run 2 data). In 
contrast, the test for carryover discussed below in Issue 8 involved 
data from both Runs 1 and 2, but excluded outliers identified using 
only Run 1 data.
    We request comment on Infineum's approach to removing outliers, 
particularly the following: (1) Use of a different statistical model to 
identify outliers (PROC GLM) than to determine emission effects (PROC 
MIXED); (2) deleting outliers one at a time; (3) identifying outliers 
after performing the assessment of carryover and comparing the emission 
impact of continuous and alternating use of Vektron 6913; and 
(4) inclusion of a vehicle type term in the PROC GLM model.

7. Exclusion of Statistical Outliers

    Determining that a specific data point or set of data is an outlier 
from a statistical point of view may or may not be sufficient evidence 
to appropriately exclude that data from an analysis. The vehicles in 
the Vektron test program were screened and pretested for 
proper operation, oil consumption and emissions. Replicate tests were 
performed at all testing points and a third test performed when the 
first two exceeded specified criteria according to the CAC Auto/Oil 
Protocol. Mechanical evaluation of FF4 at the end of the test program 
discovered problems with the EGR valve which was shown to affect 
NOX emissions significantly. At the end of the test program, 
no mechanical problems with GP1 were found. In general, based upon this 
analysis of the mechanical condition of the vehicles, should vehicles 
FF4 and GP1 be excluded from subsequent analysis? Should a statistical 
outlier be excluded if no mechanical problems could be found with the 
vehicle and the emission measurements were replicated? Should the HC 
and CO emissions of these vehicles be considered, as well as their 
emissions over the US06 and HFET test cycles, in determining whether or 
not to exclude them from the analysis?

8. Carryover Effect in the Analysis

    In the context of this test program, a ``carryover'' effect exists 
when the effect of one fuel additive persists during a subsequent test 
of a different additive, for example, treatment fuel during Run1 
affects the test of reference fuel during Run 2. Infineum believed that 
8,000 miles of operation on a specific fuel additive would be 
sufficient to eliminate such carryover, but subsequently concluded that 
the Run 2 data from vehicles fueled with the reference fuel still 
exhibited the affect of Vektron 6913. Therefore, Infineum 
chose to eliminate all Run 2 data from subsequent analysis. The model 
used to test for carryover effects specified ln(FTP NOX) as 
the response variable, accumulated-miles and treatment-order as fixed 
effects and vehicle type as a random effect.
    EPA requests comment on these points: (1) Is this model formulation 
appropriate to determine that carryover was present? Specifically, 
should the model have included a variable indicating the use of the 
Vektron additive, either continuously or in an alternating 
fashion? Also, should the test for carryover be focused exclusively on 
those vehicles in Run 2 which were operated on the Vektron 
additive in Run 1? Would there be any reason to expect carryover 
effects for those vehicles operated on the reference fuel in Run 1 and 
Vektron 6913 in Run 2? (2) Infineum states that the 
literature teaches that with a two-stage crossover design no estimate 
of direction of large carryover is possible. Is it possible to

[[Page 15240]]

determine the magnitude and direction of carryover in this type of 
study? Is this test program appropriately determined a 2 stage 
crossover design, as described in the drug testing (or other) 
literature, since the vehicles were not returned to their baseline 
conditions between Runs 1 and 2? (3) How does the use of fuel additives 
in test vehicles prior to recruitment into the test program affect 
interpretation of the test results, particularly the assessment of 
possible carryover? (4) Is it consistent to expect Vektron 
6913 to reach its full emission effect in 1,000 miles and to continue 
to affect emissions 8,000 miles after discontinuation of its use?

9. Carryover Effect and the Exclusion of All Run 2 Data

    Infineum excluded all Run 2 data from its final analysis, citing 
standard practice following detection of a large carryover effect. This 
reduces by approximately 50% the number of measures of the effect of 
the Vektron additive. As described above, Infineum concluded 
that a large carryover was observed and the reference was not repeated 
in Run 2. Therefore, Infineum believed that it was not appropriate to 
extrapolate Run 1 reference fuel data to compare to Run 2 test fuel 
data (as well as the reverse). Therefore, Infineum felt that no 
estimation of the treatment effect could be made when Run 2 data were 
included. Is it possible to include a term for carryover in the 
statistical model and continue to use all of the data to estimate the 
impact of Vektron 6913 on emissions? Is Infineum's decision 
to exclude all of the Run 2 data appropriate in the context of fuel 
additive testing, where we are not comparing the effect of two new 
factors (e.g. drugs), but the substitution of one fuel additive for 
others which are widely used? Specifically, is it appropriate to retain 
the Run 2 data for the subset of vehicles for which the test additive 
was used only in Run 2 (Vehicles numbered 1 and 3)? We also request 
comment on whether the testing of the reference additive for 8000 miles 
in Run 1 followed by the testing of the Vektron 6913 additive 
for 8000 miles simulates the situation which would occur in-use should 
the Vektron additive replace current additives?

10. Appropriateness of the Analysis on the Individual Test Results and 
the Averages

    Two options exist for analyzing the Infineum data. One 
statistically analyzes the data from all individual emission tests and 
the other analyzes the emission levels averaged across replicates at 
each test phase. Each vehicle at each test point was tested at least in 
duplicate over the FTP, HFET, and the US06 driving cycles. The need for 
a third emission test was based on test repeatability criteria 
established in the CAC Auto/Oil Protocol: if the higher emission value 
divided by the lower emission value was greater than 1.33 for HC, 1.70 
for CO, or 1.29 for NOX, a third test was run. Infineum 
believes that this procedure stabilizes the mean emission level 
sufficiently. However, in some cases where only two tests were 
performed, the two measurements differed by only a percent or two, 
while in others, the difference was nearly 30%. EPA requests comment on 
whether the statistical analysis would be appropriately performed on 
individual trial results, as well as on the average emission levels for 
each phase and which would be more illustrative of the effects of 
Vectron 6913.

11. Emissions Effects of the Vektron Additive Using the FTP 
Emissions and the US06 and HFET Emissions

    Emission data are available over three driving cycles: the standard 
EPA urban driving cycle (the FTP), the US06 high-speed, high-load 
cycle, and the Highway Fuel Economy Test (HFET). Infineum estimated the 
10% NOX emission benefit of its Vektron 6913 
additive using only emissions measured over the FTP. Should the 
emission results from these other two cycles be included in estimating 
the in-use emission impacts of Vektron 6913 or in responding 
to the other issues raised in this notice?

12. Combining Data Reflecting Continuous and Alternating Use of 
Vektron

    The Vektron additive was used continuously in half of the 
test vehicles and in every other tankful in the other half of the 
vehicles. Infineum selected these two fueling protocols to represent 
the two basic modes of potential use of the Vektron additive 
by individual vehicles in the fleet. Infineum reported no statistical 
difference in NOX emissions between the continuous and 
alternating (by tankful) use of Vektron 6913. Infineum based 
this conclusion on two types of analyses. First, Infineum used a simple 
model including only a term for vehicle type, and no treatment term. 
Based on a plot of residuals grouped by treatment (continuous versus 
every other tankful), Infineum concluded that the effects of continuous 
and alternating fuel treatments did not differ significantly. Second, 
Infineum compared the fit of two models, one including only a term for 
treatment, and another adding a term for fueling-scheme (continuous 
versus alternating). Based on a conclusion that including the fueling-
scheme term did not improve model fit, in combination with the results 
of the first analysis, Infineum pooled the continuous and alternating 
fuel treatment into a single treatment term (additive present) to 
assess the effect of additive treatment. Are the statistical analyses 
employed appropriate to support a conclusion that the two fueling 
schemes did not differ in result? Is the collapsing of data from the 
two fueling schemes into a single data set an appropriate step if no 
significant difference between them is found?

13. Appropriateness of a Single Point Estimate to be Developed from the 
FTP Emission Results or a Weight of Evidence Approach

    As indicated by the issues described above, it is possible to 
develop differing estimates of the impact of Vektron 6913 on 
NOX emissions, depending on whether Run 2 data is included 
or excluded, certain apparent outliers in Run 1 are included or 
excluded, whether the effect of Vektron 6913 used 
continuously or alternatingly is assumed to have the same effect or 
not, whether NOX emission impacts over the US06 and HFET 
cycles are considered, etc. Should we attempt to come up with the 
single best approach to analyzing the SwRI test data and develop a 
single point estimate, or where reasonable alternatives appear to 
exist, should we develop multiple estimates and then estimate an 
emission benefit for use in determining in-use emission credit trading 
purposes? If we should develop multiple estimates, how should an 
appropriate single in-use emission benefit be estimated from the range 
of NOX emission benefits resulting from the variety of 
possible statistical procedures? Should EPA be more concerned about 
ensuring that any projected NOX emission benefits associated 
with Vektron 6913 are highly likely to occur in-use (i.e., 
avoid the potential for over-estimating the benefit)? Or, should we be 
more concerned with encouraging innovative approaches to emission 
control (i.e., avoid under-estimating the benefit)?

14. Calculating the Impact on HC and CO Emissions

    Infineum found that its Vektron additive affect on 
NOX emissions was statistically significant, but that its 
effects on HC and CO emissions were not statistically significant 
(i.e., the 90% confidence interval included zero change in emissions). 
While the average change in HC and CO emissions found were smaller than 
the change in NOX

[[Page 15241]]

emissions, the average effects were not zero. If we determine the 
effect of Vektron 6913 on NOX emissions, should we 
determine that Vektron 6913 increases HC and CO emissions by 
the average amount found by the test program, or should we assume that 
the HC and CO effects are zero because the emission increases were not 
statistically significant?

IV. Conclusion

    EPA will carefully consider all comments received. We will evaluate 
these comments and other information or analyses which may become 
available, including perhaps conducting additional analyses of our own 
in arriving at our conclusion as to the emission benefits of 
Vektron 6913 as proposed for fuel additive use by Infineum. 
This conclusion will be publically available via our web site. If that 
conclusion indicates significant emission benefits could be derived 
from the use of this fuel additive, we will also prepare appropriate 
protocols for determining the extent of actual in-use on-highway fleet 
emissions benefits.

    Dated: March 7, 2001.
Robert Brenner,
Acting Assistant Administrator, Office of Air and Radiation.
[FR Doc. 01-6725 Filed 3-15-01; 8:45 am]
BILLING CODE 6560-50-U