nonthermal plasma syngas - github pages

Small Scale C1141

PartD|40

40. Small Scale Catalytic Syngas Productionwith Plasma

Adam A. Gentile, Leslie Bromberg, Michael Carpenter

The production of syngas has been desired fornearly a century as a starting point towards aneffort to achieve the synthesis of higher-energyliquid fuels. Methods that are utilized for obtain-ing syngas have undergone profound technologicaltransformations over the years. The formation ofplasma is one such concept capable of energizingmolecular chains through dielectric manipula-tion in order to excite electrons and form radicals.Many types of plasma configurations exist undera range of variable parameters such as pressure,temperature, current, power intensity, and phys-ical structure. Nonthermal plasma in conjunctionwith catalysis is a relatively new concept that takesa more subtle approach that eliminates the inten-sive energy requirement while maintaining highconversion efficiency. The technology and processpresented in this book chapter will encompassa hybrid plasma/catalysis (PRISM) system havingthe ability to ionize and reform any vapor phasespecies with a focus on short to long chain hydro-carbons. The apparatus promotes reactions thatwould otherwise be disadvantageous from eithera high activation energy standpoint or from theyield of unfavorable side products. Catalytic partialoxidation (CPOX) using a low-energy, nonthermalplasma is capable of nearly perfect conversion ofa carbonaceous starting material into clean, ho-mogeneous, hydrogen-rich syngas.

40.1 Plasma ............................................. 114240.1.1 Introduction ..................................... 114240.1.2 Types of Plasmas ............................... 1144

40.2 Partial Oxidation ReformationUsing Cold Plasma............................. 1147

40.2.1 Methane Reformation ........................ 114740.2.2 Plasma Catalysis ................................ 115140.2.3 Reactor Apparatus ............................. 1153

40.3 Cold-Plasma-AssistedExperimentation . .............................. 1156

40.3.1 Steady State Operation ....................... 115640.3.2 Transients: Startup/Shutdown . ............ 115740.3.3 Conversion and Efficiency ................... 1157

40.4 Analysis and Discussion ..................... 115940.4.1 Process Challenges ............................. 115940.4.2 Design Improvements ........................ 115940.4.3 Higher Hydrocarbon Reformation ........ 1160

40.5 Synergistic Benefits of Plasma ........... 116040.5.1 Chemical Processes ............................ 116040.5.2 Commercial Implementation .............. 1160

40.6 Conclusion........................................ 1161

References ................................................... 1162

This chapter documents work done on small- to mid-scale plasma enhanced catalysis for platform-flexiblefuel reformers in an effort to improve conventional (i. e.,internal combustion) vehicle operation as well as thecommercial standard reformation efficiency. Althoughindustry has been focusing on methane, the system per-forms as a cost-effective method in the production ofhydrogen-rich, process-ready syngas from any carbona-ceous source/feedstock.

There exist several applications of stationary andonboard hydrogen production:

� Supplementation and augmentation of hydrogeninto internal combustion engines (ICEs) for im-proved operation and pollutant elimination� Diesel PRISM fuel reformation for after-treatmentapplication� Onboard hydrogen generation for fuel cell powerproduction� Clean-burn electrical generation (turbines, gen sets)� Liquid drop-in fuel production.

© Springer International Publishing AG 2017C.S. Hsu, P.R. Robinson (Eds.), Springer Handbook of Petroleum Technology, DOI 10.1007/978-3-319-49347-3_40

PartD|40.1

1142 Part D Petrochemicals

The chapter is organized as follows: First, an intro-duction of plasma and its different forms and attributesare defined in Sect. 40.1. The primary reaction of in-terest is partial oxidation (POX) and this, along withthe micro- and macrokinetics of the PRISM process interms of theory, will be described in detail in Sect. 40.2.Section 40.3 details the experimental process from startto finish with a broad overview of operational perspec-

tive. In Sect. 40.4, we expound on a discussion of thedata including process challenges, areas of improve-ment, and future development. Some of the synergisticeffects of plasma over typical gas-phase processes canbe found in Sect. 40.5, which leads to the conclusionof how and why plasma-assisted POX in a catalyticenvironment is a viable and cost-effective technol-ogy.

40.1 Plasma

40.1.1 Introduction

There are many different paths that lead to the produc-tion plasma, which is universally known as the fourthstate of matter and also the most abundant state presentin the universe. Some generally known devices andmethods of generating plasma include the plasmatron,plasma jet or torch, spark, corona discharge, electricalor gliding arc, and glow discharge. For our purposes,plasma is regarded as a collection of freely chargedparticles in random motion that is, on average, electri-cally neutral [40.1]. Typically generated from a gaseousphase, application through an electrical field inducesa temporary condition of a plasma state regardlessof bulk medium temperature. Additional ways to ac-complish this include, but are not limited to, a radiofrequency (RF) treatment or a microwave treatment ofthe gas in a strong electromagnetic field. When utilizedto augment a reaction, plasma adds substantial benefi-cial factors to the molecular interactions and kinetics ofthe reaction at hand.

Breakdown and IonizationThe standard gas molecule is stable in an environmentwhen at or near a state of equilibrium with its surround-ings. As intensive and extensive properties change,molecules begin to lose their stability. Electron move-ment rapidly increases and ionization is achieved whenelectrons gain enough energy to remove themselvesfrom their respective orbitals unbound from their nor-mal position within or around the molecule. Ionization

+ + +

+

+

–

–

–

–

–

–Electrode

a) b)

Vrf

+

–GasPlasma

Fig. 40.1 (a) Visual of excitedion motion that makes up plasma.(b) Simple plasma discharge apparatus(after [40.1])

may also be achieved when temperatures rise beyonda distinct threshold level of the molecule. With lessdifficulty than applying extreme amounts of heat, ion-ization can be reached inside an electrical field suppliedby an adequate voltage source. This approach causesa dielectric breakdown of the gas molecule into plasmaconsisting of charged species (electrons, ions, radicals)and neutral species (atoms, molecules, and excited par-ticles). The plasma thus becomes a highly energeticstate of matter characterized by an elevated electri-cal conductivity [40.2]. Strongly bonded molecules likehydrocarbons such as methane (CH4) in a plasma-induced state have been documented to disassociate.C1H4 ! C1HŒx<4 CHCH2/ then quickly recombineinto intermediate compounds such as C2H2, C2H4, andC2H6 [40.3]. This can be achieved and maintained aslong as the methane is exposed to the electrical field.The intermediate molecules are a short-lived speciesand will reform back into methane if not convertedwhile yielding some amount of molecular hydrogen.

Generation of Plasma FieldMultiple methods exist to produce a viable plasma field,the most practical being two electrodes in close enoughproximity supplied with a voltage creating an arc ofelectricity, or plasma, through the air gap between theelectrodes. This can be seen in Fig. 40.1b using par-allel conductive plates. The power supply drives thecurrent between the conductive electrodes ionizing thegas flowing through it. Pressure plays an importantrole for the generation of the field. At pressures below

Small Scale Catalytic Syngas Production with Plasma 40.1 Plasma 1143Part

D|40.1

atmospheric and above vacuum, the electrical energyrequired to create a plasma field arc or glow dischargeis reduced to a minimum. The augmentation of the di-electric strength of vacuum by matter and pressure isthe primary function for sustaining the plasma field.

Another factor in field generation refers to the ge-ometry of the electrodes. Industrially, these can varyfrom sharp singular points to infinitely long plates toannular arrangements physically similar in appearanceto concentric pipe heat exchangers. In the case of pro-cessing gas, the annular apparatus is highly efficientand isolates the system from the surrounding environ-ment. It is difficult to classify and analyze the plasmagenerated from an external electromagnetic field be-cause of the inconsistent nature of the ionized particles.The generated plasma is on a very short time andspace scale (10�7�10�9 s) as the excitation weakens,dissipating quickly upon departure from the appliedelectrical source. If not utilized properly, the plasmawill return to a stable gaseous ground state.

Stability and EquilibriumPlasma requires constant treatment via an external en-ergy source in order to maintain its phase stability,making it a short-lived species at ambient conditions.In a continuous process, it’s wise to consider the place-ment of the plasma generating field in order to optimizeoverall system efficiency. Ionization and dissociationare a requirement for the existence of plasma, but theseinduced conditions also cause a fluctuation in equilib-rium. Plasma density, also defined as electron or iondensity (ni), is the number of free electrons per unit vol-ume. Ion density can be used to determine the degree offractional ionization (xiz) of plasma represented by theequation

xiz D ning C ni

where ng is neutral gas density. If the degree of ion-ization is � 1 then the plasma can be considered fullyionized and if xiz � 1, the plasma is weakly ionized.No matter how minute the presence of ionization inthe gas phase, plasma characteristics still show to somedegree [40.3]. Threshold levels for ionization exist dur-ing excitation of a gas-phase substance and dependingon the amount of energy applied, different chargedand neutral molecules will be produced that are ofoverall equal constituents to the original gas molecule.These excited forms consist of rotational excitation,vibrational excitation, disassociation, and full plasmaionization. Rotational and vibrational excited states aregenerally considered impractical for most processesdue to their low threshold energy (< 2 eV) and short

lifetimes (nanoseconds) [40.1]. Electrical field energyis directly proportional to electron energy responsiblefor greater molecular decomposition from the electron-impact dissociation and ionization reactions. As anexample, bonding energy of most hydrocarbons is be-tween 3�6 eV. This range expresses C�H and C�Cbond strength varying from molecule to molecule ofany hydrocarbon in general. C�H bonds are strongerand closer to 6 eV in strength than that of C�C bonds,which reach the lower interval of 3 eV in strength. Anelectron-volt (eV) is energetically equivalent to about1:6�10�19 J (Joules). The following electron-impact re-actions of methane show the threshold energy barriernecessary to achieve specific levels of plasma excita-tion [40.3]:

� Vibrational excitation

CH4 C e� ! CH4.�24/C e�

Threshold energyD 0:162 eV

CH4 C e� ! CH4.�13/C e�


� Dissociation

CH4 C e� ! CH3 CHC e�

Threshold energyD 9 eV

CH4 C e� ! CH2 CH2 C e�


CH4 C e� ! CHCH2 CHC e�


� Ionization

CH4 C e� ! CH4C C 2e�


CH4 C e� ! CH3C CHC 2e�


Methane is a highly stable compound and, as a re-sult, requires more energy to break the threshold barrierthan most other hydrocarbons. This data can be usedas an initial guide for determining the necessary en-ergy requirement for any intended plasma application.Threshold levels affecting plasma stability lead to onepossible reason as to why the plasma state is almostnever in chemical or physical equilibrium with its sur-roundings. Since the discharges are electrically drivenand weakly ionized, the applied power preferentiallyheats the mobile electrons while the heavy ions ex-change energy through collisions with the background

PartD|40.1


gas [40.1]. At or below ambient pressure, the ion tem-perature (Ti) representing the overall temperature (T) istypically much less than the electron temperature (Te).An immense amount of energy is required for neutralspecies and charged species to be in thermal equilib-rium with each other (Te D Ti D T). This is known ashot plasma and the requirement differs from element toelement where, for example, a temperature of 20 000Kmust be reached for complete thermal ionization of he-lium while around 4000K is needed for cesium [40.1].When not in thermal equilibrium (Te � Ti), the phasestate is considered cold plasma meaning that the over-all temperature (T) of the plasma is at least equal to theambient system process temperature.

40.1.2 Types of Plasmas

As stated before, there exist many different types ofplasma, both natural and conventional (Fig. 40.2). Thefocus of this chapter is on one form of conventionalplasma used in a reformation process; it is still worth-while, however, to note other types. For instance, mostplasma is naturally occurring in the universe from starsand interstellar matter. An example of hot plasma, alsoknown as thermal plasma, is the Sun’s surface or a boltof lightning. A neon light can be considered nonther-mal plasma or cold plasma. This classification dependson temperature, electronic density, and level of energyof the medium [40.2].

Conventional plasma is synthetic and every syn-thesized type carries different micro- and macroscopicbenefits for process use. Plasma is also categorized bypressure in response to the method of ionization and thestructure of the electrical field. For instance, in very lowpressure environments plasma may always be consid-ered nonthermal. The following is a list of conventionalplasmas with descriptions grouped by pressure.

Low Field Pressure (0 atm< PPlasma < 1atm)Capacitive coupled plasma (CCP) and inductive coupledplasma (ICP) (Fig. 40.3a–d) are produced by high fre-quency RF electric fields where positive ion collisionsare the dominant factor leading to electron dissociation.The bombarding ions are directly affected by the mag-nitude of frequency, typically at the MHz level, capableof causing an electron or Townsend avalanche cascad-ing the electrons into a plasma state. CCP is generallyused in material processing to enhance surface proper-ties like that of integrated circuitry via etching or filmdeposition. A substrate is present upon one of the twointeracting electrodes that are aided by a positive sheathmaterial maintaining excess negative charges and direct-ing them appropriately. ICP utilizes a coiled electrode(Fig. 40.3c,d) placed outside of the reactive zone in-

stilling an electromagnetic field that negates the directcurrent (DC) requirement necessary to reach efficientpower in CCP. For analytical methods such as spectrom-etry, ICP provides a reliable detection limit over a widerange of sample elements. The RF power source canalso be directed into the plasma gap itself as opposed tothrough the electrodes for both ICP and CCP.

Cascaded ArcHigh-density, low-temperature cascaded arc is plasmain a relatively homogeneous thermal equilibrium. Thiscreates an exception to the hot-thermal plasma clas-sification since the gas, ions, and electrons are allthermally equal (� 1 eV) at ambient conditions. Thefield is stabilized by cascaded plates with a floating volt-age potential and is typically structured cylindrically.Operated under a DC supply, the power requirement isin the lower kW range (1�10 kW). This form of plasmaand device is used in material deposition and surfacetreatment as well.

Electron Cyclotron Resonance (ECR)and Helicon Discharge

Electron cyclotron resonance (Fig. 40.3a) and helicondischarge (Fig. 40.3b) are produced by an axial vary-ing DC magnetic field sustained by electromagneticcoils. The coils surround a cylindrical chamber in whichmicrowaves are injected through a dielectric windowpropagating a rotating polarized wave. This generatesthe source plasma where it can be directed into the pro-cessing chamber. Electron cyclotron resonance (ECR)is operated with a strong magnetic field versus heli-con discharge, which is much weaker by comparison.Helicon discharge uses an RF-driven antenna coupledwith the weaker magnetic field to create a resonant heli-con wave at the plasma source where it is believed thatthe particle-wave interaction is the cause of the waveenergy transfer into the plasma. Conventional areas ofinterest for these plasma devices include semiconductormanufacturing, electric propulsion devices, or in ad-vanced medical treatment.

Glow DischargeGlow discharge is produced by a low frequency electri-cal field induced by DC or RF in the open area betweeninteracting electrodes. Once adequate voltage is met,ionization occurs releasing a vibrant light based on theionized gas. The system is usually enclosed in a non-conducting transparent tube. Known for its usage in flu-orescent lamps and neon lights as well as plasma-screentelevision devices, glow discharge plasma also has ap-plications in analytical chemistry via spectroscopy andmass spectrometry and in material surface treatment viaa method known as sputtering.

Small Scale Catalytic Syngas Production with Plasma 40.1 Plasma 1145Part

D|40.1

20

15

10

5

25

Earths sionos-phere

Flames

Fusionreactor

Focus

Fusionexperiments

Solarwind

Solarcorona

Interplanetary

Solid Si atroom temperature

λDe >1cm

Lowpressure

Glowdischarges

Thetapinches

Laserplasma

Shocktubes

Highpressure

arcs

Alkalimetal

plasmas

log Te (V)

log n(cm–3)

043210–1–2 5

�nλ3De<143

Fig. 40.2 Synthetic and universalplasma on energy (logarithmictemperature voltage equivalent)versus density plot. De D Debyelength – the distance scale overwhich charge densities can existspontaneously or more simply, howfar the electrostatic effects travel. Asan example, low-voltage and undrivenplasmas are generally a few Debyelengths wide (middle portion of figure)(after [40.1])

High Field Pressure (PPlasma > 1atm)Plasma Torch and Thermal Plasmatron

Plasmatron is a high-powered DC thermal plasma gen-erated by electric resistive heating of a turbulent flow ofgas molecules. The plasmatron provides highly control-lable electrical heating, but requires resilient electrodesto withstand corrosion from the strong arcs and extremeheat. A relatively large amount of energy is required toharness a homogeneous field in thermal equilibrium duein part to the elastic collisions of the propagating ener-gized electrons. Utilizing an annular configuration, theinput gas is injected tangentially to produce a vortexthat extends the plasma arc field inside of the ionizationzone. Water is used as a heat sink to cool and controlelectrode temperature preventing thermal damage to the

components. The high power requirement pitfall limitswidespread use of the thermal plasmatron in industry,but the plasma torch finds a niche in plasma cutting,pollution control and hydrocarbon reformation.

Corona DischargeThis form of nonthermal plasma is generated throughfinely pointed electrode tips at a voltage high enoughto initiate a conductive field, but not so high to causean electrical arc discharge to the grounded path. Al-though the corona charge can be positive or negative,the positive corona is undesirable due to the extremelyhigh energy requirement to excite positive ions. This isnot the case for a negative corona, in which electronsare much lighter and more easily ionized in a non-

PartD|40.1


RF bias

RFRF

Multipoles

Helicon

RF Antenna

ECR

Microwaves

InductiveHelical resonator

a) b)

c) d)

Fig. 40.3a–dExample sourcesof low-pressure,high-densityplasma in basicconfigurations(after [40.1])

thermal low-pressure setting. The Townsend avalanchecauses the highly energized and separated electrons tomove radially in all directions creating a plasma cloud.As the electrons move further and further away fromthe voltage point, they lose energy and eventually areno longer able to force ionization as they begin to ap-proach the grounded electrode. The field, or cloud ofelectrons, maintains ionization for as long as a volt-age is supplied. Corona discharge is a method forozone production, volatile organics (VOCs) removal,air clean-up, and water sanitization as well as for an-alytical techniques.

Dielectric Barrier Discharge (DBD)An insulating dielectric barrier is used to separate,including the space gap, the field producing elec-trodes that uses low-RF or microwave frequency froma high voltage alternating current (AC) (1�10 kV)

power source. The material used for the dielectric bar-rier ranges from polymers to glass or ceramics, allnonconductive in nature. Planar or axial configurationsmay be utilized depending on flow applications and canbe seen in Fig. 40.4. DBD is another form of nonther-mal plasma that can be generated at ambient pressureand low temperature that is not in total particle equilib-rium. Microdischarges occur at the dielectric surfacesand propagate like a glow discharge while maintaininghomogeneity, but within a more pressurized environ-ment. Originally designed for ozone generation, DBDsfind use in catalysis enhancement, surface modifica-tion of materials, emissions control, excimer lamps, andsilent-discharge CO2 lasers.

Gliding arcGliding arcs are arc discharges that initiate from thepoint of shortest distance between parallel, but diverg-

Small Scale Catalytic Syngas Production with Plasma 40.2 Partial Oxidation Reformation Using Cold Plasma 1147Part

D|40.2

High voltageelectrode

Dielectric barrierDischarge gap

Discharge gap

Ground electrode

Highvoltageelectrode

Dielectric barrier

Ground electrode

Highvoltage

ACGenerator

a) b)

Fig. 40.4a,b Configurations of different DBD plasma devices (apparatus similar to other plasma devices minus thedielectric barrier) (after [40.4])

ing electrodes. The arcs travel along the length of theelectrodes until complete breakdown of the arc occursnear the endpoints of the divergent electrodes [40.5].As the gliding arc dissipates, the encompassing tur-bulent gas maintains a glow discharge plasma state.This continuous cycle is determined by the systemparameters, which include characteristics of the elec-trical power source (AC or DC), range of current andvoltage capacity, gas type, gas velocity, electrode ge-ometry, and field pressure. Associated with a high levelof current (� 50A=arc discharge), this thermal plasma

reaches extreme temperatures (T > 10 000K) energiz-ing all the chemically active species (CAS) presentin the medium. The power requirement varies from1�10 kW based on the parameter conditions. Glidingarc can be considered a hybrid form of thermal andnonthermal plasma reaching only temporary equilib-rium with traits similar to the plasmatron and glowdischarge. Some of the many applications of glidingarc plasma include metallurgy, chemical reformation,reformation of flue gas, dissociation of CO2, and de-contamination.

40.2 Partial Oxidation Reformation Using Cold Plasma

40.2.1 Methane Reformation

Industrial practices often vary in regards to reformationtechniques as the end goal is unique to each process.Hydrocarbon reformation can be utilized as a supple-ment to larger-scale processes by making use of utilitystreams and side products that would otherwise bea detriment to overall system efficiency. For instance,when a surplus amount of steam is generated after pri-mary and secondary cooling, it can be in the engineer’sinterest to make use of it. If a steady supply of natu-ral gas is also available, then steam methane reforming(SMR) could be a source of additional revenue fora wide range of production applications, the most com-mon and basic of these being the synthesis of hydrogenor methanol. Below are the ideal reactions that accom-pany the SMR process:

� Steam methane reformation (SMR)

CH4 CH2O � 3H2 CCO

�Hrı

298K D 206 kJ=mol

� Water gas shift (WGS)

H2OCCO � H2 CCO2

�Hrı

298K D�41 kJ=mol

� Methanol synthesis

2H2 CCO � CH3OH

�Hrı

298K D�91 kJ=mol

Being diverse in its usage and easily storable,methanol typically equates into a positive revenue

PartD|40.2


source. It is possible to deploy SMR as a primary pro-cess, but would require large-capacity equipment anda high rate of production to compensate for the highcost of generating steam given SMR’s low conversionefficiency. The SMR process is well documented andcurrently used at the commercial level, but is only onemethod of methane reformation and the focus of thischapter is on partial oxidation reformation techniques.

Reformation with the goal of hydrogen productionhas seen real promise for small-scale and on-hand ap-plications like fuel cell technology or enhancement ofinternal combustion engines (ICEs). The advantagesof hydrogen include fast flame speed and high octane(which enables high-pressure operation), allowing ei-ther high compression ratios or strong turbochargingwith associated downsizing. The use of hydrogen-richgas addition to spark-ignited (SI) engines allows forhighly diluted operation. It is possible to extend thedilute-operation burn limits where nitrogen oxides arereduced to near nonexistent levels, eliminating the needfor NOx after-treatment (thus removal of the three-waycatalyst). Dilution also allows for improved efficiency,from either reduced pumping losses at light load orlower in-cylinder temperature, thus reducing heat trans-fer to the cylinder walls. Running with hot hydrogen-rich gas further improves the flame speed as well asextends the dilution limit. Faster flame speed is anothercontributor to increased engine efficiency by enablinga more isochoric combustion zone.

Hydrogen is unfortunately a difficult moleculeto store for extended periods of time. Aside froma high mass heating value, 120 kJ=g compared tomethane (50 kJ=g), methanol (20 kJ=g), and gaso-line (42:8 kJ=g), hydrogen is characterized by a lowvolumetric heating value due to its low density(2:016 g=mol): 11 kJ=l versus 16MJ=l for methanol atstandard temperature and pressure [40.2]. Volatility,flammability, and small molecule size all add to thestorage difficulties of hydrogen. Some presented solu-tions to these issues are cryogenic and high-pressurecontainment, which have some validity but are not idealdue to further increase in efficiency losses and ex-plosion hazards that are deemed high risk for mobileapplications. Methanol has been used on occasion asa storage medium for utilities like fuel cells that are typ-ically designed to operate on a source of hydrogen.

The integral step of methane reformation is theproduction of synthesis gas or syngas, a mixture of hy-drogen and carbon monoxide. Often, realities of opera-tion interfere, producing a syngas with adulterants suchas carbon dioxide, methane, water, oxygen and nitro-gen. Generally, these side constituents are unfavorableand limit overall efficiency in downstream processingrequiring minimization when possible throughout oper-

ation. The methane POX reaction and other macroreac-tions of concern that are associated with methane POX(which include the SMR reactions above) are listedbelow. The following reactions can be considered com-petitive as well as reversible in nature:

� Methane partial oxidation

CH4 C�1

2

�O2 � 2H2 CCO

�Hrı

298K D�36 kJ=mol

� Combustion reactions

CH4 C 2O2 � 2H2OCCO2

�Hrı

298K D�891 kJ=mol

H2 C�1

2

�O2 � H2OC heat

COC�1

2

�O2 � CO2 C heat

Other Possible Side Reactions� Thermal decomposition

CH4 � CC 2H2

�Hrı

298K D 122 kJ=mol

� Dry methane reformation

CH4 CCO2 � 2H2 C 2CO

�Hrı

298K D 247 kJ=mol

� CO reduction (undesired)

COCH2 � CCH2O

�Hrı

298K D�84 kJ=mol

� Boudouard reaction (undesired)

2CO � CCCO2

�Hrı

298K D�172 kJ=mol

� Methanation (undesired)

COC 3H2 � CH4 CH2O

�Hrı

298K D�206 kJ=mol

CO2 C 4H2 � CH4 C 2H2O

�Hrı

298K D 42 kJ=mol

2COC 2H2 � CH4 CCO2 C heat


D|40.2

Partial oxidation is an exothermic reaction capableof producing an ideal syngas product of a 2 W 1 ratio ofhydrogen to carbon monoxide. The reaction is able toreach temperatures in excess of 1300 ıC that need tobe levied and maintained through use of a catalyst inorder to avoid nitrogen-based reactions. Nitrogen ox-ides (NOx) are considered an emission that have, byorders of magnitude, a much greater environmental im-pact than unburned hydrocarbons and carbon oxides.Not only does NOx cause the formation of ground-levelozone, but it also has negative effects upon the humanrespiratory system. The rate of formation of NOx in-creases exponentially as temperatures rise in a lean andoxygen-rich flame zone during partial or total oxidationwith air. Nitrogen, if not removed before reformation,must remain inert throughout the process. This factoris taken into consideration along with reactant conver-sion and hydrogen yield, all of which are influenced bythermal conditions.

Thermal EquilibriumNot including plasma interaction, the factors that af-fect syngas production are reactor size, space velocity,flow dynamics, catalyst type, reactor and catalyst ge-ometry, feed material and feed ratio, system pressure,and thermal equilibrium. Exothermic processes are con-trolled not only for safety reasons, but also for systemefficiency and minimization of side products. Thermalequilibrium helps justify why the SMR reaction is inef-ficient and energy intensive while also explaining POXefficiency and NOx formation.

The tubular reactor is commonly employed in in-dustry and most often utilized for homogeneous gas-phase reactions. It is assumed that the reactants areconsumed over the length of the reactor while experi-encing some concentration variance both radially andin the direction of flow. This forces the rate of reac-tion to adjust accordingly as reactants are consumed andproducts are formed that become intermediate reactantsthemselves. In a POX reactor, the plug flow apparatusis selected to maximize production rate. The reactorcan be used with, or without, a form of packed cata-lyst; when under a turbulent flow regime, equilibriumis maintained in a quasisteady state of operation creat-ing a relatively stable, predictable thermal profile. Asthe process approaches completion under this configu-ration, different zones of the reactor experience a riseand fall in temperature due to the multitude of occur-ring reactions. This effect is more apparent in a shorteramount of reactor length using a packed-bed catalystdesign and even more so if the reactor diameter is re-duced. The thermal profile and reactions that take placeare directly proportional to the ratio of fuel and oxidizer.More specifically, a carbon-to-oxygen atomic ratio of

1 correlates to stoichiometric POX while a ratio of0.25 gives way to stoichiometric and complete com-bustion. Oxygen, within these stoichiometric bounds,is depleted quickly after injection due to its fast reac-tion rate and will not be present in the product stream.Oxygen can only be present postreactor if the systemis not functioning properly or if lean operation is beingconducted. Any ratio < 0:25 is considered lean (or fueldeficient) combustion and a ratio > 1:0 inactivates thereaction by extending the gas mixture past its flamma-bility and reactive limits.

In order to lower reaction temperature, eliminateNOx formation, and increase methane conversion, a cat-alyst is typically present inside the reactor. Currently,catalytic reformation is commonplace in the industryand as so, it is the most economically well-developedtechnique for hydrogen production. Typical POX cata-lysts include Pt, Rh, Ni, Co, Ir, Pd, or Ru metals thatare supported on metal oxides, which include Al2O3,CeO2, MgO, and TiO2 in different proportions, mix-tures, and geometric configurations [40.3, 6]. Thoughdifficult to fully model theoretically, the experimentalreaction pathways using catalysis are consistent andrepeatable. Initially, exothermic reactions of POX andcombustion are activated together to provide heat to thereactor. This spike in temperature allows the intermedi-ate products of H2O, CO, and CO2 to then immediatelyreact via the SMR andWGS reactions with unconvertedhydrocarbon as well as the hydrogen product. Due tothe continuous nature of the plug flow reactor and highgas space velocity, the residence time of the reactivespecies is short (< 1:0 s). The short residence time inrelation to the competing exothermic and endothermicreactions forces the process to reach completion that isnot necessarily preferential to the best possible results.

Thermal equilibrium tends to develop in localizedspots of the POX reactor and is contributed mostly fromgas flow convection, downstream or upstream radiation,catalytic reaction, and heat conduction of the reactorand catalyst wall [40.7]. Even through well-controlledoperation, the active species are always in a constantstate of reactive flux. The bed temperatures along thelength of the reactor help justify these claims, whichhave been seen thorough experimentation and are welldocumented in literature.

Maximization of hydrogen yield and selectivity canbe achieved through manipulation of reactor tempera-ture. Figure 40.5 shows this relationship and providesan example of thermal equilibrium influencing theproduct composition [40.7]. It is easy to see the asymp-totic increase in conversion and the perfect selectivitythat is achieved as temperatures rises in the reactor.

The syngas yield of purely two hydrogen moleculesto each COmolecule from a 1.0 ratio of C/O would the-

PartD|40.2


0.95

0.90

1.00

0.85

Temperature (K)

CH4 conversion

H2

CO

Selectivity

0.90

0.80

0.60

0.70

1.00

0.501300120011001000 1400

Fig. 40.5 Experimental POX using Rh-coated foammonolith catalyst. CH4 conversion with H2 and CO selec-tivity versus temperature (after [40.7])

oretically be perfect for downstream applications likemethanol production, but combustion is unavoidableduring POX making this idealization next to impossi-ble. It should be noted, that some CO2 is necessary formethanol production as well, and thoughmostly viewedas a negative aspect to the process, combustion can becontrolled to assist in reformation. As more oxygenis introduced to the feed, hydrogen yield tends to in-crease until a peak is reached (C/O ratio � 0:85) wherefurther addition of oxygen begins to lower hydrogenyield. Peak ratio is unique to each process with high de-pendency on reactor structure and plasma or catalysisassistance. The stability of feed ratio also contributesgreatly to thermal equilibrium and, from a mechan-ical standpoint, reactor temperature provides a quickindication of syngas quality and possible system defi-ciencies. If the reaction is controlled solely by thermalactivation then product selectivity is predictable, butwill limit maximum product potential. When plasmabecomes a dominant presence in the system, thermalequilibrium is overcome and the achievement of an evengreater efficiency is then possible [40.8].

Chemical KineticsThe reformation process can be summarized by the en-thalpy changes exhibited throughout the system. Eachreaction either imparts or removes heat in an effort toapproach completion. The POX reaction itself has rel-atively slow kinetics and cannot initiate without theassistance of energy from an external source or an inter-nal source such as combustion. Overall, the process ofPOX is still exothermic even with the multitude of asso-ciated competitive endothermic reactions. In the event

that the total net enthalpy change to the system is ap-proximately zero (�Hr

ı � 0 kJ=mol), the process canthen be considered autothermal. In order to compensate,some form of direct injection is required using water,CO2, a process recycle stream, or simply the exhaustfrom an internal combustion engine:

� Partial oxidation

CnHm C�n2

�O2 � nCOC

�m2

�H2

�Hrı < 0

� Steam reformation

CnHm C nH2O � nCOCh�m

2

�C n

iH2

�Hrı > 0

� Combustion

CnHmC�.2nCm/

4

�O2

� nCO2 C�m2

�H2O �Hr

ı � 0

� Autothermal reformation (ATR)

CnHmC�n2

�O2 C nH2O

� nCO2Ch�m

2

�C n

iH2 �Hr

ı � 0

The above global reactions give a straightforwardoverview and are the initial basis for modeling anyreformation process. Reaction kinetics can then be uti-lized to help more thoroughly map the intrinsic systemproperties for determination of syngas product compo-sition and yield. Temperature, pressure, and phase allplay a role in the mechanism determining the path-way for reactions to follow. The homogenous gas phasereactions of POX do not inherently promote the for-mation of a high-quality syngas; this problem is partlyresolved with the insertion of a durable and appro-priate supported catalyst placed within the reactivezone [40.6]. The heterogeneous catalytic environmentenhances the normal molecular pathway via active siteson the catalyst that reduce the energy barrier requiredfor stoichiometric POX and lower the amount of com-bustion heat necessary for reaction stability. In a priorstudy, the CPOX of methane on a rhodium-coated cata-lyst has been indirectly modeled based on combustion,steam reformation, water gas shift, and dry methane ref-ormation (DMR) kinetics. These equations, which arelisted below, follow the Arrhenius model and have been


D|40.2

determined using empirical methods [40.9, 10].

RCombustion D 2:119�1010 exp��80

RT

�p0:2CH4

p1:3O2

RSMR D 1:0�10�9T5 exp

��47:3RT

�

��pCH4pH2O �K�1

eq,SMRp3H2pCO

�

RDMR D 7:0�10�7T4 exp

��32RT

�

��pCH4pCO2 �K�1

eq,DMRp2H2p2CO

�

For the WGS reaction (r D steam/CO ratio)

RWGS D k�pCOpH2O �K�1

eq pCO2pH2

�

kHTSR D 1:78�1022.1C 0:0097r� 1:1364r2/T�8

� exp

��70RT

�

kLTSR D 1:74�1017.1C 0:154rC 0:008r2/T�8:5

� exp

��35RT

�

Overall thermodynamic equilibrium constant (Keq)

Keq D exp��Sı

R� �Hı

RT

�.Patm/

PNjD1 v

00

j �v 0

j :

The equilibrium constant (Keq) takes into account thereversibility of each of the process reactions. Thismust be addressed due to the variability of conditionsduring operation. As an example, the WGS reactionapproaches completion at two different temperatureranges. The high-temperature shift (HTSR) and low-temperature shift (LTSR) are performed in the rangesof 300�500 and 200�300 ıC respectively. Lower tem-peratures support better CO conversion, but the kineticsfavor an increased reaction rate at higher tempera-tures. The WGS reaction is always present, shifting theproduct composition towards thermodynamic equilib-rium between CO, CO2, H2, and water. The subsequentproduction of hydrogen-rich syngas with minimal COcontent, in particular to polymer electrolyte membrane(PEM) fuel cell application, is also based on WGS ki-netics [40.11]. Industrial-scale applications looking tomaximize hydrogen production tend to operate bothHTSR and LTSR in separate reactors in series down-stream of the reformation reactor. The use of bothshift reactors as ancillary units downstream from thereformer allows for near-total conversion of CO intoCO2 and provides additional hydrogen from steam with

further intent to isolate the product hydrogen. Thishappens through the most economical form of sepa-ration possible and can be accomplished using one ofthe several following techniques: pressure-swing ad-sorption (PSA) to separate hydrogen, PSA to separatethe CO2 followed by condensation of the remainder ofwater, and membrane or sieve separation of the hydro-gen [40.12]. With regards to the overall POX process,WGS is unavoidable and can be seen as neither bene-ficial nor detrimental to downstream application. Thisis especially true when engine combustion or electricalgeneration directly follows syngas production, as CO isalso a suitable, burnable energy source.

40.2.2 Plasma Catalysis

Low-input energy capable of producing highly activespecies has become the focal point of potential for de-velopment of more efficient plasma reactors. In plasmacatalysis, there exist a number of different configura-tions that rely upon location of the plasma field andcatalyst zone. If the placement of the catalyst zone isprior to or much further downstream of the plasma field,then the process is a two-stage system (Fig. 40.7). Hotthermal plasma imparts extreme amounts of heat en-ergy that cannot be used in conjunction with catalysisdue to the structural limitations of catalytic material.The coating and the substrate are unable to withstandthe high temperatures of the plasma field resulting inimmediate thermal degradation as well as sintering ofthe coating material. In the immediate area surroundingthe thermal plasma field, the cost of processing equip-ment and piping material would also rise in an effortto resist the high temperatures using higher durabilityequipment. This leads to applications in which thermalplasma has very limited appeal because of the ineffi-ciency from the high energy demand and the inability tointegrate into established chemical processes. For ther-mal plasma to have a role in catalysis, ample distancebetween the catalyst and plasma field is required, elim-inating most of the potential synergy of the combinedsystems.

When the catalyst zone is located inside the plasmafield or in close proximity postplasma generation, theprocess is known as a single-stage system (Fig. 40.6).In a single-stage plasma catalysis system, the effects ofplasma and catalysis that contribute to the process be-come dependent on each other. This differs greatly fromconventional catalysis due to the introduction of highlyactive species and the creation of a voltage potential andcurrent flow by result of plasma discharge across thecatalytic surface. The active species alter the status ofthe gas-phase reactions and associated reaction kinetics,which no longer ensues from the ground state. Elec-

PartD|40.2


R*

R

ET

Ea

Gas-phase

Diffusion DiffusionReaction DesorptionAdsorption

Boundarylayer

Catalystsurface

Fig. 40.6 Illustration of vibrationally excited species from a plasma-induced state accelerating a catalytic reaction. Rrepresents the gas-phase molecule at ground state (uncharged) and R� stands for the reactant in a vibrational state(charged). Ea is the activation barrier of chemisorption for R and ET is the threshold energy for the formation of R�

through a plasma-induced electron-impact reaction (after [40.3]). The thermal energy required to initiate the catalyticprocess is now much less than the standard catalytic reaction and even further so than the reaction without catalyst

a) b)

+++++++ +

+++++++ +++++

++++

++

25 mm

90 mm

High voltage

Fig. 40.7a,b Schematic drawing ofrotating arc reactor used in a priorstudy conducted by Dae Hoon Leeet al. Side view (a) and overheadview (b) (after [40.13])

trical field enhancement depends on curvature, contactangle, and the dielectric constant of the packing pellets.Higher electric field leads to higher electron densitypromoting the electron-impact reactions responsible forhydrocarbon reforming. The decomposition of hydro-carbons is attributed to ionization and electron-impactdissociation of the excited molecules. Simply through

vibrational excitation it has been noted that activationenergy is reduced in gas-phase reactions.

Physical adsorption and chemisorption are themethods where gas-phase molecules or ionized parti-cles attach to the surface of the catalyst (Fig. 40.8).Physical adsorption is slightly exothermic and is mainlyattributed to the weak Van der Waals attractive forces


D|40.2

Heat transport in solid wall

Transport ofmomentum,

energy,species

Diffusion

Adsorption,surface reactions,

desorptionThermalradiation

Gas-phasereactions

Fig. 40.8 Re-action effectsflowing througha catalyticmonolith, whichinclude transportprocesses anda microscopicview at thenanoscale level(after [40.11],reproduced withpermission ofThe Royal Soci-ety of Chemistry)

between the interactive surfaces. As the temperatureprofile surrounding the catalyst rises, the adsorption ofparticles steadily declines until critical temperatures areobtained in which physical adsorption no longer oc-curs. At this point, physical adsorption gives way to thesole influence of chemisorption. In the rate-influencingstep of chemisorption, valence forces create highly re-active instances via bond stretching of the exposed andabsorbed chemical species on the surface of the cata-lyst. Chemisorption produces heat comparable to theexotherm of the chemical reactions being performed.Atom and metastable compound formation along withelectron multiplication cascade into the catalyst bedmanipulating the normal conditions in which the cat-alyst performs. These effects are apparent and usuallyhappen under very short electrical contact times. Theheightened interaction between the surface and adsor-bate then create new bonds, forming molecules that arespecific to the catalyst and operating conditions.

40.2.3 Reactor Apparatus

In a lab setting, a reactor that imparts all of the above-listed chemical and physical effects has been con-structed. The proprietary unit, known simply as PRISMand shown in Fig. 40.9, has demonstrated successfullythe ability to reform methane gas into syngas at highefficiency and near perfect conversion using very lit-tle energy input. The process incorporates a three-stepapproach coinciding with intermediate stages that con-tribute to the final composition of the produced syngas.

At ambient or elevated temperatures, a stream of dryair or pure oxygen is introduced to a stream of naturalgas and is well mixed prior to exposure of the plasma-

generating field within the plasma head. The plasmahead is a precision-designed, stainless steel structureconsisting of a tapered rod that acts as the anode, whichis housed inside a cylindrical wall of like material (thecathode). The point of largest diameter along the lengthof the anode rod is where the plasma field initiates andcan also be considered the choke point of the flowing el-ement. The axial design makes use of the annular gap tohold steady the electrical field. This closely resemblesthat of a dielectric barrier discharge (DBD) apparatus,but without direct internal insulation while incorporat-ing more of a coronal discharge approach to the plasmatype utilized. In place of a fine-tipped point that is typ-ically seen in a coronal design, the annular region isused to grant a larger volume coverage that increasesthe duration of electrical exposure to the gas moleculesensuring sufficient ionization. The plasma chamber ofthe PRISM mimics the flow patterns of the schematicshown in Fig. 40.9 for the rotating arc reactor, but dif-fers in dimensionality and in orientation. The radialinjection points allow for a toroidal arc that furtherincreases residence time within the plasma head appa-ratus. As concentrated charges dissipate from the addedrotational turbulence, both metal erosion and degrada-tion upon the electrodes are reduced, extending theiroverall lifetime. An observation of the anode after long-term operation reveals minor impressions of spiral-typemarkings across its metal surface indicating the tangen-tial path taken by the flowing gases.

The annular chamber provides uniform exposureto the incoming gas assisted by the well-mixed effectof the turbulent flow regime promoting molecular in-teractions and ion collisions, while the toroidal flowlengthens the space and time of the overall exposure.

PartD|40.2


These factors all contribute to the formation of thehighly energized cold plasma state necessary for the ref-ormation stage of the process.

An external power source is placed above the re-actor providing voltage directly to the electrodes thatare held inside a barrier of ceramic/quartz in order tomaintain the uniform three-dimensional electric fieldthroughout the annular gap. The power supply operatesat high voltage (� 3 kV) and very low current expend-ing no more than 100W of combined processing powerat maximum feed flow rate. Voltage to the plasma headis variable based on the fluid properties, which are pri-marily defined by gas density and velocity. The PRISMsystem operates with a 12C kV compact transformerthat allows for additional reserve voltage when fluidproperties drastically change, and will adjust automati-cally without manual input control to sustain the plasmaarc field.

Thermal expansion follows quickly after the gasstream is ionized by the plasma field. From there, theflow path continues by entering the main portion ofthe reactor that houses the proprietary catalyst craftedspecifically for promoting the POX reaction. Catalystcomposition is a mostly nickel-based coating preparedupon an alumina substrate. These monolithic, cylindri-cally sculpted blocks of catalyst remain stationary ina tightly packed fashion; these extend to the inner re-actor wall.

In this chamber, the core of the partial oxidationreaction takes place. This is monitored through temper-ature readouts using multiple thermocouples drilled tothe center of each catalyst. The external body of thereactor is 2200 in length and roughly 8:500 in diameter.The inner wall of the reactor is lined by a nonconduc-tive refractory material of high alpha alumina roughly100 in thickness. This refractory participates as an insu-lator, but more importantly ensures containment of theelectrical properties that were imparted to the incomingionized gas stream. Figure 40.9 is an example schematicof an older generation model of the PRISM reactor.

Zonal ReactionsAs listed in the section of methane modeling reactionsthat occur throughout the process (Sect. 40.2.1), eachzone of the reactor and catalyst bed itself experiencea fluctuation in primary reactions when favorable con-ditions arise during each reaction. This, however, canbe manipulated through a well-controlled system. Ini-tially, the plasma zone operates as an ignition source forthe controlled combustion heating to the catalyst bed.When optimal bed temperature is achieved, the plasmafield then takes responsibility for molecular breakdownby ionization and electron dissociation of the exposedcompounds. Oxygen ionization into its atomic state and

AC

Air

Fuel

Fixed bed catalyst

Syngas

Plasmaelectrode

⎧⎪⎪⎪⎪⎨⎪⎪⎪⎪⎩

Fig. 40.9 PRISM reactor body and internals

the formation of ozone molecules are the first reactionsto occur. Not only is ozone formed, but decompositionis also apparent in the complementing stream of hy-drocarbons. These feed streams are well mixed and ina preferential syngas yielding ratio prior to introductioninto the plasma zone. Hydrogen atoms released fromC�H bond breakage and excited electrons that are ini-tially released contribute to the reactions downstreamas well as to the final syngas product. The remainingintermediate hydrocarbon compounds also react morereadily in their hydrogen-depleted configurations. Thisfirst step helps drive the reaction rate of POX up toa level sufficient for effective overall system operation.

In the subsequent portion of the reactor, the diame-ter is increased to allow expansion of the gases grantinglonger exposure to the high surface area of the catalyst.Even distribution from turbulent flow of the particlesonto the active sites of the catalyst is observed throughstability of the temperature profile. Given a consistentfeed composition at a specific ratio during steady-stateoperation, the specific zonal temperatures fluctuate ata �T near ˙1:0 ıC. This suggests all the participat-ing reactions are held constant along the length of thereactor in their corresponding zones. As stated prior,oxygen atoms are the first particles to be consumed


D|40.2

and do not appear downstream of the reactor. If theoxygen atoms are not reduced with the carbon sourceinto CO, then the atoms along with some CO willbe fully oxidized into CO2 before leaving the catalystzone. The POX reaction not only produces the syngasmixture, but also releases sufficient enough heat to sta-bilize the reactor and perpetuate the overall reactionaryprocess. Due to the fuel-rich environment and highprocess temperature, the unconverted methane tends toreact with the steam side product via the SMR reac-tion also enhancing the syngas ratio [40.14]. The finalstage falls to the WGS reaction and the reverse WGSreaction, which augment the CO2 component formingeither more hydrogen or more CO. The WGS reactionis not optimized in this configuration and purely partic-ipates as an indirectly controlled and closely monitoredside reaction. Should the syngas ratio require more COthan hydrogen or vice versa, then a downstream WGSreactor can be implemented at a low additional capi-tal cost to the system. The combination of reactionsresulting in a net exothermic process leaves the prod-uct stream hot (> 600 ıC) requiring heat exchange, ifdesired, to cool the syngas as well as condense anyresidual water.

Thermal ProfileWith the operation of the plasma head being undera cold plasma configuration, there is not a significantenough generation of heat to give cause for thermal

Back heat shield(monolith)

T(K)

T(K)

1200

700

1260

600

6

4

2

8

02520151050 30

2 slpm

6

4

2

8

02520151050 30Axial (mm)

Rad

ial

(mm

)

Axial (mm)

6 slpm

Catalyst(rhodium coated monolith)Front heat shield

(foam)

OutletInlet

Fig. 40.10 Example POX catalysttemperature profile in laminarsetting at differing flowrates. Thecatalyst coating is rhodium andthe hydrocarbon is iso-octane(after [40.11], reproduced withpermission of The Royal Society ofChemistry)

readings of this zone. Temperature measurement fora predictable profile is thus only conducted in the mainbody of the reactor housing the fixed bed catalyst,which is monitored via eight equally spaced thermocou-ples. The thermocouples are placed in predrilled holesof appropriate dimensions and sealed tightly at the pro-truding ends that are connected to the wiring relay. Thetemperature of the catalyst bed is directly linked to theexo- or endothermicity of the reactions, and also to theconditions of the catalyst, plasma head activity, systempressure, and gas flow rate. As a comparative examplethat provides an accurate visual of catalyst temperaturedistribution, this noninduced flow in a laminar to tran-sitional regime produces the temperature profile shownin Fig. 40.10.

The effect of modifying the flowrate through thecatalyst suggests that at higher flows, the reaction en-compasses a larger volume and initiates further down-stream within the catalyst itself. The PRISM differsonly slightly in flow path due to its transitional-to-turbulent regime reducing the radial variability of keyparameters.

The thermal profile shown in Fig. 40.11 representsthe consistency in occurring POX reactions down thelength of the catalyst. In actual operation where flowrate and ratio are unaltered set points, the bed tem-perature and product composition do not deviate bymore than the error present in the sensors-measurementequipment. The overall temperature drop within the

PartD|40.3


1917151311975 23 25

960

940

920

900

880

860

840

820

800

780

Temperature (°C)

Air flow rate (scfm)21

Thermocouple 2Thermocouple 3Thermocouple 4Thermocouple 5Thermocouple 6

Fig. 40.11 Exper-imental operationof a PRISM re-actor displayingaverage temper-ature readingsfor each cata-lyst zone overincreasing ratesof flow. Theoperational ratiowas adjusted onlyto the automaticfeedback controlfor the temper-ature rise in thesystem. Theinitial set point ofthe C/O atomicratio was 0:55,which increasedto 0:57 at thecompletion of theexperiment

catalyst bed between thermocouples 2 and 6 is nomore than 60 ıC translating to good thermal stabilitythroughout reactor operation. The rise and fall of zonaltemperatures correspond directly to each other and aredependent on the prior zone’s temperature, and evenmore so the active zone temperature (thermocouple 2).At each discrete flow set point, an accurate represen-tation of the PRISM’s temperature profile can be ob-served if no manipulation of flow rate had been done. Itshould be noted that the PRISM under this experimental

scheme was operated closer to a combustion-orientedratio at 0.55 of C/O species. As expressed previouslyin Sect. 40.2.1, Thermal Equilibrium, it was stated thatpeak hydrogen yield is achieved at � 0:85 C/O ratio.While this holds true, results were more than adequateciting consistent results of over 99% methane conver-sion and a� 1:8 H2/CO product ratio. It stands to showthat improvements can still be made on this alreadyhighly effective process, but are not necessary for vi-ability of the technology in the field.

40.3 Cold-Plasma-Assisted Experimentation

40.3.1 Steady State Operation

Many experiments have been conducted at a facilityin Brooksville, FL through research as well as trialand error. The PRISM system in a standalone config-uration is ready for early-stage commercialization forstable, continuous production of usable clean syngas.Another experimental-phase application has been per-formed for syngas-hydrogen enhanced combustion inspark-ignited (SI) engines. The benefits of hydrogenfuel addition during combustion are well known to thescientific community, and the difficulties of its imple-mentation have been discussed. A skid that consists ofa General Motors 8:1 l overhead valve/eight-cylinder

V configuration (OHV/V8) natural gas, turbochargedengine and a 100 kWe corresponding alternator havebeen integrated with a single PRISM unit for successfulcontinuous slipstream engine operation.

The PRISM is a relatively simple system requiringa fuel valve, air valve, power source, and control sys-tem along with some corresponding instrumentation. Incontinuous operation, the system is stable under a well-controlled apparatus of a quick response ratio feedbackloop based on system temperature. Attached are massflow controllers that dictate the feed input based onratio control and oxidizer flow rate. Adjustments canbe allocated to either set point resulting in an auto-matic change in fuel flow. The feedback control loop

Small Scale Catalytic Syngas Production with Plasma 40.3 Cold-Plasma-Assisted Experimentation 1157Part

D|40.3

is responsive to changes in catalyst bed temperature,but focused on the main zone of operation, which isalso the highest temperature zone of the process. Thiskeeps reactor conditions favorable for higher hydro-gen production and safe during long term operation.The control mapping has been optimized through manyphases of testing on the PRISM unit. Manual control in-put is unnecessary when consistent homogeneous fuelsare being used as the system will respond accordingly.This follows suit for the power supply, which operateswithin its voltage limits to always provide the elec-trically induced plasma field within the plasma head.The simplicity of the system and stability of the reac-tor make it possible for the need of only one operatoroverseeing the process. As syngas is being produced,it is then guided through heat exchange removing theformed water and lowering gas temperature to ambientlevels. After a final water knockout point, the syngasis led to a low-tolerance oxygen sensor confirming theabsence of oxygen before compression and then to stor-age of the syngas product. Excess syngas is sent to flareduring experimentation.

When added to the 8:1 l engine skid, the PRISMfunctionality remains the same; filtered air is either ac-tively fed to the reactor or naturally pulled throughthe reactor via the engine requirement. The controls inplace respond accordingly to either arrangement. In thiscase, the gen-set engine is brought to steady operationfirst, followed by the PRISM reactor, which is fully in-troduced to the engine fuel inlet. The natural gas valveis then slowly closed until the engine is no longer be-ing fed with fuel other than freshly produced syngas.From this point, system operation will continue untilshutdown. It must be noted that there is a drop in en-gine power return when fully suppliedwith only syngas,which is mainly due to the excess dilution of nitrogenfrom air being processed twice during the same systemloop. Syngas can also be introduced hot or cold, elimi-nating the need for heat exchange if deemed unrequired.The main focus of this system is to eliminate NOx pro-duction and the need for exhaust gas after-treatment.Using a five-gas analyzer, NOx levels have been rou-tinely recorded at less than 10 ppm prior to after-treatment while using slipstream syngas operation.

40.3.2 Transients: Startup/Shutdown

The startup sequence for operation of the PRISM beginswith simply turning on the power supply and openingsystem flow. The plasma field is on at all times andinitiates the process acting as a spark to ignite a verylean mixture of air to fuel. These gases are well mixedbefore introduction to the plasma head chamber. Astemperatures rise, the fuel mapping for the controls ap-

proaches a fuel-to-air ratio that is richer in operation.Above 900 ıC the methane will flood the system andbe too rich for any oxidation, thus cooling the catalystbed and stabilizing the reactor conditions. At this point,the temperature of the system reaches its peak moment� 1000 ıC and will no longer approach this temper-ature range again under steady operation. The systemis now self-sustained and very stable. Any adjustmentsto the fuel mapping at this point will only induce mi-nor changes in temperature. If flow is increased, thenthe main reaction moves toward the next catalyst zonefollowed by the controls. Given added backpressure,a gas chromatograph (GC) analysis can be processed atany point during this time. In order to shut the systemdown, simply turn off the power supply and close thefuel valve. A slight air peak may cause a minor com-bustion instance with any remaining fuel in the system,but quickly subsides as shown by the drop in measuredtemperature. The reactor retains heat well and may re-quire roughly 3 h to return to ambient levels.

In engine operation, many more transient situationsmust be considered. Startup ensues in the same manner,with normal engine startup first followed by PRISMstartup. Once the temperature change in zone 1 ap-proaches zero, syngas is being produced and can nowgive way to reduced fuel consumption on the engine’smain line. Discontinuities from the piston’s compres-sion strokes do not affect the stability of the reactor evenwhen air is being pulled through. This setup requiresminimal manipulation to operate steadily, but additionalinstrumentation would be necessary in a real-life appli-cation using a motor vehicle where stops are frequent.A few ideas are in place to alleviate these potential tran-sient issues. The most logical approach would be toimplement the lowest possible flow through the reactorduring engine idle. The system, though it retains heatwell, would fluctuate more than desired since the unit isdesigned for continuous flow and not for on–off appli-cations. If added to current motor vehicles, the benefitof the hydrogen-rich syngas can be used in a cursorymanner; this makes the PRISM advantageous in high-way situations. To shut the system down, fully openthe engine fuel valve while shutting off the reactor fuelvalve and power to the unit. After this, total shutdownis as simple as turning the ignition key.

40.3.3 Conversion and Efficiency

In order to achieve high conversion and efficiency, op-timization on a working system is vital. True optimiza-tion requires experimentation, analysis of experiments,and adjustment thereafter from theory. This is a repeat-able, stepwise process that demonstrates the versatilityof the technology and potential for performance im-

PartD|40.3


H2 2.16

H2 2.350O2 2.520

N2 2.993CO 4.453

CH4 6.893

CO2 8.030

Valve noise 6.676

Intensity (arb. units)

Retention time (min)

Fig. 40.12 GCresults of PRISMoperation withpure oxygen

provement. For instance, nonthermal plasma on its owndoes not provide enough energy or the environment toachieve any significant selectivity of hydrogen and car-bon monoxide production. On the other hand, thermalplasma is capable of converting methane into hydro-gen unassisted, but the high amount of energy requiredfor thermal plasma generation coupled with the lowmethane conversion and low hydrogen selectivity is notan economically feasible path to any commercial appli-cation. There exists a point, between nonthermal andthermal plasma, which instills the necessary amount ofenergy required to fully convert the feed stream whencoupled with other processes. Plasma does not simplyexist in one form; it is affected by intensive propertiesjust as any other phase of matter would by a changein its density or viscosity. The difference with plasmais the constant fluctuation of every property over shorttime intervals. Methane conversion into hydrogen us-ing only a pure methane feed and nonthermal plasma isclose to 6:0%, while thermal plasma with pure methaneis near a 12:0% conversion. In an environment sup-portive of POX using air and methane, thermal andnonthermal plasma systems are capable of achievinga 99:0% conversion of starting material.

Conversion weighed against the energy input to thesystem is at its peak when nonthermal plasma is usedin conjunction with catalysis. The active field of thesystem provides an ideal environment for the chemi-cal species to react. Deviation from stoichiometry of

POX by decreasing the C/O ratio enhances the selectiv-ity and yield of hydrogen (peaked curve) approximatelyaround C/O � 0:85 and T � 800 ıC. Efficiency is alsoincreased at higher temperatures. The leaner partialoxidation ratios suggest that a little combustion ini-tially assists in producing greater amounts of hydrogenthrough the side reactions (mainly steam reformation)immediately downstream in the reactor.

A number of reactions occur simultaneously alongthe length of the reactor while carrying the electricalproperties of the plasma field throughout the process.The participating reactions are combustion, partial ox-idation, steam methane reforming, and water gas shift.All of these reactions are reversible in nature, makingit difficult to surmise and predict the product out-come from so many variables. The goal is to influencethe outcome to the end user’s needs, and this canbe accomplished through manipulation of the neces-sary parameters. With a feedback control loop basedon internal catalyst temperature and reactant ratio, sta-ble reactor conditions are maintainable and adjustable.Through experimentation and analysis of these param-eters at varied set points, a database is established forspecific product yield. A production database also al-lows for early problem detection and error analysis.

Product MeasurementThe product stream is analyzed via a small slipstreamof gas into a high-precision SRI 1068C gas chromato-

Small Scale Catalytic Syngas Production with Plasma 40.4 Analysis and Discussion 1159Part

D|40.4

Table 40.1 Volume/mole percentage display of GC resultsfor air and oxygen

PRISM syngas product – GC analysis (% by volume)Compound With air With oxygenH2 30:7% 58:5%CO 15:9% 32:2%CO2 3:3% 2:9%CH4 1:3% 6:4%O2 0:0% 0:0%N2 48:8% 0:0%

Total 100:0% 100:0%

graph (GC) attached with Haysep D and Moleseive13X columns. The GC uses a thermal conductivitydetector (TCD) with the attachments to accurately mea-sure the molecular content of the syngas that the unitis receiving. An algorithm has been implemented tothe GC interface based on calibration gas at variouscompositions pertaining to the compounds of interestbeing synthesized. Based on the GC configuration, ev-ery compound known to be present in the gas product isaccounted for and returned to the user in a percent vol-ume composition. Refer to Table 40.1 and Fig. 40.12for actual experimental GC results. The GC is main-tained frequently and recalibrated on a three-monthcycle. Third-party analysis of a process-derived fuel(PDF) from municipal solid waste (MSW) reformedinto syngas from producer gas has also been performedresulting in high levels of hydrogen yield and a large

Table 40.2 GC-MS results of producer gas compositiondirectly after PRISM system reformation

Component Mole percent Weight percentOxygen and argon 1:995 2:787Nitrogen 53:749 65:723Methane 10:668 7:471Hydrogen 17:789 1:565Carbon monoxide 11:273 13:783Carbon dioxide 4:488 8:622Ethylene ND NDEthane 0:038 0:049Acetylene ND NDPropylene ND NDPropane ND NDIsobutane ND NDn-Butane ND NDtrans-2-Butane ND ND1-Butane ND NDIsobutylene ND NDcis-2-Butene ND NDiso-Pentane ND NDn-Pentane ND ND1,3-Butadiene andmethyl acetylene

ND ND

C5 Olefins ND NDHexanes ND NDHeptanes plus ND ND

percentage of tar reduction in the gas stream. Resultsfrom this experiment are shown in Table 40.2.

40.4 Analysis and Discussion

40.4.1 Process Challenges

The areas of concern that were encountered during de-sign and operation of the PRISM process are systemcontrols, modeling, catalyst durability, safety precau-tions, variable feedstock, and overall scale-up. The feed-back control loop is responsive to changes in zonal re-actor temperature where the fuel-to-air ratio set pointadjusts accordingly to balance reaction heat, firstly pre-venting catalyst damage and secondly for safety in elim-inating any potential runaway reaction. Though stable,the system must always be monitored. Plasma-assistedPOX with catalysis provides a wide range of variablesthat are difficult to model at the atomic level. This limitsthe depth of process understanding to broader general-izations using separate macrolevel modeling techniquesand theories that are well established. The catalyst issusceptible to thermal deactivation, coking, carbon de-position, chemical poisoning (typically sulfur), and de-activation after heavy usage over long periods of time.

The transient issue is minor and ideas are in place toremedy these negatives, such as timing control to maxi-mize the benefits of hydrogen supplementation.

40.4.2 Design Improvements

Startup requires short-term ignition in the form ofa modified fuel-to-air ratio. This is viewed more as anunavoidable hindrance that has a greater affect on anoperation with a high-frequency modulation betweentransient and steady-state operation. It is safe to as-sume that syngas is in full production from cold startupafter approximately one minute; this dead time is less-ened if system operation was recently shutdown dueto the substantial thermal mass of the system. If pos-sible, a more durable catalyst that has greater resistanceto sulfur poisoning and thermal decomposition wouldbe ideal for future experimentation and commercializa-tion. This would loosen the restrictions that decrease thelifespan of the catalyst bed.

PartD|40.5


The system’s flow capacity is high for the smallfootprint of the apparatus itself providing over 150 scfmof high-quality syngas per unit. Dimensions of eachPRISM unit are 1500 � 1800 � 5000, approximately. Thelimitation thus far is the capacity of the mass flowcontrollers, which reach maximum allowable flow ratebefore the reaction is extended beyond the catalyst bed.A safe assumption is that the PRISM is able to pro-duce upwards of 200 scfm of syngas. This coincideswith nitrogen displacement that reduces product out-put and system efficiency. Using pure oxygen from gascylinders is not a viable avenue for long-term systemoperation, negating the low cost benefits of the process.Pressure swing adsorption (PSA) or vacuum pressureswing adsorption are both viable methods of producing> 95% oxygen content from air in a continuous man-ner at low operating costs. The physical footprint of thesystems tend to be large, but even a small separation ofoxygen from nitrogen would provide dramatic improve-ments in production efficiency and power recovery.

40.4.3 Higher Hydrocarbon Reformation

Longevity tests showing system durability are con-ducted regularly and for months at a time. Under stableoperation, process efficiency does not diminish overtime and, in fact, results tend to improve slightly. Un-der a previous control schema, a fluctuation in syngasproduct yields and conversion is observed in the rangeof˙2% of the average data acquired over time. For ex-

ample, average hydrogen percent by volume of syngaswas 30% with low-end results being 28% and high-end results being 32% by volume. Some variance couldalso be attributed to the inconsistent composition of thenatural gas pipeline. This range of syngas compositionis observed for all components of the product streamduring short-term operation as well as month-long con-tinuous experimentation. Data have been logged andthoroughly examined to show stable production of highvolume and quality syngas. Reformation efficiency ver-sus stoichiometrically perfect conversion is high in the80% range for air operation and in the 90% range foroxygen operation. Methane conversion efficiency hasbeen recorded to > 99% at higher process flows whilegenerally in the� 98% range otherwise.

It has been deemed plausible to reform all chainsand types of hydrocarbons using the PRISM reactor.Given a vapor phase of feeds such as heptane or aro-matics like benzene in stoichiometric proportions withoxygen for POX, a product that consists of mostly hy-drogen and carbon monoxide can be achieved. Theplasma field imparts enough energy and the heated cat-alyst bed allows for activation of the POX reaction thatcould result in near-perfect conversion comparable tothe use of methane. Higher hydrocarbon reformation re-sults in a ratio of H2 to CO much closer to 1:0 than 2:0,which would inhibit downstream processing, but againcould be alleviated by a PSA unit to increase hydrogencontent. The reformation of waste has also been suc-cessfully observed at the facility.

40.5 Synergistic Benefits of Plasma

40.5.1 Chemical Processes

In gasification, a portion of potential product is lostthrough tar buildup, which consists of higher carbonchained molecules that are condensable, but only athigh temperatures up to 400 ıC. Not only is it a loss inproduction, if not properly removed before downstreamprocessing, the tars will obstruct process flow and causeeventual shutdown. These concerns can be dealt withby a plasma-induced field in a catalytic environmentsupportive of partial oxidation of the hydrocarbons intosyngas (Fig. 40.13).

The production of hydrogen is possible throughmany processes, as it is such an abundant element. Thevarious methods of obtaining hydrogen are energy in-tensive and also typically involve the destruction ofanother compound that could hold value on its own.In processes that create excess side hydrocarbons orhydrogen-containing compounds, there is little to no

value in them unless further processing is done. Thisis true for a side product such as dirty glycerol, whichis difficult to handle and costly to remove as well. ThePRISM is able to make these side products useful foradditional heating to primary processes or for a valuableside product such as methanol. Because of its low oper-ating expense, the PRISM can be utilized in any processlooking to maximize its hydrogen potential. PEM andsolid oxide fuel cells that run directly on hydrogen areanother potential beneficiary of this technology.

40.5.2 Commercial Implementation

This process has the flexibility to create numerousmarket-worthy products, allowing the user to choosetheir path of profitability. Syngas, once synthesized,can be taken down any processing path that would usethe H2 and CO molecules as building blocks for theproduct. Olefins are the raw material for most plastics

Small Scale Catalytic Syngas Production with Plasma 40.6 Conclusion 1161Part

D|40.6

Fig. 40.13 Operation and view of the ionized plasma fieldexposed to the air (precatalysis)

and can be produced in a high-pressure, temperature-controlled environment when provided with the propercatalyst. Methanol, ammonia, and diesel fuel are justsome of the many other products that can be made fromfeedstock derived of syngas.

Decreasing emissions from motor vehicles and in-creasing engine efficiency is a necessary step towardimproving air quality and decreasing greenhouse gases.Internal combustion engines for transportation consti-tute as one of the largest consumers of imported oiland are also a major source of ozone-eliminating gasessuch as NOx. A variety of potential improvements arepresently being investigated: lean-burn engines, enginesdesigned with increased compression ratios, improvedcatalyst formulations, use of close-coupled catalysts,

new types of exhaust treatment, electric and fuel-cellpowered vehicles, and alternative fuels.

One outlet for plasma reforming technology, whichalso presents a solution to the negatives of internalcombustion in transportation, is for onboard vehicularimplementation. In a fully integrated system, the gen-eration of hydrogen as syngas into an ICE allows forlower combustion temperatures, better flame speed andpropagation, increased sparkplug life, and decreasedNOx emissions during engine operation. Onboard pro-duction of hydrogen may be attractive for reduction ofcold-start emissions and minimizing engine crankingduring cold start. This is beneficial for catalyst regen-eration and posttreatment as well. The robust, compactdesign of the PRISM unit can be used with a varietyof fuels to provide a rapid response without the ne-cessity of reformed gas storage. Complete integrationof the vehicle operating system with the PRISM re-actor would call for an intuitive approach to achievea seamless, well-controlled operation. This would fac-tor in startup, shutdown, ratio control, valve timing, andadditional temperature sensors, but would otherwise notrequire further manipulation of either new or old stan-dard vehicle engine designs.

Upon total oxidation or combustion of fuel withor without syngas, large quantities of water and CO2

are produced in the exhaust. The hydrogen-rich gasmay be useful during cold start, both minimizing thecranking, as well as potentially decreasing the hy-drocarbon emissions due to cold-start ignition. In anexhaust gas recycle (EGR) process the dry reformingreaction plays a major role in altering the stoichiome-try of the system. The main goal of EGR is to lowerpeak combustion temperature in the piston chamber.This provides a smoother cycle of operation reducingNOx and ensuring maximum conversion while boostingthermal-to-mechanical efficiency.

40.6 Conclusion

Very few configurations exist, outside highly special-ized areas of use, which allow thermal plasma to beimplemented in a profit-worthy system. A low en-ergy, nonthermal plasma creates a better source forutilization of resources, providing efficient productionof high quality syngas. Nonthermal plasma also re-quires assistance from catalysis in the form of a tai-lored reaction as opposed to simple decomposition.If unassisted, gaseous carbon and hydrogen atoms ina heated environment will preferentially recombineinto methane instead of separating into hydrogen andcarbon monoxide molecules. Plasma-assisted catalytic

POX has great potential to address many environ-mental concerns while maintaining profitability for theend user. In its present design, PRISM is a viabletool for implementation into small- to middle-scalesyngas production plants. Expending roughly 100Wof energy in a robust system configuration built forextended operation, nonthermal single-stage plasmacatalysis has proven to be a viable technology for ref-ormation. Full integration of the PRISM reactor intofuture commercial endeavors is possible with onlyminor adjustments to the current benchmark opera-tion.

PartD|40


Acknowledgments. This chapter was written withthe assistance of Dr. Leslie Bromberg from MIT andMichael Carpenter of RTI International. Leslie has over40 years of plasma and reformation experience mostof which was spent in conducting research, educat-ing students, and writing a vast number of publisheddocuments for the field of plasma science and plasmaengineering. Michael has degrees in both physics andchemistry with a focus on nanoscience at NCSU. Hehas participated first-hand in key developments of thePRISM system and successful downstream processingof our fresh syngas product. His insight has been invalu-able towards the growth of our technology. We wouldlike to thank Leslie and Michael for their time andcontributions to our research and to this book chapter.A kind thank you to all previous studies on reformationand plasma technology, especially to the authors and

contributors whose works were incorporated and citedas an influence to the completion of this chapter. I’dlike to thank Shiva Wilson and Nat Mundy of FreedomEnergy Tech, LLC and Mr. Larry Bell for their tirelessefforts and dedication to the company as well as lendingtheir wisdom and expertise to the technology and theopportunity. Freedom Energy Tech, LLC has the exclu-sive license on the PRISM Plasma technology. I wouldalso like to acknowledge Chet Staron and Mike Wirth-man of TopLine Automotive Engineering, Inc. Finally,I would like to acknowledge the Chemical and Biomed-ical Engineering Department of the University of SouthFlorida for the thorough education provided to me andevery student involved with the department. The creditgoes to the well-established curriculum and one-of-a-kind faculty devoted to the wellbeing of each individualand to the engineering society as a whole.

References

40.1 M.A. Lieberman, A.J. Lichtenberg: Principles ofPlasma Discharges and Materials Processing(Wiley-Interscience, Hoboken 2005)

40.2 J.D. Rollier, A. Darmon, D. Gonzalez-Aquilar,R. Mekemeijer, L. Fulcheri, G. Petitpas: A compar-ative study of non-thermal plasma assisted re-forming technologies, Int. J. Hydrogen Energ. 32,2848–2867 (2007)

40.3 H.M. Lee, S.H. Chen: Yu. Chao, M.B. Chang, H. Chen:Review of plasma catalysis on hydrocarbon re-forming for hydrogen production – Interaction,integration and prospects (Appl. Catalysis B 85, 1–92008)

40.4 U. Kogelschatz, B. Eliasson, W. Egli: Dielectric-bar-rier discharges, principle and applications, J. Phys.IV 7(C4), 47–66 (1997)

40.5 A. Czernichowski: Gliding arc – Applications toengineering and environment control, Pure Appl.Chem. 66(6), 1301–1310 (1994)

40.6 P.K. Cheekatamarla, C.M. Finnerty: Synthesis gasproduction via catalytic partial oxidation reformingof liquid fuels, Int. J. Hydrog. Energ. 33, 5012–5019(2008)

40.7 O. Deutschmann, L. Schmidt: Two-dimensionalmodeling of partial oxidation of methane onrhodium in a short contact time reactor, 27thSymp. (Int.) on Combustion/The Combustion Insti-tute, Naples (1998) pp. 2283–2291

40.8 S. Jo, D. Lee, Y. Song: Effect of gas temperature onpartial oxidation of methane in plasma reforming,Int. J. Hydrog. Energ. 38, 13643–13648 (2013)

40.9 W.H. Chen, T.W. Chiu, C.I. Hung: Hysteresis loopsof methane catalytic partial oxidation under theeffects of varied reynolds number and Damköhlernumber, Int. J. Hydrog. Energ. 35, 291–302 (2010)

40.10 W.H. Chen, T.W. Chiu, C.I. Hung: Hydrogen pro-duction from methane under the interaction ofcatalytic partial oxidation, water gas shift reactionand heat recovery, Int. J. Hydrog. Energ. 35, 12808–12820 (2010)

40.11 O. Deutschmann: Catalytic reforming of logistic fu-els at high-temperatures, Catalysis 24, 48–82 (2012)doi:10.103g/9781849734776-00048

40.12 L. Maier, M. Hartmann, S. Tischer, D. Deutschmann:Interaction of heterogeneous and homogeneouskinetics with mass and heat transfer in catalyticreforming of logistic fuels, Combust. Flame 158,796–808 (2011)

40.13 D.H. Lee, K.T. Kim, M.S. Cha, Y.H. Song: Plasma-controlled chemistry in plasma reforming ofmethane, Int. J. Hydrog. Energ. 35, 10967–10976(2010)

40.14 S.A. Al-Sayari: Recent developments in the partialoxidation of methane to syngas, Open Catalysis J.6, 17–28 (2013)

https://dx.doi.org/10.103g/9781849734776-00048

nonthermal plasma syngas - github pages

Documents