The In-Plume Emission Test Stand: An Instrument Platformfor the Real-Time Characterization of Fuel-Based CombustionEmissions

Nicholas J. Nussbaum, Dongzi Zhu, Hampden D. Kuhns, Claudio Mazzoleni,M.-C. Oliver Chang, Hans Moosmuller, and John G. WatsonDivision of Atmospheric Sciences, Desert Research Institute, Reno, NV

ABSTRACTThe In-Plume Emission Test Stand (IPETS) characterizesgaseous and particulate matter (PM) emissions from com-bustion sources in real time. Carbon dioxide (CO2), car-bon monoxide (CO), nitric oxide (NO), nitrogen dioxide(NO2), and other gases are quantified with a closed-pathFourier transform infrared spectrometer (FTIR). Particleconcentrations, chemical composition, and other particleproperties are characterized with an electrical low-pressure impactor (ELPI), a light-scattering particle detec-tor, an optical particle counter, and filter samples amena-ble to different laboratory analysis. IPETS measurementsof fuel-based emission factors for a diesel generator arecompared with those from a Mobile Emissions Laboratory(MEL). IPETS emission factors ranged from 0.3 to 11.8, 0.2to 3.7, and 22.2 to 32.8 g/kg fuel for CO, NO2, and NO,respectively. IPETS PM emission factors ranged from 0.4to 1.4, 0.3 to 1.8, 0.3 to 2.2, and 1 to 3.4 g/kg fuel for filter,photoacoustic, nephelometer, and impactor measure-ments, respectively. Observed linear regression statisticsfor IPETS versus MEL concentrations were as follows: COslope � 1.1, r2 � 0.99; NO slope � 1.1, r2 � 0.92; and NO2

slope � 0.8, r2 � 0.96. IPETS versus MEL PM regressionstatistics were: filter slope � 1.3, r2 � 0.80; ELPI slope �1.7, r2 � 0.87; light-scattering slope � 2.7, r2 � 0.92; andphotoacoustic slope � 2.1, r2 � 0.91. Lower temperaturesin the dilution air (�25 °C for IPETS vs. �50 °C for MEL)may result in greater condensation of semi-volatile com-pounds on existing particles, thereby explaining the 30%difference for filters. The other PM measurement devicesare highly correlated with the filter, but their factory-default PM calibration factors do not represent the size

and optical properties of diesel exhaust. They must benormalized to a simultaneous filter measurement.

INTRODUCTIONGas and particle emissions from power generators, gaso-line engines, and diesel engines affect attainment of airquality standards,1 human health,2 visibility,3 and cli-mate.4 Real-world emission estimates from these sourcesare needed to improve air quality management strategies.

Engine certification emission factors (EFs) are deter-mined from chassis or engine dynamometer tests follow-ing test protocols that vary vehicle operating conditionsover time.5 Only a few vehicles or engines can be testedon dynamometers because of the time and effort required.Other methods can determine EFs for large vehicle fleets.Tunnel studies yield fleet EFs averaged over many vehi-cles, but these tests are unable to determine individualvehicle EFs and their distribution across vehicle type andage.6 Vehicle emission remote sensing systems (VERSS)and in-plume systems can measure fuel-based EFs formany individual vehicles over a wide range of real-worldoperating conditions.7,8 Fuel-based EFs are calculatedas9,10

EFP � CMFfuel

CP

CCO2� MC

MCO2� � CCO� MC

MCO� (1)

where EFP is the EF of pollutant P in g pollutant/g fuel.CMFfuel is the carbon mass fraction of the fuel, typically85–88% for gasoline and diesel fuels, and 45–50% forwood fuels. CP is the mass concentration of pollutant P ing/m3. CCO2 and CCO are the mass concentrations (g/m3)of species carbon dioxide (CO2) and carbon monoxide(CO), and MCO2 and MCO are the molecular weights ofCO2 and CO in g/mol.

VERSS9 measure instantaneous fuel-based gas andparticulate matter (PM) EFs by passing a light beam (ul-traviolet [UV], visible, and/or infrared [IR]) through anexhaust plume and measuring its scattering and/or ex-tinction.

In-plume techniques draw a diluted portion of ex-haust through a broad array of instrumentation, includ-ing the carbon content as expressed in eq 1. In-plumesystems have been mounted in aircraft to monitor emis-sions from power plants and industrial point sources,11,12

IMPLICATIONSReal-world emission rates and compositions often differfrom certification emissions because of different equip-ment, fuels, operating cycles, and test methods. Portabletest methods that take advantage of natural dilution allowfuel-based emission factors to be developed and applied toemission inventories that can be used for source modeling,receptor modeling, and air quality management. The abilityto rapidly obtain emission rates for many different combus-tion sources allows emission rate distributions, rather thansimple averages, to be used in emission models.

TECHNICAL PAPER ISSN:1047-3289 J. Air & Waste Manage. Assoc. 59:1437–1445DOI:10.3155/1047-3289.59.12.1437Copyright 2009 Air & Waste Management Association

Volume 59 December 2009 Journal of the Air & Waste Management Association 1437

biomass burning,13,14 and other aircraft.15,16 Measure-ments over the ocean have been used to distinguish emis-sions from ships from the ambient background.17 Each ofthese applications focused on identifying emissions of alarge source above a well-dispersed and homogeneousbackground.

Several mobile emission laboratories (MELs) havebeen assembled and applied to in-plume measure-ments.18–21 The Aerodyne MEL combines a tunable IRlaser differential absorption spectrometers (TILDAS), aproton transfer reaction mass spectrometer (PTR-MS), andan aerosol mass spectrometer (AMS) to sample gas and PMemissions from on-road vehicles.8,22,23 This unit was de-ployed in Mexico City and New York City in stationaryand/or chasing modes.24–28 In a stationary mode, the MELcharacterizes emissions from several vehicles driving byand calculates fuel-based EFs representative of the on-roadfleet. It can also be applied in a chase mode, in which theplume from a single vehicle is followed through a typicalroadway driving cycle that includes acceleration, cruising,and coasting. Pirjola et al.29 assembled a similar MEL withan electrical low-pressure impactor (ELPI); a scanning mo-bility particle sizer (SMPS), an ultrafine condensation par-ticle counter (CPC); and CO, CO2, and NOx analyzers toobtain real-world in-plume EFs along Finland’s highways.

Hansen and Rosen30 used a simpler roadside in-plume approach for black carbon (BC) emissions, collo-cating IR absorption CO2 aethalometer BC monitorsdownwind of a steeply graded road. The ratio of aerosolBC to CO2 was calculated for the plume from each vehiclepassing sampling inlets. Real-world BC EFs ranged over afactor of 100, from 0.004 to 1 g BC/kg carbon in the fuel.Such large variations in EFs, often accompanied by non-normal distributions, make it necessary to measure EFs ofmany units for each source type to accurately characterizefleet emission distributions.

The In-Plume Emissions Test Stand (IPETS) describedhere builds on the roadside approach of Hansen andRosen by adding a larger array of observables, but at a

fraction of the cost of assembling and supporting a full-fledged MEL. A comparison of measurements with one ofthe MELs is also reported.

MEASUREMENTS METHODS AND CALIBRATIONThe IPETS configuration is illustrated in Figure 1 withinstrument descriptions summarized in Table 1. The basecomponents are an FTIR spectrometer (Illuminator series,Midac) and two LI-840 CO2/H2O gas analyzers (LI-CORBiosciences) for gas analysis. An ELPI (Dekati), two Dust-Traks (model 8520, TSI) and a photoacoustic spectrometerdesigned by Arnott et al.31 were used for the real-timemeasurement of PM size distribution and mass concen-trations. Integrated filter samples are used for mass andsource profile analysis and to determine sample-specificcalibration factors for the continuous PM instruments.

Sample air is drawn into the IPETS manifold through2-cm inner diameter, electrically conductive silicone tub-ing (TSI) at approximately 220 L/min. Mass flow is mea-sured via factory-calibrated digital mass flowmeters (TSI4000 series) with an accuracy of �2%. The tubing lengthranges from 2 to 4.7 m, depending on the source type andfield access. Residence time in the tube is less than 2 sec.Although the large diameter conducting sample line andshort residence times minimize PM loss due to electro-static attraction and deposition, a portion of reactive spe-cies such as ammonia (NH3) might experience attenua-tion before reaching the measurement instruments. Theinlet can be positioned to reach elevated plumes such asthose from vertical exhaust pipes of heavy-duty dieselvehicles. The inlet can be placed on the road surface,protected by rubberized cable protectors, for samplingfrom low-level passenger vehicle exhausts.

Components are mounted on three handcarts andcan be operated from within a cargo van or positionednear the exhaust duct. Measurements are made at near-ambient temperatures to represent real-world PM concen-trations soon after ambient air dilution when nucleation

Figure 1. IPETS instrument layout.

Nussbaum et al.

1438 Journal of the Air & Waste Management Association Volume 59 December 2009

and condensation are expected. Temperatures are moni-tored by thermocouples exposed to ambient air and to thediluted plume at the FTIR inlet.

FTIR SpectrometerFTIR spectroscopy has been used previously to character-ize several gas concentrations in combustion emis-sions.32,33 The IPETS FTIR spectrometer draws 60 L/min todetect CO2, CO, NH3, nitric oxide (NO), water vapor(H2O), butane (C4H10), hexane (C6H14), ethylene (C2H4),nitrogen dioxide (NO2), and sulfur dioxide (SO2) concen-trations It scans through wave numbers of 1100–6500cm�1 every 1.5 sec with a spectral resolution of 0.5 cm�1

over a 10-m optical path folded within a 2-L analyticalcell. Spectral regions used for the measurement of indi-vidual gas concentrations, detection limits, and calibra-tion concentration ranges are reported in Table 2. Inter-nally mounted pressure and temperature sensors accountfor changes in air density. Calibration spectra were pre-pared using U.S. Environmental Protection Agency (EPA)-certified gases diluted with ultrapure nitrogen by an En-vironics 2020 computerized gas dilution system. Speciesconcentrations are determined using a classical least-squares (CLS) fitting technique as implemented in Auto-quant Pro software (MIDAC Corporation). Before mea-surements, the cell is purged with ultrapure nitrogen toreduce measurement interference from background air. ATeflon filter is inserted in-line to prevent contaminationof the sampling cell by particles.

NO and NO2 have many IR absorption regions coin-cident with those of H2O, but de Castro et al.33 havedeveloped methods that minimize this interference bylimiting spectral regions to those with minimal overlap.Figure 2 shows an example FTIR absorption spectrumfrom gasoline engine exhaust with NO (50 parts per mil-lion [ppm]) and H2O (5000 ppm) reference spectra. The1873- to 1876.5-cm�1 and 1898.5- to 1901.5-cm�1 re-gions show peaks for NO but not for H2O. Figure 3 showsa NO2 example with nonoverlapping regions at 1584–1588.5 cm�1 and 1597.5–1600 cm�1.

Nondispersive IR (NDIR) CO2/H2O AnalyzerThe CO2 concentration is crucial for fuel-based EFs, so aredundant Licor LI-840 CO2/H2O NDIR gas analyzer34

complements the FTIR CO2 measurement, drawing a par-allel airstream at 1 L/min with 1-sec sample durations.CO2 is quantified by the 2347-cm�1 absorption line andH2O is quantified by the 3853-cm�1 absorption line. Anin-line Teflon filter in the sample line reduces contami-nation of the cell by particles. .

ELPIThe ELPI draws 10 L/min through a unipolar coronacharger, which imparts a positive charge to the incomingaerosol. The charged particles are separated by impactioninto 12 size fractions. Impaction surfaces are isolated withTeflon supports and the accumulating charge on each ofthe substrates is measured by an array of electrometers.The measured current on each of the stages is propor-tional to the number of particles deposited on thestage.35,36 The sintered impaction plates are oiled37 tominimize particle re-entrainment and have 50% cutpoints at 0.024, 0.03, 0.056, 0.10, 0.22, 0.32, 0.59, 0.91,1.5, 2.5, 3.8, and 6.4 �m. The final stage does not effi-ciently capture particles less than 7 nm in diameter. Thefinal six stages are used for analysis, corresponding toparticles below 0.32 �m, because most engine exhaust PMemissions are in this size range.

DustTrak Aerosol MonitorThe Model 8520 DustTrak38 measures near-perpendicularparticle scattering of a 780-nm laser diode beam. A coarse(PM10) or fine (PM2.5) PM inlet is installed upstream of itsanalytical chamber to limit the size of measured aerosolparticles. There is little difference between the PM10 andPM2.5 DustTrak readings for engine exhaust PM, consis-tent with its small particle sizes. The DustTrak has a flowrate of 1.7 L/min and is factory-calibrated to the respirablefraction of standard ISO 12103-1 A1 test dust (previously

Table 1. Instrumentation of the IPETS.

Instrument Manufacturer Measurement MethodResponse Time

(sec)

Midac I-Series FTIR Midac, Costa Mesa, CA Molecular gas speciesconcentration

Dispersive IR with a mercury-cadmium-tellurium (MCT) liquid nitrogen-cooled detector.

1.5

Dekati ELPI (10 L/min) ELPI, Dekati, Finland Aerodynamic number sizedistribution of particles

Current dissipation arising fromdeposition of charged particles toimpactor substrates

5

Model 8520 DustTrak TSI, Shoreline, MN Particle mass 780-nm laser light scattering of particlestream at 90°

1

Nucleopore filter sampler Mass and chemical composition ofparticles and gases

Collection and analysis of exposedfilters

�1000

Model 4043 mass flow meters TSI, Shoreline, MN Mass flow through filter Hot wire anemometer �1Photoacoustic spectrometer Desert Research Institute,

Reno, NVEC concentration Acoustic pressure wave generated

through absorption induced gasexpansion

1

Li-Cor 840 CO2/H2O gas spectrometer LI-COR Biosciences,Lincoln, NE

CO2 (ppm) and H2O (parts perthousand)

NDIR 1

Nussbaum et al.

Volume 59 December 2009 Journal of the Air & Waste Management Association 1439

called “Arizona road dust”). The DustTrak reading is usu-ally calibrated for a specific application through linearregression with collocated filter measurements.39

Photoacoustic SpectrometerThe photoacoustic spectrometer40 measures aerosollight absorption at 532 nm by illuminating particles

with a modulated laser beam. This heats the particlesand creates an acoustic pressure wave at the modula-tion frequency. A sensitive microphone monitors theacoustic signal and its energy is proportional to theabsorbed light. Light absorption is related to elementalcarbon (EC) by mass extinction coefficients derivedfrom collocated EC measurements.31

Table 2. FTIR spectrometer spectral analysis regions, calibration ranges, and detection limits.

Species

Reference Region (cm�1)

Detection Limita (ppm) Calibration Range (ppm)Lower �1 Lower �2

CO2 723.0 750.0 12 100 4730CO 2133.5 2142.0 0.2 1.0 1005NH3 955.5 976.0 0.06 1.0 110NO 1873.0 1878.5 1 0.2 200

1880.5 1884.01898.5 1901.51926.0 1932.01934.5 1940.0

H2O 1200.0 1300.0 60 5.0 5294C4H10 3041.5 2825.5 0.05 1.0 100C6H14 3030.0 2818.0 0.2 0.2 200C2H4 958.0 936.5 0.12 0.5 20NO2 1584.0 1588.5 0.4 0.2 20

1597.5 1600.01604.0 1606.01610.5 1614.0

SO2 1112.5 1120.5 0.54 1.0 1001123.5 1134.01138.5 1148.01153.5 1164.01166.5 1172.51176.5 1185.01188.0 1197.01200.0 1209.01227.0 1236.0

Notes: aDetection limits were calculated as twice the standard deviation of the instrument response during a zero measurement (i.e., nitrogen-purged cell) overa period of 30 min.

Figure 2. Region of IR absorbance spectra for NO analysis. Figure 3. Region of IR absorbance spectra for NO2 analysis.

Nussbaum et al.

1440 Journal of the Air & Waste Management Association Volume 59 December 2009

PM Filter SystemParticles are sampled through Bendix 240 cyclones41 op-erating at 113 L/min (PM2.5) onto filters for laboratorychemical analysis,42,43 which produces source profiles forreceptor modeling.44,45 A typical filter pack consists of (1)a Teflon filter for mass by gravimetry and elements byX-ray fluorescence, followed by a quartz fiber backup fil-ter, which allows for the evaluation of potential organiccarbon (OC) sampling artifacts46; (2) a quartz fiber filterfor water-soluble ions by ion chromatography and OCand EC by thermal/optical reflectance,47–49 followed by acitric-acid-impregnated cellulose fiber filter that capturesgaseous NH3; (3) a quartz fiber filter for organic sourcemarkers50 by thermal desorption gas chromatography(GC) mass spectroscopy (MS), followed by a potassium-carbonate-coated cellulose fiber filter that captures SO2;and (4) a Nuclepore etched polycarbonate membrane fil-ter for particle morphology by scanning electron micros-copy. Filter pack configurations can be changed depend-ing on the application, and the cyclone can also beoperated at a lower flow rate to obtain a PM10 cut point.Flow through each filter is accurately monitored with adigital mass flow meter. When the filter sampler is notneeded, a bypass flow with the same flow rate as the filtersampling system is used to maintain a constant total flowrate through the system.

Data AcquisitionData are logged in real time through a serial port to alaptop Ethernet server mounted on one of the handcarts.The FTIR spectrometer communicates with the laptopthrough its A/D PCMCIA card. The ELPI and CO2/H2Odetectors use data acquisition software provided by themanufacturers. Flowmeters and the DustTrak use Labviewsoftware (National Instruments, Inc.). The FTIR spectrom-eter uses Autoquant Pro version 1.0.104 to log IR spectraand quantify gas concentrations. The data acquisitionsystem assigns a common time stamp to all measure-ments to ensure that data are synchronized at a frequencyof 1 Hz. The use of real-time displays increases data recov-ery in the field because the operator can monitor thestatus of each instrument from a single location.

Instrument CalibrationCO2 gas calibration standards ranging from 100 to 3000ppm in air are generated by an Environics 2020 gas mixer.Samples were taken approximately 1.5 m downwind of apassenger car exhaust pipe with a revved engine to pro-vide transient verified response times less than 1.5 sec.CO2 concentrations agreed for the simultaneous FTIR andNDIR measurements over a range of 350–4730 ppm forthe FTIR and 350–3000 ppm for the NDIR (regressionslope � 0.98).

The ELPI particle charging efficiency depends on par-ticle mobility diameter dm, but the size classification isbased on aerodynamic diameter da. The effective particledensity �e establishes the relationship between dm andda.51

�eCcdmdm2 � �0Ccdada

2 (2)

where Cc is the Cunningham correction coefficient and �0

the primary particle density, 2 g/m3.52 Approximately80–90% of diesel particulate matter (DPM) mass is foundin the 30- to 500-nm range.53 ELPI software converts thecurrent measured on each stage to PM mass assuming thatparticles are spherical with unit density and that dm � da.DPM particles are fractal-like agglomerates of approxi-mately spherical 10- to 30-nm diameter primary particleswith an effective mass density decreasing with decreasingparticle size.52,54

Pagels et al.36 demonstrated that charge-bearingnanoparticles deposit on the upper ELPI stages, therebycreating biases in these results. Zhu et al.55 implementedan ELPI fitting procedure52 that accounts for these arti-facts by iteratively adjusting the particle size and numberconcentrations for a bimodal distribution to match thecurrent deposited on each impactor stage. This fittingalgorithm is also used for IPETS ELPI measurements. Mar-icq et al. and Zhu et al. found fitted ELPI PM massconcentrations within �20% of corresponding filtermeasurements.

IPETS/MEL COMPARISONThe IPETS sampled emissions simultaneously with theUniversity of California–Riverside (UCR) MEL56 from astationary diesel engine in 2004 to evaluate instrumentcomparability. The UCR MEL measures regulated gaseousemissions following engine certification procedures.57,58

Its function differs from the IPETS, which is intended toquantify real-world emissions at ambient temperaturesand dilution ratios. For the MEL, the entire diesel exhauststream enters a primary constant volume sampling (CVS)dilution tunnel (Model ESU 7000, Horiba) at 142,000L/min (4000 std ft3/min). A small fraction of the dilutedexhaust is extracted from the primary dilution tunnel forcontinuous CO, CO2, methane (CH4), NOx, and totalhydrocarbon (THC) measurements. A secondary dilutiontunnel further dilutes the exhaust stream on a Teflonmembrane filter for PM collection for mass emissions. Theremainder of the diluted exhaust in the primary tunnel isvented through a bypass fan and exhaust pipe into ambi-ent air on the MEL roof; IPETS sampled from this exhaustplume. A year-2000 Kamatzu SA6D125E-2 engine (Japan)with 316 service hours, power output rated at 303 kW,and 11-L displacement was operated in a 250-kW (actualelectrical output) DEWYO model no. DF-3300K powergenerator. The test protocol followed the nonroad com-pression ignition (CI) engine certification protocol con-sisting of time-weighted sampling for 5 min each at 100,75, 50, 25, and 10% of maximum power.58 Test fuelsconsisted of military JP-8 and ultralow sulfur diesel(ULSD) fuels with average sulfur concentration of 476and 2 parts per million by weight (ppmw), respectively.Each test fuel was directly fed to the engine from thefuel drum. After fuel switches, the engine was operatedwith the new test fuel for 30–60 min at 100% load toensure complete removal of the previous test fuel fromthe engine and fuel line.

Filter pack sample durations ranged from 5 to 15 min,depending on PM concentrations estimated from the real-time DustTrak readings. Filter field blanks were collectedat the end of each test day and were analyzed to subtract

Nussbaum et al.

Volume 59 December 2009 Journal of the Air & Waste Management Association 1441

blank concentrations. These were low for all except OCbecause of adsorption of some organic vapors by thequartz fiber.46

Parallel DustTrak with PM10 and PM2.5 inlets showedlittle difference (slope � 0.99, r2 � 0.99) in these tests,consistent with other size-related diesel exhaust studies.With TSI’s Arizona road dust calibrations, the DustTrak

PM2.5 mass EFs exceeded photoacoustic spectrometer andELPI PM2.5 mass EFs by 30 and 40%, respectively (slope �1.3, r2 � 0.98 and slope � 1.4, r2 � 0.85) (Figure 4).Photoacoustic spectrometer PM exceeded ELPI PM2.5 EFsby 10% (slope � 1.1, r2 � 0.89), indicating that the defaultconversion coefficient may not apply in this case. Thesimultaneous filter packs with their PM2.5 mass and ECmeasurements allow a relationship to be derived for eachIPETS test PM2.5.59 Comparison of real-time instrumenta-tion with IPETS filters resulted in 80% overestimation offilter mass EF by DustTrak (slope � 1.8, r2 � 0.83) and40% overestimation by the photoacoustic spectrometer(slope � 1.4, r2 � 0.82).

Figure 4. IPETS PM instrument intercomparison for (a) DustTrakPM2.5 vs. photoacoustic, (b) photoacoustic spectrometer vs. ELPI,(c) DustTrak PM2.5 vs. DustTrak PM10, and (d) DustTrak PM2.5 vs.ELPI. All values are reported as g/kg fuel.

Figure 5. IPETS vs. MEL gas measurement comparison for (a)DustTrak PM2.5 vs. MEL, (b) photoacoustic vs. MEL, (c) ELPI vs.MEL, and (d) IPETS filters vs. MEL. All values are reported as g/kgfuel.

Figure 6. IPETS vs. MEL PM measurement comparison for (a)CO, (b) NO, and (c) NO2. All values are reported as g/kg fuel.

Figure 7. OC/EC from IPETS filters. Carbon evolved is categorizedas either OC1, in which the semi-volatile organic fraction is evolvedas the filter is heated to 120 °C; or OC2, OC3, or OC4, in which thenonvolatile organic fraction is evolved as the filter is heated to 250,450, or 550 °C, respectively. EC is categorized into EC1, EC2, orEC3, in which carbon is evolved as the filter is heated with oxygen to550, 700, and 800 °C, respectively.

Nussbaum et al.

1442 Journal of the Air & Waste Management Association Volume 59 December 2009

The IPETS and MEL EFs are similar for the gaseouspollutants. CO EFs were highly correlated with a near-unity regression slope (slope � 1.1, r2 � 0.99) as were NOand NO2 EFs (slope � 1.1, r2 � 0.92 for NO; slope � 0.8,r2 � 0.96 for NO2), as shown in Figure 5. The lower IPETSNO2 might be caused by many NO2 concentrations in thediluted plume that were below the FTIR detection limit.

IPETS PM2.5 EFs derived from the Teflon filter massare 1.3 times those of the MEL. DustTrak PM2.5 EFs arehighly correlated with those of the MEL, but they are 2.7times higher (slope � 2.7, r2 � 0.92) (Figure 6). PM EFsfrom the photoacoustic spectrometer and PM2.5 EFs fromthe ELPI are also highly correlated with the MEL EFs, butthey are 2.1 and 1.7 higher, respectively (slope � 2.1, r2 �0.91 and slope � 1.7, r2 � 0.87, respectively). These dif-ferences indicate that the default calibration factors arenot appropriate for diesel exhaust and that they should benormalized to the Teflon filter mass measurement. Inter-cepts for the regressions in Figure 6 are all within 3 stan-dard errors of zero.

IPETS filters present the most direct comparison ofPM EFs with MEL. The observed difference is reasonable,given that the MEL dilutes the exhaust to approximately47 °C, whereas the IPETS samples at ambient tempera-tures that ranged from 22 to 27 °C. Lower ambient tem-peratures favor the formation of nucleation mode parti-cles, primarily composed of water/sulfuric acid withcondensation of low and semi-volatile organics onto thenucleation phase.60,61 Figure 7 compares thermal carbonfractions as well as OC and EC for the tests. OC averaged32% of total carbon, with the semi-volatile fraction(�120 °C) averaging approximately 9% of total carbon. Ifthis fraction represents condensable material, then theIPETS filters still measures an approximately 20% higherEF when compared with the MEL.

IPETS precision is indicated by the replicate tests sum-marized in Table 3. Overall EFs for a combination ofoperating loads were calculated from the measurementsat individual loads with weighting factors per Code ofFederal Regulations Part 89, Title 40. The coefficients ofvariance are 0.60, 1.4, and 20% for CO, NO, and NO2,respectively. The average NO2 concentration was 1.1ppm, whereas the FTIR detection limit, defined as twicethe standard deviation of the average instrument re-sponse during a zero measurement using ultrapure nitro-gen, was 0.4 ppm. Coefficients of variation for thePM2.5 EFs are 2.3, 1.7, 8.7, and 2.2% for the DustTrak,photoacoustic spectrometer, ELPI, and Teflon filters,respectively.

SUMMARYAn IPETS was developed and calibrated to quantify com-bustion emissions under real-world operating conditions.Fuel-based EF rates for criteria pollutants and source pro-files for receptor modeling can be obtained. IPETS is moreportable than MELs, which are housed in large trucks orbuses. IPETS tubing, pumps, and data collection systemsare mounted on three handcarts and can be moved closeto various combustion source exhaust ducts.

Comparisons with a MEL used for diesel engine cer-tification tests demonstrated comparable measurementsfor CO, NO, and NO2. IPETS PM2.5 EFs were approxi-mately 30% higher than those from the MEL using filtermeasurements, about one-third of which could be attrib-uted to condensable OC at the lower IPETS plume tem-peratures. PM2.5 EFs derived from a DustTrak nephelom-eter and ELPI impactor measured 2.7 and 1.7 those of theMEL, respectively, indicating that factory calibrations forthese instruments are not appropriate for diesel exhaustand that their signals should be normalized to the collo-cated filter mass.

ACKNOWLEDGMENTSThe authors thank the UCR College of Engineering, Cen-ter for Environmental Research and Technology for theirassistance as well as the Federal Transit Administration(contract 26-70030) and the Strategic Environmental Re-search and Development Program (contract CP-1336) forfunding of this project.

REFERENCES1. Bachmann, J.D. Will the Circle be Unbroken: a History of the U.S.

National Ambient Air Quality Standards—2007 Critical Review; J. Air& Waste Manage. Assoc. 2007, 57, 652-697; doi: 10.3155/1047-3289.57.6.652.

2. Pope, C.A., III; Dockery, D.W. Critical Review: Health Effects of FineParticulate Air Pollution: Lines that Connect; J. Air & Waste Manage.Assoc. 2006, 56, 709-742.

3. Watson, J.G. Visibility: Science and Regulation; J. Air & Waste Manage.Assoc. 2002, 52, 628-713.

4. MacCracken, M.C. Critical Review: Prospects for Future ClimateChange and the Reasons for Early Action; J. Air & Waste Manage. Assoc.2008, 58, 735-786; doi: 10.3155/1047-3289.58.6.735.

5. Diesel Net Emission Test Cycles: Summary of Worldwide Engine and Vehi-cle Test Cycles; Ecopoint: Mississauga, Ontario, Canada, 2008; availableat http://www.dieselnet.com/standards/cycles (accessed 2009).

6. Pierson, W.R.; Gertler, A.W.; Robinson, N.F.; Sagebiel, Bach, J.C.; Zie-linska, B.; Bishop, G.A.; Stedman, D.H.; Zweidinger, R.B.; Ray, W.D.Real-World Automotive Emissions: Summary of Studies in the FortMcHenry and Tuscarora Mountain Tunnels; Atmos. Environ. 1989, 30,2233-2256

7. Stedman, D.H. Automobile Carbon Monoxide Emissions; Environ. Sci.Technol. 1989, 23, 147-149.

8. Herndon, S.C.; Jayne, J.T.; Zahniser, M.S.; Worsnop, D.R.; Knighton,B.; Alwine, E.; Lamb, B.K.; Zavala, M.; Nelson, D.D.; McManus, J.B.;

Table 3. IPETS replicate measurements.

CO EF(g/kg)

NO EF(g/kg)

NO2 EF(g/kg)

DustTrak EF(g/kg)

Photoacoustic EF(g/kg)

ELPI EF(g/kg)

Filter EF(g/kg)

Replicate 1 2.60 18.6 0.94 0.53 0.49 0.93 0.39Replicate 2 2.62 18.2 0.71 0.55 0.50 0.82 0.40Standard deviation 0.02 0.3 0.17 0.01 0.01 0.08 0.01Average 2.6 18.4 0.83 0.54 0.50 0.88 0.40Coefficient of variance 0.60% 1.4% 20% 2.3% 1.7% 8.7% 2.2%

Nussbaum et al.

Volume 59 December 2009 Journal of the Air & Waste Management Association 1443

Shorter, J.H.; Canagaratna, M.R.; Onasch, T.B.; Kolb, C.E. Character-ization of Urban Pollutant Emission Fluxes and Ambient Concentra-tion Distributions Using a Mobile Laboratory with Rapid ResponseInstrumentation; Faraday Discuss. 2005, 130, 327-339.

9. Moosmuller, H.; Mazzoleni, C.; Barber, P.W.; Kuhns, H.D.; Keislar,R.E.; Watson, J.G. On-Road Measurement of Automotive Particle Emis-sions by Ultraviolet Lidar and Transmissometer: Instrument; Environ.Sci. Technol. 2003, 37, 4971-4978.

10. Fraser, M.P.; Cass, G.R.; Simoneit, B.R.T. Gas-Phase and Particle-PhaseOrganic Compounds Emitted from Motor Vehicle Traffic in a LosAngeles Roadway Tunnel; Environ. Sci. Technol. 1998, 32, 2051-2060.

11. Neuman, J.A.; Huey, L.G.; Dissly, R.W.; Fehsenfeld, F.C.; Flocke, F;Holecek, J.C.; Holloway, J.S.; Hubler, G.; Jakoubek, R.; Nicks, D.K.;Parrish, D.D.; Ryerson, T.B.; Sueper, D.T.; Weinheimer, A.J. Fast-Response Airborne In Situ Measurements of HNO3 during the Texas2000 Air Quality Study; J. Geophys. Res. 2002, 107, 4436; doi: 10.1029/2001JD001437.

12. Gillani, N.V.; Meagher, J.F.; Valente, R.J.; Imhoff, R.E.; Tanner, R.L.;Luria, M. Relative Production of Ozone and Nitrates in Urban andRural Power Plant Plumes 1. Composite Results Based on Data from 10Field Measurement Days; J. Geophys. Res. 1998, 103, 22,593-22,615.

13. Goode, J.G.; Yokelson, R.J.; Ward, D.E.; Susott, R.A.; Babbitt, R.E.;Davies, M.A.; Hao, W.M. Measurements of Excess O3, CO2, CO, CH4,C2H4, C2H2, HCN, NO, NH3, HCOOH, CH3COOH, HCHO, andCH3OH in 1997 Alaskan Biomass Burning Plumes by Airborne FourierTransform Infrared Spectroscopy (AFTIR); J. Geophys. Res. 2000, 105,22,147-22,166.

14. Yokelson, R.J.; Goode, J.G.; Ward, D.E.; Susott, R.A.; Babbitt, R.E.;Wade, D.D.; Bertschi, I.; Griffith, DWT; Hao, W.M. Emissions of Form-aldehyde, Acetic Acid, Methanol, and Other Trace Gases from BiomassFires in North Carolina Measured by Airborne Fourier Transform In-frared Spectroscopy; J. Geophys. Res. 1999, 104, 30,109-30,125.

15. Karcher, B. Aviation-Produced Aerosols and Contrails; Surv. Geophys.1999, 20, 113-167.

16. Gerz, T.; Durbeck, T.; Konopka, P. Transport and Effective Diffusion ofAircraft Emissions; J. Geophys. Res. 1998, 103, 25,905-25,913.

17. Sinha, P.; Hobbs, P.V.; Yokelson, R.J.; Christian, T.J.; Kirchstetter,T.W.; Bruintjes, R. Emissions of Trace Gases and Particles from TwoShips in the Southern Atlantic Ocean; Atmos. Environ. 2003, 37, 2139-2148.

18. Cocker, D.R.; Shah, S.D.; Johnson, K.; Miller, J.W.; Norbeck, J.M.Development and Application of a Mobile Laboratory for MeasuringEmissions from Diesel Engines 1. Regulated Gaseous Emissions; Envi-ron. Sci. Technol. 2004, 38, 2182-2189.

19. Kittelson, D.B.; Watts, W.F.; Johnson, J.P.; Remerowki, M.L.; Ische,E.E.; Oberdorster, G.; Gelein, R.A.; Elder, A.; Hopke, P.K.; Kim, E.;Zhao, W.; Zhou, L.; Jeong, C.H. On-Road Exposure to Highway Aero-sols. 1. Aerosol and Gas Measurements; Inhal. Toxicol. 2004, 16(Suppl.1), 31-39.

20. Morawska, L.; Ristovski, Z.D.; Johnson, G.R.; Jayaratne, E.R.;Mengersen, K. Novel Method for On-Road Emission Factor Measure-ments Using a Plume Capture Trailer; Environ. Sci. Technol. 2007, 41,574-579.

21. Tang, U.W.; Wang, Z. Determining Gaseous Emission Factors andDriver’s Particle Exposures during Traffic Congestion by Vehicle-Following Measurement Techniques; J. Air & Waste Manage. Assoc.2006, 56, 1532-1539.

22. Zavala, M.; Herndon, S.C.; Wood, E.C.; Jayne, J.T.; Nelson, D.D.;Trimborn, A.M.; Dunlea, E.; Knighton, W.B.; Mendoza, A.; Allen, D.T.;Kolb, C.E.; Molina, M.J.; Molina, L.T. Comparison of Emissions fromOn-Road Sources Using a Mobile Laboratory under Various Drivingand Operational Sampling Modes; Atmos. Chem. Phys. 2009, 9, 1-14.

23. Jiang, M.; Marr, L.C.; Dunlea, E.J.; Herndon, S.C.; Jayne, J.T.; Kolb,C.E.; Knighton, W.B.; Rogers, T.M.; Zavala, M.; Molina, L.T.; Molina,M.J. Vehicle Fleet Emissions of Black Carbon, Polycyclic AromaticHydrocarbons, and Other Pollutants Measured by a Mobile Laboratoryin Mexico City; Atmos. Chem. Phys. 2005, 5, 3377-3387.

24. Zavala, M.; Herndon, S.C.; Slott, R.S.; Dunlea, E.J.; Marr, L.C.; Shorter,J.H.; Zahniser, M.; Knighton, W.B.; Rogers, T.M.; Kolb, C.E.; Molina,L.T.; Molina, M.J. Characterization of On-Road Vehicle Emissions inthe Mexico City Metropolitan Area Using a Mobile Laboratory inChase and Fleet Average Measurement Modes during the MCMA-2003Field Campaign; Atmos. Chem. Phys. 2006, 6, 5129-5142.

25. Zavala, M.; Herndon, S.C.; Wood, E.C.; Jayne, J.T.; Nelson, D.D.;Trimborn, A.M.; Dunlea, E.; Knighton, W.B.; Mendoza, A.; Allen, D.T.;Kolb, C.E.; Molina, M.J.; Molina, L.T. Comparison of Emissions fromOn-Road Sources Using a Mobile Laboratory under Various Drivingand Operational Sampling Modes; Atmos. Chem. Phys. 2009, 9, 1-14.

26. Canagaratna, M.R.; Jayne, J.T.; Ghertner, D.A.; Herndon, S.; Shi, Q.;Jimenez, J.L.; Silva, P.J.; Williams, P.; Lanni, T.; Drewnick, F.; Demer-jian, K.L.; Kolb, C.E.; Worsnop, D.R. Chase Studies of ParticulateEmissions from In-Use New York City Vehicles; Aerosol Sci. Technol.2004, 38, 555-573.

27. Herndon, S.C.; Shorter, J.H.; Zahniser, M.S.; Wormhoudt, J.; Nelson,D.D.; Demerjian, K.L.; Kolb, C.E. Real-Time Measurements of SO2,

H2CO, and CH4 Emissions from In-Use Curbside Passenger Buses inNew York City Using a Chase Vehicle; Environ. Sci. Technol. 2005, 39,7984-7990.

28. Shorter, J.H.; Herndon, S.; Zahniser, M.S.; Nelson, D.D.; Wormhoudt,J.; Demerjian, K.L.; Kolb, C.E. Real-Time Measurements of NitrogenOxide Emissions from In-Use New York City Transit Buses Using aChase Vehicle; Environ. Sci. Technol. 2005, 39, 7991-8000.

29. Pirjola, L.; Parviainena, H.; Hussein, T.; Vallia, A.; Hameri, K.; Aaaltob,P.; Virtanend, A.; Keskinend, J.; Pakkanene, T.A.; Makela, T.; Hillamoe,R.E. “Sniffer”—a Novel Tool for Chasing Vehicles and MeasuringTraffic Pollutants; Atmos. Environ. 2004, 38, 3625-3635.

30. Hansen, A.D.A.; Rosen, H. Individual Measurements of the EmissionFactor of Aerosol Black Carbon in Automobile Plumes; J. Air & WasteManage. Assoc. 1990, 40, 1654-1657.

31. Arnott, W.; Moosmuller, H.; Rogers, C.; Jin, T.; Bruch, R. Photoacous-tic Spectrometer for Measuring Light Absorption by Aerosol: Instru-ment Description; Atmos. Environ. 1999, 33, 2845-2852.

32. Yokelson, R.J.; Susott, R; Ward, D.E; Reardon, J; Griffith, D.W.T. Emis-sions from Smoldering Combustion of Biomass Measured by Open-Path Fourier Transform Infrared Spectroscopy; J. Geophys. Res. 1997,102, 18,865-18,877

33. de Castro, A.J.; Meneses, J; Briz, S.; Lopez, F. Nondispersive InfraredMonitoring of NO Emissions in Exhaust Gases of Vehicles; Rev. Sci.Instrum. 1999, 70, 3156-3159.

34. Li-Cor Biosciences LI-840 CO2/H2O Gas Analyzer; Li-Cor Biosciences:Lincoln, NE, 2009; available at http://www.licor.com/env/Products/GasAnalyzers/li840/840.jsp (accessed 2009).

35. Keskinen, J.; Pietarinen, K.; Lehtimaki, M. Electrical Low PressureImpactor; J. Aerosol. Sci. 2003, 23, 353-360.

36. Pagels, J.; Gudmundsson, A.; Gustavsson, E.; Asking, L.; Bohgard, M.Evaluation of Aerodynamic Particle Sizer and Electrical Low-PressureImpactor for Unimodal and Bimodal Mass-Weighted Size Distribu-tions; Aerosol Sci. Technol. 2005, 39, 871-887.

37. van Gulijk, C.; Marijnissen, J.C.M.; Makkee, M.; Moulijn, J.A. Oil-Soaked Sintered Impactors for the ELPI in Diesel Particulate Measure-ments;; J. Aerosol Sci. 2003, 34, 635-640.

38. Moosmuller, H.; Arnott, W.P.; Rogers, C.F.; Bowen, J.L.; Gillies, J.A.;Pierson, W.R. Time-Resolved Characterization of Diesel ParticulateEmissions. 2. Instruments for Elemental and Organic Carbon Measure-ments; Environ. Sci. Technol. 2001, 35, 1935-1942.

39. Chow, J.C.; Watson, J.G.; Lowenthal, D.H.; Chen, L.W.A.; Tropp, R.J.;Park, K.; Magliano, K.L. PM2.5 and PM10 Mass Measurements in Cali-fornia’s San Joaquin Valley; Aerosol Sci. Technol. 2006, 40, 796-810.

40. Arnott, W.P.; Moosmuller, H.; Rogers, C.F.; Jin, T.; Bruch, R. Photo-acoustic Spectrometer for Measuring Light Absorption by Aerosol:Instrument Description; Atmos. Environ. 1999, 33, 2845-2852.

41. Chan, T.; Lippmann, M. Particle Collection Efficiencies of SamplingCyclones: an Empirical Theory; Environ. Sci. Technol. 1977, 11, 377-386.

42. Chow, J.C. Critical Review: Measurement Methods to DetermineCompliance with Ambient Air Quality Standards for Suspended Par-ticles; J. Air & Waste Manage. Assoc. 1995, 45, 320-382.

43. Chow, J.C.; Yu, J.Z.; Watson, J.G.; Ho, S.S.H.; Bohannan, T.L.; Hays,M. D.; Fung, K.K. The Application of Thermal Methods for Determin-ing Chemical Composition of Carbonaceous Aerosols: a Review; J.Environ. Sci. Health 2007, 42, 1521-1541.

44. Watson, J.G.; Chen, L.W.A.; Chow, J.C.; Lowenthal, D.H.; Do-raiswamy, P. Source Apportionment: Findings from the U.S. SupersiteProgram; J. Air & Waste Manage. Assoc. 2008, 58, 265-288; doi:10.3155/1047-3289.58.2.265.

45. Chow, J.C.; Yu, J.Z.; Watson, J.G.; Ho, S.S.H., Bohannan, T.L.; Hays,M.D.; Fung, K.K. The Application of Thermal Methods for Determin-ing Chemical Composition of Carbonaceous Aerosols: a Review; J.Environ. Sci. Health 2007, 42, 1521-1541.

46. Watson, J.G.; Chow, J.C.; Chen, L.W.A. Methods to Assess Carbona-ceous Aerosol Sampling Artifacts for IMPROVE and Other Long-TermNetworks; J. Air & Waste Manage. Assoc. 2009, 59, 898-911; doi:10.3155/1047-3289.59.8.898.

47. Chow, J.C.; Watson, J.G.; Pritchett, L.C.; Pierson, W.R.; Frazier, C.A.;Purcell, R.G. The DRI Thermal/Optical Reflectance Carbon AnalysisSystem: Description, Evaluation and Applications in U.S. Air QualityStudies; Atmos. Environ. 1993, 27, 1185-1201.

48. Chow, J.C.; Watson, J.G.; Chen, L.W.A.; Chang, M.C.O.; Robinson,N.F.; Trimble, D.; Kohl, S.D. The IMPROVE_A Temperature Protocolfor Thermal/Optical Carbon Analysis: Maintaining Consistency with aLong-Term Database; J. Air & Waste Manage. Assoc. 2007, 57, 1014-1023; doi: 10.3155/1047-3289.57.9.1014.

49. Chow, J.C., Watson, J.G., Chen, L.A., Chang, M.-C.O., Trimble, D.;Fung, K.K.; Zhang, H.; Zhen Yu, J. Refining Temperature Measures inThermal/Optical Carbon Analysis; Atmos. Chem. Phys. 2005, 5, 2961-2972.

Nussbaum et al.

1444 Journal of the Air & Waste Management Association Volume 59 December 2009

50. Ho, S.S.H.; Yu, J.Z.; Chow, J.C.; Zielinska, B.; Watson, J.G.; Sit, E.H.L.;Schauer, J.J. Evaluation of an In-Injection Port Thermal Desorption-Gas Chromatography/Mass Spectrometry Method for Analysis of Non-Polar Organic Compounds in Ambient Aerosol Samples; J. Chromatogr.A 2008, 1200, 217-227.

51. Hinds, W.C. Aerosol Technology: Properties, Behavior and Measurement ofAirborne Particles; Wiley: New York, 1999.

52. Maricq, M.M.; Xu, N.; Chase, R.E. Measuring Particulate Mass Emis-sions with the Electrical Low Pressure Impactor; Aerosol Sci. Technol.2006, 40, 68-79.

53. Kittelson, D.B.; Watts, W.F.; Johnson, J.P. Diesel Aerosol Sampling Meth-odology; CRC E-43 Final Report; Coordinating Research Council: Al-pharetta, GA, 2002.

54. Park, K.; Cao, F.; Kittelson, D.B.; McMurry, P.H. Relationship betweenParticle Mass and Mobility for Diesel Exhaust Particles; Environ. Sci.Technol. 2003, 37, 577-583.

55. Zhu, D.; Nussbaum, N.J.; Kuhns, H.D.; Chang, M.-C.O.; Sodeman, D.;Uppapalli, S.; Moosmuller, H.; Chow, J.C.; Watson, J.G. In-PlumeEmission Test Stand 2: Emission Factors for 10- to 100 kW U.S. MilitaryGenerators; J. Air & Waste Manage. Assoc. 2009, 59, 1437–1445; doi:10.3155/1047-3289.59.12.1446.

56. Cocker D.; Sandip, S.; Johnson, K.; Miller, J.W.; Norbeck, J. Develop-ment and Application of a Mobile Laboratory for Measuring Emissionsfrom Diesel Engines. 1. Regulated Gaseous Emissions; Environ. Sci.Technol. 2004, 38, 2182-2189.

57. Control of Emissions of Hazardous Air Pollutants from Mobile Sources(40 CFR Part 86). Fed. Regist. 2001, 66(61), 17,229-17,273

58. Control of Emissions from New and In-Use Nonroad CompressionIgnition Engines. CFR, Part 89, Title 40, 2004.

59. Park, K.; Chow, J.C.; Watson, J.G.; Trimble, D.L.; Doraiswamy, P.;Arnott, W.P.; Stroud, K.R.; Bowers, K.; Bode, R.; Petzold, A.; Hansen,A.D.A. Comparison of Continuous and Filter-Based Carbon Measure-ments at the Fresno Supersite; J. Air & Waste Manage. Assoc. 2006, 56,474-491.

60. Ronkko, T.; Virtanen, A.; Vaaraslahti, K.; Keskinen, J.; Pirjola, L.;Lappi, M. Effect of Dilution Conditions and Driving Parameters onNucleation Mode Particles in Diesel Exhaust: Laboratory and On-RoadStudy; Atmos. Environ. 2006, 2893-2901.

61. Schneider, J.; Hock, N.; Weimer, S.; Borrmann, S. Nucleation Particlesin Diesel Exhaust: Composition Inferred from in Situ Mass Spectro-metric Analysis; Environ. Sci. Technol. 2005, 39, 6153-6161.

About the AuthorsNicholas Nussbaum is a former student, Dongzi Zhu is adoctoral student, and Hampden D. Kuhns, Hans Moosmuller,and John G. Watson are research professors at Desert Re-search Institute. M.-C. Oliver Chang is an air resources super-visor I with California Air Resources Board and is a formerassociate research professor at Desert Research Institute.Claudio Mazzoleni is an assistant professor in the Atmo-spheric Science Department at Michigan Technological Uni-versity in Houghton, MI. Please address correspondence to:Nicholas Nussbaum, University of Nevada–Reno, 2975Monterey Drive, Reno, NV 89509; phone: �1-775-233-5479;fax: �1-775-674-7060; e-mail: Nicholasnussbaum@gmail.com.

Nussbaum et al.

Volume 59 December 2009 Journal of the Air & Waste Management Association 1445