D-A132 161 COMBUSTION OF HYDROCARBONS IN AN ADIABATIC FLOW 1/1 REACTOR: SOME CONSIDERflTI. . (U> PRINCETON UNIV N J DEPT OF AEROSPACE AND MECHANICAL SCIENCES. .

UNCLASSIFIED I GLASSMAN ET AL. APR 75 PUAMS-1223 F/G 21/2 NL

^1

B

,-»„^',^ . • i, • L • I " i •••• • • • ' •• • ».-«.'•. ^. ^. ."- '«" ^"i,1' '^ •• *?m\w-:wy~**

I o S'- »a • g g IB

t 1= 112.0 I.I

1.25

.8

U 11.6

MICROCOPY RESOLUTION TEST CHART

NATIONAL BURIAU OF STANDARDS 1963-A

.-. -. ..•..: •• •• •• ••

* ••.-,-'ff *- _«L, -A • . - -

A •* A. _f , _J2« ut:

• • •

•

0 . ••

•

. •

• • - ' i

• \- \? V

.;

• "• V '

" / • ''.

1

- '/ s -"-'0 t •-.' ."•""-'

• $01

•'

• - » •• • - . - * - m •

• - -

-- ... •-

'-•••' i i ad l^d — - - . i

% -,- .-.—r——-/"-••- •*. -. -—•* -. "- % ' • -. -. .-. •-: J~ • « — m m v- :"—i T—w— •- - . i 1 T- • -1 1 - T —* — T -

tyS fyfr-h/t/o. I&S3 (D

CD

CO rH

Princeton University

Combustion of Hydrocarbons in an Adiabatic Flow Reactor:

Some Considerations and Overall Correlations of Reaction Rate

by

I. Glassman, F. L. Dryer and R. Cohen

Aerospace and Mechanical Sciences Report No. 1223

PROPERTY OF THE

ENGINEERING LIBRARY AEROSPACE COUECIIÖN

Department of Aerospace and Mechanical Sciences

This document has bsan approved for public xti\otam and sale; lta distribution u unlimited. ȀjP

,— p ^^

83 69 02 068 "-- - - - - ..-.. • -- * - -- * - •--- • - - - • - • - -.-••-• - . • • - .-«, •*•• .» • j.. -•••-* -

1 Presented at

Joint Meeting of the Central and Western States Sections of the Combustion Institute

San Antonio, Texas April 21, 22, 1975

:

-• i

•••

Combustion of Hydrocarbons in an Adiabatic Flow Reactor:

Some Considerations and Overall Correlations of Reaction Rate ±m

by . :-

I. Glassman, F. L. Dryer and R. Cohen

Aerospace and Mechanical Sciences Report No. 1223

:

i «1

Ofo^nt-^^f ~9\

• * - * . - •

-

i Guggenheim Laboratories

and Center for Environmental Studies

PRINCETON UNIVERSITY Princeton, N. J. 08540

> April, 1975

-

- •

• -

1 ] I T.v. * 1 iof poi -': "• • •

< 1 dutrlbi <~

9 •

.» • I • I 1 • • .• f » •••.•> I ' J • • •. ' ". » • ."

Combustion of Hydrocarbons in an Adiabatic Flow Reactor: Some Considerations and Overall Correlations of Reaction Rate

by

I. Glassman, F. L. Dryer and R. Cohen

Guggenheim Laboratories and

Center for Environmental Studies PRINCETON UNIVERSITY

Princeton, N. J. 08540

X

J* ~f& •

Dist s .1

A I. Introduction

Recent concerns about energy needs and the associated

environmental problems has again focused attention on

the rather startling fact that after burning hydrocarbons

for about a century we still do not have a thorough under-

standing of their high temperature oxidation characteristics.

The central thrust of a program at Princeton on the homo-

geneous gas phase reaction kinetics of hydrocarbons at

high temperatures is to contribute to this understanding.

Earlier work on methane and carbon monoxide oxidation

kinetics [Dryer, 1972; Dryer and Glassman, 1973; Dryer, Naeg-

eli and Glassman 1971]has been reported in the literature.

All the experimental work had been performed on the

Princeton adiabatic, high temperature, turbulent-flow

reactor (Dryer, 1972). Some recent experimental work

on this reactor, albeit preliminary, and some further

understanding of what is necessary to model complex chemical

O»1''

COP' VlNSPi '

.._« •-..--. . M . - - •_ - -- i- - - i • :

*• r w — -—

-2-

kinetic systems are thought to be of great significance

in further elucidating the hydrocarbon oxidation kinetic

process. More specifically this work which is reported

here deals with some preliminary thinking on the processes

in paraffin hydrocarbon combustion; the issue of in-

duction and steady state oxidation kinetics; the use of

global, semi-global, or overall reaction rate expressions;

and some detail measurements on ethane oxidation that

lends credance to the concepts exposed.

II, Experimental /Apparatus

Basically, the Princeton flow reactor technique

utilizes a heated cylindrical quartz duct 10 cm in

diameter through which a hot inert carrier gas flows at

velocities which yield Reynolds numbers in excess of

3500. (Figure 1). The reactor assembly is constructed

so that the reactor walls rapidly equilibrate to the

local gas temperature. Rapid mixing of small amounts

of pre-vaporized reactants with the carrier is provided

by radial injection at the throat of a high velocity

mixing inlet nozzle. Proper adjustment of carrier temp-

erature flow velocity and reactant concentrations result

in a steady, one dimensional, adiabatic reaction zone ex-

tending over a length of approximately 85 cm. Simultaneous

thermal and chemical data at discreet axial

_. __. _.. . ._^_~

• «.'.'». « » . •. 1 •••••- •• • • *>*.'••—».. • " ». •.•.'•" • " ' ". • I»

-3-

locations in the reaction zone are obtained by longitud-

inal extension of an instrumented probe. Temperature

measurements are made with a silica coated Pt/Pt-Rh

thermocouple, and gas samples are removed through a

water-cooled/expansion quenched stainless steel sampling

probe. Consistent with the long range objective of

more complex hydrocarbon oxidation studies, a sophisti-

cated gas chromotographic chemical analysis procedure

which was developed by Colket et. al. (1973), permits

measurements of all stable hydrocarbon species (in-

cluding partially oxidized compounds) as well as H2

and C"2 to 1% precision.

Several unique advantages of this approach should

be emphasized. By restricting experiments to highly

diluted mixtures of reactants, and extending the re-

actions over large distances, gradients are such that

diffusion may be neglected relative to convective effects;

thus the measured specie profiles are a direct result

of chemical kinetics only. This is in contrast to low

pressure one dimensional burner studies where diffusion

effects must be determined analytically before useful

chemical kinetic data are obtained. While this pro-

cedure has progressed significantly in its refinement,

estimation of diffusive corrections remain very difficult.

--'•-'--•- -_.

---.-.-------- - - - - !••..•.<•. --- -....--. -v - .-, ". —

-4-

Furthermore, in the flow reactor uniform turbulence

results not only in rapid mixing of the initial reactants,

but radially 1-dimensional flow characteristics. Thus

real "time" is related to distance through the simple

plug flow relations. However, the relation of a

specific axial coordinate to real time is not well de-

fined since the initial time coordinate occurs at some

unknown location within the mixing region. One would

suspect that initial mixing history might therefore

alter reaction phenomenon occurring downstream. However,

the existence of very fast elementary kinetics, which

initiate chemical reaction before mixing is complete,

permit rapid adjustment of the chemistry to local con-

ditions as the flow approaches radial uniformity. Further-

more, the large dilution of the reactants and rapidity

of the kinetics reduce the coupling of turbulence and

chemistry to the point that local kinetics are functionally

related to the local mean flow properties. This con-

clusion is experimentally supported by excellent agree-

ment of the derived chemical kinetic data with that

obtained from shock tubes and static reaction systems

at other temperatures. Agreement also substantiates that

the reactor surfaces do not significantly effect the

gas phase kinetics. Comparison of flow reactor data

from reactor tubes of significantly different surface

> - • - - - M m. L • - - • _: * •••"-- - - -•-•-• - m | . .„ ! •(' - m r'i 1> n' i' —•.•..*.„„'.,'••., .-„.'•«••,. •• m.~ m...\ m..X. *,.»., ....W-.. - - — • ...

-,—r - . - . - • - .-. • ••,-»-»•••> -» ••* - —--.-. .- -.--_. . w . . . ., u . . .. -

-5-

to volume ratio also corroborates this conclusion.

Finally and most important, the turbulent flow reactor

approach permits kinetics measurements in a temperature

range (800-1400K) generally inaccessible to low tem-

perature methods (fast flow Electron Spin Resonance,

Kinetic Spectroscopy techniques, static reactors etc.)

and high temperature techniques (shock tubes, low pressure

post flame experiments).

III. Combustion of Paraffin Hydrocarbons

Combustion of paraffins above methane has always

been thought to be complicated by the greater instability

of the higher alkyl radicals and by the great variety

of secondary products which can form. The oxidation

mechanism characteristically follows the Semenov type.

Minkoff and Tipper (1962) have reported some oxidation

mechanisms of specific hydrocarbons.

At higher temperatures most have accepted the

primary reaction in the system to be between the hydroxyl

radical and the fuel.

RH + OH + R + H20

Recent work (Dryer, 1972) has suggested that other

reactions in addition to this one were important; namely,

in fuel lean and rich combustion

RH + 0 ->- R + OH

- .- .-- - .__-. -. - . , . . • L- . . . . ^_1 ' • - - - -

-,-.-_-.-_.- _-. . . . .--, . J.1. «' »•*•'. - - - r .- i . k ; w . i

-6-

and in fuel rich combustion

RH + H •*• R + H2

It is interesting to review a general pattern for

the oxidation of hydrocarbons in flames as given by

Fristrom and Westenberg (1965). They suggest two

essentially thermal zones: the primary zone in which

the initial hydrocarbons are attacked and reduced to

CO, H2> H20, and the various radicals (H, 0, OH)

and the secondary zone in which the CO and H2 are

oxidized. The primary zone, of course, is where the

intermediates occur. In oxygen-rich saturated hydrocar-

bon flames, they suggest further that initially hydro-

carbons lower than the initial fuel form according to

0H + Cn H2n+2 * H2° + Cn K2n+1 * Cn-1 H2n-2+CH3

Because hydrocarbon radicals higher than ethyl are thought

to be unstable, the initial radical C H« ,, usually n 2n+l

splits off CH3 and forms the next lower olefinic com-

pound as shown. With hydrocarbons higher than CßH-, it

is thought there may be fission into an olefinic compound

and a lower radical. The radical alternately splits

off CH . The formaldehyde which forms in the oxidation

of the fuel and fuel radicals is rapidly attacked in

flames by 0, H, and OH, so that formaldehyde is usually

•^1*11 I I I I >^MM^1lll I I I

-i . . . •.-.-.—;—T—•—»i y^t" - r ^ - •" "" " - -•"- -•-• r-, -, -••.•« •. » : •;-.-.-» «—-—w- - - r w. ' •

-7-

only found as a trace substance.

In fuel-rich saturated hydrocarbon flames, Fristrom

and Westenberg state the situation is more complex, al-

though the initial reaction is simply the H atom abstrac-

tion analogous to the preceding OH reaction: e.g.

H + CnH2n+2 * H2 + C2H2n+l

Under these conditions the concentration of H and

other radicals is large enough that their recombina-

tion becomes important and hydrocarbons higher than the

original fuel are formed as intermediates.

The general features suggested by Fristrom and

Westenberg have been confirmed in our recent experiments.

However, this new work permits more detailed understanding

of the high temperature oxidation mechanism. As stated

earlier this work shows that under oxygen-rich conditions

initial attack by 0 atoms must be considered as well as

the primary OH attack. More importantly however, it has

been established that the paraffin reactants produce

intermediate products which are primarily olefinic and

the fuel is consumed to a major extent before significant

energy release occurs. The higher the initial tempera-

ture the greater the energy release as the fuel is being

converted. This observation leads one to conclude that

the olefin oxidation rate simply increases more appreciably

— ------ -.-.-.

•"!!'•> »1« 1 •

-8-

with temperature; i.e., the olefins are being oxidized

while they are being formed from the fuel. These con-

clusions are based on recent experimental results as type-

fied by Figures 2-6 which represent the data taken through-

out the reaction zone of ethane, propane, butane and hexane.

A summary of the intermediates formed from the

oxidation of these four paraffin hydrocarbons is most

revealing and is represented by Table I.

Fuel

ethane

propane

butane

hexane

Table I

Relative Hydrocarbon Intermediate Concentrations

ethene >> methane

ethene > propene >> methane > ethane

ethene > propene >> methane > ethane

ethene > propene > butene > methane>>pentene>ethar.

It would appear that the results in Table I would con-

tradict elements of Fristrom and Westenberg's suggestion

that the initial hydrocarbon radical C H2n+1 usuaHy

splits off the methyl radical. If this type of splitting

were to occur, one could expect to find larger concentrations

of methane. The large concentrations of ethene found

in all cases would suggest that primarily the initial

C H radical cleaves between the second and third n 2n+l

-V. - . » _^_. — .--..-. .-..,••.-•-:.. -' - .. ^ . • - —^-

• ^ • y^ i •> • n i j -j i. •. • • . • '^. ' • • • • • — • " *^^" •-••» •••-»-•«•-.- w » »»••-*-»-.: r- .- - .. - « r- *- — » - •«---w--•-—i ^- ••• */•

-9-

carbon to give ethene plus another radical. If one

assumes the abstraction process to give the initial

radical is from the second carbon atom as in lower tem-

perature oxidation process, then the formation of ethene

from this type of cleavage appears quite logical.

Thus the new reaction step appears to be

R-C2H4 - R + C2H4

(Cn H2n+1)

Notice the relative concentrations of propene found in

butane and hexane oxidation and that very little pentane

is found in hexane oxidation. These results would suggest

that next to cleavage between second and third carbon

atoms, the next most important cleavage is after the

third carbon atom. Contrary to the suggestion of Fristrom

and Westenberg, splitting off of the methyl radical,

that is cleavage between the first and second carbon,

is least important. In iso-compounds, cleavage would

probably take place between the tertiary carbon atom

and the one next to it. The dominant olefin would

probably not be ethene. In the case of 2-methyl-

pentane one could postulate that propene>butene>ethene

must be the order of the intermediates.

Figures 2-5 show clearly that experimentally there

appears to be an initial iso-energetic step in the over-

i . •—• • • • • . «_^. • ---------- -----•• ..-.•-•- 7 1 ,.;, . I • I i • « n •> ». • »—_ »—* *—•»•-•-••*

-»-.-.-.-.-.-.-. ... - - . - - - 1 • -.

-10-

all process. Of course, this step is not exactly iso-

energetic. The conversion process from paraffin to

olefin is endothermic; however, some of the hydrogen

formed during what is essentially a pyrolysis step does

react and release energy. The two reactions are com-

pensating energetically. Thus, the evidence is that

there are three distinct, but coupled zones, in

hydrocarbon combustion.

1) Following ignition, primary fuel disappears

with little or no energy release and produces

unsaturated hydrocarbons and hydrogen. A

little of the hydrogen is concurrently being

oxidized to water.

2) Subsequently, the unsaturated compounds are

further oxidized to carbon monoxid and

hydrogen. Simultaneously the hydrogen present

and formed is oxidized to water.

3) Lastly, the large amount of carbon monoxide

formed is oxidized carbon dioxide and most

of the heat release from the primary fuel is

obtained.

Each zone must have a different temperature de-

pendency and thus at different temperatures the importance

of a given step may change. Again on the basis of some

k ~ • i . r • . • i • • - . . __«_»_. » i. I J_J a . .

-—- - - -•-.-- ».-«-».-•—-—1. 1 ' •• 'V i—•—•• •••»•• M-. •.-.-.».». »—• • w-r-m »—i—c 1 : r—T-

-li-

very preliminary experimental evidence as given by

Figure 6 it is possible to put forth some interesting

speculations. The initial condition of the experiment

whose results are depicted in Figure 6 was such that not

only was it more fuel rich than the examples of

Figures 2-5 but also at a higher initial temperature.

Examination of Figure 6 reveals that the maximum

concentration of ethene is found earlier in the system.

Essentially this means that the exothermic ethene

oxidation step has become faster. This conclusion is

supported by the fact that the temperature profile in

Figure 6 rises continually and does not appear flat (iso-

energetic) throughout most of the process as found for the

conditions given in Figures 2-5.

IV. Current Chemical Kinetic Modelling

The concept of overall (global) reaction kinetics

and its use is a direct result of the complexity of

most chemical reactions and the complicated fluid mechanic-

al situation in which some knowledge of heat release or

chemical rates is necessary (ram jets, rocket engines,

gas turbines, etc.). The assumption invoked is that

the course of chemical kinetic events may be described

-a. . - - ~

-w—^—•—^—»-.« ••—• •;« .»—- -.-• .* •• .- -.- --• F r .*—7TT1—w i ; i v» *—.-•- w .

-12-

in terms of a few of the main reactants and products

(C-) in a functional relation with much of the same

form as an elementary reaction process. Typically,

the equation is of the form:

c1+c2 + c3+c4+...

with rate given by m

-[C, ] = k It [C.]"1 1 ov 1 i=l

k , the overall specific rate constant, is expressed

in the Arrhenius form

-E/RT

kov = 10 e

The n^'s are defined as the orders of reaction with re-

spect to C^, and En. is termed the overall reaction order.

a 10 and E are termed the overall frequency factor and

activation energy respectively.

The relation implies nothing about the actual'

kinetic mechanism (in terms of elementary reactions),

although the parameters in the strictly empirical

relation sometimes are governed by a single elementary

step (or a number of steps) which basically control

the rate of the chemical process. Under what circumstances

such an overall correlation is usable is largely dependent

-•- ..-.-..i..-. i_-..- ..=..— .•.-.-. Kk.il.. . . • . ...

o

•' • ' • • I »—• • " ! ' -"• •—7 • -•.-•-. »T ••----- "

-13-

on both the kinetic mechanism to which it is applied

and the physical environment in which the process is

occurring. Levy and Weinberg (1959) concluded that

the approximation is not generally applicable to

chemical measurements taken in flames; however, this

fact may not arise from the chemistry itself, but

from the physical structure and diffusive character

of the flames studied. Where a particular rate-

determining step or sequence in the true chemical

reaction mechanism occurs and the physical circum-

stances of the application are similar to those in

which the expression was derived, the overall approxi-

mation is a valid and vastly simplifying idea. How-

ever, extension of such a correlation to experimental

conditions outside the range studied should never be

done without some reservation.

Considerable global modelling efforts have

been attempted for carbon monoxide and these have been

reviewed by Dryer (1972)and Howard et.al.(1973) Methane

ignition and oxidation kinetics have also been expressed

in this manner, often in conjunction with developing

detailed mechanisms. Many of these studies were also

<* reviewed by Dryer (1972). Avery (1953) is the only

research which has attempted global modelling of a

higher paraffin oxidation (Butane). Global modelling

.. .. . _ ... ... _ ------- - - - -

^—^^^^^^^^^^^^^"^^^^V^^^^^^^^^V^»

-14-

has the potential to describe only spatial energy re-

lease and reactant/final product concentrations. However,

prediction of more detailed parameters such as inter-

mediate species is not possible. Edelman and Fortune

(1969) have proposed a "quasi-global" modelling pro-

cedure for paraffin hydrocarbons as an approach to

supplying these details. This technique approximates

the paraffin hydrocarbon oxidation as a global reaction,

k • nr.O+fn+llh

2

ov CnH2n+2 +n 02 f nCO+(n+l)H2 (I)

with a rate given by

-tCnH2n+2]=[CnH2n+2]a[02]

bk (II)

and combines these expressions with a number of elementary

reactions from the hydrogen/oxygen and carbon monoxide/

oxygen reaction mechanisms. Values for a and b were

assigned (a=0.5,b=l) and k„„

kQV = 1.8xl09[(T/1111)-0.5]T P0'2 exp[-13,700/RT]

was determined from results of an analytical study

of propane ignition kinetics (Chintz and Bauer, 1965).

It was suggested that this mechanism could be applied

in the temperature range 800-3000K. The assumption

^ . m - - - . . • • m . - - - . . a ^V.

i—- • t ••• • - -.—i—• -.' "."".. >*. ". " ^ i". ". "..' • ""• • 'T •—• " •' • •—T—T""

-15-

that a quasi-global model based on propane character-

istics was applicable to higher paraffins was presumed

to be substantiated by shock tube ignition aelay ex-

periments performed by Nixon et. al. (1964-1967). These

experiments identified some similarity in the functional

behavior and the order of magnitude of ignition delay

times for propane/and n-octane/oxygen mixtures.

The quasi-global model thus developed has

undergone some revision, as more ignition delay and

elementary reaction data became available, primarily

in the values of elementary rate constants for hy-

drogen-oxygen and CO-oxygen reactions and the expression

for k . Recent applications quote k as

DM

-•;

-0.815 k =6x10 P ov T exp[-12,200/RT] , Edelman et.al.(1972)

or

kov=5.52xl08 P~0-825 T exp[-12,200/RT] , Englenan et.al. (1973)

In accordance with earlier work (Marteney, 1970), Hammond

and Mellor (1971) suggested

C H +[a + b] O 3 b 2 4 2

ov aCO+b H20 (la)

.- -^- • , - -_ . , . _ IllWnll ,;••«•• ±m*.m • - im

—7-7" • • - - - - - -. . . ,--.--, ... -.- .

-16-

to replace expression (I) in preliminary analytical

studies of gas turbine emissions. However, Bowman,

in comments to Edelman et.al. (1972) has shown that

"infinite" quasi-global kinetics do not offer any sig-

nificant advantages over the partial equilibrium ap-

proach to prediction of NO emissions, and he con-

cluded that these approaches were far inferior to

detailed or quasi-global kinetics interpretations

of methane oxidation.

It should be re-emphasized that Engleman et. al.

(1973) is the only study which has attempted to

experimentally corroborate the Edelman model with

other than shock tube ignition delay results, and

this study was restricted primarily to a comparison of

experimentally measured and analytically predicted NO

emissions. No other fundamental experimental in-

formation to which the semi-global equation might be

compared has been available previously, and thus

there has been no challenge or corroboration of its

applicability. The studies to be reported here are

the first to supply some insight to this question.

i . . . . . . - . - - i • •

«iii ,-••:••».-.• L •"-

-17-

V. Global Modelling of the Hydrocarbon Systems - Particularly Ethane.

Clearly a new global mechanism is needed for the initial

olefin formation phases of the hydrocarbon oxidation process.

The oxidation of methane was studied in some detail by

Dryer (1972) and Dryer and Glassman (1973), and encouraged

the belief that the rate of reaction of hydrocarbon could

be expressed by a simple global expression of the form of

Equation II.

The rate of reaction in the post induction phase of

the lean methane oxidation was experimentally found to be

well described by the overall expression

13.2+.2 0.7 0.8 -[CH ] = 10 exp (-48,40011200) [CH ] [0J

4 4 I RT

It should be noted that the parameters of this equation

are significantly different than those found by investigators

who have studied the induction (ignition delay) phase of

this reaction in shock tubes [e.g. Skinner & Ruehrwein

(1959), Asaba et.al.(1963), Higgin and Williams (1969)]

and flow reactors [Lloyd (1946), Mullins (1953), Nemeth

and Sawyer (1969)]. These studies predicted the rate of

reaction to be inhibited by the concentration of methane.

For example Seery and Bowman (1970) empirically correlated

. . . . __ ._• _ _

— —pi—;—: z =—v •• »—;•*-•»•—• • • . ' | -—*-T"w~* t— r- r— v- w- » • •—•—-—-—-r——--•—: —\ c—-• •; •—»—* —- —-•

-18-

the ignition delay time as

Reaction Rate

-18 0.4 -1 6 a T = 7.65x10 exp[51,400/RT][CH4] [02]

However, Figure 7 shows a comparison of the overall rate

constant derived by Dryer (1972) with results calculated

from parameters predicted by the detailed analytical

studies after Bowman (1970). An analytical overall

rate constant was calculated from

kQV = - [CH4]/([CK4]°-7[02]

0-8)

Clearly there are two phases of this reaction which are

not modelled by the same global parameters. Indeed,

the experimental flow reactor data of Dryer (1972) show

similar behavior (Figure 8).

Qualitative comparison of ignition delay character-

istics of propane (Hawthorne & Nixon, 1966) and the

m rate of disappearance of propane in the post induction

phase of its oxidation (Figure 3) are also in disagree- ]

ment. It is important to remember that the semi-global .

1§ modelling effort reviewed earlier was derived primarily •

from ignition delay data, and yet this model purports to i

i# 1

-• 1 i

1

< |

• •9:

t mi» -..-. .. rti

"." : •.•••••«—•——?—•—-—• -- • •—•—• • . • • i— • • • i • •! • i» ! • i •—»,«.-'- v i—.- • i- • - w -

-19-

describe finite disappearance rates of fuel during the

combustion process.

Indeed, the induction (ignition) phase of the hydro-

carbon oxidation with the exception of fuel rich methane

combustion is generally very short relative to the

post induction reaction. In fact, in some practical

situations, the ignition kinetics may be totally pre-

cluded. For example, in turbine combustors, recircula-

tion of hot, partially burned gases provides stabilization

of the combustion. These hot gases contain partially

oxidized reactants, combustion products, and some re-

active centers, all of which may significantly effect

the initial chain branching mechanism leading to ig-

nition. Thus partial stirring may significantly reduce

or even eliminate the ignition phase chemistry. The

post induction phase of the reaction will remain largely

unaffected, and it is this chemistry v/hich is important

to prediction of energy release rate and emissions. The

turbulent flow reactor is particulary suited to study

of this phase of the reaction and can in addition provide

some information about ignition chemistry in cases where

it must also be considered.

Furthermore Equation II appears inadequate for de-

scribing the rate of disappearance of initial fuel in

------~~- ^-.-- -_ - « . .....

I». 1. |i. l.l 1 • n «. IH'H'^ •.'•'^" L'."«!'.!,'.' '.'

-20-

• I • " — 1 1 = • C-

higher paraffin oxidation reactions. If one conducts

an experiment at constant temperature and large con-

centrations of oxygen, Equation II would predict

a "(CnII2n+2) = Constant (CnH2n+2)

Thus a plot of

log (cnK2n+2) vs' lo<? <cnH2n+2)

should yield a straight line of slope a. Figure 9 is

such a plot of flow reactor data on ethane oxidation.

The non-linearity of the curve suggests that ethane

disappearance rate cannot be described by a functional

relation of the form of Equation II. Similar behavior

is also observed for the oxidation of propane and

butane (Figures 3 and 4). Thus it appears that Equation

II does not in general predict the rate of disappearance

of a paraffin hydrocarbon in the post induction phase

of its oxidation. Further, Figure 2 as discussed pre-

viously suggests that the paraffin reactant produces

intermediate products which are primarily olefins and

the paraffin is consumed to a major extent before sig-

nificant energy release occurs. If the kinetics followed

_ ^— — »—J i •- . _^ . . . - . . . i

• .^fl .-- - • - - •« .- •—. - - - • ---.-.-•- --..-. . ...

-21-

a mechanism represented by Equation I, there would be

considerable energy release proportional to initial re-

actant disappearance.

In contrast, the following correlation was found

to give a good representation of some extensive data

taken with ethane:

-[C2H6]M[C2H6]0 -rc2Hj)a

An example of how well these data fit this correlation

is given in Fig. 10 for the experimental data shown in

Fig. 2. For nine/ethane/air runs the average value of

the slope, determined by the least-squares method, is

a=0.975*0.05. There is, however, a small temperature

rise in each run which contributes to the slope.

Taking the temperature rise into account, the fuel

dependence is reduced to the 0.8 0^0.0 3 power of the

amount of ethane reacted (Fig.11) and the activation

energy is about 33 kcal/mole. Experiments were then

made in which the initial oxygen concentrations and

equivalence ratios were varied 6 . 06<4><0 . 79) . A global

correlation of the form

~[C2H6] = kE([C2H6]o-[C2H6])0*8[C2H6]

b[O2]C (HI)

was attempted for all the data. It was found that "b"

could not be defined and "c" was negligibly different

from zero. The final correlation (Figure 12) was found

• /

.-...- .....-.---.-.- .-.....-..,-

•22-

to be

7.18±0.33 . , _ _ + . -[C2Hg]=10 exp[(-32,900^1530)/RT]

[(C2H6)0-(C2H6)]0'8

As yet, no detailed mechanism has been developed to ex-

plain this correlation. However, similar correlations

also have been qualitatively recognized for propane

and butane-oxygen rich reactions.

VI. Conclusions

The applicability of kinetic measurements obtained

in induction (ignition) phase measurements such as those

made in shock tube and those obtained in the post in-

duction phase by flow reactors has been clearly delineated -

particularly with reference to hydrocarbon oxidation.

The usefulness and limitation of global kinetic modelling

has also been carefully analyzed.

Experimental results show that there appears

to be three stages in the paraffin hydrocarbon oxidation

process that must be considered if there is to be global

modelling. The new element is a well defined initial

step most easily recognizable in the temperature range

of the flow reactor. This step is the relatively

isoenergetic conversion of the paraffin to the olefin.

•

. .- . _ —» •••••,«'• •• i- . •> *—•- «-••«—' i - : _..-*-*-»-^

»»»>>!'.• • i • • i • i • i • I • • I I .'.••'••. • • ' ' '

-23-

Extensive work with ethane indicates that the con-

version rate is proportional to the amount of ethane con-

verted. Future work will attempt to explain this fun-

ctional dependence of this phase of the overall paraffin

hydrocarbon oxidation process.

VII. Acknowledgement

This research was supported by the Air Force Office

of Scientific Research under Princeton University Grant

AFOSR-74-2604 with Capt. Lloyd R. Lawrence, Jr., as

technical monitor. This support is gratefully acknowledged.

*j 1^_ - ^_» ••'-•*- »i m' It n ' - ' - »Mi'iii» m,MJL,Z.,H..>mAi.',wM...„mfii.tit^

. j -. «• — , - .-.-- - - - - -. - -V .-W-7-- ~-»—^—- . -T -•» -W-W- W •- »-*----•- - - _-

-24-

VIII. References

Asaba, T., Yoneda, K., Kakihara, N. and Kikita, T. (1963) "A shock Tube Study of Ignition of Methane-Oxygen xMixtures" Ninth Symposium (International)on Combustion, Academic Press, London, p. 193. j

Avery, W. H., Appleby, W. G., Meirbott, W. K., and Sartor, A. F. (1953). "The Decomposition of n-Butane in the Presence of Oxygen", J.A.C.S. 7_5, 1809.

Bowman, C. T., (1970) "An Experimental and Analytical ( Investigation of the High Temperature Oxication Mechanisms of Hydrocarbon Fuels", Combustion Science and Technology, 2,161.

Chintz, W. and Bauer, T., (1965) "An Analysis of Nonequili- brium Hydrocarbon Air Combustion" WSS/CI paper 65-19. j

Colket, M. B. Ill, Naegeli, D. W. Dryer, F. L., and Glassman, I. (1973) "Flame Ionization Detection of Carbon Oxides and Hydrocarbon Oxygenates"»Environmental Science and Technology 8_, 43.

Dryer, F. L., (1972) "High Temperature Oxidation of Carbon Monoxide and Methane in a Turbulent Flow Reactor" AFOSR Scientific Report TR-72-1109.

Dryer, F.L. and Glassman, I., (1973) "The High Temperature Oxidation of CO and CII4 " Fourteenth International i Symposium on Combustion, The Combustion Institute, Pittsburgh, p. 987.

Dryer, F. L., Naegeli, D. W., and Glassman, I., (1971). "The Temperature Dependence of the Reaction CO+OH=C02+H", Comb, and Flame 17, 270. _•;

Drysdale, D. D. and Lloyd, A. C, (1970) "Gas Phase Reactions of the Hydroxyl Radical" Oxidation and Combustion Reviews, Elsevier Publishing Co., Amsterdam, 4, p. 157.

J Edelman, R. B., Fortune, 0. F. (1969) "A Quasi-Global Chemical Kinetic Model for Finite Rate Combustion of Hydrocarbon Fuels with Application to Turbulent Burning and Mixing in Hypersonic Engines and Nozzles", AHA Paper 69-86.

• 4

MM*^^^^^W-I-1M^.

-25-

Engleman, V. S. , Bartok, W. , Longwell, J. P. and Edelman, ", R. B. , (1973) "Experimental and Theoretical Studies of N0X Formation in a Jet Stirred Combustion" Fourteenth International Symposium on Combustion, The Combustion Institute, Pittsburgh, p. 755.

Fristrom, R. M. and Westenburg, A. A., (1965) Flame Structure, j McGraw Hill, New York.

Hammond, D. C. Jr., and Mellor, A. M., (1971) "Analytical Calculations for Performance and Pollutant Emissions of Gas Turbine Combustors" Comb. Sei, and Tech.4,101.

Hawthorne, R. D. and Nixon,A.(1966) "Shock Tube Ignition Delay Studies of Endothermic Fuels" AIAA J.4_, 513.

Higgin, R. M. R. and Williams, A., (1969) "A Shock Tube Investigation of the Ignition of Lean Methane, and in Butane Mixtures with Oxygen" Twelfth Symposium *#| International on Combustion, The Combustion Institute, p. 579.

Howard, J. B., Williams, G. C., Fine, D. H. (1973) "Kinetics of Carbon Monoxide Oxidation in Postflame Gases", Fourteenth -« International Symposium on Combustion, The Combustion Institute, Pittsburgh, p.975.

Levy, A. and Weinberg, F. J., (1959) "Optical Flame Structure Studies: Examination of Reaction Rate Laws in Lean Ethylene - Air Flames" Combustion and Flame '* 3,229.

Lloyd, P., (1946) "Combustion in the Gas Turbine" Proc. Sixth International Congress for Applied Mech., Paris.

Martenay, P. J., (1970) "Analytical Study of the Kinetics !

of Formation of Nitrogen Oxide in Hydrocarbon - Air Combustion" Comb. Sei. Tech.1,461.

Mellor, A. M., (1972) "Current Kinetic Modelling Techniques for Continuous Flow Combustors" Emissions From Continuous Flow Combustion Systems, Plenum Press, New York, pp. 23. -

Minkoff, G. I., and Tipper, C. F. H., (1962) Chemistry of Combustion Reactions Butterworth & Co., Ltd. London.

. ... . • .'• . - - - - • j. La »I « 1 »T . 1 a 1 - 1 . »»•

r~~

-26-

Mullins, B. P., (19DJ) "Studies on the Spontaneous Ig- nition of Fuels Injected into a Hot Air Stream IV. Ignition Delay Measurement on Some Gaseous Fuels at Atmospheric and Reduced Static Pressures" Fuel 3_2, 211 and 34 3.

Nemeth, A. and Sawyer, R. F., (1969) "The Overall Kinetics of High Temperature Methane Oxidation in a Flow Reactor" J. Phys .Chem. 7_3_, 2421.

Nixon, A., et.al., (1964-1967)"Vaporizing and Endothermic Fuels for Advanced Engine Applications" Shell Develop- ment Company Reports, APL-TDR-64-100, Parts II and III, AFAOL-TR-67-114 Part I.

Seery, D. J. and Bowman, D. T., (1970)"An Experimental and Analytical Study of Methane Oxidation Behind Shock Waves" Combustion & Flame 14_, 37.

Skinner, G. B. and Ruerhwein, R. A. (1959), "Shock Tube Studies of the Pyrolysis and Oxidation of Methane" J.Phys.Chem.63,1736.

»j

!

-*

N

hi CO

? or ÜJ iii m H -> < UJ H X ce

o *z 1- UJ o > < Ü LJ

>-, _l CD 2 LÜ CD CO <

hi o: OQ O Z> l- H o rr < o ÜJ H Q: O Id ~3

FIGURE I

' • -'—'•!• • •" '" I

0

0 0

1.0

0.8

0.6

w 04

ö LÜ Q_

0.2

ÜJ Ü oc UJ Q_

LJ _l O

A—A *_ V "* + + A—A-^

A fc.

I cm « 0.91 msec <f> = 0.078

TOTAL CARBON

CjHs /

TEMPERATURE-. >T\7\

* ^ X _ _ A C? H4

0-* /C0n y°\ -

30 40 50 60 70 80 90 100

2000

I080

i.' I060 L-

P

I040 \

I020

1000

0.03

0.02

0.01

v-^O*

/

^a

J. Jo^i-D ^-O'

60 70 80 30 40 50 60 70 80 90 DISTANCE FROM INJECTION (cm)

CHEMICAL COMPOSITION OF SPREAD ETHANE-AIR REACTION

I00

•

FIGURE 2

>» a fc '• W..t .», i:,m.lk:,.m..±~.

—7 ^ ^—^ • •" •—-—- -. - . - - ^-—" —; "»i •* "— - wm— . - - - •

(>0 3UniVH3dW31

S3D3dS lN3DU3d 3H0l^J

o < Ld CC

< i

LÜ

O CC Q.

Q < LU CC CL (/)

o

o o

6 ÜJ X o

FIGURE 3

^ * - * - « ^ « - . _ . - « - r _ - - . - *••-.*. - — *t !~ 1 m» irft irr* «. «... ..».-».- k.Aw-4>*>-* .- - ..... .* *^_*_i-.J

v-« m • • r* " • • ; •• •• - v • • - •* • » • • • - 7 - :• -„w ' * ~w T--;-!-"; w; *~ - --.-..

1.40+

1.20

L TOTAL CARBON

1.00

w 0.80 ü Lb! GL </)

I- 0.60 Ld O ÜC Ul °- 0.40» LÜ

r C4H|0X4

I cm« 1.15msec <£ = 0.104

V \ noo

TEMPERATURE

0.20

\ A •-•—Trr- \ZJooo»

Ld Q:

1050 < a: ÜJ

C2H4x2-\^ ^O^ H,0- .^;^°' \/ 2U A ^o-

,j.

.6:

*s~—^ «V^Pn^ «^.fc^ö*^" o—-"U' ^^-D-**'*LJ "* \ <y \

0.04 - H

CH

C2H6x2^. D

0,021- \q-^D^0"0- +

0 40_ 50

*~\

°^CH20

CH£H0x2-' )

60 70 80 90 100 DISTANCE FROM INJECTION (cm)

CHEMICAL COMPOSITION OF SPREAD BUTANE-AIR REACTION

120

FIGURE 4

«• - - - - -

JP ...... . , . 'i •« • : • i •—•—«T-«—' ' •* • * • »— • • - - . - -

0.14

0.12

0.10

0.08

co 0.06 UJ

CJ Id Q. CO 0.04 -

UJ ü

uS 0.02 OL

UJ _l o

I cm=0.95 msec <£= 0.039

- 1020 * UJ c: -

1000

980 H

90 100

0

8

0.008 -

0.004 -

30 40 50 60 70 80 90 DISTANCE FROM INJECTION (cm)

100

- -

CHEMICAL COMPOSITION OF SPREAD HEXANE/AIR REACTION FIGURE 5 m

_ - — - -: - - - -t-i -.- * • •• - • - -» - - ..-..- . .*.-..

_________ _— . I . . I ... I

(>0 3UfUV_3dW31 O in O )P in CJ o p; • = — — O o

— — Al o 1 1 1 1 —~ CM

""• __:

• • • r\ O X n> o o j t h-

A + • ca x o -^ x ___< 0 G o • 1 \ rJN j It / /Ml / + • o oa O^ x o D o o

1 \8 \\ 1 / / 1 II / / <

1 I 1 G LU

/ u_

+ __.•_>•<> X 0 0<3 O G O N

m 1 J. \ \ \l / / W

0 /

G __ ° w 1 3 \ \ 1 / / 1 1 I / ^ LU

+ b • o<o_ix o j—-_w o a E _- 1 __ \ \W / £ If

+ i±j • o$a t> R ä l| \ h\ / ° 1/

o \

I a I 0N

(c

:TH

A

% + S •*<>• t> *? o _>-D \— 1 H A i\\ / J rt\ r^\ ü a + x • oo • 0 <u. oo CM G __ O-i <r

1 / MAX Tf + X iO ->• <•

1 H / \\/\\ Jl

CO \ \ x \ 00 * ii -_ —J o «* a — Q_ \ §, \ _> QL

^ CO • + b x •> ca <JD o <-> a O v '

1 g « / AW M \ \ - ££ + V ? x 0# <• <3 G • O D O > W / \\ A /i \

•K X _><__•• 0 G O __ ° 1 / 1 \\ W / Jo \ 0< H

• + X _> 0» CD <] D-» o _tr- —

+ X £• l> 0» <• G r-~

CO C/J 5 O

Q_ / / _> 1 \\ /W 1 i\ _> + x #% -£. > <*<)<• -G o

o f / i a \ 1/ w 1 ° \ 40 + X c 1 ___> <# CO G o 1

i / 1 * y^i W i \ -J <

+ X 0} CM 0 <3 <!• «G0 o o o r 1 / *o x \/ \\ \\ 1?

+ x o> n <f> <c> <» <• a / / T 8 _• ./\ \\ W I

\ o 1 • 1

M _

1 1

LU X

0 ffX •©-_ u—o t> <» o"c

I I I I I J

I o

c 5 O O O O C *- CM c > _t c _ 00 ID <_- CM o o

-odd d d d

S3l03dS lN3__3d 3101N

FIGURE 6

-~l » - 1 • — * w T—-V-" • " " - — -*—rv."." • • -*-----»-- — - T

8.5

0

0 $

8.0-

7.5

7.0 y «/>

«r o ^ 6.5 o E \ m E ^ 6.0

S

O

5.5

5.0

4.5

4.0 0.3

XCH4 xo2

<f> To

+ 0.048 0.192 0.5 1400 o 0.096 0.64 3.0 1400 D 0.048 0.192 0.5 2000

Kovs[cH4][cH4]""T [02]"

\ \ \ 4840011200 i\\ * K•=IO,3-2±-2e RT"

- EXPERIMENTAL

1.0

COMPARISON OF ANALYTICAL AND EXPERIMENTAL OVERALL DISAPPEARANCE RATE OF METHANE

FIGURE 7

: •• J _ . . . . _

1».». >—»—:—i—i- .• •—v- «.- i—-. i. •',—71—r •' •—i—:—*—•—'—:~~~ .

J

\

S

4.3

_ 4.1 i u 0> to

«1

IO 3.9 i

E o a>

2 37

> o * 3.5 o

e> 3

3.3

3.1

E=48 kcal/mole

a = 0.7 b = .8

CONDITIONS: = 0.5%

X°o2 =20%

_ d[CH4] Kov = ^ [CH4]0-7[02].B

1 .81 .83 .85 .87 .89

J^0~(K-') .91 .93 .95

DETERMINATION OF E FOR REACTION OF ClVC

FIGURE 8

~ - •-

,——^—.—-.-.•• r-> , • w—i «-7^ • < • • • • " •

•P

s

S5

£ Ü

o E T

I0'6|-

<0 x (SJ

Ü

10'

o o

o

o a

6> a o a

D — o

— o CONDITIONS

OF FIGURE 2 ° o

O D

-

ü 5 % WATER ADDED O D

O

o

o o

8 1 1 1 l i i i

I0"8

[c2H6]~ (mole cm"3) io-7

RELATION OF ETHANE DISAPPEARANCE RATE ON ETHANE CONCENTRATION

FIGURE 9

- - • • ^

•t'1.1'. ". ». •. --. w~- —•--•- ---•-•- - - - " ' ._" V" •" • " " " • - • • " - - -. -

g l

E 0) o E

to x Ü

1019-1027 K /

IO"6

8

7-36% REACTION

y 6 - y 4

&

/

o/

2

LEAST - SQUARES SLOPE = 0.977

<r = 0.023

/

i

CONDITIONS OF FIGURE 2

III 1 8 IO-8

(CC9Hc]n - [C2Hß]) -(mole cm'3) '2' '6J0

00

o o

DEPENDENCE OF REACTION RATE ON EXTENT OF REACTION

FIGURE 10

i • - - • - . . . . - - - . . . %. »__« •—• • ^_ i . - - . , - - • • . l . . I . • .,.;.!. _ - - _ -

—i- - ••- -—•—. ••I'IMMIII»'|,»'I'.,H . «—r—i :•-.'- •- .— —-. • • • T ,

CO

00

I

(D V)

ro I e o a> o £

x (M

Ü

1020 ± 2K SLOPE = 0.81

10 r6

8

6

2-.6 1

4 6 8 I0"8 2

([C2Hfi3n-[C?Hfi]) -(mole cm"3) '2n6J0 '2"6

DEPENDENCE OF REACTION RATE ON EXTENT OF REACTION

FIGURE

* - • - M • . , ^_

0.6

Vt\

o 0

0.4 o

o (VI

to E ü

o £

0.2

I 0 l

•z.

? en o o UJ

CD

3

-0.2

-0.4

-0.6

-0.8

o

o o ° \o <?poo o

-X v ° o°Q °°

% <ö o^o

°o <o

<b

*\ o o oo\.

° \~ o

o >v

LEAST-SQUARE FIT \

"d C^He] = IOAeE/RT([C2H6]0-[C2H6])a8

108 POINTS E= 32.9kcal/mole,crE: = 1.53 A=7.I8, a-A = 0.33

1 I I 1 0.96 0.98 1.00 1.02

lOOO/T-dC1)

1.04 1.06

ETHANE/OXYGEN OVERALL REACTION RATE INDUCTION REGION

950-1050 K FIGURE 12

- '- m.

•

T&m S^WLfi'K^.

*V « «KB

L«*#

»"/ft ••' ' •

••if • C .';»•'«

* 1

1^1 "v

- \JL t

v • • ' • ' *•

".-•v •.

* ^ »• • >.• t "'. "••''•5;r>. rK*!lj?S

,$j

•' ' ».1 jk«

vir- '.v-r-c.

l44. .<•'"

Wi\

j

>•