AFAL-TR-87-049

AD:

o0
vFinal Report
for the period

June 1905 to

May1987

Kinetic Mechanisms for Ionization and Afterburning Suppression

November 1987

Authors: D. P. Weaver T. Singh

Approved for Public Reieass

IDistribu,on Is unlimited. The AFAL Technical Services Office has reviewlx thi3 report, and It Is releasble to the Netloiaa Technical Information Service, where it w!' be available to the general public, including !oreign nationals.

DTIC
JAN 28 1988! Air Forco

Astronautics
Laboratory

Air Force Space Technology Center Space Division, Air Force Systems Command Edwards Air Force Base, California 93523-5000

NOTICE When U.S. Government drawings, specifications, or other data are used for any purpose other than a definitely related Government procurement operation, the fact that the Government may have formulated, furnished, or in any way supplied the said drawings, specifications, or other data, is not to be regarded by implication or otherwise, or in any way licensing the holder or any other person or corporation, or conveying any rights or permission to manufacture, use, or sell any patented invention that may be related thereto.

FOREWORD This final report documents the Air Force Astronautics Laboratory (AFAL) in-house study of kinetic mechanisms for ionization and afterburning suppression. AFAL Project Manager was Dave Weaver.

This report has been reviewed and is approved for release and distribution in accordance with the distribution statement on the cover and on the DD Form 1473.

DAVID P. WEAVER Project Manager FOR THE COMMANDER

LRENCE P. QUIIRN-Chief, Aerothermochemistry Branch

ROBERT L. GEISLER Deputy Chief, Proputsion Analysis Division

.,ErURI'

IASSIPICATION OF TH-IS

AGE

REPORT DOCUMENTATION PAGE

Ia. RE.PORT SECURITY CLASSIFICATION

FormApproved
OMI N~o. 0704-0181

UNCLASSI FIED
2a 'SECURITY CLASSIFICATION AUTHORITY

lb. RESTRICTIVE MARKINGS ________________________ 3. DISTRIBUTION /AVAILABILITY OF REPORT

;'AiNG CHEULEApproved DCLASIFIATIN 2b IrOW 2b ~OW~AI.N~ OC~ASIFCATQN CHEULEunlimited. 4. PERFORMING ORGANIZATION R[PORT NUMBER(S)

for Public Release, -Astributlon is

S. MONITORING ORGANIZATION REPOR T NUMBER(S)

AFAL-TR-87-049
68. NAME OF PERFORMING 0111GANI1ZAT'N

Air Force

6b. OFFICE SYMBOL (if aggiliable)

7a. NAME OF MONITORING ORGANIZATION

Astronautics:Laboratory
6c. ADDRESS (City. Statv, ndZIP Cod)

DYCR 1b. ADDRESS (City,

State, and ZIP Code)

Edwards AFB CA 93623-5000

So. NAME OF FUNDING/ISPONSORING ORGANIZATION
Sc. ADDRESS (city, State, and ZIP Cod)

8b. OFFICE SYMBOL (if applicable)

9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER

PROGRAM

10. SOURCE OF FUNDhING NUMBERS TASK PROJECT

ELEMENT NO.

NO.

NO

WORK UNIT

ACCESSION NO.

11I TITLE (Include Security Cawisficition)610F

3500T

KINETIC MECHANISMS FOR IONIZATION AND AFTERBURNING SUPPRESSION (U)

12. PERSONAL AUTHOR(S)

Weaver-, David P., and Singh, 1'.

.3a. TYPE OF REPORT

Final
16 SUPPLEMENTARY NOTATION

IFROM

1l3b TIME COVERED TO

Ui

14DATE

0F REPORT (VearMonth,Dy

87/11

'PAGCON

31

17.
FIE._D

COSATI CODES
GROUP SUB-GROUP

18. SUBJECT TERW (Continue on revu'e if necesary and identify by block number)

21
19 ABSTRACT

02
tinue

Inhibition, Ions, Chemical Kinetics, Potassium, Hydrogen/ Oxygen Flame, Rate Constants, Fuel.Rich, Fuel Lean, Kinetic
nn

by block number)~

.~

Curent vrylit e 'a exists detailing the ch a-";-'-chansm asscite wtl te ddtin of suppressants such as O4)or(W4TlCand their impact on the ,Oroductipr of ions in either typical laborator-1tam6- e. solid propellant oxhus ws. On an empirical basis, potassium-bearing salts (K jN 2 , etc.) have been found to be among th3 most general and 2S0 4 most effective suppressants. The addition of such alkali metal salts to the plume gases, however, has been observed to increase ion/electron conicentrations, which in turn may significantly affect the radar cross section and reduce the motor's performance (e.g., by producing a lower specif'ic impulse).This report summarizes the results of a recent analytical investigation into the influence of the addition of m salts on both the inhibition effects and the produeiion
elecitrons in a g/'Air flame A;7
20 DISTRIBUTION / AVAILABILIT I OF ABSTRACTg3UNCLASSIFIED/UNLIMITED 03 SAME AS RPT 22a NAME OF RESPONSIBLE INDIVIDUAL
.______

TIC USERS

David P. Weaver

21 ABSTRACT SECURITY CLASSIFICATION I HNZr 22b. TELEPHONE (Include Area Code) 12,';. OFFICE SYMBOL

(805) 275-5656

1_DYCR
SECURITY CLASSIFICATION OF THIS PAGE

DD Form 1473, JUN 86

Previousedltfonsore obsolete.

Block 19 (continued):

An eleven step reaction mechanism was used for computing the structure of the base hydrogen-oxygen flame. This mechanism has been shown to give good agreement between the computed results and the corresponding experimental data. An additional sixteen step mechanism was added to simulate the influence of potassium inhibition and Ionization. The computed concentration profiles of [H] and [OH] show good agreement with experimental data reported in the literature. The influence of KOH addition on the laminar burning velocity, peak OH mole fraction, the rate of fuel and oxidizer decay, the rate of production of product species (including K+ and e- ) are presented. Analysis shows that the inhibition effect of K-salts ( decrease in burning velocity, peak H and OH mole fraction ) increases as thG mole fraction of KOH increawes in the unbumt mixture. The peak electron mole fraction increases as the inhibition effect of potassium salts increase for both fuel-rich and fuel-lean mixtures. The effect of variation in the rate constants for potassium reactions has been studied. Based on this analysis, two simplified reaction mechanisms ( a five step mechanism and an eight step mechanism ) are suggested foi the inhibition effects.

Ao0es8lon For

TIS GRI DTIC TAB

Just I f I cat,I

C1
_. ....

Unarnnounoed

[1

\N

Distrlbution/ Availability Cedos Avalr . anAIor Dist SPACIal

Sii

TABLE OF CONTENTS

iNTRODUCT ION INHIBITION EFFECrS OF KOH CHEMISTRY OF FLAME SUPPRESSION STUDIES RELATED TO IONIZATION IN FLAMES COMPUTATIONAL MODEL CHEMICAL KINETIC MECHANISM IO IZATION EFFECTS SUMMARY AND CONCLUSIONS REFERENCES

1 2 3 6 8 9 11 22 23

LIST OF FIGURES Figure 1 2 3 4 5 6 7 8 9 10 11 12 13 Caption Peak (e-) vs KOH Mole Fraction Peak (e-) vs KOH Mole Fraction Peak (e-) vs KOH Mole Fraction Burning Velocity vs (KOH) Burning Velocity vs (KOH) Burning Velocity vs (KOH) Peak e- Concentration vs Rate Constant Peek (e-) vs Reactiun Rate Peak (e-) Concentration vs Rate Constant Peak (e-) vs Reaction Constant Burning Velocity vs Potassium Added Mole Fraction e- vs Potassium Added Flame Temperature vs Potassium Added Page 14 14 15 15 16 16 17 18 18 19 21 21 22

iii

INTRODUCTION The work described in this report stems from the phenomenon in rocket and missile plumes that results from the reignition of the fuel-rich exhaust products mixing with the ambient atmosphere. The secondary flame zone, called afterburning, occurs some distance downstream in the exhaust plume, and generates temperatures typically higher than that of the nozzle exhaustplane. Therefore, afterburning is a significant source of radiation associated with the rocket in the infrared (IR), visible, and sometimes ultraviolet (UV) spectral regions. Such radiation creates or augments a variety of related general problems: launcher location, early warning, target evasion, and temporary blindness or loss of night vision. Over the past two decades a variety of substances have been added to rocket and gun propellants in an effort to eliminate or reduce afterburning (or gun muzzle flash), and a variety of compounds have been found to be effective to a greater or lesser extent. However, the physical/chemical mechanisms of afterburning suppression are not completely understood. A suppressant effective in one propellant may not work in another, and the degree of effectiveness cannot be accurately predicted. On an empirical basis, potassiumbearing salts (K 2 SO4, KNO 2 , KCI, etc.) have been found to be among the most general and most effective suppressants. The addition of such alkali metal salts to the plume gases, however, has been observed to increase ion/electron concentrations, which in turn may significantly affect the radar cross section and reduce the motor's performance (e.g., by producing a lower specific impulse). The description of a rocket exhaust as a supersonic jet of hot fuel-rich gases is deceptively simple. The generalized physical phenomenon that it repre.snas, reactive turbulent mixing, is extremely complex. It is not the objective of this report to attempt to solve this general problem or the specific example of rocket afterburning. To approach an understanding of afterburning suppression, it is necessary to assume a phenomenological model based on the available data. As a first approximation, it is possible to identify two aspects of the problem and treat them independently; the first aspect being the gasdynamics of the flow, and the second being the reaction chemistry of the flow. In this perspective, afterburning and its suppression are conventionally explained as follows. The exhaust gases diffuse and mix turbulently with the

atmosphere. Most of the energy released from the chemical reactions taking place is from hydrogen (from the rocket) and oxygen (from the air) producing water. These reac-ions are very fast so that the energy is released in a short time period, which in the flow of the plume also means in a relatively small, localized volume of space. This, In turn, causes the temperature In that volume to increase and create what is visually Identified as the afterbuming flame zone. Afterbuming therefore occurs or does not occur depending on a dynamic balance between competing effects: the diffusion and dilution of the plume flow, and the speed of the oxidation reactions. The radical species that mediate the oxidation of hydrogen (H and OH) are normally In high concentrations In the exhaust plume coming out of a rocket. If these could be significantly reduced, the overall reaction time would be Increased. The gases would be more diffused and the volume over which tne reaction energy is released would also be increased (nonlinearly) which dictates that the temperature rise would be less. These two conditions will also tend to increase the reaction time, which delays the overall reaction and allows the exothermicity to be dissipated. In the case of potassium, it is hypothesized that radicals are depleted by a set of chain branching reactions involving KOH. Other consistent "explanations" of afterbuming suppression exist, notably those involving the participation of suspended particulate material, but the forgoing sequence of events is the most plausible. Currently, very little data exists detailing the chemical mechanisms associated with the addition of suppressants such as K2SO4 or KNO3 and their impact on the production of ions in either typical laboratory flames or "real world" propeliants. This report summarizes the results of a recent analytical investigation into the influence of the addition of potassium salts on the production electrons in a H2 / Air flame. The kinetic mechanism reported for the addition of KOH to the H2 Air flame has been detailed in an earlier report (Ref. 1).

INHIBITION EFFECTS OF KOH Potassium containing salts such as K2 SO 4 and KNO 3 have been known to be capable of inhibiting flames (Refs. 2-12). Alkali metal salts have been added to propellants/gun powders in an effort to reduce or eliminate afterburning in the rocket exhaust or to reduce muzzle flash in gun firing. Tailoring of such propellants for specific plume and performance characteristics 2

has been done largely by empiric:al variation of additives in extremely costly testing. rh!s approach ;s critically limited by the lack of fundamental knowledge of the underlying chemical mechanisms controlling the suppression process. (This was highlighted at a workshop on suppression kinetics of rocket afterburning and gun muzzle flash held at the US Army Ballistic Research Laboratory, Aberdeen Proving Ground, MD, June 10-12,1986). This section of the report reviews recent work in the chemical kinetics of afterbuming and the influence of additives on the production of ions in such flame systems. CHEMISTRY OF FLAME SUPPRESSION The afterbuming of the rocket exhaust products, principally H2 and CO, is a major cause of high plume signature and other related problems. Secondary combustion 'afterbuming) takes place when the sequence of chain branching chemical reactions (Ref. 2) H+0 O+H and the chain propagating reaction
2 2

- OH+O , OH+H

OH+H

2 -

H2 0+H

produces active tree radicals H, OH and 0 faster than these radicals can be removed by chain terminating recombination reactions such as H+H+M= H2 +M OH+H+M= H2 0+M The burning of carbon monoxide, which takes place concurrently with that of hydrogen, tends to propagate the combustion sequence. The principal reaction (Ref. 12) controlling the oxidation of CO CO +OH - C0 2 +H simply replaces one active radical by another radical. Potassium compounds have been dem',nstrated to be effective suppressants of exhaust afterburning for double base propellants fr a considerable range of thrust levels (200 N-12 kN) (Refs. 2-4). There is no agreement on the inhibition mechanisms of potassium salts (Refs. 5-9).

Hynes, Steinberg and Schoefield (Ref. 5) measured OH and Na concentrations in 10 oxygen rich flames by saturated. and low powered- laser Induced fluores"ence techniques. The flames were seeded with NaCi solutions using an ultrasonic nebulizer. A reaction mechanism of 20 pctential reactions of Na, NaO2 and NaOH in H2 /02/N2 flames las been postulated to explain the observed /calculated species profiles in the flame. The concentrations of major species H2 0, 02 were assumed equal t,) the equilibrium values. The concentrations of H and 0 were calculated from the measured OH concentration and assuming balanced chemistry for H+0
2

- OHO

O+H2-OH+H OH + H2 - H2 0+ H A simplified scheme made of four dominant reactions involving Na was able to predict the experimental data with sufficient accuracy: Na + H20
-

NaOH + H

Na + 02+ M - NaO 2 + M NaC2 + H = NaO + OH NaOH + OH = NaO + H20 Heimerl (Refs. 6-9) has reported studies on the suppression of muzzle flash (similar to the afterburning phenomena in rocket exhausts ) by the addition ot potassium salts. Veiy little data concerning alkali salts suppressants at the elementary chemical level in actual gun systems are available. Results from several key labovatory experiments were reviewed for an alkali suppression mechanism and an elaborate mechanism (of up to 40 reactions) for suppression of H2 /0 2 /N2 flames seeded with K-salts was suggested. This mechanism has yet to be experimentally verified. Friedman and Levy (Ref. 10) employed an opposed jet diffusion flame of CH4 and 02 in their experiment. They added atomic potassium to the fuel flow up to about 6% of the fuel and did not observe any quenching of the flame. They proposed that gaseous KOH is the species responsible for flame inhibition. it 4

might act through reactions such as:

KOH+H = H2 0+K KOH + OH a H20 + KO McHale (Ref. 11) studied afterbuming suppression using solid rocket propeliants formulated with known amounts of selected potassium salts. For K2SO4, KHCO 3 and K2 C2 0 4 , he observed the intensity of 4.09 11 m CO2 bands to decrease by a factor of 15 relative to the unsuppressed case. The addition of KBF 4 to the rocket propellant did not tchange the intensity of CO bands. The computation of the equilibrium species distribution for effective suppressants species (K2 SO4 ,KHCO 3 ,K2C 2 0 4 ) showed that either K or KOH, or both are important species containing potassium. The equilibrium distribution for KBF 4 (which is not an effective suppressant of the afterbuming) showed that neither KOH nor K are important potassium containing species. Relying on the results of Friedman and Levy (Ref. 10), McHale concluded that KOH is the important suppressant species. Cohen and Decker (Ref. 13) performed shock-tube experiments to determine the effect of KOH, K2 SO 4 and KNO 3 on the inhibition of H2/02 explosions. Inhibition and suppressioti were observed in experiments involving KOH (presumed in the gas pnase), but not involving K2SO4 or KNO 3 , both in the form of particulates. This work confirms McHales (111 interence about the importance of KOH and indicate that KN0 3 or K2SO 4 routinely employed as flash suppressants must first be converted physically or chemically before being effective. Jensen, Jones and Mace (Ref. 3) studied the flame inhibition by adding potassium in the form of potassium dipivaloylmethane, to a premixed H2 /0 2 /N2 flame. They monitored the temperature profile and hydrogen atom profile in the recombination region above the flame. They observed a decrease in hydrogen atom concentrations with increasing amount of suppressant. Jensen et. al. (Rtf. 4) assumed a two step suppressant mechanism: K+OH+MKOH+M

KOH+H . H2 0+K and by fitting the observations, they deduced an exoresson for thew-e rate

coefficients.The analysis of 10 fuel lean H2 /0 2 /N2 flames seeded with NaCI by Hynes et al. (Ref. 5) showed that the reaction K + OH + M - KOH + M does not make significant contributions to the analogous reactions of Na salts. The value of the rate constant used for the backward reaction for this case used by Jensen et al. (Ref. 3) seems to be in error according to Hynes et al. (Ref. 5). There appears to be a contradiction between the experimental results of Friedman and Levy (Ref. 10) and Jensen et al. (Ref. 14). The experiments may be sufficiently different and sufficiently complex that their results are correct in different operating ranges (Ref. 6). Recent work by Hynes, Steinberg and Schoefield (Ref. 5) shows that the reaction mechanism could be quite complex (20 potantial reactions). Heimerl (Ref. 6) has suggested a mechanism consisting of 40 reactions for H2 /0 2 /N2 flames seeded with potassium salts. These studies do not take into account the ionization reactions which may be occurring simultaneously in H2 /02/N2 flames seeded with K-salts. The feasibility of these mechanisms needs to be theoretically and experimentally investigated and will to be discussed in a later work. STUDIES RELATED TO IONIZATION IN FLAMES A number of studies have been reported on the rate of chemi-ionization in the laboratory flames. Most of these studies have been performed for hydrocarbon flames. Carabetta and Porter (Ref. 14) measured concentration of positive ions in CO/0 2 /N2 flames seeded with Na, K, Cs salts. Electrical probes were used for measuring positive ion concentrations. Peters and Van Tiggelen (Ref. 15) determined the overall rate of chemi-ionization in CH 4 /0 2 /N2 flames. They also used electrical probes for monitoring the concentrations of positive ions. Traverse and Williams (Ref. 16) have discussed in detail the theory and use of electrical probes in flame plasmas. Maclatchy and Forsman (Ref. 17,18) have demonstrated the use of a double probe (one probe acts as a reference probe and the other acts as the test probe) for measuring ion density in propane air flames seeded with NaOH. Cousins and Jensens (Ref. 19) have reported the computation of ionization leve!s in the rocket exhaust. The basic combustion mechanism for H2 /CO/air flames was combined with the following reactions assumed to govern the production and removal of the charged species:

K++e*+M-

K+M

(1)
(2)

K+ + C"-, K +C

K+ +C1"+M a KCI+M K + H2 0 a KOH+ H K+OH+MKOH+M

(3) (4) (5)

The rate at which the reverse step proceeds for reaction 3 depends on the specific concentration of KCI and is linked to K concentration through the reaction: K+HCI ,KCI+H (6)

Reaction 6 was found to be close to equilibrium. Electron attachment and detachments are assumed to be affected through the following reactions (Ref. 19): HCI + e-= CI" + H Cl+e- +M- CI-+M (7) M8)

Reaction 7 was found to be the controlling reaction for the ratio [Cli]/[es] and was close to the stoichiometric balance in a rocket exhaust. Cousins and Jensen's (Ref. 19) calculated values of e" concentration were found to be dependant on hydrogen atom concentrations, which in turn were linked to the concentration of other radicals. The following bi-molecular reactions were found to be nearly balanced: OH+H O+H
2 2

= H2 0+H = OH+H

(9) (10) (11) (12)

OH+OH - H 2 0+O H2 0 + CI - HCI + OH

7
?*

The computed concentration of hydrogen atoms was found to be above equilibrium by a factor of approximately 30 at the nozzle exit. In addition, this concentration was found to be above equilibrium by a factor of 3-4 in the hottest, near stoichiometric, region ot the flame. The calculations fo; the charged spe ies were done for the exhausts of two rocket motors. Tho first motor cor'ained K, Na and Cl salts as fuel impurity species, while the seccnd motor had K, Na and Ca salts for the fuel impurities. The noise power, expressed in reiative db per 1 Hz bandwidth, were nomputed for the motor exhaust under a set of specified conditions consistent with some known data firings. The computed results were found to agree within about t 15 db with the experimental observations reported in an unpublished work (Ref. 19). The differences were attributed to uncertainties in the specific rate constants for radical recombination reactions and the ionization reactions. It was further noted in these studies that high electron concentrations are achieved in the hottest part of the exhaust and that microwave radiation traversing these plume sections are heavily absorbed. Further, charged species concentrations including e- concentrations in the outer parts of the exhaust are all well above the expected equilibrium levels. Finally, electron attachment reactions in the outer, oxygen rich regions dominate. This fact is true in particular for composite and double base propellants and should be taken into account. Some of the reactions deteiling these conditions are presented below: OH+e +M = 0H1-M 1"20 + e- = 0H- + H 02+e+M = 02 +M" 02- +OH = OH- +02 (1+) (14) (15) (16)

The rate coefficient for all of these reactions have suostantial uncertainties. COMPUTATIONAL MODEL

The experimental program needs to be supported by a computational model. A computer code for simulating the composition p,'ofile and temperature profile in one dimensional flat flame was obtained trom the Sandia National Laboratory (Ref. 20). This pi ,gram was developed by Kee, Grcar, Smooke and

Miller of the Sandia National Laboratory. The program accounts for the finite rate chemical kinetics and molecular transport. Finite difference approximations are made to discretize the governing conservation equations on a non-uniform grid from the cold boundary to the hot boundary. The Newton method Is used for solving the boundary value problem. Global convergence of this algorithm Is aided by Invoking the time integration procedures when the Newton method has convergence difficulties (Ref. 20). CHEMICAL KINETIC MECHANISM An eleven step reaction mechanism (Table 1) is used for simulating the base H2 /02/N 2 flames. This mechanism has been shown to give good agreement between the computed results and the corresponding experimental data (Refs. 21,22). This mechanism has been used by Zacharia and Smith (Ref. 23) to simulate sulfur chemistry in rich H2 /0 2 /SO2 flames, and by Coffee to simulate) H2 /N2 0 flames (Ref. 24). Added is a 15 step kinetic mechanism to describe the influence of potassium inhibition (equations 12-26).

TABLE 1, Pre Exp -A 1.2E17

1.5E07

Rac t. 2.
3.

Temp Exp Act Enrg* b _E -. 0.91

2.00 1.60

H+O 2 =OH+O 0 + H2 = OH ,+H

OH + H2 - H2 0 + H

16422.
7542.0 3295.0

4. 5.
6.

20H = O + H2 0 H + H + M = H2 + M
H + OH + M = H2 0 + M

1.0E08 i.5E09 2.OE18

2.2E21

1.14 -1.00
-2.00

0.0 0.0
0.0

7. 8.
9.

H+0 2 +M=HO 2 +M H 0-H02 = 20H

H + HO 2 = H2 + 02

2.0E18 1.5E14
2.5E13

-0.80 0.00
0.00

0.0 1003.0
692.0

10. 11. 12. 13 14. 15. 16. 17, 18. 19. 20. 21. 22. 23. 24. 25.

O+H0 2 =OH+0 2 OH + HO2 = H0 + 0 2 H 1 -KOH = K + H2 0 K+OH+M=KOH+M K+02+M=KO2 +M H2 +'KO= H +KOH H K0 2 =OOH +KOH KO + H2 0 = OH + KOH H + K02 = KOH + O H+K0 2 =K+H0 2 K + H02 = KO + OH O + KO 2 = KO + 0 2 KO + KO =K - KO2 H +KH = K + H2 KH + OH= K + H2 0 KH + 0 2 =K+H 2
2

2.0E13 2.0E13 1.0E13 0.54E21 0.50E18 0.72E13 1.0E13 1.0El 3 1.9E12 1.9E12 1.9E12 1.9E12 1.3E13 1.0E13 1.0E13 1.0E13
1.0E13

0.00 0.00 0.00 -1.00 -1.00 0.00 0.00 0.00 0.00 0.03 0.00 0.00 0.00 0,00 0.00 0.00
0.00

0.0 0.0 1984.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0,0 0.0
0.0

26. H + K02 =KH + 0

*The rate coeificients are expressed as calories units

k=ATbexp(-E/RT) in cm-mole-sec-

The associated thermodynamic data and transport data for H2 /0 2 /N2 species have been compiled by Kee et al. (Ref. 25) and were used for these computations. The thermodynamic data and tile transport data for K-species have been taken from the work of Heimeri (Ref. 7). For each species the 10

molecular mass, the Lennard-Jones collisional diameter (in Angstroms), the .ennard-Jones potential well depth (in Kelvins), the dipole moment (in Debyes), the polarizabillty (in cubic 'Angstroms), and 14 thermodynamic parameters for use in the polynomial fits of Gordon and McBride (Ref. 26) have been compiled. The forward "rate constants fo" H2 /02/N2 reactions have been taken from Ret. 22. The rate data for'K-reactions have'been taken from Refs 6 and 7. The backward rate constants are calculated from the forward rate constants and the equilibrium constants (based on asso!:-aed thermodynamic data). There Is a large degree. of uncertainty associated with the rate constants involving Kreactions. The rate constant for the reaction K + OH + M= KOH + M has been estimated to be 5.4E20 T- 1 cm 6 /mole 2 /sec by Jensen, Jones and Mace. (Ref. 3). This result was derived from their temperature and composition profile measurements in a H2 /0 2 /N2 flame seeded with K-salt and supported on a flatflame burner. The recent measurements of Husain et al. (Ref. 27) show this value to be 4.7E21,rr 1 .28 cm6/mole 2 /sec. We used a value of 5.4E20 T- 1 for the rate constant of tnis reaction for baseline calculations and varied it up to 5.4E21 T- 1 in-our computations. The rate data for the reaction K + 02 + M = K0 2 + M has been measured over the temperature range of 300-800 K. Silver et. al. (Ref. 28) recommond a value of 0.478E20 T- 0 . 6 while Husain et al. (Ref. 29) recommend a value of 0.5E1 8 T-1 cm 6 /mole 2 /se- for this reaction. Hynes et al. (Ref. 5) recommend a value of .72E18 T-1 for the corresponding analog reaction (Na + 02 + M NaO 2 + M) of Na system . A value of 0.5E18 T- 1 was used for the rate constant of this reaction for baseline computations and has been varied up to 0.5E21T' 1 . The rate constant for H + KOH = K ,. H2 0 was estimated to be 1.0E13 exp(-19 8 4 /RT) by Jensen and Jones (Ref. 3) and this value has been used in our computations. Very little data is available for the rest of the K-species reactions. Their rates havb been estimated to vary from 1.OE1 1 to 1.OE1 4 cm3 /mole/sec Ref. 7). IONIZATION EFFECTS The addition of potassium salts to flames leads to a varying degrees of ionization depending on the temperature and concentration of K-salts present. A summary of the literature related to chemi-lonization in flames has been presented. For this work, the sensitivity of various ionization reactions has been 11

studied by adding one reaction at a time to the 26 step reaction mechanism shown In Table 1. For potassium salts, the primary reactions governing the production and removal of charged species are presented below (Ref. 19): K++e'+M= K+M K+ Cl K+CI (17)

(18)
(19)

K++CI'+M= KCI+M

For the addition of KOH, the reaction K+ + e- + M K + M would be most significant. In this reaction the rate constant is equal to 1.44E22 T-1 with an uncertainy factor of 30 (Ref. 1,,). The computed peak electron concentration mole fractions determined by the addition of Table 1 reaction 17 to the 26 stop reaction mechanism are summarized in Table 2. These computations were performed for various equivalence ratios ranging from fuel-rich to fuel-lean conditions,

12

TABLE 2

,2Q2L.N2/KQOHBRto

P.iAkXI.Er
042.38

Sumrn Yell CM

.50/.10/.39/.00001 .49/.100.39/.005 .49/.10/.39/.010 .494.100191.0115 .49/..101.391.020

1.189E-10 1.210E-07 2.400E-07 3.250E-07 3.790E-07 0-0.79

257.7 236.9 216.4 196.4 176.7

.251.16/.59/.001 .25/.16/.59/.005 .25/.16/.59/.010

1.800E-08 1.61 OE-07 3.360E-07 0=1.28

167.5 149.7 128.1

.35/.14/.51/.00001 .35/.14/.51/.0001 .35/.14/.51/.001 .35/.14/.51/.005 .35/.14/.51/.01 .34/.13/.50/.015 .34/.13/.50/.020

2.650E-09 2.800E-08 4.890E-07 4.880E-06 8.800E-06 1.010E-05 1.020E-05

262.4 261.6 259.0 244.9 226.7 205.3 189.9

From Table 2 and Figs 1 - 3, it is seen that peak electron mole fraction increases as the mole fraction of KOH increases in the unburnt mixture for both fuel rich and fuel lean mixtures. This trend Is opposite that exhibited by burning velocity, as shown in Figs 4 - 6, which is seen to decrease as the KOH mole fraction increases in the unburnt mixture. Thus, it is seen that as the inhibition effect of K-salt (seen as a decrease in burning velocity and peak OH mole fraction) increases, the peak mole fraction of electron concentration increases.

13

0.00

0.01 KOH Mole FrwUtion

0.02

Figure 1. Peak fe-I vs KOH Mole Fraction

4.

.12
'1 0.79

0.001

0.003

0.005

0.007

0.009

KOH Mole Fraction

Figure 2. Peak re-I vs KOH Mole Fraction

14

12

~10-

01.23

0~ 0.00

0.61
KOH Mole FMation

0.02

Figure 3. f.flk re-I vs KOH Mole Fraction

f200160 0.2.'"

0.00

0.01

0.02

Figure 4. Burling Velocty vs [KOH)

15

S~170
t160

S1150

I1301
*a

140

.79

120 0.000

0.002

0.004

0.006

0.008

0.010

KOM se Fmsetiw

FIgw S. fUMi

Vfocite VS EKOM

260

24O

180 0.00

0.01 KOH Mole Fraction

0.02

Figure 6. Burning Velocity vs (KOH)

As was discussed earlier, the rate constants for the K-reactions have a high level of uncertainty. The sensitivity of various key reactions and their influence on the peak electron mole fraction was examined by changing the pre-exponential factor of the individual rate constants one at a time. The results for these reactions are given below: REACTION - K*+e'+MaK +M 16

Fig. 7 shows the sensitivity with respect to rate constant for the reactin K+ + + e" + M . K + M. As the pro-exponential factor for this e"action Is

increased by three orders of magnitude, the peak electron mole fraction

increases by a factor of about 8.5. The rate of change for peak electron mole fraction is smaller as the rate constart pre-exponential ?iatr Is increased from 1 .0E23 to 1.0E24, as compared to the Increase from 1.0E22 to 1.0E23.

4Q-7
30-7 /1263,0

SO -M 0,70
0e+O 21 22 23 24 25

,oqlO(A)-K*+ *- +Ma K +M-

Figure 7. Peak a-

gCo(AntfrKL v. Rat Constant

REACTION H + KOH a K + H20 From Fig. 8 it is oeen ihat the computed peak electron mole fraction increases as tie prm-exponential factor for this rate consfant Is Increased by two orders of magnitud& for various concentration levels of KOH at 0 - 2.38. The increase in tt, fate 'on.tant for th, reaction H + KOH - K + H20 also results in a decrease in the burning velocity. This means an incresse in the inhioition effect.

17

NZO 40 4.67/1/&76

XKOH,*.02

J'o

I
10

10,01

11

I2

13

14

L.gl(A).-H KO - K 10 Figure 8. Pek Me-! vs R1eation F81te REACTION K + OHMUKOH + M Fig. 9 shows that the computed peak electron mole fraction again increases with the rate constant increase for the reaction K + OH + M - KOH + M. These results are shown for various concentrations of added KOH. The burning velocity decreases (the inhibition effect increases) as the rate constant for this reaction increases.

6H2XX/N2XKHoa so.~4.76/l/3.76 j40 320 130 10 0-2.38

KH02 OH15

.'XoH~05

XKOH..005
19 20 21 22

LoglO (A)-K+OH*M-KOH*M

Figure 9. Peak e- Concentration vs Rate Conatant 18

REACTIOH

K , 02 + M a K02 + M

Fig. 10 shows t

the computed pek electron mole fraction decreases

d. The burning velocity

sliCy as the rate conMtant for this reaction ISIncOMe

incr,-ms as the rate constant Incoreases. This Is e*.,valent to a decrmee In the effect of Inhibition n the rae constant noreaset.

S 40 0

,-W0

SQ. .3sXKOH..o0s
i 7117 XKOH-.005.23 KOmO

10 16

, 17

18

19
Mu KOi +M

20

Lql*O(A)-K0O

Figure 10. p1ak je.! vs Reaftion Constant From the above analysis It can be seen that the peak electron concentration is proportional to the Inhibition effect. The inhibition effect is identified by a decrease In burning velocity with an accompanying decrease In the peak OH mole fraction of flames to which potassium containing salts have been added.

19

In the above analysis it can be seen that the chemical kinetic mechanisms controlling flarnre Inhibition can be modelled and connected to those responsible for Ion production in a flame seeded with potassium salts In low concentrations. As can be seen, the desired chemical kinetic influence in the flames of interest can be obtained by selectively manipulating the available radical pool to control and distribute the released heat and to simultaneously minimize the production of Ions. In order to Illustrate this point graphically, a series of computations were performed using the 26 step reaction mechanism shown in Table 1. To this mechanlim was added the reaction K+ + e" + M - K + M. This reaction is most significant in its Influence In the reaction scheme when KOH is added to the flame. Table 3 presents results for equilibrium mole fraction of e-, flame speed, and flawe temperature for the cases where potassium is added as 1) atomic potassium or 2) KOH. These results are shown graphically in Figs. 11 - 13. TABLE 3 K Percent Temperature,K e- Equilibrium Mole Fraction* Flame Speed, cm/sec (Initial K as) KOH K 236.7 163.3 106.7 88.3 236.7 175.0 125.0 83.3 30.0 30.0

(Initial K as) KOH K 2017 2024 2032 2017 1990 1963

K 0.0 3.429 6.571

KOH 0.0 3.857 2.857

0.0 1.0 2.0 3,0 4.0 5.0 7.0

2050

1934

7.190 9.095

1.286 1.238

Mole Fraction Approximately E- 10.

20

a 12w

300a

v.KmKOH

11001
0 248 Porout Potomham Added

Figure 11. Hurning Velooity va Potmasium Added

I102-

2om Potesekam Added

Figure 12. Male Fraction a. vs P~otassiulm Added

21

2060

2040 : 2020 2000

S1980

Ka

"

1960

1940
1920 -",

Percent Potassium Added

Figure 13. Flame Temperature vs Potassium Added As the concentration of potassium is increased from 0 to 5% in the unburnt fuel mixture, the equilibrium adiabatic flame temperature increases from 2017 K to 2050 K. The corresponding equilibrium mole fraction of free electrons ( e-) changes from approximately E-1 0 to 9.095E- 7 . If the potassium is present bound as KOH, then the equilibrium adiabatic flame temperature decreases fr,m-n 1990 K to 1934 K as the percentage of KOH in the unburnt fuel mixture increases from 1 to 5. The corresponding equilibrium mole fraction of free electrons peaks at 3.857 E-7 for the addition of 1% KOH and then decreases. Hence the ability to control collectively the inhibition effect and th e- concentration. SUMMARY AND CONCLUSIONS -The chemical kinetic mechanisms controlling flame inhibition and the production of irins in a hydrogen-oxygen flame seeded with potassium has been modelled. An 11 step kinetic reaction model was used for computing the structure of the base hydrogen-oxygen flame. Fifteen additional reactions for the addition of KOH were used to simulate the inhibition effects. Additional individual reactions were added to this mechanism one at a time to evaluate the degree of ionization inouced by the introduction of potassium salts to the unburned fuel mixture. -The laminar burning velocity and peak mole fraction of H and OH decrease as the mole fraction of pctassium seed increases in the unburnt 22

mixture. This observed trend was sinilar for both fuel-lean and fuel-rich flames. -The peak electron concentration was found to be proportional to the
potassium salt inhibition effect. -Selective manipulation of the available radical pool can be used to distribute the released heat within the flame for the purpose of inhibition while at the same time minimizing the production of ions (in this case free electrons). -The particular mechanism used in the process of flame inhibition can be the key in limiting ion production. The use of KOH as an inhibitor can provide successful inhibition at 3% suppressant addition, while simultaneously controlling electron production when compared to the use of K as the inhibitor. -The suggested mechanism can ba used to evaluate the specific increase in electron concentration with suppressant addition. Evaluation of the optimum degree of suppression and specific level of electron production can be performed. References: 1. Singh, T.S., Weaver, D.P., and Eversole, J.D. Afterburning Suppression Kinetics, AFRPL-TR-87-001, Air Force Rocket Propulsion Laboratory, Edwards Air Force Base, California, March 1987. 2. Jones, G.A. and Mace, A.H.C., "Secondary Combustion Suppression in Rocket Exhaust", ProceedingS of the WorkshoD on the Chemical Suppression of Rocket Exhaust and of Gun Muzzle Flash, US Army Ballistic Research Laboratory, Aberdeen Proving Ground, MD., June 11-12, 1986. 3. Jensen, D.E., Jones, G.A., and Mace, C.H.," Flame Inhibition by Potassium"' Journal of the Chemical Society Faraday Transactions, I, Vol. 75, pp. 23772385, 1979. 4. Jensen, D.E. and Jones, G.A. "Kinetics of Flame Inhibition by Sodium ", JLour..l Lof.h Chemical Society Faraday Transactions, I, Vol. 78, pp. 28552866, 1982. 5. Hynes, A.J., Steinberg, M., and Schofield, K. "The Chemical Kinetics and Thermodynamics of Sodium Species in Oxygen-Rich Hydrogen Flames 23
",

Journall of Chemnical Physics, Vol. 80, pp. 2565-2597, 11984.

6. Helmerl,,J..M., Kelle9r,G.E.

n Kilngenberg, G. MuzzIle Fash, Ktnticg'a, M~ijin. Rpot 1/85FrunhferIntltute 'fur:Kurzzeltdyrnamlk, Ernst-MachInstitut, Wel Am Aen a 95

7., Helmerl, J. M.' An Advanced Flash Supnjrbsion -Network rInvolng"'Alkali, BA1,BRL-TR 2622, US Army. -Ballistic Researc h Laboratory, Aberdeen Proving Ground, MD., December 1984.' 8. Heimerl, J. M., Muzgia..e lagh and.Alklali Salt' Inhbibition fto p Elemengtary Kinetic Point of Viewt, BRL-TR-02479', USArmy .B*Rlllstic .ResearhLaotry Aberdeen Proving Ground, MD., March 1983.

9. Keller, G.E., An Evaluation of MduzZle Flash Pr,,diygtieModj1';BRLMR03318, US Army Ballistic Research'Laborator ,y,Aberd~een Provilng Ground, M~I., November, 1983. 10. Friedman, R. and Levy, J.B. of. Oppose~d-Jdt M kethane-.Air Diffusion Flames. The Effect of alkali Metal Vapors and Organic Ha'lides", Cgmbtstion and Flame, Vol. 7, pp. 195-201, 1963.
"Inhibition

11., McHale, E.T. "Flame Inhibition by.Potassium Compounds", Combustion and

Flame, Vol. 24, pp. 277-279, 1975.

12. Singh, T. and Sawyer, R.F. "CO reactions in the Afterflame; Region 'of Ethylene/Oxygen and Ethane/Oxygen Flames", 13th Symposium (International) on QomIbutioUn, the.Combustio'n Institute, Pittsburgh, Pa., pp 403-41 6, 1971. 13. Cohen, A. and Decker, L. " Chemical Mechanisms for Secondary Flash Suppression" 18th Symposium (International) onJCombustion, the Combustion Institute, Pittsburgh, Pa., p.225-231,1981. 14. Carabetta, R. and Porter, R.P. " Absolute Positive Ion Concentration Measurement in Flames with Langmuir Probes", 12th i3LmoQ.i~urn (International) on Combustion, the Combustion Institute, Pittsburgh, Pa., pp 423435, 1969. 15. Peters, J. and Van Tiggelen, A. "Experimental Determination of the Rate of the Chemi-lIonization Process ", 12th SyMposium (International) onl 24

Combustion, the Combustion Institute, Pittsburgh, Pa., p. 437-446, 1969. 16. B.F.L. Travers and H. Williams, " The Use of Electrical Probes In Flame Plasmas ", Tenth Symposium (international) on Combustion, the Combustion Institute, Pittsburgh, Pa., pp 657-672, 1965. Novel Electrostatic Probe for 17. Maciatchy, 0.5 and Forsman,J.W. Measuring the Ion Density in a Flame", Combhustion and Flame Vol.53, pp.4148, (1983).
"A

18. Fontijn, A. and Vree, P.H. "Cheml-7ionlzationiin the Reactions between C2174 gen .Mi ,xture", Eleventh Symposium Ox.y and Atomic Nitrogen - AtomiclO Institute, Pittsburgh,. Pa., p. 343, Combustion the Combustion, on (Ir'ternational) .1967. 19. Cousins, J.E. and Jensen, D.E. "On the Computation of Ionization Levels in Rocket Exhaust Flames", Combustion and Flame, Vol. 52, -pp. 111-125, 1983. 20. Kee, R.J., Grcar, J.F., Smooke., M.D., and Miller, J.A., A Fortran Prog~am for Modelling Steady Laminar One-Dimensional Premixed Flamel', SAND858240, Sandia National Laboratory, Livermore, California, December 1985. 21. Warnatz, J. "Calculations of Laminar Flat Flame 11i: Flame Velocity and Structure of Freely Propagating Hyd('rogen-Oxygen and Hydrogen Air Flames", Physical Chelmistry, Vol. 82, pp. 643-649 ,1978. 22. Warnatz, J., 18th Symposium (internaticnfl) on Combus6tiony, the Combustion Institute, Pittsburgh, Pa., pp 369-384, 1981. 23. Zachariah, M.R. and Smith, 0.1. "Sulphur Chemistry in Rich H2/0 21S02 Flames", ~jqjjaj Canadian and Western States Sectio~n of the Combustion

Institute, WSSCI-86-7, Banff Canada, April 1986.

24. Coffee,T.P. "Kinetic Mechanism for Premixed Laminar, Steady State

Hydrogen/Nitrous Oxide Flames", Combustion and Flame , Vol. 65, pp. 53-60
1986. 25. Kee, R.J., Wartanatz, J., and Miller, J.A. &Eortran Computer Code Package for the Evalumidon of Gas-Phalse Viscosities. Conductivities. and Diffusion Coeff 'cient, SAN D83-82009, Sandia National Laboratory, Livermore, Ca~ifornla, 25

March 1983. 26. Gordon, S. and McBride, B.J. Conmuter Pragram for Cislculations of CompleX Chemical Equilibrium Compositions. Rockqt PErfgrrnance-. Incident anid Raflectatd Shodck and Chgpman..Jougjuet Detonations, NASA-SP-273, 1071 (1982 Program Version). 27. Husain, D., Plane, J.M.C., and Xiang, C.C. "ADirect Kinetic Study of the Reaction K+OH+He=KOH+He by Time-Resolved Molecular ResonanceFluorescence +Spectroscopy*. Journal of the Chemi3cal Socigty. F.araday

Iransacins 2,Vol 80,

pp. 1465.,1483, 1984.

28. Silver, J.A.,, Zahniser, M.S., Stanton, A.C. and Kolb,C.E. "Temperature Dependent Termolecular Reaction Rate for Potassium are! Sodium Superoxide Formation", Twentieth Symposium (international) on Qomnbustlon, the Combustion Institute, Pittsburgh, Pa., pp 605-612, 1984. 29. -HusAin, D. arid Plane, J.M.C.' "Kinetic Investigations of the Third Order Rate Processes between K+02+M by Time-resolved Resonance 'Absorption Spectroscopy", ~Journal of the Chemical Society. Faraday Transactions 2, Vol. 48, pp 1175-1.194, 1982.

26