Solid-Fuel Regression Rates and Flame Characteristicsin an Opposed Flow Burner

S. C. Shark,∗ C. R. Zaseck,∗ T. L. Pourpoint,† and S. F. Son‡

Purdue University, West Lafayette, Indiana, 47907

DOI: 10.2514/1.B35249

Anopposed flowburner is examinedas an instrument to screenandcharacterize fuel before full-scale hybrid rocket

testing. This device requires small amounts (∼10 g) of solid fuels, and it can save time andmaterial in early phases of

fuel characterization. Although impinging jet configurations have been investigated in the past, the full range of

operation of these systems in terms of hybrid rocket motor flowfield conditions has not been fully explored. The

regression rate, flame structure, and flame temperature in an opposed burner configuration is investigated, and an

analysis to relate the results to hybrid rocket applications is developed. Hydroxyl-terminated polybutadiene,

dicyclopentadiene, and paraffin are investigated via an opposed flow burner over an oxidizer mass flux range of 4 to

25 kg∕s∕m2. Results show solid-fuel regression rate sensitivity to laminar and turbulent flow regimes. Aluminized

hydroxyl-terminated polybutadiene regresses ∼34% slower than neat fuel in the opposed flow burner. Infrared

spectroscopy reveals peak flame temperatures of the samples tested, ranging from 1850 to 2100 K. Although the

opposed flow burner is not a perfect representation of hybrid rocket motor operation, it may prove useful in small-

scale screening of fuels.

Nomenclature

a = regression rate coefficient, mm∕sB = blowing coefficientCf = friction coefficientCfo = nonblowing friction coefficientD = oxidizer supply nozzle exit diameterG = mass flux, kg∕s∕m2

n = regression rate exponentRe = Reynolds number_r = regression rate, mm∕sx = position along fuel grain length, mμ = viscosity, kg∕m∕s

I. Introduction

S OLID-FUEL combustion with a gaseous or liquid oxidizer, asseen in hybrid rocket motor configurations, is governed by the

solid-fuel pyrolysis and is generally diffusion limited [1]. It ischallenging to study the instantaneous combustion dynamics andsurface pyrolysis in the hybrid rocket motor configuration due todifficulty viewing the combusting fuel surface. Hybrid rocket motorcombustion is traditionally limited by low solid-fuel regression ratesand combustion efficiency. Some approaches to these issues considerpolymeric fuels, low melt temperature fuels, and energetic fueladditives [1]; however, assessing the efficacy of these materialscurrently requires significant experimental effort. Understandingthe mechanisms behind the solid-fuel pyrolysis and combustionis important to improving both regression rate and combustionefficiency in hybrid rocket motors.Hybrid rocket propellant performance can be quantified with ex-

perimental static hybrid rocket motor evaluation. However, promisingfuels and additives can be expensive,maynot be commercially available,

or may be in low supply. Visualizing the burning surface and flame zoneof the fuel in a hybrid rocket motor configuration can also be difficult.Slab burner experiments are useful, but fuel sample quantities are stillmoderate and visualization of the burning surface often proves difficultdue to viewing window fouling. A small-scale (less than 10 g of fuel)screening method for evaluating the combustion performance of hybridfuel combinations that is analogous to strand burn experiments used forsolid propellant development could be useful [2].An opposed flow burner with a solid-fuel and gaseous oxidizer

could be useful for investigating both diffusion flames and solid-fuelpyrolysis. The opposed flow burner has an inherently different flamezone structure, and it typically operates at mass flux and pressurelevels an order of magnitude lower than the typical hybrid rocketmotor crossflow configuration. However, a relative comparisonbetween fuels can be made in order to compare combustion per-formances. Regression rate characteristics may also be comparablebetween configurations in terms of oxidizer mass flux. The re-gression rate and flame zone could also be examined directly andnonintrusively to reveal combustion characteristics without large fuelquantities and complex windowed configurations. The effects of fueladditives on combustion performance and flame structure could beinvestigated directly. If a simplified flowfield comparison can bemade between the opposed flow burner configuration and the hybridrocket motor configuration, then the performance results may becomparable. The opposed flow burner may prove to be a usefulscreening and characterization tool for hybrid solid fuels especiallyif its operation could be related to hybrid rocket motor operationtheoretically and empirically.To understand the relationship between opposed flow burner and

hybrid rocketmotor configurations, both inert and reacting flowfieldsmust be understood to compare the flow characteristics with andwithout surface blowing. Opposed or stagnation-point flows havebeen used to study impingement flow characteristics since the 1950s[2,3]. Analytical solutions have been established for the Navier–Stokes equations applied to a jet impingement onto an infinite planewithout surface blowing for both laminar and turbulent flow regimes[3]. Hybrid rocket motor port flow is normally considered to beturbulent; thus, the turbulent flow impingement is most relevant forcomparison in the opposed flow burner configuration. A solution tothe equations ofmotion governing the turbulent jet impingement ontoa wall was presented by Glauert [4] using the classic Blasiusempirical formula for the turbulent friction in a pipe but applied to thestagnation-point flow. Watson [5] and Nakoryakov et al. [6] alsoapplied this approach to investigate radial jet impingement and toexamine the near and far flowfield characteristics, respectively. These

Received 11November 2013; revision received 30May 2014; accepted forpublication 4 June 2014; published online 12 September 2014. Copyright ©2014 by the American Institute of Aeronautics and Astronautics, Inc. Allrights reserved. Copies of this paper may be made for personal or internal use,on condition that the copier pay the $10.00 per-copy fee to the CopyrightClearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923; includethe code 1533-3876/14 and $10.00 in correspondence with the CCC.

*Ph.D. Student, School of Aeronautics and Astronautics, 500 Allison Rd.Student Member AIAA.

†Research Associate Professor, School of Aeronautics and Astronautics,500 Allison Rd. Senior Member AIAA.

‡Professor, Mechanical Engineering, 500 Allison Rd. Senior MemberAIAA.

1675

JOURNAL OF PROPULSION AND POWER

Vol. 30, No. 6, November–December 2014

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9

impingement flow characteristics are expected to be related toopposed flow burner operation.The diffusion flame of coaxial impinging gaseous fuel and

oxidizer jets has also been studied and is thoroughly discussed inliterature [7]. The same theories have also been applied to thediffusion flame of gaseous oxidizer jet impingement onto a burningcondensed fuel [8]. The flowfield bulk strain rate or stagnation-pointvelocity gradient was normally used to establish a correlationbetween oxidizer jet velocity and condensed fuel mass flux.Although the regression rate correlates well with the strain rate, mostof these studies focused on using the experimental results to validatenumerical simulations of this complex flowfield [9–11]. Theresearchers were not concerned with the implications of these resultsin terms of hybrid rocket motor operation. Regression rates aretraditionally correlated with oxidizer mass flux for hybrid rocketmotor operation, and it is of interest to correlate opposed flow burnerresults with measured motor data at similar mass fluxes.Recently, researchers used an opposed flow burner apparatus to

characterize solid fuels for hybrid rocket motor applications [12–14].Risha et al. burned small pellets, 10 mm in diameter, of hydroxyl-terminated polybutadiene (HTPB)-based fuels in an opposed flowburner apparatus at various flow rates and pressures [12].Arrhenius-typepyrolysis equations and existing hybrid rocket motor theories based onconvection anddiffusionwere employed to derive semiempiricalmodelsto describe the solid-fuel regression rate. A laminar flowfield wasassumed near the stagnation point, andArrhenius-type expressionswereused to account for the heterogeneous reactions. Multiple models werepresented, showing agreement on the order of �10% with the ex-perimental results.Younget al. [13] investigated fuelswith andwithoutvarious additives in an opposed flowburner apparatuswith thehopesofcharacterizing the fuel combinations before using them in a hybridrocket motor configuration. Regression rate results were presented as afunction of oxidizer jet velocity. Young et al. also investigated solidoxidizers burning in gaseous fuels. Regression rates were correlated asa function of bulk strain rate [14]. Most results from these studiesfluctuate within�10% of the mean, but increased oxidizer velocitiescaused increased regression rates. Neither Risha et al. [12] nor Younget al. [13,14] reported correlations of opposed flow burner regressionrates to oxidizer mass flux. It also appears that only a laminar flowfieldwas applied in the opposed flowconfigurationexperimentally basedonoxidizer jet velocities and images of the burning pellets. The hybridrocketmotor port flowfield is typically assumed to be turbulent [1], andit is important to apply the same flowfield in the opposed flow burnerconfiguration for a relative comparison.The primary objective of this paper is to examine the regression

mechanisms, flame structure, and flame temperature of solid fuels inan opposed flow burner configuration in relation to hybrid rocketmotor operation. Hydroxyl-terminated polybutadiene, dicyclopenta-diene (DCPD), and paraffin are examined as solid fuels with gaseousoxygen. Also, HTPB loaded with 10 wt% 80 nm Al particles wasexamined to observe the effects of metal particle addition in thisconfiguration. Laminar and turbulent flow regime effects on re-gression rate, flame structure, and peak flame temperature in theopposed flow burner configuration are examined. Comparisons aremade between the opposed flow burner configuration and hybridrocket motor configuration, and an analysis is presented to relate thetwo experiments.

II. Experimental Methods

A. Opposed Burner Apparatus

Figure 1 shows a schematic of the opposed flow burner apparatusbased on the design in [14]. The 7.5-mm-diam nozzle outlet wasplaced 10 mm above the fuel surface for ease of measurement for theexperiments presented in this paper. The convective heat transfer forthis ratio of nozzle height above the fuel surface to the nozzlediameter in an opposed flow apparatus has been observed to beindependent of nozzle spacing from the impingement plane [15].Each fuel pellet sample was held in the stainless steel pellet holder.A spring loaded Omega LD620-25 linear variable displacementtransducer (LVDT) was placed through the base of the holder seated

against the bottom of the fuel pellet. An extra spring was also placedbelow the fuel pellet to increase the force on the base. A 30-gaugetungsten wirewas placed over the pellet holder and fuel pellet sampleto resist the force of the spring, and to keep the top of the pellet in linewith the top of the holder. As a result, the top surface of the fuelsample was restricted from moving while consumed during com-bustion with the oxidizer jet. The tungsten wire is a small fraction ofthe burning surface and does not appear to significantly affect thepyrolysis, and it remains at the surface of the burning sample duringoperation. The bottom surface traveled at the regression rate of thefuel sample. An oscilloscope was used to record the LVDT outputvoltage at a 5 kHz sampling frequency, which was converted toposition data. The LVDT tracked the displacement of the bottom ofthe fuel pellet with time. The position data were numericallydifferentiated via the technique presented by De Levie et al. todetermine the instantaneous regression rate of a fuel sample [16]. Thesteady-state portion of the regression rate is reported in this paper.Oxygen was supplied via 15.2MPa (2200 psi) 99.5% pure oxygen

size-K cylinders and was routed to the opposed burner through acopper tubing feed line. An Omega FMA-A2317 flow meter wasemployed to measure the flow rate of the gaseous oxygen. Theoxygen flow ratewas varied from 7 to 50 standard liters per minute tocompare the HTPB, DCPD, and paraffin pellets. The oxidizer massflux was calculated by dividing the oxidizer mass flow rate by thenozzle exit area, which corresponded to a range of 4 to 25 kg∕s∕m2.The HTPB and DCPD fuel samples would pyrolyze during thecombustion process, whereas a portion of the paraffin samples wouldmelt and collect on the pellet holder. The fuel pellets and the pelletholder were weighed before and after each paraffin opposed burnertest. For this study, it is assumed that the totalmass loss from the pelletand the holder accounted for vaporized or combusted fuel. Theremaining pellet mass loss on the holder accounted for molten fuelthat was stripped from the surface.

B. Optical Flame Zone Measurements

The opposed burner apparatus can support a variety of opticalinstrumentation to study the exposed flame zone. A standard speed(30 frames per second) Cannon XL2 3CCD video camcorder wasused for flame zone visualization. The HTPB and DCPD sampleswere recorded with an f number of 11 and an exposure of 1∕1000 s.The paraffin samples were recorded with an f number of 11 and anexposure of 1∕100 s.A Spectraline ES100 spectrometer with an SS100 scanner was

used to find the peak temperature of the flame zone. The spectrometerwas used to measure spectral radiation intensities in the infraredrange for wavelengths from 1.3 to 4.8 μmwith a resolution of 22 nm

Fig. 1 Opposed flow burner apparatus.

1676 SHARK ETAL.

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9

at a sampling frequency of 1320 Hz. The scanner was swept over theflame zone at a spatial frequency of 10 Hz. The spectral intensitieswere fit to gray body intensity vs wavelength relations (Planck’s law)to determine flame temperature. The resulting 20% highest tem-peratures were filtered to include blackbody fits with R2 > 0.95.

C. Material and Handling

Hydroxyl-terminated polybutadiene, DCPD, and paraffinwere thefuels investigated in this study. A great deal of hybrid rocketcombustion literature focuses on HTPB-based fuels. Hydroxyl-terminated polybutadiene was chosen as the baseline fuel for thissystem. Dicyclopentadiene is an experimental binder investigated forboth composite solid propellants and hybrid solid fuels, and it canbetter protect sensitive particles from decomposing when exposed toair or water vapor [17]. Dicyclopentadiene has the potential to protectsensitive energetic particles that can increase the solid-fuel regressionrate. Paraffin wax has a regression rate three to four times higherthan HTPB, which makes it of interest as a fuel. The HTPB/DCPDvendors, mixing procedures, and casting procedures used in thisstudy are outlined in [17].The paraffin wax used in this study was purchased from Sigma

Aldrich (part 76241, lot BCBG1345V). It has a melt temperature of50–52°C. The paraffin pellets weremelted at 100°C on a hotplate andthen poured into 12.7 mm outer diameter straws. The straws werethen submerged in ice water to consistently cool the samples andprevent additive settling.Nanoaluminum was also investigated as an additive in this study.

The nanoaluminumwas purchased fromNovacentrix,with a nominaldiameter of 80 nm and 79% active content, which was sized by thetechnique described by Mang et al. [18]. All nanoaluminum wasstored in an inert and low-moisture environment before use.The thermal diffusivity and conductivity of each fuel were

measured using the transient plane source (TPS)method described byFlueckiger et al. [19]. The TPS sensor was placed between the twohalves of a sectioned fuel pellet. Power was applied to the sensor for30 s to increase the sensor temperature and, via conduction, thesample temperature. The thermal diffusive properties of the materialcan then be derived from the transient temperature response of thesensor [19]. Twenty milliwatts of power caused a 2 to 3.5 K increasein pellet temperature and produced repeatable results. The mean ofthree experimental runs and a 95% confidence interval for each fuelwere obtained.

III. Results and Discussion

A. Flowfield Comparison

Hybrid rocket motor and opposed flow burner operations can berelated through a comparison of the theoretical convective heattransfer governing the solid-fuel regression rates. The opposedburner flowfield is related to that of a hybrid rocket motor in terms ofthe traditional convective heat transfer theory. For this analysis,kinetics are assumed to be fast enough such that combustion isdiffusion limited and radiation is assumed negligible. Figure 2 showsa representation of the classic hybrid rocket motor flame zone,adapted from [1]. This idealized flame zone is based on the flamesheet theory presented by Marxman and Gilbert [20] and Marxmanand Woodridge [21] using convective heat transfer properties.Figure 3 shows a similar flame zone representation for the opposedflow burner. Within Marxman’s classic hybrid regression analysis,the regression rate is given by

_r � 1

ρf

Cf2GB (1)

where ρf is the fuel density,Cf is the blowing friction coefficient,G isthe mass flux, andB is the blowing coefficient. Marxman andGilbert[20] and Marxman and Woodridge [21] presented an empiricalrelation between �Cf∕Cfo� andB in a crossflow configuration,whereCfo is the nonblowing friction coefficient. They used this empiricalformula to complete the regressionmodel presented for hybrid rocket

motors. It is assumed that this empirical relation does not hold true foran impingement type flow, since the correlation was based oncrossflow data. For simplicity, it is noted that Cfo is proportional toCf in terms ofB, and combining this statement with Eq. (1) produces

_r ∝Cfo2G (2)

In this treatment, it is assumed that impinging flow and transverseflow differ in Eq. (2) by the nonblowing friction coefficient, since theboundary-layer growth is directly dependent on this parameter. Forthe hybrid rocket motor configuration, the Blasius correlation for thenonblowing friction coefficient in turbulent flow along a flat plate isused:

Cfo2� 0.0296Re−0.2 (3)

where Re is the hybrid rocket motor Reynolds number defined by

Re � Gxμ

(4)

For the hybrid rocket motor, G is the port mass flux, x is a givendistance along the fuel grain, and μ is the freestream viscosity [1,2].Combining Eqs. (2–4) produces

_r ∝ x−0.2G0.8 (5)

which reveals that the hybrid rocketmotor solid-fuel regression rate isdependent on the position along the fuel grain port and portmass flux.

Fig. 2 Hybrid rocket motor flame zone schematic.

Fig. 3 Opposed flow burner flame zone schematic.

SHARK ETAL. 1677

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9

For stagnation-point flow, we assumed that the boundary-layerthickness was constant near the impingement zone and that Cfo wasrepresented by the classic empirical formula for the turbulent frictionin a pipe [3,5,6]. As described in Sec. I,

Cfo2� 0.0225Re−0.25 (6)

was used to account for the boundary-layer growth for a turbulentstagnation-point flow,whereRe is the opposed flow burner Reynoldsnumber defined by

Re � GDμ

(7)

for an opposed flow burner, G is the nozzle exit mass flux, D is thenozzle exit diameter, and μ is the freestream viscosity [3,5].Combining Eqs. (2), (6), and (7) results in

_r ∝ D−0.25G0.75 (8)

Acomparison of Eqs. (5) and (8) reveals that the opposed flow burnersolid-fuel regression rate is theoretically less dependent on the nozzleexit mass flux than the hybrid rocket motor configuration, and it ismore dependent on nozzle exit diameter than is the hybrid rocketmotor regression rate on position along the fuel grain.

B. Solid-Fuel Thermal Properties

The thermal diffusivity and conductivity were measured for eachpellet as described in Sec. II.C. The results are presented in Table 1.The percent increase over the baseline neat HTPB is also reported forboth properties. Dicyclopentadiene has a 31% higher thermaldiffusivity than HTPB, but it is only 2% higher thermal conductivity.This property indicates that the DCPD will heat up quicker thanHTPB, but it will not reach a significantly higher temperature. SinceHTPB has a lower decomposition temperature than DCPD, thehigher thermal diffusivity may not increase the regression rate ofDCPD [17]. The paraffin fuel has an 11% higher thermal diffusivitythan HTPB and an approximately 40% higher thermal conductivitythan HTPB and DCPD. The higher thermal conductivity will allowthe paraffin fuel pellets to increase temperature more quickly than theHTPB and DCPD fuel pellets. This result indicates that the paraffinfuel will reach its melt temperature faster than the other fuels willreach their decomposition temperature, which will cause the paraffinto regress more quickly than the other fuels. The HTPB loaded with10wt%80 nmAl has approximately a 25%higher thermal diffusivityand conductivity than neatHTPB. This result indicates that theHTPBloaded with 80 nm Al will heat up more quickly and reach a highertemperature than neat HTPB. However, as discussed in the nextsection, the 80 nmAl addition does not increase the regression rate ofHTPB in the opposed flow burner configuration.

C. Regression Rate Results and Flame Zone Visualization

Regression rate measurements and video imaging were analyzedfrom neat HTPB, DCPD, and paraffin experiments for oxidizer massfluxes ranging from 4 to 25 kg∕s∕m2, which correspond to a nozzleoxidizer velocity of 300 to 1900 cm∕s. In this configuration,25 kg∕s∕m2 was the maximum mass flux that would allow fuelignition. Figure 4 demonstrates the nominal burning behavior ofHTPB and DCPD samples. Each HTPB or DCPD test lasted ∼20 swith 3–6 s of steady-state combustion. Paraffin samples generally

burned for 5–7 s with steady-state combustion that lasted for 1.5–2.5 s due to the high regression rate. The uncertainty of the LVDTmeasurement was�0.0142 mm. Following the uncertainty analysisby Coleman and Steele [22], the uncertainty of the regression ratemeasurements is the uncertainty of the LVDT divided by the steady-state burn time.Figure 5 shows the regression rate results for neat HTPB, DCPD,

and paraffin samples. A regression rate slope change around anoxidizer mass flux of 7 kg∕s∕m2 in the DCPD and HTPB trends isobserved. The slope break occurs at an oxidizer nozzle exit Reynoldsnumber of ∼2500, and it falls in the 2000 to 3000 range thatcorresponds to the laminar to turbulent transition range for pipe flow[3]. At the higher oxidizer mass fluxes (greater than Re ∼ 2500), theflowfield is turbulent at the oxidizer nozzle exit, and at the lowermassfluxes (less thanRe ∼ 2500), the flowfield is laminar. Figure 6 showsa flame zone comparison for HTPB in the proposed laminar and

Table 1 Solid-fuel thermal properties (95% confidence interval shown)

Material Average density, g∕cm3 Average diffusivity, mm2∕s Increase above HTPB, % Average conductivity, W∕m · K Increase above HTPB, %

HTPB 0.900� 0.031 0.118� 0.001 0 0.245� 0.001 0DCPD 0.997� 0.006 0.155� 0.015 31 0.258� 0.009 2Paraffin 0.935� 0.039 0.131� 0.002 11 0.353� 0.004 40HTPB/nAl 1.003� 0.027 0.148� 0.015 26 0.307� 0.007 25

Fig. 4 Typical LVDT displacement trace.

Fig. 5 Regression rate trends for fuels in laminar and turbulent flowregimes: maximum UHTPB � UDCPD � 0.0047 mm∕s, and maximumUparaffin � 0.0095 mm∕s.

1678 SHARK ETAL.

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9

turbulent flow regimes. In the laminar flow regime, the visible flametends to uniformly curl upward in streamlines away from the fuelgrain surface in the opposed flow burner configuration. In theturbulent flow regime, the visible flame tends to stay almost parallelto the fuel grain surface in this configuration and then disperse intothe shear layer with the ambient environment in a nonuniformmatter.The change in flow structure is indicative of a laminar to turbulenttransition [3]. The convective heat transfer characteristics differfor each flow regime and, in turn, directly affect the solid-fuelregression rate.Similar DCPD and HTPB regression rate trends (Fig. 5) are

observed in the turbulent regime. Both HTPB and DCPD are cross-linked polymers that should have similar pyrolysis behaviors, andtherefore similar regression and combustion characteristics. In thelaminar regime, the regression rate trends for DCPD and HTPB haveapproximately the same slope, but they have an offset in regressionrate magnitude. A carbonous combustion product layer builds uponthe DCPD fuel pellets during combustion in the laminar regime. Thethick carbon layer likely restricts diffusion, insulates the surface, andtherefore inhibits the regression rate. This layer is not observed forHTPB (C7.3H10.3O0.1) fuel combustion, since it contains less carbonmolecularly than DCPD (C10H12) [23,24]. The oxidizer velocity inthe turbulent regime was sufficient to remove the carbonous layer,and it allowed the DCPD fuel pellets to regress similarly to the HTPBpellets. The paraffin fuel pellet regresses approximately 350% fasterthan both DCPD and HTPB. This is consistent with hybrid rocketmotor literature, where paraffin regresses ∼300–400% faster thanHTPB under gaseous oxygen [25].Though the specific entrainment phenomenon is likely not present

in the opposed burner configuration, paraffin regression is stilldominated by the melt layer shedding via interfacial drag. Vaporizedparaffin accounted for 26.5� 6.2% of the paraffin mass removal,whereas the remaining paraffin was shed from the pellet due tomelting. Opposed burner paraffin shedding is proportional to en-trainment observed in hybrid rocket operation, where vaporizationand entrainment is observed to be responsible for ∼21 and ∼79% ofmass loss, respectively, at lowmass fluxes [26]. From these results, itis clear that the opposed burner does capture the mechanical masstransfer that is imperative for fast paraffin regression. In both theopposed flow and crossflow systems, the primary mechanism formass removal is interfacial drag. In crossflow systems, waves formedthrough Kelvin–Helmholtz instabilities shear due to interfacial dragand form droplets that are entrained in the flow. However, drag in theopposed flow system causes the molten paraffin to shed from thesurface. The mechanical mass transfer mechanism accounts forapproximately 70–80%of the totalmass transfer in both systems. The

paraffin fuel stripping in the opposed flow burner may be re-presentative of an upper limit to the entrainment mechanismpostulated by Karabeyoglu et al. [25].Figure 7 shows a comparison of HTPB, DCPD, and paraffin fuel

sample combustion in the opposed flow burner. The HTPB andDCPD flames are similar in intensity and structure. The paraffinflame is less intense, and the melt layer shedding is clearly visible.There is no apparent difference in regression rate trends in both flowregimes for paraffin. The lack of transition suggests that the meltingand shedding of the fuel is the primary mechanism of fuel regression,and it may be less affected by flow structure.The regression rate coefficients and exponents for the HTPB,

DCPD, and paraffin sample are shown in Table 2 along withexamples of typical hybrid rocket motor results from literature.However, no DCPD/gaseous oxygen (GOX) motor tests have beenconducted. Data for DCPD and decomposed 90% rocket-gradehydrogen peroxide (RGHP) aer reported for comparison [17]. Allregression rate exponents are lower than the theoretical 0.75 forconvective heat transfer controlled regression in an opposed flowburner configuration: approximately 30% lower for HTPB andDCPD, and 40% lower for paraffin. Approximately a 25% lowerexponent than the theoretical 0.8 for convective heat transfer con-trolled regression for HTPB in a hybrid rocket motor configuration isobserved in some hybrid rocket motor literature [27]. As expectedfrom the analysis previously presented, the empirical opposed burnerregression rate exponents are lower than that of hybrid rocket motors.However, the opposed burner exponents are ∼5–15% lower thanreported empirical exponents for HTPB and DCPD in hybrid rocketmotor configurations. The differences are likely due to scaling effectscaused by the small pellet samples, such as a nonuniform flowfieldabove the surface and heat conduction to the sample holder. The lowregression rate exponent of paraffin likely indicates that paraffinregression in the opposed burner is dominated by the shedding ofmolten paraffin. Karabeyoglu et al. postulates that entrainment ismore dependent on material properties than on local mass flux [28].The regression of paraffin in the opposed flow burner may have

Fig. 6 Laminar vs turbulent (nozzle exit) HTPB flame structure at 5.5and 12.5 kg∕s∕m2, respectively.

Fig. 7 Neat fuel flame structure comparison at 12.5 kg∕s∕m2.

SHARK ETAL. 1679

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9

a similar dependence on material properties. The regression rate dataobtained for DCPD and decomposed RGHP in a hybrid rocketconfiguration show a similar trend observed for the HTPB empiricaldata. Other mechanisms, such as radiation and heterogeneousreactions, are not directly dependent on mass flux, but they can stillaffect fuel regression [12,29,30]. Thesemechanismsmay be themaincause of the difference between theory and empirical data. There isalso more surface blowing in the hybrid rocket motor configurationdue to the higher regression rates, which may increase the mass fluxdependence.Figure 8 shows the regression rate results for HTPB loaded with

and without passivated 80 nm Al in the opposed flow burner. Theaddition of the nanoaluminum reduces the regression rate in thisconfiguration by approximately 35%. Figure 9 shows an image theflame zone of HTPB with 10 wt% 80 nmAl. Aluminum particles areobserved being ejected from the flowfield above the pellet. Thealuminum particle combustion may be occurring too far away fromthe pellet surface to add enough energy to the system to increase the

regression rate. The HTPB surface temperature during pyrolysis isapproximately 750 K, and 80 nm Al ignition occurs above 1000 K[30,31]. The Al particles will not ignite at the surface and have a lowresidence time before being removed from the combustion zone. Theejection of the aluminum particles may be transporting thermalenergy from the flame zone and causing the observed decrease inregression rate of the fuel with nano-aluminum (nAl) addition.Accumulation of aluminum/HTPB combustion products forming atthe surface was observed, which may have also contributed to theregression rate reduction. In traditional hybrid rockets, nanoscale Aldoes increase the regression rate of HTPB. It is postulated thatregression rate trends for additive particles that rely on the traditionalhybrid rocket motor residence times cannot be accurately resolved inthe opposed burner. However, if a fuel additive causes a regressionrate increase in the opposed flow burner configuration, it likelyreleases energy near the fuel surface and may cause a significantincrease in a hybrid motor configuration.

D. Flame Temperature

Figure 10 shows a temporal scan of the spectral intensity as afunction of wavelength for HTPB combustion. The spectral intensitycurve remains relatively consistent over the course of the burn timefor HTPB, and it has R2 > 0.95 when fitted to the spectral intensitycurve for a gray body emitter. Similar results were seen for the DCPDsamples. Temperature is calculated from this fit. A temporal scan ofthe spectral intensity as a function ofwavelength for paraffin is shownin Fig. 11. The paraffin samples do not emit the same levels ofradiance as the HTPB and DCPD samples, and they do not showagreement with the blackbody spectral intensity curves likely due tomuch lower amounts of soot. In addition, the combustion productspecies emit relatively high levels of radiance at wavelengths of 3.5and 4.4 μm, which are near the emission frequencies of H2O andCO2, respectively [32]. The poor gray body fit likely indicates a lackof carbon in the combustion products. Due to these inconstancies,

Table 2 Regression rate coefficient and exponents for opposed flow burner and hybrid rocket motorb,c

Laminar Turbulent

a n a n

Opposed flow burner HTPB 0.157 0.19� 0.07 0.082 0.49� 0.005DCPD 0.097 0.23� 0.06 0.067 0.56� 0.013Paraffin 0.552 0.34� 0.04 0.463 0.44� 0.007

Hybrid rocket motor HTPB [27] N/A N/A 0.049 0.61DCPD [17]a N/A N/A 0.057 0.50Paraffin [25] N/A N/A 1.868 0.62

aDCPD correlation for 90% RGHP.b_r are in millimeters per second, and Gox are in kilograms per second per squared meter.cN/A denotes not applicable.

Fig. 8 Regression rate trends for HTPB loaded with and without 10 wt

% nAl and maximum U � 0.0047 mm∕s.

Fig. 9 Flame structure for HTPB loaded with 10 wt% nAl at12.5 kg∕s∕m2.

Fig. 10 Spectral intensity vs wavelength during the steady-stateburning of a HTPB sample.

1680 SHARK ETAL.

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9

the spectral intensity curves for paraffin could not be accurately usedto calculate temperature.Figure 12 shows a nominal temperature trace for a HTPB sample.

Temperature bands observed in the trace correspond to spectralemissions across the opposed burner diffusion flame zone. Theaverage peak combustion temperature is of interest for this study, asindicated in Fig. 12. Figure 13 shows the average peak temperaturesfor two samples of HTPB, DCPD, and HTPB with 10 wt% nAl forboth the laminar and turbulent flow regimes. The error barscorrespond to the 95% confidence interval of the data [22]. Thetemperatures range from 1850 to 2100 K, and they do not varysignificantly between each regime or from sample to sample. Thetemperatures measured in the opposed flow burner are ∼40% lowerthan the adiabatic flame temperature of typical GOX/hydrocarbonbased flames (∼3600 K). The temperature difference most likelycorresponds to the outer flame region that envelops the primary flamezone at a lower temperature due to the diffusion of product species.This layer most likely contains a significant amount of soot thataffects the radiance and can result in a lower temperature. The fuelpyrolysis is similar for HTPB and DCPD, since regression rates aresensitive to oxidizer mass flux but peak flame temperature is not. Thenanoaluminum addition was expected to increase the flametemperature, but the low resonance time in the flame zone preventedan increase. If the peak average temperature could be significantlyincreased, then an increase in regression rate may be observed, sincethe convective heat transfer would increase from the flame zone to thecondensed fuel surface.

IV. Conclusions

This paper examines the regression rate and flame characteristicsof various solid fuels in an opposed flow burner configuration.Specifically HTPB, DCPD, and paraffin are examined under animpinging gaseous oxygen flow. The effects of metal particleaddition on regression rate in the opposed flow burner configurationare studied with a mixture of HTPB loaded with 10 wt% 80 nm Al.The regression rate is dependent on the flow regime of the flame zonein the opposed flow burner. A transition point in the regression ratetrends occurs as the oxidizer nozzle transitions from a laminar to aturbulent flow regime. In theory, the regression rate exponent for theopposed flow burner configuration should be lower than the hybridrocket motor configuration due to the difference in flowfieldstructure. A lower regression rate exponent was observed in theseresults than theorized. However, both opposed flow and hybridmotorconfigurations have lower exponents than expected for convectiveheat transfer alone (25–30%). The paraffin fuel has a ∼350% higherregression rate than the HTPB fuel in the opposed flow burner that isthe same relative regression rate increase seen in hybrid rocket motorregression rate trends, revealing that the opposed flow burner doescapture the mechanical mass transfer mechanism of paraffinentrainment via the shedding of the molten paraffin. This result ispartly expected, since the paraffin has a 40% higher thermaldiffusivity than the other fuels, and it will reach its melt temperaturemore quickly than the other fuels will reach their decompositiontemperature. A 35% reduction in regression rate was observed forHTPB loaded with 10 wt% 80 nm Al compared to the neat HTPB.This is unexpected, since the 80 nmAl addition increases the thermalconductivity and diffusivity of HTPB, and increases in regressionrates in aluminized fuels have been reported in literature. Thereduction in regression rate is likely a result of the aluminum particlecombustion occurring far enough from the pellet surface to have noeffect while transporting thermal energy away from the flame zone.The average peak flame temperature for the HTPB, DCPD, and

HTPB loaded with 10 wt% nAl ranges from 1850 to 2100 K. Nodistinct change in average peak flame temperature was observedbetween the fuels investigated or for varying oxidizer mass flux. Theflame temperature independence from oxidizer mass flux indicatesthe HTPB and DCPD have similar pyrolysis behavior and that thenanoaluminum addition does not increase the flame temperature inthe opposed flow burner configuration.The opposed flow burner has shown the potential to be useful as a

screening tool to depict regression rate trends for polymeric and lowmelt temperature fuels as a function of oxidizer mass flux in aturbulent flow regime. The opposed flow burner also allowsvisualization of the flame zone and flame temperature measurement.

Fig. 11 Spectral intensity vs wavelength during the burning of aparaffin sample.

Fig. 12 HTPB pellet temperature (Temp.) scan.

Fig. 13 Average peak temperatures for HTPB, DCPD, andHTPBwith10 wt% nAl samples for laminar and turbulent regimes.

SHARK ETAL. 1681

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9

The opposed burner may not be useful in depicting a regression ratetrend for additives unless they can react adequately near the fuelsurface. Further experimentation will need to be conducted to fullyqualify the opposed burner as a screening tool.

Acknowledgments

The authors would like to thank the U.S. Air Force Office ofScientific Research for financial support under contract numberFA9550-09-1-0073, the National Science Foundation GraduateResearch Fellowship Program under grant number DGE-1333468,and the NASA Space Technology Research Fellowship programunder grant number NNX11AN53H. The authors would also like tothank the research conducted at Pennsylvania State University thatinspired this work, Jay Gore at Purdue University for the use of hisequipment, and Stephen Heister at Purdue University for hisguidance.

References

[1] Humble, R. W., Space Propulsion Analysis and Design, 2nd ed.,McGraw–Hill, New York, 1995, pp. 145–182, 497–515, 631–637.

[2] Young, G., Koeck, J. J., Conlin, N. T., Sell, J. L., and Risha, G. A.,“Influence of an Electric Field on the Burning Behavior of Solid Fuelsand Propellants,” Propellants, Explosives, Pyrotechnics, Vol. 37, No. 1,Feb. 2012, pp. 122–130.doi:10.1002/prep.v37.1

[3] Schlichting, H., Boundary Layer Theory, 7th ed., McGraw–Hill, NewYork, 1979, pp. 371–403.

[4] Glauert, M. B., “The Wall Jet,” Journal of Fluid Mechanics, Vol. 1,No. 6, 1956, pp. 625–643.doi:10.1017/S002211205600041X

[5] Watson, E. J., “The Radial Spread of a Liquid Jet over a HorizontalPlane,” Journal of Fluid Mechanics, Vol. 20, No. 3, Nov. 1964,pp. 481–499.doi:10.1017/S0022112064001367

[6] Nakoryakov,V.E., Pokusaev,B.G., andTroyan,E.N., “Impingement ofanAxisymmetric Liquid Jet on a Barrier,” International Journal of Heatand Mass Transfer, Vol. 21, No. 9, 1978, pp. 1175–1184.doi:10.1016/0017-9310(78)90136-9

[7] Kee, R. J., Coltrin, M. E., and Glargourg, P., Chemically Reacting

Flows: Theory and Practice, Wiley, Hoboken, NJ, 2003, pp. 246–297.[8] Krishnamurthy, L., “A Flame Sheet in the Stagnation-Point Boundary

Layer of a Condensed Fuel,” Acta Astronautica, Vol. 1, Nos. 5–6, 1974,pp. 711–736.doi:10.1016/0094-5765(74)90029-0

[9] T’ien, J. S., Singhal, S. N., Harrold, D. P., and Prahl, J.M., “Combustionand Extinction in the Stagnation-Point Boundary Layer of a CondensedFuel,” Combustion and Flame, Vol. 33, Jan. 1978, pp. 55–68.doi:10.1016/0010-2180(78)90045-7

[10] Seshadri, K., and Williams, F. A., “Structure and Extinction ofCounterflow Diffusion Flames above Condensed Fuels : ComparisonBetween Poly (Methy1 Methacrylate) and Its Liquid Monomer, BothBurning inNitrogen-AirMixtures,” Journal of Polymer Science, Part A:Polymer Chemistry, Vol. 16, No. 7, 1978, pp. 1755–1778.doi:10.1002/pol.1978.170160726

[11] Lindstedt, R. P., Luff, D. S., and Whitelaw, J. H., “Velocity and Strain-Rate Characteristics of Opposed Isothermal Flows,” Flow, Turbulenceand Combustion, Vol. 74, No. 2, 2005, pp. 169–194.doi:10.1007/s10494-005-4130-6

[12] Risha, G. A., Harting, G. C., Kuo, K. K., Peretz, A., and Koch, D. E.,“Pyrolysis and Combustion of Solid Fuels,” AIAA Paper 1998-3184,1998.

[13] Young, G., Risha, G. A., Miller, A. G., Glass, R. A., Connell, T. L., andYetter, R. A., “Combustion of Alane-Based Solid Fuels,” InternationalJournal of EnergeticMaterials and Chemical Propulsion, Vol. 9, No. 3,2010, pp. 249–266.doi:10.1615/IntJEnergeticMaterialsChemProp.v9.i3.50

[14] Young, G., Roberts, C., and Dunham, S., “Combustion Behavior ofSolid Oxidizer/Gaseous Fuel Diffusion Flames,” AIAA Paper 2012-1127, 2012.

[15] Gardon, R., “Heat Transfer Characteristics of Impinging Two-Dimensional Air Jets,” Journal of Heat Transfer, Vol. 88, No. 1, 1966,

pp. 1–7.doi:10.1115/1.3691449

[16] De Levie, R., Sarangapani, S., and Czekaj, P., “NumericalDifferentiation by Fourier Transformation as Applied to Electrochemi-cal Interfacial Tension Data,” Analytical Chemistry, Vol. 50, No. 1,Jan. 1978, pp. 110–115.doi:10.1021/ac50023a029

[17] Shark, S. C., Pourpoint, T. L., Son, S. F., and Heister, S. D.,“Performance ofDicyclopentadiene/-BasedHybridRocketMotorswithMetal Hydride Additives,” Journal of Propulsion and Power, Vol. 29,No. 5, 2013, pp. 1–8.doi:10.2514/1.B34867

[18] Mang, J. T., Hjelm,R. P., Son, S. F., Peterson, P.D., and Jorgensen,B. S.,“Characterization of Components of Nano-Energetics by Small-AngleScattering Techniques,” Journal of Materials Research, Vol. 22, No. 7,2007, pp. 1907–1920.doi:10.1557/jmr.2007.0230

[19] Flueckiger, S., Voskuilen, T., Pourpoint, T., Fisher, T. S., and Zheng, Y.,“In Situ Characterization of Metal Hydride Thermal TransportProperties,” International Journal of Hydrogen Energy, Vol. 35, No. 2,Jan. 2010, pp. 614–621.doi:10.1016/j.ijhydene.2009.11.001

[20] Marxman, G., andGilbert,M., “Turbulent Boundary Layer Combustionin the Hybrid Rocket,” Symposium (International) on Combustion,Vol. 9, No. 1, 1963, pp. 371–383.doi:10.1016/S0082-0784(63)80046-6

[21] Marxman, G. A., and Wooldridge, C. E., “Fundamentals of HybridBoundary-Layer Combustion,” Heterogeneous Combustion, editedby Wolfhard, I., and Glassman, H. G., AIAA, New York, 1963,pp. 485–521.

[22] Coleman, H. W., and Steele, W. G., Experimentation and Uncertainty

Analysis for Engineers, Wiley, New York, 1999, Chap. 3.[23] Cruise,D.R., “Theoretical Computations of EquilibriumCompositions,

Thermodynamic Properties, and Performance Characteristics of CPropellant Systems,”Naval Air Weapons Center Rept. TP-6037, ChinaLake, CA, 1979, pp. 36–53.

[24] Corpening, J. H., Palmer, R. K., Heister, S. D., and Rusek, J. J.,“Combustion of Advanced Non-Toxic Hybrid Propellants,”International Journal of Alternative Propulsion, Vol. 1, Nos. 2–3,2007, pp. 154–173.doi:10.1504/IJAP.2007.013018

[25] Karabeyoglu, A., Zilliac, G., Cantwell, B. J., Dezilwa, S., andCastellucci, P., “Scale-Up Tests of High Regression Rate Paraffin-Based,” Journal of Propulsion and Power, Vol. 20, No. 6, 2004,pp. 1037–1045.doi:10.2514/1.3340

[26] Kim, S., Lee, J., Moon, H., Sung, H., and Kim, J., “Liquefying SolidFuel for Hybrid Rocket Motor,” 47th Joint Propulsion Conference,AIAA Paper 2011-5823, 2011, pp. 1–11.

[27] Chiaverini, M. J., Serin, N., Johnson, D. K., Lu, Y., Kuo, K. K., andRisha,G.A., “RegressionRate Behavior of HybridRocket Solid Fuels,”Journal of Propulsion and Power, Vol. 16, No. 1, 2000, pp. 125–132.doi:10.2514/2.5541

[28] Karabeyoglu, M. A., Cantwell, B. J., and Stevens, J., “Evaluation ofHomologous Series of Normal-Alkanes as Hybrid Rocket Fuels,” 41stAIAA Joint Propulsion Conference, AIAA Paper 2005-3908, 2005,pp. 1–42.

[29] Farbar, E., Louwers, J., and Kaya, T., “Investigation of Metallized andNonmetallized Hydroxyl Terminated Polybutadiene/Hydrogen Per-oxide Hybrid Rockets,” Journal of Propulsion and Power, Vol. 23,No. 2, March 2007, pp. 476–486.doi:10.2514/1.22091

[30] Chiaverini,M. J.,Hartingl,G.C., Lu,Y.,Kuol,K.K., Peretz,A., Jonesft,S., Wygleft, B., and Arves, J. P., “Pyrolysis Behavior of Hybrid RocketSolid FuelsUnderRapidHeatingConditions,”AIAAPaper 1997-3078,1997.

[31] Huang, Y., Risha, G. A., Yang, V., and Yetter, R. A., “Effect of ParticleSize on Combustion of Aluminum Particle Dust in Air,” Combustion

and Flame, Vol. 156, No. 1, Jan. 2009, pp. 5–13.doi:10.1016/j.combustflame.2008.07.018

[32] Modest, M. F., Radiative Heat Transfer, 2nd ed., Academic Press, NewYork, pp. 304–338.

E. KimAssociate Editor

1682 SHARK ETAL.

Dow

nloa

ded

by P

EN

NSY

LV

AN

IA S

TA

TE

UN

IVE

RSI

TY

on

May

30,

201

5 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/1.B

3524

9