
                 UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                NATIONAL VEHICLE AND FUEL EMISSIONS LABORATORY
                             2000 TRAVERWOOD DRIVE
                           ANN ARBOR, MI  48105-2498


									     OFFICE OF
									AIR AND RADIATION


March 1, 2013

MEMORANDUM

SUBJECT:	Test Program to establish LDV Full Useful Life PM Performance

FROM:	Rafal Sobotowski  -  USEPA/OTAQ/ASD

--------------------------------------------------------------------------------
TO:		EPA Docket # EPA-HQ-OAR-2011-0135
--------------------------------------------------------------------------------


Description of the Full Useful Life Program

      In order to establish the feasibility of the particulate matter (PM) emission standards proposed in the Tier 3 rule, the U.S. EPA conducted a test program to measure PM emissions from Tier 2 light-duty vehicles.  The test program was designed to measure PM emissions from late model year vehicles that represented a significant volume of annual light duty sales and included vehicles that ranged from small cars through trucks, including a cross section of technologies from port fuel injected (PFI) engines to gasoline direct injection (GDI) engines.  In addition, one PFI vehicle was included with known high oil consumption. 

      Seventeen model year 2005-2010 Tier 2 Bin 4, 5, and 8 vehicles were tested at the U.S. EPA National Vehicle and Fuel Emissions Laboratory (NVFEL) facility.  A summary of their characteristics is provided in Table 1. They included ten cars and seven trucks. Fifteen of these vehicles had accumulated 102,000-124,000 miles prior to the launch of the test program.  One vehicle had accumulated 37,000 miles and another accumulated 21,000 miles. Three cars and one truck were equipped with Gasoline Direct Injection (GDI) engines.  Twelve of the seventeen test vehicles had previously been used in the Department of Energy (DOE) V4 Program.  The remaining five vehicles were recruited in southeastern Michigan. 
      
      The twelve vehicles acquired from the DOE V4 Program were selected to represent a broad cross section of some of the highest sales vehicles in the U.S. market for model years 2005-2009.  These vehicles had originally been purchased by DOE with odometer readings ranging from 10,000-60,000 miles, placed in a mileage accumulation program and operated over the EPA Standard Road Cycle on a test track or on mileage accumulation dynamometers to 110,000-120,000 miles.  Immediately prior to inclusion in the EPA PM Test Program, the test vehicles were serviced per the manufacturer's published service schedule and maintenance procedures and underwent engine oil aging over a distance of 1,000 miles accumulated over the EPA Standard Road Cycle to stabilize engine oil contribution to PM emissions.
      
      One vehicle (vehicle K) tested from the DOE V4 program was suspected of having atypically high oil consumption and had only 37,000 miles of mileage accumulation. Vehicle K was a duplicate of Vehicle C and was determined to consume two and one half times the average oil consumption of vehicle C and three vehicles of the same make, model and model year which were tested within the DOE V4 program. Vehicle K did not accumulate 120,000 miles because it was removed from the mileage accumulation program after the atypical oil consumption was confirmed.  Vehicle K was tested in this program to investigate the potential impact of high oil consumption on PM emissions.  Three recruited test vehicles were selected because they used GDI technology and one GDI equipped vehicle was obtained from the DOE V4 Program.  Vehicle Q was recruited for another test program but was included in this program because it was a GDI technology vehicle.  Otherwise, an attempt was made to only recruit vehicles approaching the 120,000 mile useful life level.  All of the recruited test vehicles were thoroughly inspected, but otherwise tested as received.
      
      All vehicles were tested on an E15 fuel with RVP, aromatic content, sulfur content, T50 and T90 of 9.1 psi, 23.8 vol%, 7 ppm, 160F and 311F, respectively.  The properties of this fuel approximated those of a projected E15 market fuel.

      The test program included three cold start and three hot start UDDS tests and three US06 tests conducted on each vehicle.  FTP results were calculated for gaseous and PM emissions by applying the cold-start and hot-start weighting factors to the complete cold and hot UDDS results, respectively.  This eliminated separate analysis of the typically very low concentration FTP phase-2 gaseous and PM emissions samples and represented one method proposed within 40 CFR 1066 for increasing sample integration of measured gaseous and PM mass.  During these tests, triplicate PM samples were collected in parallel on PTFE membranes and single (composite) PM samples were collected on primary and secondary quartz filters for Thermol Optical Reflectance (TOR)  Organic Carbon (OC) and Elemental Carbon (EC)  PM speciation analysis.  Additional quartz filters were collected to determine the contribution of gas-phase artifact to the OC collected on the quartz filter samples.  The compositing of quartz filters over three repeats of each test was done to enhance the precision of subsequent OC/EC thermogravimetric measurements.  Single background (dilution air) PM samples were also taken during each emissions test.  Weekly tunnel blank and field blank PM samples were also collected. 
 
      The following parameters were measured: NOX, NMHC, NMOG, alcohols, carbonyls, CO, CH4, CO2 and fuel consumption and PM mass as per the 40 CFR 1065 and the proposed 40 CFR part 1066 test procedures.  Limited exhaust HC speciation was also performed.

      PM composition was determined from filter samples taken on both quartz filters and PTFE membranes.  PM compositional analyses include determination of the contribution of elemental and organic carbon to PM mass, elemental analysis via energy dispersive x-ray fluorescence (EDXRF), and sulfate analysis via ion chromatography.  
Table 1. Vehicles tested as part of EPA's Light-Duty Vehicle PM Test Program 
Vehicle Make, Model and Designation
Model Year
Certified to Emissions Standard
Odometer at Start of Program, miles
Fuel Delivery[a]
Used in DOE V4 Program?
Honda Civic
A
2009
Tier 2/Bin 5
121,329
PFI
Yes
Toyota Corolla
B
2009
Tier 2/Bin 5
120,929
PFI
Yes
Honda Accord
C
2007
Tier 2/Bin 5
123,695
PFI
Yes
Dodge Caliber
D
2007
Tier 2/Bin 5
114,706
PFI
Yes
Chevrolet Impala
E
2006
Tier 2/Bin 5
114,284
PFI
Yes
Ford Taurus
F
2008
Tier 2/Bin 5
115,444
PFI
Yes
Toyota Tundra
G
2005
Tier 2/Bin 5
121,243
PFI
Yes
Chrysler Caravan
H
2007
Tier 2/Bin 5
116,742
PFI
Yes
Jeep Liberty
I
2009
Tier 2/Bin 5
121,590
PFI
Yes
Ford Explorer
J
2009
Tier 2/Bin 4
121,901
PFI
Yes
Honda Accord
K
2007
Tier 2/Bin 5
36,958
PFI
Yes
Ford F150
L
2005
Tier 2/Bin 8 
111,962
PFI
No
VW Passat
M
2006
Tier 2/Bin 5
102,886
TGDI
No
Chevrolet Silverado
P
2006
Tier 2/Bin 8
110,898
PFI
No
Manufacturer's Development Vehicle[b]
N
PC
None (Tier 2/LEV II Prototype)[*]
120,011
TGDI
No
Saturn Outlook
O
2009
Tier 2/Bin 5
123,337
GDI
Yes
Cadillac STS4
Q
2010
Tier 2/Bin 5
21,266
GDI

No
Notes:
[a] PFI is Naturally aspirated, port fuel injected; GDI is Naturally aspirated, gasoline direct injection; TGDI is Turbocharged, gasoline direct injection
[b] Manufacturer's developmental vehicle.  Vehicle used a spray-guided GDI fuel system with a centrally-mounted injector and turbocharging.  Emissions were targeted at Tier 2 Bin 5 or better. 


Original Particulate Matter Results (Fall 2011)

      The original particulate matter test results measured in this program are plotted in Figure 1 and Figure 2, for the FTP and US06 test cycles, respectively.  In the figures, the vehicles are divided into two groups: PFI and GDI.  Within each group they are listed in the sequence of increasing CO2 emissions on the FTP test cycle.  The colored bars shown in the figures represent the means by test cycle and vehicle, and include accompanying standard deviations.  The data points indicate the average emissions for each individual vehicle test.  The number of tests, mean PM emission rates, and standard deviations by vehicle are tabulated in Tables 2 and 3.  All reported tests passed the Federal Register proportionality requirements for particulate sampling light-duty gasoline exhaust[iv].  Vehicles I and M are not included in Figure 3 because the tests did not meet the proportionality requirements on the FTP cycle. 
       
      As shown, the original PM emission rates for PFI vehicles were quite low on the FTP cycle, but were very high on the US06 for several PFI vehicles.  The original results on the US06 showed that several of the PFI vehicles had very high PM emission rates, which tended to be the higher-weight trucks. 

                                       
Figure 1. Original composite FTP PM emission results. Individual test results are marked with data points, average test results by vehicle +- 1 standard deviation are displayed with the bar graphs. 
                                       
Figure 2. Original US06 PM emission results. Individual test results are marked with data points, average test results by vehicle +- 1 standard deviation are displayed with the bar graphs. 
                                       
                                       
PM Sampling Contamination Issue

      Upon review of the elemental analysis results, high concentrations of silicon (Si) were discovered on some of the PTFE filters. Corresponding Si emission rates in mg/mile are reported in Tables 2 and 3, and displayed in Figure 3 and 4 for the FTP and US06 cycles. The highest Si emission rates were all measured on the US06 cycle, and were more than an order of magnitude higher than the highest Si emission rate reported on the FTP cycle. The Si emission rates on the US06 exhibited substantial variability, spanning four orders of magnitude among the different vehicles. The highest Si emission rates occurred predominately on the trucks. Each of the vehicles with high Si emission rates also had high PM emission rates on the US06. 
      

       Table 2. Original FUL FTP composite PM and Si results by vehicle
                                    Vehicle
                                     Tests
                               PM mean  (mg/mi)
                                 PM sd (mg/mi)
                                Si mean (ug/mi)
                                 Si sd (ug/mi)
                                       A
                                       3
                                     0.27
                                     0.11
                                      9.5
                                      2.8
                                       B
                                       3
                                     0.22
                                     0.09
                                      5.1
                                      3.5
                                       C
                                       3
                                     0.18
                                     0.07
                                      5.6
                                      1.4
                                       K
                                       3
                                     0.93
                                     0.23
                                      3.6
                                      0.7
                                       D
                                       3
                                     0.45
                                     0.10
                                     11.4
                                      0.8
                                       E
                                       3
                                     0.14
                                     0.03
                                      2.9
                                      1.3
                                       F
                                       3
                                     0.11
                                     0.06
                                      4.8
                                      1.3
                                       G
                                       3
                                     0.36
                                     0.05
                                     11.6
                                      3.6
                                       H
                                       3
                                     0.40
                                     0.22
                                     13.8
                                      4.9
                                       I
                                       -
                                       -
                                       -
                                       -
                                       -
                                       P
                                       3
                                     1.22
                                     0.12
                                     22.0
                                      2.5
                                       J
                                       3
                                     0.11
                                     0.10
                                     10.2
                                      5.8
                                       L
                                       3
                                     0.81
                                     0.14
                                     16.8
                                      5.6
                                       N
                                       3
                                     2.55
                                     0.74
                                     47.9
                                      4.2
                                       M
                                       -
                                       -
                                       -
                                       -
                                       -
                                       O
                                       2
                                     4.72
                                     0.30
                                     84.6
                                      3.5
                                       Q
                                       1
                                     7.15
                                       -
                                     73.0
                                       -
                                       
       Table 3. Original FUL US06 Composite PM and Si results by vehicle
                                    Vehicle
                                     Tests
                               PM mean  (mg/mi)
                                 PM sd (mg/mi)
                                Si mean (ug/mi)
                                 Si sd (ug/mi)
                                       A
                                       2
                                     0.76
                                     0.35
                                     18.4
                                      0.2
                                       B
                                       3
                                     2.05
                                     0.41
                                     225.6
                                     61.9
                                       C
                                       3
                                     1.05
                                     0.36
                                     101.5
                                     36.4
                                       K
                                       3
                                     2.84
                                     1.08
                                     257.5
                                     128.7
                                       D
                                       3
                                     7.39
                                     3.16
                                    1447.5
                                     605.0
                                       E
                                       3
                                     0.46
                                     0.36
                                     10.1
                                      7.4
                                       F
                                       3
                                     1.61
                                     0.09
                                     243.6
                                      5.8
                                       G
                                       2
                                     5.84
                                     1.10
                                     866.7
                                     350.2
                                       H
                                       3
                                     2.04
                                     0.64
                                     295.0
                                     84.6
                                       I
                                       2
                                     24.44
                                     19.26
                                    4349.7
                                    3939.1
                                       P
                                       3
                                     29.46
                                     2.10
                                    4777.3
                                    1119.4
                                       J
                                       3
                                     41.93
                                     5.97
                                    7441.1
                                    1038.2
                                       L
                                       3
                                     8.82
                                     1.21
                                    1298.3
                                     18.5
                                       N
                                       3
                                     2.37
                                     0.25
                                     133.1
                                     19.1
                                       M
                                       3
                                     37.42
                                     18.59
                                     516.8
                                     250.3
                                       O
                                       3
                                     4.75
                                     1.03
                                     858.4
                                     247.7
                                       Q
                                       3
                                     11.13
                                     5.98
                                     629.9
                                     374.2
                                       
                                       
Figure 3. Original composite FTP silicon emission results. Individual test results are marked with data points, average test results by vehicle +- 1 standard deviation are displayed with the bar graphs. 
                                       
Figure 4. Original composite US06 silicon emission results. Individual test results are marked with data points, average test results by vehicle +- 1 standard deviation are displayed with the bar graphs. 

      Sources of contamination were investigated in the test cell used for this program at the NVFEL facility.  The investigation discovered that silicone elastomer transfer tubes (pictured in Figure 5) were used to transfer the tailpipe exhaust from the vehicle to the constant volume sampler (CVS) system.  Silicone is a polymer that contains Si, C, H, and O atoms, which can break down at high temperature and exposure to lubricating oil[v]. The silicone tubes were suspected as the source of the high measured Si and PM rates on the US06.  Maricq et al. (1999) confirmed that silicone couplers can contribute to high concentrations of ultrafine particles measured under high exhaust temperatures[vi]. 
                                       
                                       
   Figure 5. Silicone exhaust transfer tubes used in initial NVFEL testing 

      A follow-up investigation program was conducted to verify the effect of silicone sampling lines on PM emissions and to assess its magnitude on the PM emission rates. Vehicle J (Ford Explorer) was selected for the investigation follow-up test program because it was the vehicle with the highest recorded PM and Si emission rates on the US06 test cycle. Three US06 tests were conducted with the silicone rubber transfer lines, and three tests with stainless steel transfer lines. Figure 6 graphs the PM emission rates recorded from the 2009 Ford Explorer. The average PM emission rates were 10.28 mg/mile with the silicone transfer tube and 0.92 mg/mi with the stainless steel transfer line. The Si emission rates with the silicone transfer tube were on average 0.91 mg/mile, but were only 0.003 mg/mile with the stainless steel transfer line. Using the results from the follow-up Ford Explorer US06 tests, almost all the PM measured on the US06 was coming from the silicone transfer tube. As suspected, the measured elemental silicon was only a portion of the silicone break down products that contributed to the particulate mass on the PTFE filter samples. In the case of the follow-up investigation, the elemental Si only contributed ~ 10% of the increase in measured PM. The follow-up study confirmed that silicone elastomer exhaust connections exposed as wetted surface area to hot engine exhaust can contribute substantially to particulate matter filter sample mass from CVS samples. 
                                       
Figure 6. US06 PM results from follow-up investigation on silicone elastomer transfer tubes
                                       
Criteria for Test Removal

      The follow-up investigation demonstrated that samples contaminated excessively by silicone transfer lines were no longer valid and need to be excluded.  The elemental Si emission rates were used to identify tests with excessive silicone contamination.  The follow-up test program estimated that silicone lines can contribute up to 10 times as much PM as measured elemental Si, yielding a PM:Si ratio of 10.  The PM:Si ratio likely varies test-to-test and vehicle-to-vehicle, and the PM:Si ratio for most tests appears to be smaller than 10.  By applying a 10:1 PM:Si ratio to the US06 Si data in Table 3, the silicone transfer tube would contribute more PM than was measured for nine of the sixteen vehicles, including Vehicle J (Ford Explorer). 
      
      All vehicles that had elemental Si emission rates greater than 300 ug/mi (0.3 mg/mi) were excluded.  This limit was selected to assure that none of the high PM emission rates on the US06 were caused by silicone contamination.  This criteria excluded eight vehicle tests on the US06: one PFI car (Vehicle D), five PFI trucks (Vehicle G, I, P, J, L) and two of the GDI vehicles (Vehicle  Q, O).  Using this criterion, the silicone transfer tube cannot contribute more than 2.84 mg/mile in the any of the remaining US06 PM averages, because all the remaining US06 PM values are less than 2.84 mg/mi (Vehicle K), notwithstanding the remaining Si rates are below 300 ug/mi. 
      
      The FTP tests had minor contributions of silicone contamination compared to the US06 cycle.  This is expected due to the smaller engine loads and lower exhausts temperatures experienced on the FTP cycle.  None of the FTP composite emission rates had Si emission rates above  100 ug/mile.  Assuming the 10:1 PM to Si ratio is an upper bound, the maximum silicone PM contribution on the final FTP results were 0.84 mg/mile, using the maximum FTP Si emissions recorded from Vehicle O.  

Supplemental Test Program (Summer 2012)

      The light-duty trucks with the four highest US06 PM and Si rates (J, I, L, and P) were retested at the NVFEL laboratory with stainless steel exhaust transfer lines on the same test fuel as the original test program.  The results for both the FTP and US06 from the stainless steel tests were used to replace the original data. 
      
      The final set of PM results (which includes all the retests) is graphed in Figure 7 and Figure 8 for the FTP and US06 test cycles, respectively.  The final average PM emission rates, number of tests and standard deviations are presented in Table 4.  The quality control criteria of 300 ug/mile appears to be a reasonable cutoff for excluding vehicle tests from the sample for demonstrating compliance.  Vehicle I had the highest PM emission rates in the final US06 data set which was tested with the stainless steel transfer line.  The US06 cycle PM emission rates from four vehicles (D, G, M and O) were removed due to silicone contamination and were not retested.  The FTP results from three of these vehicles are reported.  The FTP results from Vehicle M were removed due to proportionality requirements as discussed earlier.
                                       
Figure 7. Final composite FTP PM emission results. Individual test results are marked with data points, average test results by vehicle +- 1 standard deviation are displayed with the bar graphs. 
                                       
                                       
                                       
                                       
Figure 8. Final US06 PM emission results. Individual test results are marked with data points, average test results by vehicle +- 1 standard deviation are displayed with the bar graphs. 
                                       
         Table 4. Final PM results for the composite FTP and the US06
 
                                 Composite FTP
                                     US06
Vehicle
                                       n
                                 Mean  (mg/mi)
                                  SD (mg/mi)
                                       n
                                 Mean (mg/mi)
                                  SD (mg/mi)
                                       A
                                       3
                                     0.27
                                     0.11
                                       2
                                     0.76
                                     0.35
                                       B
                                       3
                                     0.22
                                     0.09
                                       3
                                     2.05
                                     0.41
                                       C
                                       3
                                     0.18
                                     0.07
                                       3
                                     1.05
                                     0.36
                                       K
                                       3
                                     0.93
                                     0.23
                                       3
                                     2.84
                                     1.08
                                       D
                                       3
                                     0.45
                                     0.10
                                   Removed 
                                       E
                                       3
                                     0.14
                                     0.03
                                       3
                                     0.46
                                     0.36
                                       F
                                       3
                                     0.11
                                     0.06
                                       3
                                     1.61
                                     0.09
                                       G
                                       3
                                     0.36
                                     0.05
                                  Removed  
                                       H
                                       3
                                     0.40
                                     0.22
                                       3
                                     2.04
                                     0.64
                                     I[1]
                                       2
                                     1.36
                                     0.08
                                       2
                                     3.31
                                     0.81
                                     P[1]
                                       3
                                     0.18
                                     0.06
                                       3
                                     1.83
                                     0.81
                                     J[1]
                                       3
                                     0.10
                                     0.04
                                       3
                                     0.27
                                     0.01
                                     L[1]
                                       3
                                     0.39
                                     0.13
                                       3
                                     2.13
                                     1.76
                                       N
                                       3
                                     2.55
                                     0.74
                                       3
                                     2.37
                                     0.25
                                       O
                                       2
                                     4.72
                                     0.30
                                   Removed 
                                       Q
                                       1
                                     7.15
                                       -
                                   Removed 
Notes:
[1]Retested vehicles at NVFEL laboratory with stainless steel 


Discussion

      The silicone contamination issue had a very significant impact on the results of the US06 cycle, illustrated most dramatically with Vehicle J, which reduced the US06 PM emissions from over 40 mg/mile with the silicone contamination to under 1 mg/mile with the stainless steel transfer line. The silicone contamination on the FTP cycle was minor and no tests were removed or replaced due to silicone contamination. 

      In the final data set, all but one of the PFI vehicles had FTP PM emission rates under 1 mg/mile, and all PFI vehicles had emissions below the proposed 3 mg/mi FTP standard.  On the US06 cycle all of the PFI vehicles were well in compliance of the proposed US06 standards, and all tests were below 4 mg/mile. 
      
      The GDI vehicles all had higher FTP emissions than the highest tested PFI vehicle. In comparison to the proposed 3 mg/mi standard, only Vehicle N, the spray-guided OEM vehicle, had average FTP emissions below the standard. More testing on GDI vehicles is needed to evaluate their PM emissions on the US06 cycle, but the current testing shows that GDI technology (Vehicle N) can perform comparable to PFI engine technology on the US06 cycle.  
      
--------------------------------------------------------------------------------

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

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

