Characterization of Emissions from Handheld Two-

 stroke Engines

 AS3A

 Peter A. Gabele

 U.S. Environmental Protection Agency
 MD-46, Research Triangle Park, NC 27711

 ABSTRACT

 Despite their extremely high organic and particulate matter emission rates, two-stroke
 engines remain among the least studied of engine types. Such studies are rare because
 they are costly to perform. Results reported in this paper were obtained using a facility
 that shares exhaust emission sampling and analytical systems with an adjacent
 automotive emissions test laboratory. This enabled a comprehensive examination of
 emissions from three hand-held, two-stroke, engines. In addition to the determination of
 routine regulated pollutant emission rates, organic emissions are speciated and particulate
 matter emission rates are measured. Each engine was tested using two fuels:  1990
 Baseline Gasoline and a reformulated gasoline. One  engine was ran for an extended
 period to assess durability effects on exhaust emissions, and another was operated at a
 leaner air-fuel ratio setting to examine enleanment effects. In the durability testing,
 organic and particulate matter emissions actually appear to decrease slightly with
 increased operating time. In the enleanment test, organic and carbon monoxide emission
 rates decreased significantly, as expected, with the increase in air-fuel ratio. Large
 decreases in particulate matter also ensued. Fuel selection strongly influenced
 composition of organic emissions. With reformulated gasoline, gaseous toxic and ozone
 potential emissions were consistently lower compared to emissions with the Baseline
 fuel.

 INTRODUCTION

 Small gasoline nonroad engines (below 19 kilowatts) contribute about five percent of the
 summer volatile organic compounds (VOCs) in the 19 nonattainment areas nationwide '.
 In July 1995, U. S. Environmental Protection Agency (EPA) finalized the first federal
regulations affecting these engines as a first step toward reducing their emissions. The
 regulations, commonly known within the industry as "Phase 1", took effect in the 1997
 model year and apply to gasoline engines below 19 kilowatts. The most common
 applications which use these engines are lawn mowers and other garden equipment.

The string trimmer, after the lawn mower, is the second most common small  gasoline
nonroad application in the United States, A typical string trimmer is powered by a two-

-------
 stroke gasoline engine that has a hydrocarbon (HC) emission rate approximately six times
 that of the typical four-stroke lawn mower engine1. Unburned HC emissions from string
 trimmer engines are extremely high due to fresh mixture (fuel-air) short-circuiting the
 cylinder during scavenging.  Carbon monoxide (CO) emissions also tend to be elevated
 but are comparable with the high levels measured on small four-stroke engines 1>2>3>4.

 String trimmers and other small two-stroke engines are categorized as small spark-
 ignition nonroad handheld engines.  Proposed Phase 2 standards for these engines are
 currently being considered and would reduce HC and nitrogen oxide (NOJ another 80
 percent below the Phase 1 standards. These standards would be effective beginning in
 2002.

 Small, hand-held engines have an impact on the atmospheric environment that is
 accounted for in current emission inventories. But perhaps a more cogent concern is their
 potential as a source for air toxic exposure to operators.  Laborers in the landscape
 industry frequently operate these devices for extended periods, thus exposing themselves
 to high concentrations of exhaust gases over prolonged periods of time. Since the
 exhaust gases consist of large fractions of unburned gasoline, there is a likelihood that
 workers are being adversely exposed to benzene, 1,3-butadiene, and other possible toxic
 compounds contained in gasoline. Toxic compounds produced during combustion may
 also present a hazard.

 Toxic emissions from mobile sources are defined as benzene, 1,3 butadiene,
 formaldehyde, acetaldehyde, and particulate poly-aromatic hydrocarbon (PAH)
 emissions. All except the particulate PAH emission rates were determined in this study
 by speciating the organic emissions from the engines. The speciation was carried out to
 assess ozone potential of the emissions since the study was sponsored by EPA's North
 American Research Strategy for Tropospheric Ozone (NARSTO).  Specifically, this
 study develops speciation profiles from nonroad, small two-stroke utility engines. Only a
 few papers describe such characterization efforts for small two-stroke engines 5*6'7 because
 the studies are rare and costly. Facilities that characterize emissions from roadway
 vehicles are best equipped to economically perform the same analyses on small engine
 emissions.

 The results presented in this paper can be used to perform exposure assessments to
gaseous toxic emissions (aggregate toxic minus the particulate PAH emissions) and to
model source contributions to ozone pollution from small two-stroke engines. The
 studies, which characterize exhaust emissions from new and in-use, two-stroke string
trimmer engines, were carried out at EPA's mobile source emissions laboratory in
Research Triangle Park, North Carolina. Exposure and ozone impacts from similar two-
stroke engines can be inferred from the results in this study because the organic
composition of their emissions can be assumed to be similar. The impact of reformulated
fuel on emission rates and composition is examined by testing two fuels: a 1990 baseline
gasoline (RFA) and a reformulated gasoline (RFG).

-------
 EXPERIMENTAL METHOD

 Test Equipment

 The test cell has an eddy-current dynamometer that can absorb up to four horsepower.
 An engine-dynamometer controller consists of two separate units: a dynamometer
 controller that varies excitation current to the dynamometer maintaining speed or torque,
 and a throttle controller that controls the engine's throttle to maintain desired torque,
 speed, or throttle position. The engine-dynamometer controller is interfaced to a
 computer that commands second-by-second speed, torque, or throttle position called for
 by the test cycle.

 Engine exhaust is directed to a constant volume sampler (CVS) via a large hood near the
 engine's exhaust outlet (see Figure 1). Dilution air from the engine room forms an
 envelop around the exhaust and the diluted exhaust mixture is drawn into the CVS at a
 rate of about 470 SCFM. Heated sample lines are used to collect diluted exhaust gas
 samples for aldehyde, real-time, and bag analyses. The bags are first analyzed for
 regulated emissions (HC, CO, and NOX), then a sample is drawn  off for HC speciation.
 Particulate matter (PM) emissions, currently unregulated for small gasoline nonroad
 engines, were also collected and measured.
          Background Sample Line
                                                                     Real-Time
                                                                     Analysis Bench
 * Heated Sample Line (HSL)
Figure 1. Sampling system schematic.

During the test, data needed to calculate emission rates from emission analyzers, the
engine-dynamometer controller, and the CVS are brought into a computer in real-time
(one sample per second). This enables the calculation of real-time or modal regulated
emissions following the test.

-------
 Test Procedure
 Three engines (see section below for a more complete description of the engines) taken
 from string trimmers were selected as test engines for the study. One was an in-use
 engine that had been used by a consumer for about one year before it was tested, the other
 two were new.  Of the two new engines, one was examined at a leaner (higher air-fuel
 ratio) setting in addition to its "as received", richer, setting. The other new engine was
 operated over an extended period of time to examine durability effects on emissions.

 Table 1. Test matrix showing number of replicates for each test condition.
| RFA
RFC

J1088 tests
Ryobi (rich)
Ryobi (lean)
Weed Eater
Homelite
RFA
32h
48h
RFC
48 h

4
4
6
4
2
X
X
5
X
X
X
X
X
X
X
X
X
X
X
X

C2M tests
Ryobi (rich)
Ryobi (lean)
Weed Eater
Homelite

4
7
6
3
9
X
6
6
X
X
1
X
X
X
4
X
X
X
3
X
Back-to-back composite two-mode (C2M) duty cycles were used to accumulate operation
time on the engine. The C2M is a composite of the 2-mode, SAE J1088 recommended
practice, that was adopted as the federal certification cycle for handheld engines.

Exhaust gas emissions were generated by operating the engines using the 2-mode version
of the SAE J1088 and the C2M duty cycle. Replicate tests were conducted throughout the
study except for tests on the Weed Eater for the 32 hour durability checks. Table 1 shows
the test conditions examined using both duty cycles and the number of successful
replicates performed in each case. The RFA and RFG fuels were alternated after every
three tests.

The SAE J108S duty cycle - The SAE J1088, or the certification duty cycle, is
conducted by taking exhaust gas samples while the engine is operating at two steady-state
modes or conditions.  The first condition is defined as 100 percent of rated torque and 85
percent of rated speed; the second is at idle or no load. After all engine and sampling
system readings have stabilized for the mode to be tested, the sample is taken for a 600
second period.

The C2M duty cycle - The engine is run through two modes (same modes as in the
J1088) while emissions are sampled continuously. The time spent in each mode is
proportional to the weight given the mode in the certification procedure: 90 percent is

-------
 spent at full load and 10 percent at idle. The C2M has a duration often minutes and it is a
 quasi-steady-state duty cycle since emissions are collected during the transient that occurs
 when shifting from mode 1 to mode 2. Instead of having four (two samples and two
 background) bags to analyze following the test, the C2M has only two (a sample and a
 background) bags. This reduces the cost and simplifies the analysis component. During
 the C2M test, the engine is maintained at rated speed except at idle when it decreases to
 the lowest speed inside a smooth operating regime.

 Gaseous and Participate Emissions Measurement

 Exhaust emission rates were determined for HC, CO and NOX using standard bag
 sampling, analytical, and calculation procedures8. They were compared for QC purposes
 to emission rates obtained using integrated results from the real-time sampling and
 analysis system. The non-methane organic gaseous (NMOG) emissions were determined
 by taking the sum of the individual non-methane HC emissions following analysis with a
 gas chromatography-flame ionization detector (GC/FID), then adding in the oxygenate
 and aldehyde emissions. Samples for GC/FID analysis were collected in 60-L Tedlar
 bags from the CVS dilution system and from room air (background). The GC speciation
 methodologies were those used in the Auto/Oil Air Quality Improvement Research
 Program9. Integrated GC-FID peak measurements were compared to the HC
 measurements obtained using the standard FID procedure for QC purposes.

 Aldehydes were sampled through a heated sample line (110 C) and collected on
 dinitrophenylhydrazine (DNPH) -coated silica gel cartridges. Two cartridges were drawn
 during each test; one from the exhaust gas and one from the background. The aldehyde
 samples, trapped on the cartridge as individual DNPH aldehyde derivatives, were then
 analyzed by high-performance liquid chromatography9.

 Particulate matter was isokinetically sampled using two straight stainless steel probes that
 extend from the rear of an 20.3 cm diameter, 3 m long dilution tube.  One of the sample
probes was connected to a PM10 and the other to a PM2 5 cyclone (University Research
Glassware, Carrboro, NC). Particles exiting the cyclones were collected onto 47 mm
diameter, 2.0 urn pore size Gelman Zefluor filters that were equilibrated for 24 h before
being weighed.

Emissions* Reactivity

Ozone-forming potential of the exhaust VOC is based upon the incremental reactivity
approach developed by Carter and Atkinson 10.  Application of this concept has led to the
development of the Maximum Incremental Reactivity (MIR) method, that  expresses the
reactivity of exhaust NMOG and CO from the test engine n. The method assigns a
specific reactivity level to each of the organic species and to CO, and calculates a
reactivity-weighted emission (RWE) rate for each specie by multiplying the specie's
specific reactivity times its emission rate. The RWE rate for an exhaust sample is

-------
 obtained by summing all of the individual RWE rates for each specie present. The RWE
 rate, expressed in units of grams ozone per kilowatt-hour, is a useful measure of the
 ozone potential of the CO and organic emissions in urban atmospheres where the VOC-
 to-NOx ratios are relatively low (approximately 6). The specific reactivity of the organic
 emissions can easily be obtained by dividing the organic RWE rate by its emission rate.

 Test Engines and Fuels

 Three two-stroke engines were used in this study: two new and one in-use engine. A
 description of the engines is given in Table 2. The in-use  engine, a Homelite (produced
 by John Deere), was used regularly around a heavily landscaped, one acre home-site for

 Table 2. Engine descriptions.
Engine ID
Ryobi
(TN4031UB24
RA:EM)
Weed Eater
(VPW021UB2
4RA:EM)
Homelite
(SH2025UB24
RA)
Model #
725REZ

UT#2060
1C
Age & Time
In- Use
(years)
New
New
4yr,
In use 1 year
Displace-
ment
(cc)
3 Ice
24cc
25ec
Rated Power
@rpm
0.94 hp@6500
0,65 hp@7000
0.75 hp@7500
Max. Test Torque
@rpm
0.44ft-lb@5910
0.43 ft-lb@5 170 (rich)
0.61 ft-lb@6230
0.45 ft-lb@6205
Bore/
Stroke
35 mm/
32.5 mm
34.5mm/
26.4 mm
33,3 mm/
28.6 mm
about one year before being tested.  It complies with 1995-98 California regulations
pertaining to utility, lawn and garden equipment. The Ryobi engine, also certified to the
1995-98 California regulations, was tested first.  In its initial test series, the engine was
operated "as received". However, the uncharacteristically high organic emission rates that
ensued may have been caused by an inadvertent turning of the mixture adjustment screw.
In the final test series, the carburetor (mixture adjustment screw) was adjusted to a leaner
setting in order to lower the organic emission rates. The carburetor was adjusted so that
the engine continued to operate smoothly without misfire, and was able to be tested at
approximately the same power settings used before the engine's combustion was made
leaner. The other new engine, a Weed Eater (produced by Poulan), was tested over a
particularly long period of time to determine durability effects on emissions. This engine
was certified to the 1997 California regulations.

Two test fuels were examined in this study: 1990 Baseline Gasoline designated RFA, and
a reformulated gasoline (California certification  gasoline), designated RFG. Compared to
the RFA gasoline, the RFG has lower levels of olefm and aromatic compounds,  lower T90
(temperature at which 90 percent of the fuel boils-off), lower Reid vapor pressure (RVP),
lower sulfur, and a presence of oxygenate (about 10 percent MTBE). A more detailed
description of the fuels is given in Table  3.

-------
 Table 3. Fuel descriptions.
Fuel Type

Specific Gravity
Sulfur, wt %
Benzene, ppmC%

Aromatics, vol. %
Olefins, vol. %
Paraffins, vol. %
MTBE, vol. %

Research Octane No.
Motor Octane No.
Octane Index

Carbon, wt %
Hydrogen, wt %

Reid vapor pressure, psi
Distillation, C IBP
10%
50%
90%
EP
RFA

0.7469
0.0315
2.11

31.8
11.5
56.7
0.1

92.2
83.8
88.1

13.3
86.7

8.65
35
51
102
165
216
RFC

0.7394
0.0045
1.19

27.0
5.1
55.1
12.6

96.6
87.3
92.0

13.8
86.2

6.65
40
58
94
150
193
RESULTS AND DISCUSSION

Regulated and Particulate Matter Emissions

The routinely gathered emission rates, including those for regulated pollutants, are given
in Table 4. In general, emission rates varied more between engines than between fuels.
The Ryobi engine is treated as two separate "engines" because the two mixture settings
that were tested represent two entirely different engine configurations. The standard
deviations included in Table 4 provide a measure of the data repeatability in the study.
Two of the test engines, the Weed Eater and the Homelite, comply with the Phase I
federal emission standards for Class IV handheld engines. These Phase I standards for
HC5 CO, and NOxwere 241 g/kW-h, 805 g/kW-h, and 5.36 g/kW-h, respectively. The
Ryobi (Ryobi-rich) engine's emissions exceeded the standard when it was operated "as
received". However, it is possible that the mixture adjustment screw had been
inadvertently turned to a richer setting prior to testing. Its emission rates came close to
certification levels when the air-fuel mixture screw was adjusted to a leaner setting
(Ryobi-lean) for additional testing. The dramatic difference in emission rates between the
lean and rich mixture settings illustrates the important effect that air-fuel ratio has on
emissions. Increasing the air-fuel mixture away from stoichiometry results in decreased
emissions of hydrocarbons and CO, but increased emissions of NOX. That classic trend is

-------
 apparent here and it is noted that PM emission rates also decrease substantially with
 increasing air-fuel ratio.
 Table 4. Regulated emission rates (standard deviations in parentheses).

Ryobi-rich RFA
J1088
C2M
Ryobi-rich RFC
J1088
C2M
Ryobi-lean RFA
J1088
C2M
Weed-Eater RFA
J1088
C2M
32-h
48-h
Weed-Eater RFC
C2M
48-h
Ilornelite RFA
J1088
C2M
Ilomelite RFC
J1088
C2M
HC
g/kW-h

614
(71)
649.5
(32)

604
(90)
664
(21.8)

265
(40)
290
(60)

170.5
(9.4)
185.4
(9.3)
172.3
176.6
(4.1)

180.0
(5.6)
130.2
(6.9)

173.9
(11.8)
208.1
(24.1)

157.9
(31.4)
204.7
(25.7)
CO
g/kW-h

980
(17)
1018
(12)

961
(79)
980
(30)

646
(71)
633
(70)

527
(22)
518
(17)
497
486.0
(63.4)

488
(14)
300.5
(69.0)

377.4
(38.3)
401.9
(29.7)

315.1
(134.3)
357.2
(97.4)
NOx
g/kW-h

0.02
(.05)
0
(0)

0
(0)
0
(0)

2.2
(1.5)
3.4
(1.2)

1.6
(0.9)
1.1
(0.8)
2.0
2.3
(0.7)

1.0
(0.7)
1.5
(0.5)

2.4
(1.0)
2.1
(0.6)

2.8
(0.5)
2.4
(2.2)
CO2
g/kW-h

772
(150)
835
(8.4)

867
(166)
720
(53)

976
(88)
1074
(61)

889
(50)
930
(36)
993
947
(63)

1018
(71)
1057
(53)

1176
(41)
1250
(139)

1237
(57)
1458
(89)
NMOG
g/kW-h

572
662
(4.1)

503
673
(31)

249
238
(0.6)

192.6
188.0
(6.2)
163.8
163.9
(18.5)

186.9
(8.6)
136.4
(4.0)

159.4
(10.1)
197.7
(11.7)

167.5
208.6
(29.8)
PM10
mg/kW-h

39.4
(1.6)
42.2
(0.3)

46.9
(8.8)
39.2
(.5)

17.9
(2.2)
18.8
(3.3)

15.7
15.2
(0.8
11.4
12.8
(3.3)

15.0
(0.9)
11.1
(1.3)

9.0
(1.5)
8.0
(2.4)

10.3
(1.0)
11.2
(3.8)
PM2.5
mg/kW-h

35.5
(2.8)
38.5
(2.7)

25.9
(8.6)
36.6
(1.8)

16.9
(3.3)
18.0
(4.1)

16.3
15.2
(0.8)
16.0
12.7
(3.2)

15.1
(0.9)
11.0
(1.3)

9.7
(2.9)
8.0
(3.0)

10.8
(1.4)
11.4
(3.2)
This trend associated with leaned-out combustion is observed in an examination of the
durability data reported on the Weed Eater engine. Here there are slight decreases hi HC
and CO, accompanied by slight increases in NOX emission rates as run-time for the engine
increased. The decreases in PM rates are even more apparent. All this suggests that the

-------
engine's air-fuel ratio was increasing slowly as the engine aged. This enleanment could
occur if deposits began obstructing passageways in the carburetor over time.

Both PM10 and PM2.5 emission rates were measured during testing. Emission rates are
about equal for both implying that the particles being generated are predominately less
than 2.5 microns. This is not to say that the engines were not spraying oil during the tests.
An aerosol spray was emitted that impacted on the transfer tube and tunnel walls.
Virtually none of this aerosol made it to the cyclones which were situated at the end of
the tunnel. The interior surfaces of the cyclones were surprisingly dry following a test.

The NMOG emission rates vary both above and below the reported HC emission rates
because they are determined using a different method. Integrated non-methane readings
from the GC/FID are added to oxygenate and aldehyde emissions also determined using a
GC/FID.
Use of RFG fuel resulted in lower HC and CO emission rates in eleven of twelve head-to-
head comparisons. Most are statistically insignificant due to variability of the data, but
overall there appears to be a trend. Particulate matter emission rates are more of a mixed
bag with no trend emerging due to usage of the two fuels.

Table 5. Gaseous toxic emission rates.

Ryobi-rich RFA
J1088
C2M
Ryobi-rich RFG
J1088
C2M
Ryobi-lean RFA
J1088
C2M
Weed-Eater RFA
J1088
C2M
32-h
48-h
Weed-Eater RFG
C2M
48-h
Homelite RFA
J1088
C2M
Homelite RFG
J1088
C2M
1,3-buta-
diene
g/kW-h

0.6
0.6

0.5
0.7

0.5
0.5

0.4
0.3
0.3
0.3

0.3
0.3

0.2
0.4

0.4
0.3
Benzene
g/kW-h

14.3
14.9

6.5
9.5

6.5
5.8

4.1
3.8
3.4
3.7

2.5
1.9

3.2
4.0

1.8
2.2
Form-
aldehyde
g/kW-h

0.6
1.5

0.9
1.6

0.5
1.0

0.5
0.4
0.3
0.5

0.6
0.5

0.8
1.1

1.6
1.6
Acetalde-
hyde
g/kW-h

0.2
0,7

0,2
0.4

0.2
0.4

0.2
0.1
0.1
0.2

0.1
0.1

0.3
0.4

0,4
0,4
Gaseous
Toxics
g/kW-h

15,7
17.7

8.1
12.2

7.7
7.7

5.2
4.7
4.1
4.6

3.5
2.7

4.5
5.9

4.1
4.5

-------
Gaseous Toxic Emissions
Gaseous toxic emissions from these engines are dominated by benzene. Benzene
emissions are closely related to benzene levels in the gasoline. Benzene emission rates
with RFG, which contains about half the benzene contained in an equal measure of RFA,
are nearly half the levels with RFA, Generally, formaldehyde emission rates are higher
with RFG. This is consistent with results from other studies that have examined
emissions using the same two fuels.12 Most oxygenates will produce some increase in
aldehyde emissions due to fragmentation of their carbon-hydrogen-oxygen groups that
closely resemble aldehydes. Formaldehyde emissions with MTBE occur when its
methoxy group is broken off during combustion.

The gaseous toxic emission rates from engine-to-engine and from the two fuels vary in
the same direction as the HC emission rates. Both the gaseous toxic and HC emission
rates are ordered as follows: Weed Eater < Homelite < Ryobi-lean < Ryobi-rich,  and RFG
<  RFA. Higher emissions of unburned fuel (HC) translate to higher emissions of
benzene, the principal component of the gaseous toxic pollutants.

Table 6. Ozone potential emissions and related measurements.


Ryobi-rich RFA
J108S
C2M
Ryobi-rich RFG
J1088
C2M
Ryobi-lean RFA
J1088
C2M
Weed-Eater RFA
J1088
C2M
32-h
48-h
Weed-Eater RFG
C2M
48-h
Homelite RFA
J1088
C2M
Homelite RFG
J1088
C2M
MIR
g O3/kW-h


1976
2388

1490
1909

909
891

732
721
650
622

641
454

616
774

594
681
Specific
Reactivity


3.35
3.52

2.86
2.76

3,51
3.63

3.68
3.69
3.80
3.65

3,29
3.21

3.74
3.81

3.44
3.18
Principal MIR Fractions
Paraffin
Olefm
Aromatic
CO

19%
17%

20%
23%

17%
16%

15%
16%
14%
-15%

16%
19%

15%
15%

15%
18%
22%
20%

19%
20%

26%
25%

24%
26%
24%
26%

24%
26%

26%
27%

32%
29%
55%
59%

52%
50%

52%
53%

56%
53%
55%
52%

52%
48%

54%
52%

45%
45%
3%
2%

3%
3%

4%
3%

4%
4%
4%
4%

4%
4%

4%
3%

3%
3%

-------
 Ozone Potential Emissions

 Ozone potential emissions are a function of the organic and CO emission rates, and of
 their reactivity. The reactivity of the organic emissions (specific reactivity) varies with
 their composition, and for a given engine configuration is largely dependent upon the
 type of fuel being burned. In every case the specific reactivity is less with RFG than with
 RFA because it contains fewer reactive hydrocarbon compounds. The specific reactivity
 of the two fuels themselves is 3.47 for RFA and 2.82 for RFG. Fuel selection with these
 two-stroke engines influences the emissions' reactivity more than with four-stroke
 engines tested in a separate study using the same two fuels.12 In that study the average
 specific reactivity was reduced only two percent with RFG compared to an average
 decrease of over 14 percent here. Organic emissions resemble the fuel more with the two-
 stroke engines because combustion is less complete resulting in high unburned fuel
 fractions being emitted. Figure 2 shows the carbon number distributions for HC
 emissions from both engine types burning RFA and from the RFA fuel itself.  The
 distributions illustrate the particularly close similarity between RFA composition and
 composition of the two-stroke HC emissions.

 Lower organic and CO emission rates obtained with RFG team with the lower specific
 reactivities to produce lower ozone potential emission rates. On average, ozone potential
 emission are reduced over 16 percent by using RFG.
Figure 2. Carbon number distributions for 4-stroke, 2-stroke, and RFA hydrocarbons.

-------
 CONCLUSIONS

 The study examines emissions from three handheld, two-stroke engines. Results are from
 a rather limited test matrix that may not lead to representative emissions from such
 engines nationwide. However, composition of the emissions, which is similar from
 engine-to-engine despite large differences in emission rates, probably can be assumed to
 be approximately the same for other small, two-stroke engines burning similar fuels.
 With this understanding, the following conclusions drawn from the study and relative to
 the two-stroke engines tested are:

    Specific reactivity of organic emissions was significantly decreased when fuel was
    switched from RFA to RFG.
    RFG was effective in lowering the ozone potential emissions because of lower
    specific reactivity and lower HC and CO emission rates.
    Emission rates are extremely sensitive to air-fuel ratio which can change as an engine
    is operated for extended periods.
    Particulate matter emission rates tend to decrease with increases in air-fuel ratio.
    RFG resulted in sharply lower benzene, but somewhat higher formaldehyde emission
    rates.
    Gaseous toxic emissions are dominated by benzene which follows trends for total HC
    emission rates.
    Particles emitted are predominately less than 2.5 microns in diameter.
    Composition of organic emissions resembles the composition of the fuel more than
    with four-stroke engines.

Acknowledgments

We thank the members of our research team who conducted the tests and analyzed
samples. Special thanks go to Jerry Faircloth, Versal Mason, and Mike Kirby for their
hard work in the engine laboratory. Colleen Loomis, Christy Pack, and Angela Farinacci
are  also acknowledged for their diligent work with the gas and liquid chromatographic
equipment.

Disclaimer

The information in this paper has been funded by the Environmental Protection Agency.
It has been subjected to Agency review and approved for publication. Mention of trade
names or commercial products does not constitute endorsement or recommendation for
use.

-------
References

1. Nonroad Engine and Vehicle Emission Study, EPA-21A-2001,  Office of Mobile
   Sources, U.S. Environmental Protection Agency , November, 1991 (report available
   on request to the author).

2. White, J.J.;  Carroll, J.N.; Hare, C.T.;  Lourenco, J.G.; SAE Paper 910560, Detroit,
   February 1991.

3. Sun, X.; Brereton, K.; Morrison, K.; Patterson, D.; SAE Paper 952080, Milwaukee,
   September 1995.

4. Brereton, G.J.; Morrison, K.; Chishty, H.A.; Schwartz, G.; Patterson, D.J.; SAE Paper
   941806, Milwaukee, September 1994.

5.  Hare, C.T.; White, J.J.; SAE Paper 911222, Yokohama, Japan, October 1991.

6.  Hare, C.T.; Carroll, J.N.; SAE Paper 931540, Pisa, Italy, December 1993.

7. Hare, C.T.; White, J.J.; SAE Paper 901595, Milwaukee, September 1990.

8. Federal Register, 40 CFR Parts 9 and 90, Final Rule, 60(127), July  3, 1995.

9. Siegl, W.O.; Richert, J.F.O.; Jensen, T.E.; Schuetzle, D.; Swarin, S.J.; Loo, J.F.;
   Prostak, A.; Nagy, D.; Schlenker, A.M.; SAE Paper No. 930142, SP-1000, October,
   1993.

10. Carter, W.P.L.; Atkinson, R.J.; Environ. ScL Technol, 21, 864-880., 1987.

11. Lowi, Alvin Jr.; Carter, W.P.L.; SAE Paper No. 900710, March 1990.

12. Gabele, P.A.; J. Air & Waste Manage., 1997. 47:945-952.

-------
  NEiL-RTP-HEASD-oo-049    TECHNICAL REPORT DATA
 1. Report No.
     EPA/600/A-00/012
                   3. Recipient's Accession No.
 4. Title and Subtitle
 Characterization of Emissions from Hand-held Two-stroke Engines
                   5. Report Date
                                                                           6.  Performing Organization Code
 7. Author(s)
 P.A. Gabele
                   8. Performing Organization
                   Report No.
 9.Performing Organization Name and Address
 Source Apportionment and Characterization
                   10. Program Element No.
                                                                           11. Contract/Grant No.
 12.Sponsoring Agency Name and Address
 National Exposure Research Laboratory
 Office of Research and Development
 U.S. Environmental Protection Agency
 Research Triangle Park, NC 27711
                   13. Type of Report and Period
                   Covered
                   14.Sponsoring Agency Code
 15. Supplementary Notes
16. Abstract
Despite their extremely high organic and participate matter emission rates, two-stroke engines remain among the least studied of engine types. Such
studies are rare because they are costly to perform. Results reported in this paper were obtained using a facility that shares exhaust emission sampling and
analytical systems with an adjacent automotive emissions test laboratory. This enabled a comprehensive examination of emissions from three hand-held,
two-stroke, engines. In addition to the determination of routine regulated pollutant emission rates, organic emissions are speciated and paniculate matter
emission rates are measured. Each engine was tested using two fuels: 1990 Baseline Gasoline and a reformulated gasoline. One engine was run for an
extended period to assess durability effects on exhaust emissions, and another was operated at a leaner air-fuel ratio setting to examine enleanment
effects. In the durability testing, organic and particulate matter emissions actually appear to decrease slightly with increased operating time. In the
enleanment test, organic and carbon monoxide emission rates decreased significantly, as expected, with the increase in air-fuel ratio. Large decreases in
particulate matter also ensued. Fuel selection strongly influenced composition of organic emissions. With reformulated gasoline, aggregate toxic and
ozone potential emissions were consistently lower compared to emissions with the Baseline fuel.
                             17.  KEY WORDS AND DOCUMENT ANALYSIS
A. Descriptors
Two-stroke engine
Small engine emissions
Toxic emissions
Ozone precursors
B.  Identifiers / Open Ended
Terms
C.  COSATI
18. Distribution Statement
RELEASE TO PUBLIC
19. Security Class (This
Report)
UNCLASSIFIED
                                                        20. Security Class (This
                                                        Page)
                                                        UNCLASSIFIED
21. No. of Pages
                               22.  Price

-------