lable at ScienceDirect

Atmospheric Environment 107 (2015) 307e314

Contents lists avai

Atmospheric Environment

journal homepage: www.elsevier .com/locate/atmosenv

Impact of biodiesel and renewable diesel on emissions of regulatedpollutants and greenhouse gases on a 2000 heavy duty diesel truck

Kwangsam Na a, *, Subhasis Biswas a, William Robertson a, Keshav Sahay a,Robert Okamoto b, Alexander Mitchell b, Sharon Lemieux c

a California Air Resources Board, Monitoring Laboratory Division, 9528 Telstar Ave., El Monte, CA 91731, USAb Industrial Strategies Division, 1900 14th Street, Sacramento, CA 95811, USAc Emissions Compliance, Automotive Regulations and Science Division, 9528 Telstar Ave., El Monte, CA 91731, USA

h i g h l i g h t s

� Effect of biodiesels (BD) and renewable diesel (RD) on regulated emissions (RE) was investigated on a chassis dynamometer.� BD showed higher degree of emission reductions for PM, THC, and CO than RD.� BD showed significant increases in NOx emissions for 50% or higher blends.� BD and RD impacts on CO2 and N2O emissions were of lower magnitude relative to other RE.

a r t i c l e i n f o

Article history:Received 4 August 2014Received in revised form17 February 2015Accepted 23 February 2015Available online 24 February 2015

Keywords:BiodieselRenewable dieselRegulated pollutantsElemental carbonChassis dynamometer

* Corresponding author.E-mail address: kna@arb.ca.gov (K. Na).

http://dx.doi.org/10.1016/j.atmosenv.2015.02.0541352-2310/Published by Elsevier Ltd.

a b s t r a c t

As part of a broad evaluation of the environmental impacts of biodiesel and renewable diesel as alter-native motor fuels and fuel blends in California, the California Air Resources Board's (CARB) Heavy-dutyDiesel Emission Testing Laboratory conducted chassis dynamometer exhaust emission measurements onin-use heavy-heavy-duty diesel trucks (HHDDT). The results presented here detail the impact of biodieseland renewable diesel fuels and fuel blends as compared to CARB ULSD on particulate matter (PM),regulated gases, and two greenhouse gases emissions from a HHDDT with a 2000 C15 Caterpillar enginewith no exhaust after treatment devices. This vehicle was tested over the Urban Dynamometer DrivingSchedule (UDDS) and the cruise portion of the California HHDDT driving schedule. Three neat blendstocks (soy-based and animal-based fatty acid methyl ester (FAME) biodiesels, and a renewable diesel)and CARB-certified ultra-low sulfur diesel (CARB ULSD) along with their 20% and 50% blends (blendedwith CARB ULSD) were tested. The effects of blend level on emission characteristics were discussed ong$km�1 basis. The results showed that PM, total hydrocarbon (THC), and carbon monoxide (CO) emis-sions were dependent on driving cycles, showing higher emissions for the UDDS cycles with mediumload than the highway cruise cycle with high load on per km basis. When comparing CARB ULSD tobiodiesels and renewable diesel blends, it was observed that the PM, THC, and CO emissions decreasedwith increasing blend levels regardless of the driving cycles. Note that biodiesel blends showed higherdegree of emission reductions for PM, THC, and CO than renewable diesel blends. Both biodiesels andrenewable diesel blends effectively reduced PM emissions, mainly due to reduction in elemental carbonemissions (EC), however no readily apparent reductions in organic carbon (OC) emissions were observed.When compared to CARB ULSD, soy- and animal-based biodiesel blends showed statistically significantincreases in nitrogen oxides (NOx) emissions for 50% or higher biodiesel blends. The 20% blends of thebiodiesels showed no statistically significant effect on NOx emissions on any cycle. In contrast, renewablediesel slightly decreased NOx emissions and the degree of reduction was statistically significant for 50%or higher blends over the UDDS cycle, but not at the 20% blends. The highway cruise cycles did not showa statistically strong NOx emission trend with increasing blend level of renewable diesel. Biodiesel and

K. Na et al. / Atmospheric Environment 107 (2015) 307e314308

renewable fuel impacts on two greenhouse gases, CO2 and N2O emissions were of lower magnitudewhen compared to other regulated pollutants emissions, showing a change in their emissions withinapproximately ±3% from the CARB ULSD.

Published by Elsevier Ltd.

1. Introduction

Biodiesel is a widely recognized renewable and alternative fuelfor diesel engines which meets the U.S. Environmental ProtectionAgency (EPA) registration requirements for fuels. The use andproduction of biodiesel has been growing for the past decade and itis projected to continue with the trend as governmental programsand regulations promote the use of increasing levels of biofuel inthe transportation sector to reduce greenhouse gas emissions(when a complete carbon life cycle is considered). For instance, theCalifornia Governor's Executive Orders S-1-07 (Low Carbon FuelStandard, LCFS) and S-06-06 have spurred biofuel development inCalifornia by establishing targets for biofuel production and usewhile pursuing 10% reduction in the carbon intensity of California'stransportation fuels by 2020 (Farrell and Sperling, 2007). In addi-tion, the Federal Renewable Fuel Standard (RFS) under the EnergyIndependence and Security Act of 2007 mandated that a minimumof 500 million gallons of biomass-based diesel fuel must be used inthe United States in 2009. On August 6, 2013, the EPA released itsfinalized volume and percentage standards for four renewable fuelcategories e cellulosic, biomass-based diesel, advanced biofuels,and total renewables e under the 2013 RFS program. The EPAproposes the use of 16.55 billion gallons of renewable fuels in thetransportation sector, corresponding to a percentage standard of9.74% (ratio of renewable fuel volume to gasoline or diesel volume)(US EPA, 2013a).

A number of studies have suggested that biodiesel can lead tohigher NOx emissions (Karavalakis et al., 2009; Cheung et al., 2009;Durbin et al., 2011; Hajbabaei et al., 2012). The potential for increasein NOx emissions has been recognized as a major concern to theapplication of biodiesel. This concern has resulted in a number ofearlier studies to measure NOx emissions from various biodieselmixtures (Durbin et al., 2000; US EPA, 2002; McCormick et al.,2006; Sze et al., 2007; Durbin et al., 2011).

Renewable diesel is a synthetic diesel fuel with no oxygen andaromatic compounds, which is produced based on the hydroge-nation of fatty acids. In order for some consistency, renewablediesel will be regarded as a diesel substitute from a renewablefeedstock which are not esters. Chemically, it involves direct cata-lytic hydrogenation of plant oils and animal fats which mainlyconsist of triglycerides, into the corresponding alkanes. Sideproducts of this hydrogenation include propane and gasolinecomponents. The process has input flexibility to convert anyvegetable oil or animal fat into an aliphatic product with charac-teristics similar to Fischer-Tropsch synthetic diesel. Though thereare differences with each company, renewable diesel is usuallysubjected to thermal hydrotreating or pyrolysis-rapid thermalprocessing. Unlike the yellow transesterified biodiesel, the productis a clear and colorless hydrocarbon fuel, with a good Cetanenumber and better combustion properties than even petroleum-based diesel. Engine dynamometer emissions testing using 100%NExBTL renewable diesel reported reducing NOx emissions by5e18% and PM by 5e28% compared to CARB-certified ULSD(Hajbabaei et al., 2012). Most of the studies on vehicle emissionshave been generally focused on biodiesel impact using an enginedynamometer (McGill et al., 2003; Fang et al., 2008; Hajbabaei

et al., 2012). However, little investigation has been done exam-ining whole vehicles on Chassis Dynamometer on the effect ofblend levels of renewable diesel fuels on regulated emissions,global warming gases, and chemical composition of particulatematter in terms of a chassis dynamometer.

The California Air Resources Board's (CARB) Heavy-Duty DieselEmission Testing Laboratory conducted exhaust emission mea-surements from in-use heavy-duty vehicles on a chassis dyna-mometer with the purpose of comparing the regulated andgreenhouse gas emissions impact of biodiesel and renewable dieseland blends to CARB ULSD over a range of driving cycles. To achievethe goal of this study, a renewable diesel fuel and two biodieselswith different feedstocks (soy- and animal-based diesels) wereused.

According to 2012 data compiled by R.L. Polk for the DieselTechnology Forum, more than 28% of all trucks registered in theUnited States (i.e., 2.5million of the 8.6million trucks) are equippedwith diesel particulate filter (DPF) (Polk, 2013). It is expected thatthe number of DPF-equipped vehicles will continue to increase.Even though the truck tested in this study lacks of exhaust after-treatments and categorized into a fading technology vehicle, thisis an important class of diesel vehicles from an emission standpoint, as the majority of the emissions will be contributed by highemitters in the near future. In addition, it will provide a referencefor comparison and evaluation on the impact of biodiesel andrenewable diesel on regulated emissions between not onlycontrolled and uncontrolled vehicle emissions, but also chassisdynamometer and engine dynamometer tests. This study wasconducted as part of a broad evaluation of the environmental im-pacts of biodiesel and renewable diesel as motor fuels in California.

2. Experimental procedures

2.1. Testing facility

This study was conducted at the California Air Resources Board'sHeavy Duty Emission Test Laboratory (HDETL) in Los Angeles whichis equipped with a heavy-duty chassis dynamometer. The chassisdynamometer (Schenck-Pegasus unit) is driven by a direct current(DC) 447 kW (600HP)motor that can absorb up to 492 kW (660HP).It utilizes a single 182.9 cm (diameter) roller and has the capacity tosimulate inertial weights from 2268 kg to 45,359 kg (5000 to100,000 lbs). The chassis dynamometer cell is equipped with an45.7 cm (diameter) full exhaust flow dilution tunnel (constant vol-ume sampler, CVS). The total flow in the CVS during the testing was70,792 LPM (2500 SCFM) maintained by a critical flow venturi. Thedilution air is filtered through a HEPA filter and carbon cartridges toremove particles and gaseous impurities. The mixed flow (flow ofvehicular exhaust and dilution air) is fed to a train of gaseous and PMinstruments. The emission testing system is based on the Code ofFederal Regulations Section 40, Part 86, and Subpart N.

2.2. Vehicles and cycles

The details of the test vehicle are described in Table 1. Thisvehicle was a Freightliner truck equipped with a 2000 C15

Table 1Description on vehicle tested.

Transmission Odometer (km) Inertia weight (kg) Engine displacement (liter) Certification standard (g (kW-h)�1)

NMHC NOx CO PM

Manual 54 K 26,646 (Cruise)19,895 (UDDS)

14.6 1.609 5.364 20.115 0.134

K. Na et al. / Atmospheric Environment 107 (2015) 307e314 309

Caterpillar engine with no exhaust after-treatment devices andexhaust gas recirculation (EGR). For this test, two driving cycleswere employed: a lightly loaded Urban Dynamometer DrivingSchedule (UDDS) and the more aggressive 80 km per hour (50 mileper hour) CARB heavy heavy-duty diesel truck (HHDDT) cruisecycle (CRC, 2007) at a higher test weight. The UDDS refers to U.S.EPA's cycle developed for chassis dynamometer testing of heavy-duty vehicles (US EPA, 2013a,b). A randomized test matrix wasused to minimize contributions from other factors such as diurnaland temporal effects (Durbin et al., 2011).

Table 2Test fuel properties.

Properties Blend stock

CARB (ULSD) Soy B100 An B100 R100

Kinematic viscosity(40 �C), mm2/s

2.7 4.2 4.4 2.5

Physical distillation,T90, �C, Max

360 350 348 286

2.3. Fuels tested

Four blend stocks used to formulate the test fuels for this studywere: (1) soy-based biodiesel, (2) animal-based biodiesel, (3)renewable diesel, and (4) CARB-certified ultra-low sulfur diesel(CARB ULSD). Of these fuels, the CARB ULSD was used as thebaseline for testing. The CARB ULSD was obtained from a Californiarefinery. Two types of biodiesels from different feedstocks wereselected to provide a range of characteristics of biodiesel in terms ofCetane number and degree of saturation. The biodiesels wereprovided by the National Biodiesel Board, while the renewablediesel used was supplied by Neste Oil Inc., and it is known as ahydro-treated vegetable oil and or animal fat derived (NExBTL)diesel fuel. For this study, the biodiesel and renewable diesel wereblended with the CARB ULSD from a single production lot indifferent blending ratios. The neat soy-based biodiesel (Soy B100),neat animal-based biodiesel (An B100), and neat renewable diesel(R100) were blended with CARB ULSD at levels of 20% (Soy B20, AnB20, and R20) and 50% (Soy B50, An B50, and R50). Test fuelblending was conducted using gravimetric method. Each test fuelwas stored in an individual drum with its headspace filled withinert nitrogen to prevent oxidation. The Soy B100, An B100, R100,and CARB ULSD fuels were analyzed in triplicate for the propertiesunder the ASTM International standard ASTM D6751 for Soy B100and An B100 and D975 for R100 and ULSD. In addition, R100 cetanewas measured using the IQT test, which is more accurate for highcetane fuels. Details of the fuels analyzed are provided in Table 2. Inthis table, the straight CARB ULSD, Soy B100, An B100, and R100 arepresented. It is noted that carbon density is the highest for theCARB ULSD showing 86% of the total mass of the fuel, followed byR100 (85%), Soy B100 (77%), and An B100 (76%).

Specific gravity 0.827 0.881 0.881 0.773Flash point, �C, min 148 169 164 146Cetane number 55.8 47.7 57.9 72.3Cloud point, �C �6.6 0.4 12.5 �27.1Sulfur CONTENT,

ppm, Max4.7 0.7 2.0 0.3

Nitrogen, ppm <1.0 2.9 NRb 1.3Carbon (wt. %) 86.1 76.7 75.9 84.8Hydrogen (wt. %) 13.7 12.0 12.2 15.1Oxygen (wt. %)a 0.2 11.3 11.9 3.0Total aromatics (wt. %) 18.7 NR NR 0.4

Soy B: Soy-based biodiesel; An B: Animal-based biodiesel; R: Renewable diesel.a The highest probable oxygen content is determined by subtracting the sum of

theweight percentages of Carbon, Hydrogen, and Nitrogen from 100% and assumingthat any remaining elements are present in negligible amounts.

b Not reported.

2.4. Emissions measurement

Emission data measured include oxides of nitrogen (NOx), par-ticulate matter (PM), total hydrocarbons (THC), carbon monoxide(CO), organic carbon (OC), and elemental carbon (EC). Methane(CH4), Carbon dioxide (CO2) and Nitrous Oxide (N2O) were alsomeasured as representative of global warming gases. All data re-ported here are tunnel blank corrected. The CVS tunnel blanks werecollected over the same UDDS cycle duration (35 min) as used forthis study. PM and NOx measured in the CVS tunnel blanks were onaverage 1.178 mg km�1 and 0.469 mg km�1, respectively.

In this study, CH4 emission data are not reported because the

CH4 concentrations measured from exhaust were consistentlylower than those found in background air. PM was collected on47 mm Teflon filters by using a particle sampling unit (PSU, Horiba,Ann Arbor, MI, USA). The sampling condition was maintained at47 ± 5 �C in accordance with 40 CFR 1065 guidelines. PM masscollected on these filters was weighed using UMX2 Mettler Toledomicrobalance (Mettler-Toledo LLC, Columbus, OH, USA) with areadability of 0.1 mg. A Horiba exhaust gas analyzer (MEXA 7200D,Horiba, Ann Arbor, MI, USA) was used for CO, CO2, THC, and NOx

analysis. N2O was collected in a ~4 L Teflon bag from a CVS inte-grated sample bag and quantified with a Fourier Transform Infraredspectrometer (FTIR, Thermo Scientific Nicolet 6700, Pittsburgh, PA,USA) within eight hours of sampling. CO and CO2 were quantifiedwith non-dispersive infrared (NDIR) technique. THC and NOx wereanalyzed with flame ionization detector and chemi-luminescencedetector, respectively. In this study, regulated emission data ob-tained from direct sampling from the CVS tunnel was used becausebag sampling caused a storage loss by gas adsorption to the samplebag. To ensure data quality, performance of gas analyzer and CVStunnel were checked every week including linearization checks,NOx efficiency test, and propane recovery check. OC and EC sampleswere collected on pre-fired quartz filters and analyzed by CARB'sSouthern Laboratory Branch of Monitoring Laboratory Divisionusing the DRI Thermo/Optical Carbon Analyzer Model 2001,following the IMPROVE_A protocol. Their analytical methods aredescribed elsewhere (CARB, 2011).

2.5. Statistical analysis

TheData Analysis Tool providedbyMicrosoft Excel 2003was usedto conduct t-test for thedata set. Thepurposeof this test is to look intohow different fuel blend and blend levels significantly affect the

K. Na et al. / Atmospheric Environment 107 (2015) 307e314310

emission of regulated pollutants. For this test, two-tailed distributionand two-sample assuming unequal variances were selected becausethe distribution is not symmetrical. Test results with a p-value lessthan or equal to 0.05 were considered statistically significant for thistest. Prior to applying the t-test to the data set, the outlier test wasperformed in terms of the Interquartile Range (IQR) because thenumber of data is not sufficient (i.e., six to eighteen points) to verifyGaussian distribution. Any data that was less than or more than 1.5times the IQR was considered an outlier.

3. Results and discussion

3.1. NOx emissions

Fig. 1(a) shows NOx emissions in grams per mile obtainedindividually from varying blend levels on two different drivingcycles: a two consecutive UDDS cycle and the 80 km h�1 averagespeed CARB HHDDTcruise cycle (also called highway cruise cycle inthis study). The x-axis lists fuels, blend levels, and test cycles. Theerror bars represent the standard deviation of the mean. As can beseen in this figure, more NOx is emitted over the highway cruisecycle with high load (inertia dyno weight (kg) of 26,646) than theUDDS cycle with medium load (inertia dyno weight (kg) of 19,895)independent of biodiesel feedstocks and their blend levels. It isobserved that the highway cruise cycle generates more exhaustheat than the UDDS cycle. The average exhaust temperature was

Fig. 1. Emission rates for regulated pollutants, NO

measured as 262 ± 22 �C for the highway cruise and 193 ± 8 �C forthe UDDS cycle when using CARB ULSD fuel. The exhaust temper-ature was monitored at a position approximately 1.0 m away fromthe tailpipe. Higher NOx emissions for the highway cruise cycle thanthose for the UDDS cycle may be due to its higher load and com-bustion temperature on grams per mile basis.

When compared to CARB ULSD, soy-/animal-based biodieselblends generally showed a trend of increasing NOx emissions withincreasing blend level. In contrast, renewable biodiesel blendsshow a trend of decreasing NOx emissions. Table 3 provides thepercent emissions differences relative to CARB ULSD for each bio-diesel and renewable diesel blend with shaded numbers in thistable representing emissions differences that are statistically sig-nificant. Values with a negative sign represent a decrease in regu-lated pollutants relative to CARB ULSD.

Both soy-/animal-based biodiesel blends show statistically sig-nificant increases in NOx emissions for 50% or higher blend levelrelative to CARB ULSD fuel. For example, Soy B50 and An B50 showNOx increases of approximately 10% and 5%, respectively, for theUDDS cycle. This increasing effect of biodiesels on NOx emissions isconsistent with previous studies (Karavalakis et al., 2009; Cheunget al., 2009; Hajbabaei et al., 2012). The degree of unsaturationand Cetane number of biodiesel are reported to be associated withNOx emissions (Knothe, 2005; Pham et al., 2013). That is, the higherthe proportion of unsaturation and the lower the Cetane number,the higher NOx emissions. Neat soy-based biodiesel typically

x, PM, THC, and CO on grams per km basis.

Table 3Percent changes in emissions compared to the CARB ULSD fuel.

Driving cycle Fuel THC CO NOx CO2 PM

UDDS � 2 Soy B20 5.6 �9.4 �4.2 �3.0 �33.3An B20 �14.0 �11.1 3.3 1.3 �40.6R20 �10.1 �0.7 �0.6 �0.3 1.8Soy B50 �2.7 �25.1 9.5 �3.0 �60.7An B50 �32.1 �24.8 4.9 1.6 �50.0R50 �19.9 �9.3 �2.4 �3.0 �21.4Soy B100 �36.0 �32.8 20.2 �2.6 �79.0An B100 �54.2 �40.5 13.7 2.1 �79.8R100 �23.4 �14.8 �6.6 �3.1 �29.4

50 mphHighwayCruise

Soy B20 6.5 �15.9 �1.9 �2.0 �24.5An B20 �14.0 �8.4 0.4 1.1 �16.0R20 0.9 �3.3 0.9 0.2 �10.1Soy B50 �6.9 �25.9 6.4 �1.5 �51.7An B50 �25.8 �24.2 6.7 2.3 �37.7R50 �7.5 �8.6 �1.2 �1.4 �18.3Soy B100 �36.4 �41.7 17.5 �0.6 �67.6An B100 �56.3 �43.6 17.1 3.1 �69.5R100 �13.3 �18.8 �3.7 �2.6 �24.8

Note: shaded numbers represent statistically significant % changes for p � 0.05 at95% confidence interval.Soy B: Soy-based biodiesel; An B: Animal-based biodiesel; R: Renewable diesel.

K. Na et al. / Atmospheric Environment 107 (2015) 307e314 311

comprises more unsaturated compounds (~83%) and lower Cetanenumber (~48) as compared with neat animal-based biodiesel beingtypically composed of ~52% unsaturated compounds and Cetanenumber of ~58 (Hoekman et al., 2012). This difference generallyresults in higher NOx emissions for soy-based biodiesel than foranimal-based one. However, no significant difference in NOx

emissions between Soy B100 and An B100 is observed in this study.For renewable diesel blends, only neat renewable diesel, R100 isshown in this study to have a statistically significant reduction inNOx emissions over the UDDS cycle.

3.2. PM emissions

Fig. 1(b) shows PM emissions rates in grams per kilometer forCARB ULSD, soy-/animal based biodiesel, and renewable dieselfuels and blends. Among the regulated pollutants measured in thisstudy, PM shows the greatest g km�1 reduction when using thebiodiesel and renewable diesel blends. Both test cycles show anoticeable decrease in PM emissions with increasing blend levels ofbiodiesels and renewable diesel. However, PM emissions are higherfor the UDDS cycle than the highway cruise cycle regardless ofdiesel feedstocks and their blend levels, indicating duty cycledependence. It is noted that the degree of PM reduction for the soy-and animal-based biodiesel blends when compared to CARB ULSDis apparently larger than the corresponding reduction in PMemissions found for the renewable diesel blends. Based on t-test,the animal-based biodiesel blends show statistically similar re-ductions (at 95% confidence interval) to the corresponding soy-based biodiesel blends in PM emissions except for the blend levelof 50%. For renewable diesel when compared to CARB ULSD sta-tistically significant reductions in PM emissions are observed for50% (R50) or higher blends over the UDDS cycle and for R100 overthe highway cruise cycle. As seen in Table 3, when compared withCARB ULSD, PM emissions for Soy B100 and An B100 are signifi-cantly reduced by 79% and 80% for the UDDS, and 68% and 70% forthe highway cruise, respectively. R100 also shows significant re-ductions of 29% and 25% for the UDDS and the highway cruise,respectively.

To gain further insight into the reduction of PM emissionschanges in the major components of PM, OC and EC wereinvestigated. Results are shown with respect to the blend levels

for the UDDS cycle in Fig. 2. As seen in this figure, EC emissionsmarkedly decrease with increasing blend levels for both bio-diesels and renewable diesel. In contrast, OC emissions arerelatively consistent over all fuels including CARB ULSD. Thisindicates that PM reduction is attributed primarily to decreasingEC emission. EC emissions reductions of the soy-based biodieseland animal-based biodiesel blends are considerably larger thanthose of the renewable diesel blends. For Soy B100 and An B100,EC emissions are reduced by 81% and 80%, respectively whileR100 only shows a 20% reduction. However, the neat renewablediesel shows slightly higher OC emissions reductions than thetwo neat biodiesels.

3.3. THC emissions

THC is defined herein as the combination of methane andgaseous non-methane hydrocarbon. As seen in Fig. 1(c) and Table 2,soy-/animal-based biodiesels and renewable diesel blends whencompared to CARB ULSD show a trend of decreasing THC emissionswith increasing blend levels over both test cycles. The exception tothis trend is that soy-based biodiesel blend, Soy B20 is higher thanCARB ULSD in THC emissions for both driving cycles. However thereason for this experimental result is unknown. Soy-based bio-diesel blends show a different degree of reduction in THC emissionsfrom those of animal-based biodiesel and renewable diesel. Themagnitude of THC reduction for the animal-based biodiesel blendsis statistically significantly higher than the corresponding re-ductions found for the blends of soy-based biodiesel and renewablediesel by approximately 10%e30% and 10%e50%, respectively. Aslisted in Table 3, for the UDDS cycle, Soy B100, An B100, and R100show statistically significant reductions in the THC emissions by36%, 54%, and 23%, respectively. For the highway cruise cycle, twoneat biodiesel fuels, Soy B100 and An B100 offer statistically sig-nificant reductions in THC emissions by 36% and 56%, respectively,but neat renewable diesel, R100 did not show a statistically sig-nificant reduction in THC emissions.

3.4. CO emissions

Fig. 1(d) shows CO emissions rates for blends of biodiesel andrenewable diesel and CARB ULSD fuel. The figure shows that bio-diesel and renewable diesel blends when compared to CARB ULSDresults in a trend of increasing reduction in CO emissions withhigher blend levels for both driving cycles. This is consistent withanother study (Cheung et al., 2009). CO emissions are shown to bedifferent according to feedstocks and driving cycles. As expectedfrom less transient operation, CO emissions are less for the highwaycruise cycle than UDDS cycle regardless of biodiesel feedstocks andtheir blend levels on g$km�1 basis. This implies that duty cycleeffects are larger than fuel effects for CO emissions. Biodiesel blendsare found to be more effective than CARB ULSD and renewablediesel blends in reducing CO emissions for both cycles. However, nostatistically significant differences in CO emissions were foundbetween the soy-based and the animal-based diesel blends. As seenin Table 3, for all blend levels and both test cycles, both soy-basedand the animal-based diesels show a statistically significantreduction in CO emissions relative to CARB ULSD. CO emissions ofR20 and R50 are generally similar to those found on CARB ULSD forboth test cycles. However, R100 provides a statistically significantreduction in CO emissions for both test cycles.

3.5. Greenhouse gases emissions

The percent differences in emissions rates between biodiesel/renewable diesel and CARB ULSD on two greenhouse gases, CO2

Fig. 2. Emission rates for OC/EC obtained from different feedstocks and different blend levels of biodiesel and renewable diesel fuel for the UDDS driving cycle.

K. Na et al. / Atmospheric Environment 107 (2015) 307e314312

and N2O, are observed to be of lower magnitude than for otherregulated pollutants emissions for both UDDS and the highwaycruise cycles, showing a change in emissions of less than ±3%. Inaddition, no readily discernable trends were observed between thetest alternative fuels and CARB ULSD. It should be noted that onlytailpipe GHG emissions were analyzed in this study. To determine afuel's impact on climate change, a full fuel lifecycle analysis of GHGemissions is necessary.

As seen in Fig. 3, CO2 g km�1 emissions are lower for the high-way cruise cycle than the UDDS cycle regardless of biodiesel feed-stocks and their blend levels (p < 0.05 at 95% confidence interval)though the magnitude of reduction in CO2 emissions is minor. CO2emission pattern is opposite between the biodiesel blends andrenewable diesel blends. When switching fuel from CARB USLD toanimal-based biodiesel blends, CO2 emissions slightly increasewith higher blend levels with no statistical confidence. In contrast,renewable diesel blends show a decreasing trend in CO2 emissionswith increasing blend level. Small (approximately 3%) but statisti-cally significant increases in CO2 emissions for renewable diesel

Fig. 3. Carbon dioxide (CO2) emissio

blends are observed for 50% or higher biodiesel blends for the UDDScycle only. Calorific values for biodiesel (37e41 MJ/kg) are reportedto be lower than conventional diesel (42e46 MJ/kg) (Radich, 2013).As a result, more biodiesel must be consumed to propel a dieselvehicle the same distance than conventional diesel, emitting moreCO2 due to less energy content of biodiesel. However, this theorydoesn't seem to be always correct for real-world vehicle drivingsituations. Conventional diesel (~86 wt.%) contains more carbonthan biodiesel (~77 wt.%), which means that there would be apossibility for conventional diesel to form more CO2 than biodieselfor the same amount of fuel burnt. This conflicting property causesan ambiguous difference in exhaust CO2 emissions between bio-diesel and conventional diesel. Lower CO2 emissions for biodieselthan for conventional diesel were also observed in the compre-hensive US EPA study on biodiesel impacts on exhaust emissions(US EPA, 2002). We also observed lower CO2 emissions for soy-based biodiesel only. However, the difference from the CARBULSD is less than 3%. Therefore, it seems that our test results fallwithin the variation of entire set of results observed in literature

n rates on grams per km basis.

Fig. 4. Nitrous Oxide (N2O) emission rates on grams per km basis.

K. Na et al. / Atmospheric Environment 107 (2015) 307e314 313

(US EPA, 2002).Fig. 4 shows N2O emissions with respect to biodiesel feedstocks,

blend levels, and UDDS cycle. Comparing CARB USLD to the twoFAME biodiesels and renewable diesel blends shows no statisticallysignificant impacts on N2O emissions even though animal-basedbiodiesel blends and renewable diesel blends show a small in-crease in N2O emissions over higher blends. In addition, no statis-tically significant differences in N2O emissions are found betweenfuel blends with an exception that N2O emission of Soy B50 issignificantly larger than that of An B50.

4. Conclusions

The potential impacts of biodiesel and renewable diesel blendson regulated emissions were investigated using a Freightliner truck(2000 C15 Caterpillar engine) with no exhaust after-treatmentcontrol device. Four types of fuels, (1) CARB-certified ultra-lowsulfur diesel (ULSD) baseline fuel, (2) soy-based biodiesel, (3)animal-based biodiesel, and (4) renewable diesel were evaluated.Emissions from the four different diesel fuels were evaluated atblend levels of 20%, 50%, and 100%. The emission tests were con-ducted on a chassis dynamometer over two different test cycles.The major results of this study are:

(1) Soy- and animal-based biodiesel blends when compared toCARB ULSD showed statistically significant increases in NOx

emissions for 50% or higher blend biodiesel blend on bothtest cycles. The 20% biodiesel blends NOx differences did notexceed the testing variability in this study on any cycle.

(2) Emissions between soy- and animal-based biodiesel blendswas observed despite the difference in the degree of unsa-turation and cetane number. Soy B50/B100 and An B50/B100showed NOx increases of approximately 10%/20% and 5%/14%,respectively for the UDDS cycle. In contrast, renewable dieselblends showed a trend of decreasing NOx emissions withincreasing blend levels. However, only R100 showed a sta-tistically significant reduction in NOx emissions over theUDDS cycle.

(3) When compared to CARB ULSD, soy- and animal-based bio-diesel blends offered statistically significant reductions in PMemissions for all ranges of blends and both test cycles. SoyB20/B50/B100 and An B20/B50/B100 showed PM reductionsof approximately 33%/61%/79% and 41%/50%/80%, respec-tively for the UDDS cycle. Renewable diesel blends showedstatistically significant PM reductions for 50% or higher blendlevel for both test cycles. The PM reducing effect was higherfor soy- and animal-based biodiesel blends than for renew-able diesel blends.

(4) PM reduction was mainly due to decreasing elemental car-bon (EC). The reducing effect of the two biodiesels on ECemissions was much larger than the effect of renewablediesel. On the other hand OC emissions did not significantlychange over biodiesel and renewable diesel blend levels.

(5) Compared to CARB ULSD, soy-/animal-based biodiesels andrenewable diesel blends showed a trend of moderatelydecreasing THC emissions with increasing blend levels forboth test cycles. The magnitudes of THC reductions for theanimal-based biodiesel blends were higher than the corre-sponding reductions found for the blends of soy-based bio-diesel and renewable diesel by approximately 10%e30% and10%e50%, respectively.

(6) For all blend levels and both test cycles, both soy-based andthe animal-based biodiesels showed statistically significantreductions in CO emissions relative to CARB ULSD. Unlikebiodiesel blends, for renewable diesel blends, only R100provides a statistically significant reduction in CO emissionsfor both test cycles.

(7) The difference in the emission rate between biodiesel/renewable diesel fuels and CARB ULSD on two greenhousegases, CO2 and N2O emissions were observed to be of lowermagnitude when compared to other regulated pollutantsemissions for both test cycles, showing a change in theiremissions less than ±3% relative to the baseline CARB ULSD.

(8) Lower CO2 emissions for soy-based biodiesel only wereobserved despite its lower heating value than CARB ULSD.However, the test results fell within the variation of entire setof results observed in literature.

K. Na et al. / Atmospheric Environment 107 (2015) 307e314314

Disclaimer

The statements and opinions expressed here are solely the au-thors’ and do not represent the official position of the California AirResources Board. The mention of trade names, products, and or-ganizations does not constitute endorsement.

Acknowledgments

The authors wish to thank George Gatt, Thomas Ladzinski, RalphRodas, Richard Ling, Peter Wong, Oliver Chang, Shaohua Hu, JornHerner, and Tao Huai for their critical support. Additionally, theauthors are grateful to the National Biodiesel Board and Neste Oilfor donating the fuels for this project.

References

California Air Resources Board (CARB), 2011. Procedure for Organic Carbon andElemental Carbon (OC/EC) Analysis of Vehicular Exhaust Particulate Matter(PM) on Quartz Filters.

Cheung, C.S., Zhu, L., Huang, Z., 2009. Regulated and unregulated emissions from adiesel engine fueled with biodiesel and biodiesel blended with methanol.Atmos. Environ. 43, 4865e4872.

Coordinating Research Council (CRC), 2007. Heavy-duty Vehicle Chassis Dyna-mometer Testing for Emissions Inventory, Air Quality Modeling, SourceApportionment and Air Toxics Emissions Inventory. CRC Report No. E55/59.

Durbin, T.D., Collins, J.R., Norbeck, J.M., Smith, M.R., 2000. Effects of biodiesel,biodiesel blends, and a synthetic diesel on emissions from light heavy-dutydiesel vehicles. Environ. Sci. Technol. 34, 349e355.

Durbin, T.D., Miller, J.W., Johnson, K., Hajbabaei, M., Kado, N.Y., Kobayashi, R.,Vogel, C.F.A., Matsumura, F., Wong, P.S., 2011. Final Report for the CE-CERTEngine Testing Portion for the CARB Assessment of the Emissions from theUse of Biodiesel as a Motor Vehicle Fuel in California Biodiesel Characterizationand NOx Mitigation Study. Final Report Prepared for CARB. California Air Re-sources Board (CARB). available at: http://www.arb.ca.gov/fuels/diesel/altdiesel/.

Fang, T., Lin, Y.-C., Foong, T.M., Lee, C.-F., 2008. Reducing NOx emissions from abiodiesel-fueled engine by use of low-temperature combustion. Environ. Sci.Technol. 42, 8865e8870.

Farrell, A.E., Sperling, D., 2007. A Low-carbon Fuel Standard for California Part 1 andPart 2. Final report for The California Energy Commission. California EnergyCommission. http://www.energy.ca.-gov/low_carbon_fuel_standard/.

Hajbabaei, M., Johnson, K.C., Okamoto, R.A., Mitchell, A., Pullman, M., Durbin, T.D.,2012. Evaluation of impact of biodiesel and second generation biofuel on NOxemissions for CARB diesel fuels. Environ. Sci. Technol. 46, 9163e9173.

Hoekman, S.K., Broch, A., Robbins, C., Ceniceros, E., Natarajan, M., 2012. Review ofbiodiesel composition, properties, and specifications. Renew. Sustain. EnergyRev. 16, 143e169.

Karavalakis, G., Stournas, S., Bakeas, E., 2009. Effects of diesel/biodiesel blends onregulated and unregulated pollutants from a passenger vehicle operated overthe European and the Athens driving cycles. Atmos. Environ. 43, 1745e1752.

Knothe, G., 2005. Dependence of biodiesel fuel properties on the structure of fattyacid alkyl esters. Fuel Process. Technol. 86, 1059e1070.

McCormick, R.L., Williams, A., Ireland, J., Brimhall, M., Hayes, R.R., 2006. Effect ofBiodiesel Blends on Vehicle Emissions. Milestone Report, NREL/MP-540e40554,October.

McGill, R., Storey, J., Wagner, R., Irick, D., Aakko, P., Westerholm, M., Nylund, N.,Lappi, M., 2003. Emission Performance of Selected Biodiesel Fuels. SAE Paper2003-01-1866.

Pham, P., Bodisco, T., Stevanovic, S., Rahman, M., et al., 2013. Engine performancecharacteristics for biodiesels of different degrees of saturation and carbon chainlengths. SAE Int. J. Fuels Lubr. 6, 188e198.

Polk, R.L., 2013. New Technology Clean Diesel Trucks with Near Zero EmissionsMake up 28% of All Trucks on U.S. Highways. Diesel Technology Forum (DTF)available at: http://www.dieselforum.org.

Radich, A., 2013. Biodiesel Performance, Costs, and Use available at: http://www.eia.gov/oiaf/analysispaper/biodiesel/.

Sze, C., Whinihan, J.K., Olson, B.A., Schenk, C.R., Sobotowski, R.A., 2007. Impact ofTest Cycle and Biodiesel Concentration on Emissions. SAE paper 2007-01-4040.

United States Environmental Protection Agency US EPA, 2002. A ComprehensiveAnalysis of Biodiesel Impacts on Exhaust Emissions. EPA 420-P-02e001,October.

United States Environmental Protection Agency (US EPA), 2013a. RegulatoryAnnouncement. EPA-420-F-13e007, January 2013.

United States Environmental Protection Agency (US EPA), 2013b. DynamometerDrive Schedules available at: http://www.epa.gov/nvfel/testing/dynamometer.htm#vehcycles.