AFOSR

TR

76-

07 9

COMBUSTION

STUDIES OF FUEL-RICH PROPELLANT

Merrill K. King

August,

1976

FINAL REPORT

Kinetics and Combustion Group Atlantic Research Corporation 5390 Cherokee Avenue Alexandria, Virginia 22314

TR-PL-5520

OCT 19

1976

Contract No. F44620-71-C-0124

Approved for Public Release; Distribution Unlimited

Air Force Office of Scientific Research Building 410 Boiling AFB, DC 20332 ATLANTIC RESEARCH CORPORATION AL EXANORIA.VIRGINIA - 22314

iiI

Qualified requestors may obtain additional

copies from the Defense Documentation Center, all others should apply to the National Technical Information Service.

I I I I I
Conditions of Reproduction Reproduction, translation, publication, use and disposal in whole or In part by or for the United States Government is permitted.

I
AIR F'ruC OV'FtC CF SCIENTIFIC RESEARCH (AFSI NOTICE OF TLAZSMITTAL TO DDC Ihis techiLcl report ha ... a reviewed and is approvcd for publ ;, 1 AvVi A A 1IJO-12 (7b). DiLLributiou is wllmited. A. D. BLDSE Technical InforLntion officer

4CESSiON 1w
ITIS te K N

COMBUSTION STUDIES OF FUEL-RICH PROPELLANTS

7.1

<A

Merrill K. King

August, 1976

FINAL REPORT

Kinetics and Combustion Group Atlantic Research Corporation 5390 Cherokee Avenue Alexandria, Virginia 22314

TR-PL-5520

I
i
Contract No. F44620-71-C-0124 Approved for Public Release; Distribution Unlimited Air Force Office of Scientific Research Building 410 Bolling AnI, DC

20332

OCT 19 1976

, I

-K,
----

"- . . . .

UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE (WheunDote Entered)

REPORT DOC)4fTATIOt4 PAGE

REPOR. GOVTe ACCESSION NO.

READ INSTRUCTIONS

BEFORE COMPLETING FORM

3. RIECIPW" CATALOG NUMBER

Combustion Studies of Fuel-Rich Propellants&

uV31

~ue7

O N T OI N A MDNA ME C 4~~~~~~~~~~~~~~~

r n f o e CO A7EN

f c )

R T 10 7 L S

3)fd f

J *u r7

6rt

5 PESRRINGIORNIST

ATEN

NAo tAND E

ADDoRESt0)RGA

LMET

RJC.T

Alnticeseforchu~ Copreaton

6813D8it

his FrepOrtfpeofSentf r~easeveral tak R aiSme a761tfin Bild~ min ng highl maneim-oae fomuaton fo ai-aumete and MOfIORN-cnann
LtTO DD~~~~~

pGNYNAE6aDrtS~iculaeeta from
FINV65I

fighyboo-lae holt~ I.SCRTCLS.(futi e-rich RLS

SECUNITY CLSSF s

SOES

UNC43 LASFE

C ASIF

~IC

AGE

mN' O

de

ntADING

JUNCLASSIFIED
sECuT C"ASSIFICATION Or THIS PAGr, D.a Etro.d,

20.

of the motor ejecta were studied e their f;nition/combustion characteristics were compared to those for pure boron. The minimum surrounding temperature for ignition for the two materials was nearly identical but the ignition delay time was considerably lower for the motor eJecta material. Burning times for comparable size particles of each type were equal within experimental error. In the size range of interest (2 t2 10 micron diameter), the burning was observed to be characterized by a d-1aw (kinetics-limited mbusto rather than the d 2 -law (diffusion-limiLed .,ombustion) observed with larger particles. A model of boron ignition treating the Inhibiting -ffcr of a boric oxide coatinfz was develoned rd found to give good agreement with boron particle ignition data in dry gases. A detailed model of boron-oxvgen-nitrogen dust cloud flames including this boron narticle Ignition model was developed and used for prediction of flame speeds in such clouds as a function of nnmrrtis paramet'rs; again, good agreement was found with the limited available eyperimental data. I Thermodynamic screening of possible highly-magnesium-loaded fuel-rich propellants was carried out and immeral candidate castable and pressed formu lations for air-augmented rocket applications were defined. Several of thpqs candidates~were s\ formulated aharacterized , in terms of their safety and ballistcproperties. Experiments were performed to identifv the characteristics of the exhaust products r.oasm.tin from combustion mf-the pgopelants in the absence of air. These experiments indicated that considerable magnesium, with a significant fraction of it in the gas phnqe -IS produced. Alr4 th the presence of such quantities of magnesium gas should result in excellent combustion of the product streams emanating from these propellants with air in a ramburneqr. Motor tests were plugging conducted with four ofthese with Identification ofozzle and residue formation formulations, sopossible problem areas#,however, thesd tests did indicate likely approaches to resolution of these problems. Connected-pipe afterburning tests were conducted with the three most promising formulations. 4 These tests indicate that very high afterburning efficiencies may be achieved under extremely adverse operating conditions where state-of-the-art high-boron-loaded formulations 9 frP.* 1 ineffiS cient afterburning.

I ! I I I

I
UNCLSIFIED
SI[CURgTY CLASSIFICATION OF THIS pAGIE(%honaa ffot ,,..dW

I7

I
TABLE OF CONTENTS

I
I. II. INTRODUCTION .............................................. BORON COMBUSTION STUDIES .................................. A. B. C. Collection and Analysis of Condensed-Phase Exhausts of Boron-Loaded Fuel-Rich Propellants ....... Combustion of the Condensed Exhaust.................. Analytical Modeling of Boron Single-Particle Combustion and Boron Cloud Flame Speeds ..............

Page
I 3

10 I0 19 19 24 24 -5 43 50 5C

III.

MAGNESIUM PROPELLANT STUDIES .............................. A. B. Background and Purpose ............................... . Thermodynamic Screening .............................. 1. General Criteria................................. 2. Castable Compositions .......................... 3. Pressed Compositions ............................ 4. Candidate Formulations Selected.................. Propellant Formulation and Characterization .......... 1. Preliminary Formulation Work on Magnesium-Loaded Castable Fuel-Rich Composite Propellants (Feasibility Assessment) ......................... 2. 3. Formulation of Candidate Defined in Thermodynamic Formulation of Candidate Defined in Thermodynamic Castable Compositions Screening Effort ....... Pressed Compositions Screening Effort .......

C.

59 63 76 80

D. E.

Propellant Combustion Product Characterization ....... Single-Particle Burning Studies ...................... .101 1. Magnesium........................................ .105

I
F. G.

2.

Magnesium Nitride ................................

105
106

Theoretical Effect of Magnesium Vapor on Plume

Ignition .............................................

Combustion of Selected Candidates in Motors .......... 1. Castable Compositions ........................... 2. Pressed Compositions ............................ Connected-Pipe Afterburner Tests ..................... 1. Castable Compositions ...........................
2. Pressed Compositions ............................

112 112 124 124 124

141 144

H.

IV. V.

REFERENCES ................................................ BIBLIOGRAPHY - Papers and Reports Resulting From This Effort ....................................................

146

!_

LIST OF TABLES Page 1 2 Motor Firing Conditions ............................................. Thermodynamic Equilibrium Computations ............................ Chemical Analysis of the CE (Weight Percent) .................. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Heating Values for Various Types of Practical Fuel-Rich Propellants ................................................... Maximum Volumetric Heating Values Obtainable With Castable Magnesium Systems Examined ........................... Optimized Volumetric Heating Values out of Various Possible Magnesium Pressed Compositions ...... ............ ....... .... Candidate Compositions Selected for Further Evaluation ........ Summary of Preliminary Formulation Work to Define Suitable Magnesium-Loaded Composite Propellant Systems .................... Formulation MKI Mixes to Optimize Particle Sizes .............. Formulation MKI Pot Life and Mechanical Property Optimization. Formulation MK2 Mixes ............................................. Formulation MK12 Mixes ............................................ Tensile Properties of Final Candidate Formulation............ Hazard Properties of Final Candidate Formulations ............. Summary of TKI Properties ......................................... Summary of TK2 Properties......................................... Summary of TK3 Properties ..................................... Results of Calibration Tests for Optical Absorption Measurement of Magnesium Vapor Density........................... Results of Absorption Measurements for Four Uncured HandTamped Formulations Burning in Argon at One Atmosphere ........ Results of Absorption Measurements for Formulations MKI, MK2, and MK12 Burning at One Atmosphere in Argon.................... Results of Sample Test of MK2 Burned at 500 PSI in a Closed Bomb ................................................... Nascent Magnesium Yields in Closed Bomb 500 PSI Combistion Tests ......................................................... 92 i' 94 100 102 1+2 55 54 61 69 67 .'8 69 10 71 77 78 4 6

*1

-,

LIST OF TABLES (Continued) Page 23 24 25 26 Kinetic Rate Expressions unse for Hydrogen and Carbon Monoxide Combustion Reactions in Plume Ignition/Combustion Calculations. Grain Configurations and Sizes Used in Motor Tests of Castable Formulations ............ ............................ Residue and Throat Closure Characteristics of Motor Firings of Castable Formulations ...................................... Parameters Measured Continuously During Each ConnectedPipe Ramburner Test ................................... .......

109 115 116 132

LIST OF FIGURES I 2 Burning Times (tb) of Various Particles ....................... Comparison of Fuel Flow Requirements for Given Operating Design Point for MKI versus State-of-the-Art Boron Formulation ................................................... Volumetric Heating Value Versus Primary Motor Flame Temperature for NC/PEG400/Magnesium System .......................... Volumetric Heating Value Versus Fraction in Condensed-Phase Combustibles for NC/PEG400/Magnesium System ................... Heating Values Versus Magnesium Loadings for NC/PEG400/ Magnesium Systems ............................................. Heating Values Versus Magnesium Loadings for NC/NG/ADN/ Magnesium Systems ............................................. Thermodynamic Calculations for NC/DOA/AP/Mg Systems ........... Heating Values Versus Magnesium Loadings for NC/DOA/AP/ Magnesium Systems ............................................. Heating Values Versus Magnesium Loadings for HX730/TEGDN/ HMX-AP/Magnesium Systems ...................................... Thermodynamic Calculations for HTPB/HMX/AP/Magnesium Systems.. Heating Values Versus Magnesium Loadings for HTPB/i14X-AP/ Magnesium Systems ............................................. Heating Values Versus Metal Loadings for HTPB/HMX/Magnesium/ Aluminum Systems .............................................. Heating Values Versus Boron Loadings for HTPB/HMX-AP/ 15

22 27 28 29 31 33 34 35 37 38 40

3 4 5 6 7 8 9 10 11 12 13

Magnesium/Boron Systems .......................................

41

K'

LIST OF FIGURES (Continued) Page 14 15 l Volumetric Heating Value Versus Flame Temperature for Four Potential Pressed Composition Systems ......................... Volumetric Heating Value Versus Moles of Gas/IO Grams of Propellant for Four Potential Pressed Composition Sv trms..... Volumetric Heating value Versus Percent Heating Vnlui, in Condensed-Phase Primary Motor Products for Four Pcrontial Pressed Compositions .......................................... Volumetric Heating Value Versus Metal Loading for Four Possible Pressed Magnesium Composition Systems ................. Total Volumetric Heating Value Versus Metal Loading for Various Candidate Pressed Propellant Systems .................. Reduced Volumetric Heating Value Versus Metal Loading for Various Candidate Pressed Propellant Systems .................. Flame Temperature Versus Air/Fuel Ratio with Condensed-Phase Combustibles from Primary Not Burning (500K Air) 100 psi ..... Comparison of Fuel Flow (Mass) Requirements for Given Operating Design Point for Candidate Formulations vs S.O.A. Boron Formulation ............................................. Comparison of Fuel Flow Volumetric Requirements for Given Operating Design Point for Candidate Formulations vs S.O.A. Boron Formulation ............................................. Burning Rate Data for Magnesium-Loaded Fuel-Rich HTPB Formulations Made During Preliminary Feasibility Assessment ........ Burning Rate Data for MKI ..................................... Burning Rate Data for MKl2 ..................................... Burning Rate Data for MKI2 .................................... Burning Rate Data for TKI ..................................... Burning Rate Data for TK2 ..................................... Burning Rate Data for TK3 ......................................... Magnesium Vapor Atom Number Density Experiment Schematic ...... Optical Absorbtion Apparatus .................................. Theoretical Magnesium Vapor Absorption vs Number Density ...... Theoretical Variation of Fraction of Gas-Phase Products Which are Magnesium with System Temperature ................... 45 46

48 4) 51 52 56

17 18 19 20 21

57

22

58 62 73 74 75 81 82 83 85 86 90 97

23 24

25 26 27 28 29 30 31 32 33

i
* iv

K!

LIST OF FIGURES (Continued) Page 34 ,5 Closed Bomb Propellant Combustion Product Test Apparatus ...... Thermodynamically Predicted Gram-Moles of Nascent Magnesium as a Function of Temperature for MKI, MK2, and MK12 ........... Particle Size Distribution of Carbon Particles Collected From 500 psi Bomb Combustion of Propellant MKI ................ 7 3S AC . 41 42 43 44 45 46 47 48 49 50 Magnesium Nitride (35-40 micron) Burning Times ................ Results of Plume Calculations for Exhaust of 15/35/50 HTPB/HMX/Mg Formulation Mixing with 500'K at P 1 Atmosphere... Primary Motor Assembly ............................................. Throat Closures in Motor Firings of MKI, Residue Levels in Motor Firings of MKI, MK2, MK2, and IKl2 ........ and MKl .......... 98 103 104 107 ill 113 117 118 121 123 125 127 130 137 138 140 142

Velocity of Cases Away from Surface of Burning Propellant versus Pressure .................................................... Alternate Possible Explanation for Increase in Residue Level with Increasing Pressure .................................... Pressure-Time Traces for Two Motor Tests with TKI ............. Schematic of Primary Motor-Ramburner Test Hardware ............ Mixing/Flameholding Aid used in this Study.................... Efficiency Data - MKI - Test 401, 401A and 402...... ......... Efficiency Data - MK1 - Test 451................................. Efficiency Data - MK 2
-

Test 403 .............................

Efficiency Data - TKI - Test 460.................................

I I I I I

I
'I

.......... .. . ....

, A

I I
I
S'' 'ION I INTRODUCTION

Recent years have witnessed a large development of fuelrich propelhnt technology to produce formulations for primary rockets In air-breathing propulsion systems. Two types of systems have been especially

prominent, air-augmented rockets and external-burning missiles or projectiles. In an air-augmented rocket, the incompletely combusted products of a fuelrich primary motor formulation are mixed and burned with ram air inside

I
i

a secondary chamber at a pressure equal to or less than the air inlet recovery pressure. In external burning, the fuel-rich primary motor

exhaust is injected directly into the free-stream around the missile and burns at atmospheric pressure (sea level or altitude). Both types of

systems primarily utilize composite solid propellants with ammonium perchlorate oxidizer: thus non-metalized formulations typically produce

compounds consisting of C, H, 0, N, Cl, and possibly F, while metalized systems also produce B, Al, Mg, or combinations of these metals. An essential prerequisite of success in air-breathing applications is rapid and efficient combustion of the primary exhaust in air. tion. It is of interest to spell out here the conditions for this combusIn ducted propulsion, the pressures in the secondary range from

15 or 20 atm down to perhaps half an atmosphere depending upon the flying altitude and Mach number, and the residence times typically are a few milliseconds. In external burning, the pressures range from 1 to 0.1 Thus it is

atm, and the residence times are a fraction of a millisecond.

easily seen that these air-breathing systems often place stringent demands upon preignition and combustion times of the primary exhaust fuel products I with air. Boron is a particularly attractive ingredient for propellants for air-breathing rocket applications due to its high heating value. Generally, it is desirable that the amount of boron in the fuel-rich

3 I

*
I

propel lant be r.:,'ir'i:-,d, eff ic iencv constra l!:t

,ect ,.

to processability and primary motor ejection upper limit of boron loading loading levels, of suffici,,iit form wit h lized. In is

prac tlcal
IT't

;ipproximateiv -erv little

idi::er. aiir in

V" t e,'d

p, r
ir:i!;

by weiVilt. tot,-

At these

boron Thtis,

is m,t

i'

primarv motor,

due to in

1 ,ik

1),, horon must be burned *r ifor its

pirt ;,:ilate t it

the secondarv (o:n ,ist at

heating value

addition, to be low.

these hti :h l,, dioics, These factors,

primary motor flame

temperatures

tend

c-mbined with the low residenc, in

times availal'le

for mixing, had led

ignition,

and combust ion processes efficiency in problems,

the secondarv ,'hamber tinder orrat ,isia or 1,ess. tif years of this
ini

to afterburning

particularl

conditions which result

afterburner pressures of 50 two-and-one-l

The objective of the first w,rk was to characterize material

the nature of the primary exhaust

condensed-phase

from fuel-rich motors containing boron, and

to study its ignition infor-

and combustion characteristics, especially as compared to the known mation for pure boron.

Results of this phase of the program were presented entitled "Studies of the

in detail in an Interim Scientific Report

Ignition and Combustion of Boron Particles for Air-Augmented Rocket Applications" (AFOSR-TR-75-0043, October, 1974) and will be covered only relatively briefly in this report. During the final two-and-one-half years of the program, emphasis was placed on magnesium-containing fuel-rich propellant systems.

Theoretical screening calculations were performed to identify promising compositions, both pressed and castable, several of which were subsequently formulated. These formulations were then characterized in terms of combusin terns of the In

tion of the basic propellants themselves, as well as combustibility of

their fuel-rich products with air in a ramburner.

addition, basic studies of the composition of the exhaust products were carried out along with single particle combustion studies of magnesium and magnesium nitride (two of the more important exhaust products). The

magnesium propellant work will be described in the boron work in this report.

somewhat more detail than

*2

I
9
A.

SECTION II
BORON COMBUSTION STUDIES

COLLECTION AND ANALYSIS OF CONDENSED-PIIASE LOADED FUEL-RICH PROPELLANTS In this phase of the program,

EXHA1'UTS OF BORON-

condonsed-phiise exhaust fuel-rich propellants

products

from two different

highly-boron-loaded

were collected and analvzed in terms of both physical and chemical characteristics. In addition, combustion of the condensed-phase exhaust products Finally, modeling stodies des-

(CE) was studied in a flat-flame burner. cribing the

ignition of single boron particles in hot gas streams and

propagation of flames in bor,- oxvgen-nitrogen dust clouds were performed. Two propellants were studied. perchlorate-binder oft formulations Both were composite ammonium amounts (almost 50'.)

containing large

boron and a few percent of magnesium and aluminum.

ARCADENE (Atlantic

Rusearch Corporation trade name) 256A is a carboxy-terminated polyhutadiene

formulation and ARCADENE 280, a polyester formulation. The boron used in

both formulations was Kawecki Company material with nominal average particle

diameters of 3 20 1:m or more. C, H, 0, N, C1,

.m.

However, the diameters range widely from about I to

Thus both formulations are made up of the following elements: B, Mg, Al. In addition to these, the 256A formulation

contains a small amount of iron, because it is catalyzed by n-btityl ferrocene. Two motors were fired with each formulation into a large cylindrical tank (5.2 m long, 2.9 m in diameter) which was first evacuated The

and then filled with argon to 0.35 atm pressure prior to each run.

rocket chamber pressure and the tank pressure were measured continuously throughout each run. Data pertaining to the four motor tests are given in Table I.

Each propellant grain was weighed before the test; as was the residue
inside the motor after the test. Comparison of the third and

remaining

fourth columns of Table 1, in which the respective weights are listed,

13
,

Ir
rI
-0 CL

A c~U) a)

'.0;

-4

'

-4

41

&~

I4 -4 4 C 14

~
0 on0

og

~I~to

en

-K

-44

shows that the ejection efficiency ranged from 93.8% in Run 4 to 99.4% in Run 2. During all the tests, the tank pressure rose continuouslv The low final value recorded in

from the initial to the final value.

Run 2 can be ascribed to the fact that the burning time was about twice that of other runs, thus allowing more cooling and condensation during the run. pressure. After each test, the tank filled with argon was kept closed As shown, each propellant was tested at a high and a low

tor about one hour so the CE would settle in a large (5.4 m 2 surface area) metal trough at the bottom of the tank. The trough was then emptied Thus, the CE

into a bottle by remote action without opening the tank. samples were not exposed to air while hot.

Thereafter, ordinary precautions

were observed to avoid undue exposure to air and humidity (e.g., the samples were stored in corked vials in desiccators), but no attempt was made at absolute exclusion of atmospheric oxidants. A set of two thermodynamic equilibrium computations, "A"

and "B", was made, corresponding to each experimental motor run. results are given in Table 2.

The

Computations "A" took into account most

species which can reasonably be expected to appear in the products. Computations "B" omitted BC and BN, two species which are very important

thermodynamically, but which are generally suspected not to appear in reality for kinetic reasons. The theoretical characterization of the CE relevant to this report can by summarized as follows: found in the CE. (a) All of the boron will be

The fraction of the CE which is boron (any form) is

about 50% for either propellant regardless of the type of computation ("A" or "B"). (b) Most of the boron (roughly 90%) is

found in three

solid species:

B, B4 C, and BN.

The question of the realistic distribution

of boron among these three, i.e., the relative merits of computations "A" and "B", must be resolved by experiment. (When full equilibrium

is assumed ["A"], most of the boron in the exhaust is predicted to be solid B4 C and a significant fraction solid BN.) (c) The remainder of

15

CI

I
0-4

00

00 a, )

0-

00
4 1. -4

IT
-4

C)
-14

I
-0 -I
Is

"X

a
C

c
'

CONJcc

0-4

nI

the boron is in the CE in it, form of boron-oxygen-hydrogen species. some of these are rather volaile, all will be condensed at the room tempt,rature. (d) The r':nmiindr of the CE consists largely of MgAl ind MgAl
2 04, 204

While

;ind MgClI2 in computation.; "A",

MgCl 2 , and C(s) in computations

The

:o] I, ti I 0:udCT',;ed exhaust was examined under an optical up to 5OOX. The untreated CE consists of The tumbling of the

microscope with magnification:

very large, but only loosely agglomerated clusters.

clusters with soft (plastic) balls breaks up the mass of the CE into smaller particles, ranginv from about 3 to 50 microns in diameter. Several distinctive features were observed under the microscope. The

bulk of the CE is dark and apparently amorphous, but quite a few crytals are also visible. The crystils are either white or colorless. The dark or possibly B2 0 3

material could be carbon and Larbonous material, B4C.

elemental boron,

The light material culd be any and all of the following:

(colorless), B2 0 3 -H2 0 (triciinic crystals, white), MgCI 2 (colorless hexagonal Mg Cl .H 2 0 (white), MgAl,04 (colorless), and BN (white hexagonal). Densities of the CE samples were determined by suspension in inert liquids having sever.l different specific gravities. For CE

collected from Sample 1, all particles were found to have specific gravities of less than 2.2, with some being even less than 1.8, but with most having a specific gravity of about 2.0. After washing in a hot mixture of methyl

and ethyl alcohol, a standard procedure for reroval of B 203 from boron

powder, the residue was found to have a specific gravity greater than 2.3. For the purpose of comparison, two suspensions made with Lhe "off-theshelf" boron powder (Kawecki Co. material used in all ARC propellants). were checked for specific gravity; virtually all of the powder has the

specific gravity of about 2.3, i.e., distinctly higher than the unwashed CE, but perhaps slightly lower than the washed CE. The handbook value Thus the 2.3 value

for the specific gravity of pure boron is 2.32 - 2.35.

for the commercial powder is very reasonable, especially since the material is known to contain about 2% B2 03 , which has a lower specific gravity.

*7

.. 7=;

--

.'v

"

Our conclusion is that in the CE, low-density substances either coat or are firmly attached to solid B (and B 4 C, if any). The

iwshing removes these substances and leaves a residue consisting largely i,(sp. gr. 2.32 - 2.35),and possibly graphite (sp. gr. 2.26), B 4 C (sp. g:. .54) or BN (sp. gr. 2.25). (-p. gr. 1.85), B2( 3 '311 20 The possible low-density species are: i. gr. 1.49), and MgCI2 .61110 (sp. B2 03

gr. 1.36):

gSince

the MgCI. 2 without any crystalline water has the specific gravity of 2.33. carbon is insoluble in alcohol, the fact that the washed CE contains

little low-density material suggests that the untreated CE contained little ir no amorphous carbon (sp. gr. 1.8 to 2.1). This has to be interpreted the full computation on

in the light of the thermodynamic computations:

Run I including B4 C and BN in the products, predicts no solid carbon; if B4C and BN are "suppressed," the thermodynamic computations predicts This comparison suggests that computa-

the CE to contain about 7% carbon.

tions "A" nay be somewhat more realistic.

The CE samples from all four motor runs were analyzed chemically for boron. The analysis was performed by Ledoux and Company, The CE

Teaneck, New Jersey, on a subcontract from Atlantic Research.

samples were first separated into the soluble and the insoluble portions hy boiling water, and then each of the two portions was analyzed for total boron. The results of the chemical analysis are summarized in Table 3, which lists (a) fraction of the CE which is boron (any form), (b) division of the CE into soluble and insoluble portions, and (c) In each case, we

fraction of the total boron in each of these portions.

have listed the thermodynamically predicted values along with the experimentally determined ones. The species assumed insoluble for the purpose

of comparison of the computation with the experiment are listed in Table 3. Detailed discussion of Table 3 Report
11)

is presented in the Interim Scientific It is concluded that

referred to in Section I of this report.

according to both the experimental data and either one of the two computations, only about 50% of the total CE is boron. Beyond that, one can see

!8

mdmI

2i

r.

-.7

OC

0.

LC

~
0-

rn

-Tr

C-

00

r-,

'-4

10
..

oi

0 z
4

..

.......

.-

.'

that three out of four questions which can be answered by the results of the chemical analysis, namely the amount of baron in the CE, the fraction of the CE which is insoluble, and the amount of boron in the insoluble portion, are predicted reasonably well (order of 10%) by the full thermo-

fless

dvnamic computation (column "A" in Table 3).

Computations "B" are sompwhat

successful, but it appears that they also give a rough idea of the Indeed. it is possible that the assumptions

composition of the CE.

inherent in the theoretical comparison, e.g., the freezing of the exhaust equilibrium, introduce more of an uncertainty than the difference between the two sets of computations. Only an extensive chemical analysis of

several species (Mg, B4C, C, and others) could resolve the problem of the detailed composition of the CE. B. COMBUSTION OF THE CONDENSED EXHAUST As described earlier, the ball-milled CE consists of many small particles with sizable agglomerates interspersed. Prolonged ball-

milling and sieving through a 20 i'm mesh screen produced a fine powder consisting of particles having a fairly uniform size distribution with average diameters of 3 to 4 pm. By careful screening between 30 and 35 lim,

it was also found possible to separate coarse powder samples consisting

gCombustion
tn!
2-4 )

of firm agglomerates which could be treated as single particles. of the coarse CE samples was studied by a single-

particle technique, utilizing a gas burner, which was previously developed and extensively used at Atlantic Research for combustion studies of alumberyllium, ( '- ' ) and boron. ( 7 ) Since particles smaller than

10 or 20 Vm in diameter are difficult to handle and observe singly, the gashurner technique was adapted for a study of particle clouds. The

newly developed technique was used to study ignition and combustion of the fine CE powder. Details of the experimental procedures were presented and will not be repeated

in the aforementioned Interim Scientific Report, (1) here.

The single-particle technique was used for a study of both the ignition and the subsequent combustion of a 30
-

40 jim sample collected

10

___

in Run 4.

The gas burner was run on propane-oxygen-nitrogen mixtures.

The burner gases consisted of 1.4 - 20% 02P most observations having been made with about 20% 02, 15 - 20% H 2 0 , and 10 - 12% CO2 ; the remainder was mostly N2 . Gas temperatures were varied from 1850 to 2300"K. Velocities

of burner gases were about 1000 cm/sec, and those of the carrier (helium) about 100 cm/sec. Quantitative ignition and combustion parameters of neat boron powder in the 30 to 40 Vm range are well-known from our previously published work. ( 7 ) Since it Is of obvious interest to compare the combustion

characteristics of the CE to those of boron, the relevant results of that work shall now be briefly stated. The entire history of the boron particles

(d = 35 pm) in the burner gases (2240*K, 19% O29 16% H 2 0) can be divided into three phases: a dark pre-ignition period (ca. 8 msec), followed by

a peculiar two stage combustion, the durations of the two periods being
-.

4 msec and t2 = 11 msec.

The first combustion stage begins when The first stage is generally

the particle temperature is about 1850K.

believed to correspond to slow combustion, impeded by a layer of liquid

gcontrolled
t2 i tb .

B2 03 .

Most of the mass of boron is consumed during the rapid second stage, by the gas-phase diffusion of oxidants toward the particle;

thus t2 approximates the total diffusion-combustion time of the particle, The minimum temperature of the burner gases necessary for ignition

of particles was found to be 1880 + 20*K. The CE agglomerates, 30 "m < d < 40 pm, were found to burn as single particles: there is no separation of particle trajectories.

Thus, we can make a quantitative comparison of the CE combustion to combustion of boron. When the CE, 30 pm < d < 40 pm, is burned in the same
=

burner gases as the 35 pm boron (Tg

2240-K, 19%02, 16%H 2 0;

see previous

paragraph), the following differences are observed: 1. The CE "particles" give well-defined burning times, but there is no sign of two-stage combustion. 2. The average burning time of the CE is tb
=

10 + 1.5 msec.

I
* 11

IV

I I
3
Iparticles.
In addition to the test at constant gas temperature and constant mole fraction of 029 two series of experiments were run in which 3. The CE ignites substantially lower in the burner gases than the boron: it is estimated that the preignition delay of the CE is 30 to 50% lower than for boron

these two parameters were varied.

Variation of temperature at constant

mole fraction of 02 (ca. 0.2) showed that the minimum gas temperature necessary for ignition of the CE is 1850 + 250 K. Variation of 02 content

at the approximately constant gas temperature of 2200 to 2300*K shows that the CE continues to burn, although very slowly, even at the lowest mole fraction of 02 which we tried, 0.014. Since the burning time of a particle is a strong function of its diameter, and since we have no accurate value for the average diameter I of the CE, the two values, tb = 10 msec for the CE and tb . t 2 = 11 msec for boron, must be considered the same within the experimental error. The cloud technique for combustion of small particles, described in detail in the Interim Scientific Report, ( I ) yields average preignition delays (ti) and burning times (tb). Both t i and tb are approx-

imate because in a burning cloud one cannot define clear points of onset and cessation of combustion. Furthermore, since the preignition delay

necessarily overlaps the carrier-gas mixing time, one cannot obtain cloud ignition temperatures from measured preignition delays as in single-particle work; one can only determine relative ignitabilities of two clouds hy

comparing the preignition distances in two replicate experiments. Two CE samples, from Runs 1 and 4, have been studied by this technique. The results obtained with these two samples are indistinguishable

from each other. The CE samples used in the cloud experiments consisted of the material which had passed through a 20 wm mesh screen. the gas burner operates in such a way that the material However, since

is subjected to

I
*

prolonged swirling with plastic balls, there is no assurance that the

12

i
I
particle size distribution injected into the burner gases is the same as in the original (sieved) sample. Therefore, the unburned material at a point Microscopic

was sampled after injection into the burner gases, i.e.,

between the injection orifice but below the point of ignition.

examination of the sampled material revealed that the particle diameters

were in the range between 2 and 7 pm.

The average diameter was about 3.5 Um.

In addition to the CE, two other samples were studied by the cloud-combustion technique for the purpose of comparison: an amorphous

boron sample, supplied by U. S. Borax, Inc., stated to be 98% pure; and the material supplied by Kawecki, Inc. The Borax powder particle

diameters are in the narrow range of 1.0 + 0.2 im. The Kawecki powder has a wide distribution of particle diameters, ca. I to 20 pm, with the

number average of 3 um. and 280 propellants.

The Kawecki boron is used both in ARCADENE 256A

The minimum gas temperature necessary for ignition of the particle cloud was found to be the same, within experimental error for the Borax sample and for the CE: 1980 + 20K. (This is approximately 100 to 150C

higher than the analogous limiting temperature for 30 to 40 micron particles of either the CE or crystalline boron, as mentioned earlier.)

3 3 3 3

Measurements of the minimum gas temperatures necessary for ignition of powders, which we shall refer to as "temperature limits" from here on, have thus resulted in two conclusions: (a) temperature

Slimits

for clouds of small (< 10 lim) particles are 100 to 150*K higher

than for large 30 to 40 jm particles; (b) regardless of the particle size, temperature limits are the same for the CE and for boron. The cloud combustion technique has also been used for burning time measurements of the CE, of the Borax boron, and of the Kawecki boron. The burning time tb was taken to be equal to the length of the burning cloud divided by the linear velocity of the burner gases. Since the luminosity

of the cloud both develops and fades out gradually, an approximate visual judgement concerning the length of combustion must be made. The

velocity of the burner gases can be calculated from known (metered) input gas flow and known flame temperature; it can also be obtained by measurement

I
FJ

13

I I
of single-particle velocities. usually agree within about 257'. The calculated and ! he measured values It is estimated that !he ombined uncer-

f
I

tainties of the cloud length and particle veloitv may cnceivably lead to an error in determination of tb of up to a factor of two, but most th values are probably more closelv defined than th.at. The average rvsu lts of three cloud-comb-,t ion measurements are shown in Figure 1. were: The burner gas properties in all three experiments A single point for the The

T = 22400K, X(O 2 ) = 0.19, X(H 2 0) = O.1f.

tb of a spherical boron particle, d = 37 ,m,is given in the same plot.

5
I

:Aingle-particle point is a small extrapolation of an accurate measurement (see Table VI of Ref. 8) in a somewhat different burner gas: T = 2280*K, X(02) = 0.23, X(H 2 0) = 0. Even with the extrapolation, the values of both

tb and of d are no doubt much more accurate for the single-particle point than for the cloud measurements. In view of the uncertainties in the cloud

measurements, one should be careful about assigning a specific numerical value to the slope, n, in Figure 1. However, an inspection of the figure

shows that even if the cloud data are in error by a factor of 2, n will still be closer to unity than to a value of 2, predicted by single-particle 9 10 ) ' diffusion theories.(

In addition to experiments run with burner gases having a constant, relatively high mole fraction of oxygen, X(0 2 ) = 0.2, cloud experiments were made with decreasing values of X(O 2 ), both with the CE and with the Borax boron powder. kept at about 2300*K. Gas temperatures in all the tests were It was found 0.14. 0.14 and

Only visual observations were made.

that both powders burned about equally vigorously down to X(O 2 ) The combustion brightness decreased rather sharply between X(O2 )

0.12, and then very gradually down to about X(O 2 ) = 0.05 with both powders. At very low mole fractions of oxygen, 0.015 to 0.05, the CE powder burned very

3
3

weakly, and the boron powder hardly at all. The single-paiticle experiments indicate that the burning times tb of the CE are about the same as for boron. This appears a priori

reasonable, particularly if the combustion is diffusion-limited, but even if it is kinetics-limited, provided the limiting reaction is the same,

1
I

14

BURNING TIMES (t01 AT TGAS =2240K. X

=0.2

10

Fo

.00

1.0

____ ____Q>

Ii

U.S. BORAX
KAWECKI

CE-i , CE-4
<>SPHERICAL SINGLE PARTICLE 1010

Figure 1.

Burning Times (tb

of Various Particles

15

I.W

as seems likely.

Higher volatility of some expected CE ingredients should Thus, if the CE is rich in

accelerate the burning rate of particles. spec ies such as Mg, MgCI 2 ., !aster than boron. and Fe,

one should expect it if it

to burn somewhat

On the other hand,

contains large amounts of B4 C is known to burn

B C (and possibly BN), more slowly than boron, difficulty.

it may burn more slowly:

and BN has been observed to ignite only with the change in th of boron caused by adulter-

In either case,

ation is probably not sufficient to be clearly demonstrated by our experiments. C. ANALYTICAL MODELING OF BORON SINGLE-PARTICLE COMBUSTION AND RORON CLOUD FLAME SPEEDS The model(I I ) (originally developed under Internal Research and Development funding) of the ignition of boron particles treating the removal of a boron oxide layer from the surface of the boron particle thro-igh numerical integration of difference equations describing the

generation and removal rate of boric oxide has been revised to more accuratelv treat the various processes involved. This new model yields predicted

i'nition times and minimum ambient temperature requirements for particles studied bv Ma-ek (7) in dry gas streams which are in good agreement with his measured data. An atttapt to treat the effect of water gas on boron

part icle ignition as being a diffusion-limited reaction of the water with the horic oxide coating to form gaseous HBO 2 was unsuccessful. Effects

of v riius parameters on boron particle Ignition were studied with thin; mode I. times, cxide thickness was found to moderately affect predicted this effect decreasing with increasing particle size. ignition

The effects

of total pressure and oxygen mole traction on minimum ambient temperature required for ignition and on itnition time were found to be fairly complex. At low oxygen mole fractions, lacreasing total pressure leads to more. difficult Ignition, true. while at hiv'i oxvgen mole fractions, the reverse.'

At low total pressurt , ig'ni t ion was found to be independent of txgcli at high pressure, increasing oxygen mole fractions led

mole fraction:

to easier, faster ignition.

Initial For tile case of ttre,

particle radius effects are also somewhat temperature

complex.

surrounding radiation

equal to the gas temperaIncreases with

minim'iim ambient gas temperature required for Igni:tion particle size f*,r particle radii below 10 to

decreasing

15 microns. excCpt Igni t ion

Ignition time generally decreases winh decreasing particle size, wiyere ,ffect.d by the ignition limit dependency on pirticle

size.

time Is

predicted to increase strongly with decreasing gas temperature. predicted

The same type of dependency on surroundings radiation temperature Is in the region of ambient gas temperatures

which are marginal for ignition.

The same equations used in the numerical analysis have also been used in a stability analysts to define conditions (e.g., ambient temperature) required for ignition. Excellent agreement was found

between results obtained with the numerical analysis and the stability .1nalysis.

(12)) from this eff6i: A copy of a technical paper (1)generated

Interim Scientific Report. A model for prediction of boron-oxygen-nitrogen dust clouds

Is included as part of the earlier referenced

treating the boron particle ignition and combustion processes in detail, as described immediately above, was subsequently developed. The prpblem
I

was formulated as a doubl.e-eigenvalue

problem with numerical integtati.n

of mass, species and energy difference equations in combination with difference equations describing the ignition and combustion processes through ignition and combustion zones. This model was found to yield

predicted flame speeds in reasonable agreement with the limited boronoxygen-nitrogen dust cloud flame speed data available.
In addition, a vastl.y siiplified closed-for-m flame speed expression was developed. found to exceed Predicted flame speeds from this model wer,

those predicted

from the detailed model by 0 to 130% It was shown that two of the major tend

over the range of parameters examined. error-prodtiring approximations

associnted with the simplified model while a third

to drive predicted

flame speeds down,

tends to drive them up, Over the range

leading to a large degree of

cancellation of these errors.

of condition.s examined,

It

appears that the simple closed-form model can

17

a
t
be used with a correction factor dependent upon Nusselt Number and pressure

tto

obtain reasonable estimates of boron-oxygen-nitrogen dust cloud flame speeds. Effects of various parameters on predicted flame speed were ;L-idied. Predicted flame speeds were found to increase with increasing

initial temperature, decreasing initial particle radius, increasing initial oxvgen mole fraction, decreasing pressure, increasing particle loading (on fucl-lean side of stoichiometric), increasing post-flame emissivity, decreasing initial oxide coating thickness, and decreasli g Nusselt Number. A copy of a paper
(1 3 )

generated from this modeling effort is also included

as part of the Interim Scientific Report. ( I )

i" I ! !

I I I I
* 18

ilt

i
SECTION lI

MAGNESIUM PROPELLANT STUDIES

A.

BACKGIROUND AND PURPOS, The potential performance advantages oI air-.iugmented

rocket systems over conventional rocket systems, reslt ing from using "free" air as an oxidizer replacing a large fraction ot the relatively low

energy oxidizer which must be carried along in a conventional propellant (displacing fuel) have long been recognized. In the past, fuel-rich pro-

pellants centered around carbon, aluminum, and boron have been considered as fuels for such systems, while magnesium has been neglected. been two ma or reasons for this neglect. There have

First, magnesium has a lower

theoretical heating value (an important parameter for air-augmented rocket applications) than the other three materials. formulations present practical g-containing l curing problems: Second, castable magnesium the art of preparing

propellants with satisfactory physical properties was,

therefore, not well-devei'ped at the inception of this program. Closer examination of the situation, however, Indicates that there are other factors dictating further evaluation of magnesium as an invredient in fuel-rich air-o .gmented rockets. eitail the
Oiestion

First, consider in more

of theoretical performance. Listed in Table 4, are

theoretical performance parameters for typical highly-boron-loaded fuelrich formulations, aluminized formulations, carbon-loaded formulations, formulations containing no metal fuel additive, and magnesium-loaded formulations. (20/1 For operation under conditions where the air-to-fuel ratio is high

or greater), the mass and volumetric heating values are important As may be seen from Table

parameters in ranking theoretical performance.

4, the state-of-the-art boron systems have significantly higher theoretical heating values than any of the other systems. Although the 50% carbon

formulation gives higher theoretical heating value than any of the other nonboron formulations, attempts to develop good afterburning propellants with that carbon loading have been unsuccessful, 30% carbon-loaded systems

*,19

.. . ... ... . . .. S . . . .. . ... 1 L.

Cr) LO u

(a L)

CV t U

L4UA

a-)

UA

-U

C) LL

IL

CL

4.
2 4 "-I0 2
i

C
2

I-ui

4j

< 1

.jL

LU

LU

02

representing the highest successful loadings to date.

These may be seen

to compare closely in heating value with the potential magnesium systems listed at the bottom of the table. Aluminized fuel-rich propellants and

mn-metalized formulations within practical composition limits are seen to be inferior in performance. among practical formulations, Thus, in terms of theoretical heating value only the boron systems clearly outrank the

listed magnesium systems. In addition, magnesium offers some important advantages over even boron for air-augmented rockets under certain conditions. For very low

air/fuel ratio applications, it may be shown that magnesium-loaded systems will provide higher theoretical performance than systems containing the higherheating-value additives due to a lower oxygen requirement (per unit heat release) of the magnesium metal. This is demonstrated in Figure 2, via (developed

a comparison of a magnesium-loaded formulation designated MKI

in the course of this program) with a state-of-the-art boron-containing formulation. In development of this plot, various air-augmented rocket

design conditions (configuration, Mach Number, altitude, drag coefficient) are chosen, which, with e use of the boron formulation, would require

operation at the air/fuel ratios plotted on the abscissa of this figure. The required fuel flow rate for each of these conditions is then calculated.

fformulation

The MKI formulation is then substituted for the state-of-the-art boron and the required fuel flow with the MKI fuel for operation The calculational procedure

at the same design condition calculated.

employed is a very simplified one, with a number of approximations and assumptions, basically involving setting the air flow equal for the two cases, and calculating the air/fuel ratios for each fuel which will yield equal values of vacuum sonic gross thrust/air flow rate (SA). The ratio

of the required MKI fuel flow rate to the required state-of-the-art boron formulation fuel flow rate is then plotted against the operating air/fuel ratio for the boron formulations. (This somewhat involved method of presentation is used to avoid security classification problems.) calculated both on a mass basis and a volume basis. I This ratio is

As may be seen, as

the operating air/fuel ratio decreases, the ratio of required flow rate of

21

>z
w
w0

0 w w.
__

Iz
---

0IA 0
0 w

E0
-~

04
w

LL

00 J0 w

cE<

-2
ot
CO LL tA

<

0~-9

O0~

<

ocr

-4

__

__x

00

00$-

'UL LL W

to -

0 -r

E 0

<

<~

MK1

to state-of-the-art

boron formulation also decreases.

A value of 1.0

,,r this ratio signifies a hreak-even point between use of these fuels
ktn

theoretical

thermodynamic performance considerations alone),

with

the MK1 magnesium system being superior at lower air/fuel ratios (which tvnd to be associated with high thrust coefficient requirement,). As

showni,

the MKi

formulation is

favored for volume-limited systems in appliboron formulation would be required to is favored for weight-limited system

cations where the state-of-the-art );#perate at air/fuel <3.7, ir boron air/fuel <4.4. while it

I
3
ered.

In choosing fuel-rich propellants for air-augmented rocket ;vstems, parameters other than theoretical performance must also be considFor instance, with the aluminum formulations, there are severe (thus intractable) nozzle deposition problems caused by formation of A]2 03 With boron formulations,

1ar,

slag in the primary motor (fuel generator).

materials cost and requirements for elaborate pretreatment of the boron raise considerable difficulty regarding large scale production. Most

iportant,

however, is the question of efficiencv of combustion of the fuelof little use if the heat cannot be

rich products exhausting from the primary motor into the afterburner chamber. Hi.gh theoretical heating, value is rleased - n' inc.

through combustion of this fuel-rich stream with air in the limited of the afterburner. Boron and carbon have been found to be quite

difficult to afterburn,

leading to severe degradation of delivered performance high air/fuel ratio, low air

'inder conditions of low afterburner pressure,

Stmperature, or low afterburner residence time (small combustor or large

3 3

thr-at-to-chamber diameter ratio), cannot be used.

particularly

if

a flame-holding aid

This problem should not be encountered with magnesium.

In this regard, two positive trends become apparent upon inspection of thermodynamic data presented later in this report. niti'neslum loadings, 50 to 607, First, even high well

give high primary chamber temperatures,

in excess of 2000K, suggesting that these formulations would burn well .1nd
give good

primary ejection.

Second,

very substantial

fractions of

total magnesium are predicted to appear as the metal vapor in the primary

1
Ui

products, so secondary combustion should present no problems; indeed, it should be exceptionally vigorous at all pressures. Thus, it appears that

magnesium-loaded fuel-rich propellants will be advantageous In terms of combustion efficiency under conditions of low afterburner pressure, high air/fuel ratio, low air temperature, low afterburner residence time (small combustor or large throat-to-chamber diameter ratio), particularly if a flame-holding aid cannot be used. In addition, the high degree of reactivity

of primary motor exhausts from fuel-rich magnesium propellants should make them attractive for external burning rocket applications, where the burning pressure is, of course, low; available residence time for useful heat release is short; and use of a flame-holding aid is impossible. Accordingly, the objective of this phase of the program was the identification, formulation, and testing of promising highly-magnesiumloaded fuel-rich propellants suitable for use as air-augmented rocket propellants. Major tasks included: (a) thermodynamic computations to (b) experimental formulation and

identify promising candidate formulations;

ballistic evaluation of selected compositions; (c) characterization of the combustion products of these formulations; (d) single-particle combustion studies of condensed-phase fuel products produced by burning of these formulations in the absence of air; (e) theoretical study of the effects of magnesium vapor on plume ignition; (f) testing of these formulations in motor configurations to characterize their motor ballistics and ejection properties; and (g) connected-pipe afterburning tests quantifying the afterburning properties of the products of combustion of the selected formulations. B. THERMODYNAMIC SCREENING General Criteria There are several characteristics of highly-magnesium-loaded propellants important to their selection as promising candidates which can be quantified through thermodynamic calculations. First, it is generally

1.

desirable to maximize the propellant heating value (per unit mass for use

g
IN7

in weight-limited systems or per unit volume in volume-limited systems).

Next, in order to insure high efficiency of ejection of the fuel stream from

24

! d

the primary motor to the afterburner and to improve ignitability of this fuel stream as it mixes with air, it is desirable that the primary motor flame temperature be high. (With boron systems, it has been found that

primary motor flame temperatures of at least 2000 to 2200 * K are necessary for even marginal performance.) reaction of combustibles Finally, it Is obviouslv desirable that the

leaving the primary motor which are in the gas phase

rather than a condensed phase be maximized in order to minimize particle ignition difficulties and burning time. A useful figure of merit for

comparison of potential formulations which contains the first and last parameters is the "reduced volumetric heating value" defined as the product of the total volumetric heating value and the fraction of the heating value not contained in condensed-phase primary motor products. Screening calcula-

tions to select promising candidate castable and pressed compositions have been carried out during the course of this program. 2. Castable Compositions The following classes of castable propellant systems have been considered in this thermodynamic screening effort: (1) Nitrocellulose (NC)/TEGDN/Polyethylene Glycol (PEG400)/ Magnes ium (2) Nitrocellulose/Nitroglyerir,(NC),'Adiponitrile Magnes ium (3) Nitrocellulose/Dioctyl Adipate (DOA)/Ammonium Perchlorate (AP)/ Magnes ium (4) Polyester Resin (HX 730)/TEGDN/HMX-AP/Magnesium HTPB/HMX-AP/Magnesium HTPB/HMX/Magnesium/Aluminum HTPB/HIX-AP/Magnesium/Boron (ADN)/

(5) (6) (7)

(HTPB refers to a standard hydroxy terminated polybutadiene binder system.) The first two systems are double-base systems, the next is a composite-modified double-base the fourth is a nitro-plasticized composite propellant system,

g
It

-25-

and the last three are composite propellant systems utilizing a hydroxy terminated polybutadiene binder. In the first five systems, the only metal

fuel is magnesium, while mixed metal systems are considered as the sixth and seventh systems. Results of thermodynamic computations for these systems The parameters of interest

are presented in various forms in Figures 3 - 13. plotted in these figures are:

(1) Volumetric Heating Value of the Formulation (assuming complete combustion with excess air to H 2 0 (gas), CO 2 (gas) and the condensed-phase metal oxides) (2) (3) Primary Motor Flame Temperature Percent heating value contained in Condensed-Phase Combustibles Leaving the Primary Motor (4) (Volumetric Heating Value) x (1 - Fraction Tied Up in Condensed-Phase Combustibles Leaving the Primary Motor) [Referred to subsequently as Reduced Volumetric Heating Value] With the first type of system (NC/TEGDN/PEG/Mg), it appears that volumetric heating values up to 520 BTU/in 3 can be obtained at primary motor flame temperatures in the range of 2300 to 2400K. Heating values are maxi-

mized by elimination of TECDN from the system, with only slight accompanying degradation of primary motor flame temperature. Results of thermodynamic

calculations for the NC/PEG/Mg system are shown in Figures 3 - 5, for PEG/NC ratios of 1.0 to 2.0 (covering the practical formulation range) and magnesium

Sloadings

of 30 to 75 percent by weight.

In Figures 3 and 4, volumetric heat-

ing value is plotted against primary motor flame temperature and against the percent heating value contained in condensed-phase combustibles leaving the primary motor. (For these cases, the condensed-phase combustibles are all

carSon, with all unburned magnesium existing as an extremely combustible vapor.) As may be seen, increasing magnesium loading at a fixed PEG/NC ratio

3 I

increases heating value and primary motor flame temperature up to about 50-557 magnesium level - at higher magnesium levels, flame temperature drops rapidly. At a fixed magnesium level, increasing PEG/NC ratio increases heating value,

26

In
C4

IC
CIz z 9

I 1I1I

Ln 0M

0
WLLJW
E) 0

C-

I-4

mw ( wL

I 3~

w
w
2

z z

<. 0< o
LA.
w~I

z3

z
0
0

>-

z
0

(Euo/n19) 3flVA !DNIIV3H

27

w-

ILIL)

CL
0_ 0___ ZA-~~4 0

>*

K
0f -1

1:ID
c

~iO'co

4>
CM
00D

w LA
NL

Z
L.

40I

n0

to

&n

In

In

(ul/n~g) 3niVA !)NIILV3H

28

060

V'

Cuu

w
0.

zUL
AN

'<S

IE
___ __

in
_nlV

-0
_

0
_L

_inI3

___ ___

__

___

___

___

__

__29

but decreases primary motor

flame temperature.

Figure .

shows that the per-

centage of heating value tied up in condensed-phase combustibles (relatively difficult to afterburn) is independent of magnesium loading for a 1/1 PEG/NC

ratio, and actually decreases with increasing magnesium loading for PEG/NC ratios of 1.5 and 2.0, up to approximately 65-70 percent magnesium loading. For a fixed volumetric heating value, the percentage tied up in condensed phase combustibles increases with increasing PEG/NC ratio. It may be con-

cluded from these calculations that theoretical heating value may be maximized without increase of heating value tied up in condensed-phase primary motor exhaust products and with maintenance of adequate primary motor temperatures by increasing magnesium loading up to approximately 65 weight percent, with the optimum relative amounts of nitrocellulose and polyethylene glycol being determined by tradeoffs between heating value and percent heating value densed-phase combustibles emanating from the fuel generator. In Figure in con*3,

total volumetric heating value and volumetric heating value reduced by the fraction of that value in condensed-phase combustibles leaving the primary motor are plotted against magnesium loading for the three PEG/NC ratios examined.

Total heating value increases monotonically with magnesium loading up to 75 percent magnesium (the highest level examined) while the reduced heating value maximizes at approximately 65-70 percent magnesium loading. indicate that a heating value of approximately 520 in an NC/PEG/Magnesium system with approximately 60 which should afterburn well
-

Figures
3

3 to 5

540 BTU/in

can be obtained

65 percent magnesium,

(primary motor flame temperature in excess of

2200'K, approximately 20 percent of potential heating value in condensed phase combustibles entering the afterburner). Theoretical volumetric heating values and reduced heating values are plotted against magnesium loading for the Nitrocellulose/Nitroglycerin/ Adiponitrile/Magnesium system (30/63/7 ratio for the in Figure b. first three ingredients)

This is a very hot system, with a theoretical primary motor As a

flame temperature of 2460*K even at 70 percent magnesium loading.

result there are no condensed-phase combustibles predicted in the primary exhaust stream at magnesium loadings of less than 50 percent, and only 3.9

'1

30

cIn

41~ w

I0
I-<.

D~ I

< S1

I0

and 8.9 percent of the total heating value appears in such prducts at 60 and 70 perce:it magnesium loading. Thus, although the total heating value .3 potential of this system is less than the preceding -;v,tem (.70 -495 BTU/in at 60 - 65 percent magnesium loading) the reduced heating value is somewhat higher (460 versus 420 BTU/in 3). With the third type of system considered (NC/DOA/AP/Mg), it 3 was found that theoretical volumetric heating values of 500 to 520 BTU/in can be achieved at flame temperatures of 2200K or less for a wide range of compositions, with these values dropping off to 440 to 480 BTU/in flame temperature requirement is raised to 2400'K.
3

as the

The fraction of combustibles

leaving the primary motor as condensed phases (mostly carbon) is strongly influenced by the ammonium perchlorate level, level. increasing with decreasing AP

Results of thermodynamic calculations for a series of practical formu-

lations in which the nitrocellulose and dioctyl adipate content were held at 8 and 12 percent respectively, and the relative amounts of ammonium perchlorate and magnesium were varied are shown in Figures 7 and 8. As the percentage

of magnesium is increased (and AP correspondingly decreased), heating value increases, primary motor flame temperature decreases, and percent heating value tied up in condensed-phase combustibles leaving the fuel generator increases. It appears that a heating value of 500-520 BTU/in 3 is attainable at a flame temperature of 2200-2300K, with about 15 percent of that value in condensed phase combustibles, not markedly different from the NC/PEG/Mg system potential.

3Magnesium 3 I

Results of thermodynamic calculations for HX730/TEGDN/IHMX-AP/ systems are presented in Figure 9. In all of these calculations, Three binder levels were

the ratio of HX730 to TEGDN was held at one-to-one. considered.

In addition, cases where all of the oxidizer was HMX and cases

where part of the HMX was replaced by ammonium perchlorate were considered. As may be seen, the total heating value increases monotonically with increasing magnesium level, is essentially independent of binder level, and decreases with substitution of AP for HMX. The reduced volumetric heating value (total

32

Ww

.-

1~~
zo Dw
w

00>:

u
z
wA

0
00

:
a.

L)

NW In

*z

-00

CM

0r

00

w
_ _ _

33
_ _

u-

oo
_____ ______ __-0~.

%n/i

wCL___

>

ww
<

10

N~
ui/nlg)~~

-zlASNI3

-34

u.9

0
0~

D0
C,

:re
0

E)

<

CL

In ww

_____0

____

I(

61z
ui/n~) !)N.Lu3 3niA

.35

heating value minus the heating value tied up in condensed-phase combustibles leaving the primary motor), on the other hand, decreases with increasing

binder content (increased carbon), and Is essentially independent of replacement of 10V with AP. Primary motor flame temperatures for all cases As may be seen, total heating values of
3

examined were in excess of 2400"K. '00J to 520 BTU/in

3

and reduced heating values of 430 to 440 BTU/in

may be

obtained with this type of system at 60 to 65% magnesium loading. A large number of thermodynamic calculations have been performed for the HTPB/HMX/AP/Mg system; only a brief summary of these will Among the variables examined were percent binder, the

he presented here.

effect of interchange of the HMX and AP oxidizers, and magnesium loading. Typical results are plotted in Figures 10 and 11. Two binder levels, 10%

and 15%, representing the probable range of minimum level giving acceptable propellant processing characteristics were examined. As may be seen from

Figure 10, the volumetric heating value versus primary motcr flame temperature characteristics are nearly identical for the two binder levels; however, the higher binder level does result in approximately 5 percent more of the heating value being tied up in condensed-phase combustibles at any given total heating value. Substitution of small amounts of AP for HMX (experi-

mentally shown to improve propellant burning characteristics) causes a drop in flame temperature for a given heating value, chiefly because more magnesium must be used to achieve a given level of heating value when AP is substituted for HM. than AP.) (HMX is a considerably more energetic oxidizer

The bend to the right of the two curves for AP-containing

systems on the right-hand side of Figure 10, is caused by the temperature dropping sufficiently in these cases to cause some condensation of magnesium,

thereby tying additional potential heating value up in condensed phase products. Except for this effect, occurring only at the higher heating values,

the substitution of AP has very little effect on the percent heating value contained in condensed-phase combustibles at a given total heating value. As may be seen, increasing the magnesium loading increases both the theoretical volumetric heating value and the percent of that value which is tied up in condensed-phase combustibles. However, primary motor temperature is nearly

36

Ud

Ic
________~4 _ _

'

0__

z>
4K4

A.

oz

6 (n

A.w

w
C) A.

to D.

00

< x m

P0

P C4

en

3~~~

zAV
0 0

NL3
CL

ILC

L37

IIJ
____en_

LLI-

117

>

31

0..

IL

<

00~~

Y4

(ut/ntg) 3rflVA!)NtLVIH

38

independent of the magnesium loading tip to about 50 too 55 percent due to the

highly exothermic reaction, Mg(g)

4 CO - MgO(s)

C(s)

- at higher loadings,

the CO is exhausted and primary motor temperature decreases drastically with further increases in magnesium content. Thus it appears that tht maximum i4,J/MX/AP/Mg) is about
of

practical magnesium content in this type of system 60 percent.

At this loading, theoretical volumetric hatin. ",iljes

approx-

imately 550-560 BTU per cubic inch are attainable, with maximum reduced volumetric heating values of 420-440 BTU pe cubic inch.

Results of thermochemical calculations for two mixed-metal systems (magnesium/aluminum and magnesium/boron) are presented in Figures 12 and i I. Boron and aluminum both have higher heating values than magnesium (10636 BTU/lb

3 g 3 3

for magnesium, 13361 BTU/lb for aluminum, and 23,900 for boron) - however, both are ejected into the afterburner as relatively hard to burn particles. In Figure 12, heating value and reduced heating value are plotted against total metal loading for 157, HTPB binder formulations with HMX oxidizer for all-magnesium, 5/1 magnesium/aluminum, and 1/1 magnesium/aluminum cases. As

might be expected, the theoretical heating values increase with substitution of aluminum for magnesium; however, it may also be noted that the maximum obtainable reduced volumetric heating values are essentially independent of Magnesium/Aluminum ratio. With boron/magnesium mixed metal systems, the

picture is similar except the reduced heating value actually decreases monotonically with substitution of boron for magnesium. A summary of maximum volumetric heating values obtainable with each of the castable systems examined is presented as Table 5. With

the exception of the mixed-metal systems, the various systems examined give similar maxima, 520 to 550 BTU/in 3 for total volumetric heating value and 430 to 470 BTU/in 3 for reduced volumetric heating value. In terms of total heating value, the HTPB/|[M(-AP Magnesium system appears best, while the NC/NG/ADN/Magnesium system yields the highest reduced volumetric heating value. The mixed metal systems permit attainment of higher total volumetric

heating values, but at the expense of increasing percentages of the heating value being tied up in condensed-phase combustibles (which are relatively

I!3

w
RJ z

<

4z

<~

U)
U

z C,

Li~z
___ __ ____

___

Son

o~~
a w

En0
D~~n9 pNIb 3fTV

< C

-e

"t

zo0
u*

M.~

nU. :zin

0u0
0 w-

'z ~u
w
0

t.

#A-

41

C
.00

Ln
I

--

:3

>

~0

7a

z~

I Cc

a m

u0

-D
FH u

2
-

I 3

-CP

~cIL
wP c

fQ

-E
a XX

U)x >IjR
QX
Q

LUN

CIC
MM

( u

42

hard to burn) entering the afterburner.

The tradeoff for optimum delivered

performance is obviously a function of the severity of the afterburner environment, with maximum total heating value being desired for after-

burner operating conditions conducive

to complete burning of condensed-

phase materials and maximum reduced heating value being desired for condiLions where condensed-phase combustibles cannot be Ignited and burned. 3. Pressed Compositions Similar thermodynamic screening calculations were carried out for a large number of potential pressed composition systems: (1)
(2) (3) (4) Pb(NO 3 )
-

Mg

(CH 2 )n
-

Ca(NO 3 )2 - Mg Sr(NO 3 ) 2 - Mg Ba(NO 3)2 - Mg

(CH

(CH 2 ) n (CH 2 )n

(5) (6)

KNO 3 - Mg - (CH 2 )n NaNO3 - Mg - (CH2 )n

NH 4 NO 3 - Mg Oxallc Acid

(7) (8) (9)

(CH Mg

Tartaric Acid - Mg Sucrose AP HMX

-

(10) (11)

Mp, (C')
n

Mg -

(12)

Mg - (CH 2 )n
6

(13)

(NF4 )2 SiF

Mg -

(CH2 )n

(14)

NF 4 BF 4 - Mg NF 4 ASF Mg

(CHI)

(15) (16)

(CH 2 )
2)

Teflon - Mg

-(CH

(17)

Delrin - Mg

43

(18) (19)

Polyvinvlidene Fluoride - Mg XXX

-

Mg

(CH, ) 1 9

-XXX

= Proprietarv ARC Oxidizer

(20)

XXX - Mg - Al - (CH,) . n (CH2 ) as a pressing aid, At

Within each system containing

levels of this Ingredient from 1 to 5 weight percent were examined.

each level, magnesium content was varied over a considerable range, representing a wide variation in fuel/oxidizer ratio. Detailed results of all rather, detailed

of these calculations will not be presented herein:

restults for a few typical systems will be presented and discussed, and results for the other systems will then be summarized.
Detailed results for the Mg - Teflon - 5%(CH2) , Mg - Teflon
-

I "(CH,)

, Mg n in Figures 14

Polyvinylidene Fluoride, and Mg-elrln systems are presented 17. In Figure 14, volumetric heating value is plotted

against primary motor flame temperature, with the percent metal loading for each point designated. As metal loading is increased, in each case, It

flame temperature decreises while volumetric heating value increases. is particularly interesting Lco note the shapes of the curves with the

existence of regions where volumetric heating value Increases very slowly with large decreases in flame temperature, separated by a region where volumetric heating value increases sharply with only a small sacrifice In flame temperature. Using this crlte'ia -. !one, it would seem that the most

promising composition from these systems would be those at t!e top cft the sharply rising parts of the curves. order listed above, Thus, for the four systems, in the

it would appear that the optimal magnesium loadings

(would be about 60 to 65%, 65 to 707, 607 and 60%. Since pressed composition of these types tend to be poor in terms of gas production, and since gas production is needed to break

3 3

up clinkers and give efficient propellant ejection from a motor, plots of volumetric heating value versus moles of gas produced per 100 grims of propellant, as plotted in Figure 15, are also of Interest. there is, As may be seen,

in general, an optimum metal loading in terms of gas production,

44

-4

--

co c.'J

CNj

c'j

0A

00
LL. Li. Li.

I
0r

wA w

>

WL
,-

~ ~.

C)

00

C)C
Ln c., J L

T~

_i

40 fl 40

Cl

LAI

(D L)

Cj

<

U. CL

z
U)

LL

Ln
0

0
0n

C~

UJ

UJ

-J

'a

a.

(ELnn

f1A~N13

I3~l0

3 3 In>L

-4'- L

it h

-prod(Iuct.

p>

ton dropping r ap Id I y withI hb, on I v modest in cre as v in

f urt her

In creaistvs in
.
()n

metal

1load ing

uncoiipin i cd

heat in p va It lie)

th is rder

has is,

opt Hu:m. magnesium

loading s lot
601 to 16,

the four systems,

657. 2. 5 to 55/7

in the

! i>,t( 1
agiinst product s

<x

Aeaott L,

',

to 60,',

aind 65, ted

In Figure 1 n t hreat ing va1lue I r t lie. ,-ame in

volumetric

heating vaIlue' Isi ',ot

condensed -phase A ,,.i in,

combust ible primary motor in the curves

four systems. loadings

knees appear

I nd icat irn
increased

opt ima 1 m,.etal

based on hal ancing off

the benefit of

heat ing Value wi th (1! thalt

Increasing metal

loading against less l oading.

ready aval la1 hi I i t upon In this

heaNt ing value with

*

increased metal

Based

r itot ion ' rder

,

the opt Imal metal to he about Thus, it

load ings for the 60 to 65%,

four svst ems

the samnk' and 65"' load ings lead

appear

55 to 60'Z,

52.5 to 55',

appears that select ion of opt imal metal the three criteria

or the,;,

sys tems on the basis of

discussed above

o --iirIV

similair conclusions.

The optimal metal

loadings are:

S yt .- em
MgMg Mc
-TeflIon -57'

Optimal Metal Loading (Wt7%)

(CH 2) /)' (CH 2) L60Z ^,657 5 5"% 0 17,

Teflon

Po1 vvi nylIidene Fluoride

-

1elrin In Figure

total and reduced volumnetric heating value

3 3expe)-(t
niot

aire 1plot ted against percent magnesium loading for each of the four sytremIs di1sceussed above. For each system, total volumetric heating valule incyoases
ais one woul d

moot on ical iv wi th metal

.

loading over the range examined,

The reduced volumetric heating value curves, however, loadings approx imately 3 to 4%, belo OW

-how maximal

oc( urring at metal above ;is opt ImaI

igphit f f st I,

hose l isted
.

on the has is

of the criteria iprev ious lv (Iiscu ssed

'Th 1 s

niot unex pert cd,

s inCe' the opt imalI meta IIoad ing ShOn 1(

ht- that

at which the reduced volumetric heat ing valu te maximizes, but

higher valu te at which

ins tead some

Its decav with further increase In metal

loading has not become very large.

47

-. 1

0
cc

I
U, U,

a. I

r-.r

z
CM'
-Jr

Lnw
__L

___L__
0o4
in_
_

e;
_

$
0

l
-

zz

0
U.
w

zI
.>
Wn in w C4J

UWj

.0c

U,

Ll

U,

U,"U

(ui/nIa3) 3n-1YA 9NI.LVIH :WIHvniAOA

48

FT

LL)
0
4i L-

w
0
4
LA-

>0
>
m

>

z
-

Lij z J
-)

>
i U
'

04
)
-7

4i

< 0

w U
W
U')

>

~Z
LL
4i

z
wi

w.
C71

> 4i 0 6% a'

a:

4i w 6
a'

> 0
a'

4 i Ld 6 a

0
w a' ft

0
w
0

-o

2(

LAm

(5

C__

ul

%A

0
-.

Z
-n

Unq

ui/II3

ONloH::U1w

-O

-nV

\49

Th t hr

-sensstudiedin this screenijug effort

showed

simi lar character jot ics in general and will not be discussed in detail . A summary of resulits cxaminied
*

for 17 of t he pressed composition sv.-t'fl15

in Lterms

)f Total Volumetric Heating Value and Reduced VolumcLr i( LoadIng are presented in Figures 18 and 19.

fihat ini- Vil uc versous Mtii

(Three ot the'

nitrate Syste'ms, are not included due to their similarity t,)

*

jn the sodiu

itrate

p0

tarssiurm n it rate

and ammonI inm nitrate svs tems.

,

S ince m in imiz ing uise oif thle press ing aid

(Cil.,)

,was

found to give optimal

resulIt s thLermidvnalc(A liv,, (as, shown in F igure 17) I" 1level ofi this ingredilent are shown.

only results for the

in thle only system containing mixed

-io t a is

(igcs l evelI
.

ium and aluminum)

results are presented onlv for thle 10(V

iIr Li num

Table 6 gives a summary of optimized volumetric heating

vailies avail able from all1 of thle piressedi compositions t hermod'riari I all v s, rteentd. The svstems con ta In ing theit in term,- of total heatinog propr ietajry AEC o)x id i zer (XXX) are cl ear I , super i or

value and, with thle except ion of thle exotic this tine),

fluorine salt systems (which are not realistic cand idates at are also superior In terms o? appear attract ive reduced heat ing value.

Other systems which both of which

incl[ude Mg/Teflon/ (CH,) nand MG,/ItX/((CH,) n,

apipear respect able in terms of both heat ing value parameters.

~.Candidate_ Formulations

Selected !live and Ohrke

At the end of thle thermodynamic screening phas;es, rastahie formuliat ions des ignated 'MK] presse,;d forrnl.at ions desi,gnated TKI eVA I Uat I on. Latetr,
, ,

nK2, TK2
,

?4K6, MKI 1, and KK1 2

and TK3 were selected for further in making MK6

,

When di ff icrilty was encountered

It

was,

replaced bv a similar formulation MK6A. Detailed

(bitt with lower sol Ids loading)

designated

romposit ions of the-se formu lat ions are not presented here,

in ordler t hat thi s report may be uincIass if ied , but general chiaracter!ist ics are presented in lablte 7. All of the formul at ions contain in excess of 50% metal (7 containing magnesium, I containing a 211 MG/Al alloy, and I The magnesium systems have volumetric 3 heating values of 490 to 570 BTU/in with reduced volumetric heating values containing a 5/1 Mg/Al alloy).

Is

Mq/AI/ XX

CH 2

Mg/

4 8/CH

Z)Mg/N

F/Cm F AsF

55

0 555

60

058

METAL~MI/XAI

LODNG(CghIDret

WIN3/H
Prop Ln LODIGs wegh prcnt

51

0 NM

w
>

x I

( V
Z

U y-<U
J

X
I

~0

Z
z ZZ 0

~~

00

w
z-

1:

--

I 40

LL

--

AC
z L-

0'0

Xu

-y*

LA

II

04

in

'000,

40

in0 (E uo/nIIe) 3n-iVA

DNliV3H
)Ikj13vqfljOA

a3:flG3

in

I5

IL.
(A

O.J

M0 -

4'

00 ~ 061: 01~~

~6

~ 6,

d 6

6N

CM

Co4

~NN

di

ww
g ODED ~

Q D~) Swow CC Z 2 (a

~~ODI~tL

D~ Z < to

E-D>

F-

cc E

o
-c4 .

0 zo z
x>~
F-w co

cc 00

uuJ

I ID I IL

~OO0Z

~0

53

a)Q

r14

-4

-4

n..

-- r

t'

-I

a'

-.1

4~

C.)L
04

~W
00

a~o

~~

r-

r,

0I

~ -44 ~ ) -

r N c-

Ic~ (

Q)C

I-.J
-L

<I

A0 -4 0)~~~Ha) 0
1
J

0N0 -. . 0n 0 0-~O -1 IN 0 N C

A Lw~~

-40

a)

54

o"

W:. i"

of 410 to 490 BT/in 3 , wh lle

the heating values of 590 to 750 BTU/in 3 .

alloy ,vstemg: vg/Al have volimetric All of the castable compositions (for 1000 psia combustion) in excess

have theoretical flame temperatures of 2450'K, to 2400 0 K.

while the pressed compositions have flame t.,mpcratures of 2340

An important factor in determining whether or not condensedphase combustibles emanating from the fuel generator (primary motor) will burn well in the afterburner is the curve of flame temperature in the absence of such burning versus air/fuel ratio, since, given a minimum

temperature required for ignition of the condensed-phase combustibles, this determines over what air/fuel ratio range ignition can occur., (A

small range indicates that if ignition does not occur in a very limited fraction of the afterburner volume, it will not occur at all, while a large range provides a large fraction of the afterburner volume suitable for ignition.) Such curves are presented for the original five candidate

castable formulations, the three candidate pressed compositions, a~d a typical highly-boron-loaded formulation in Figure 20. As may be seen, the

limits are considerably broader for the magnesium-loaded formulations. For example, if a 2000*K temperature is required for particle ignition, this is provided only in an air/fuel ratio envelope of 2/1 to 5/1 for the boron formulation, but for all air/fuel ratios below approximately 8/1 for the magnesium systems (6/1 for the mixed metal system). In Section Ill-A, the development of plots [Ratio of Required Flow Rate for a Candidate Formulation to Required Flow Rate for a State-ofthe-Art Boron Formulation] was discussed. As stated there, under the highly

simplified anlaysis used, this procedure basically involves fixing air flow rate and then calculating the air/fuel ratios for the formulations being compared which will yield the same air vacuum sonic specific impulse (S A)
values for the two fuels. Results of such ctilcu.ations with

K1,

MK2'

MK12,

T&:, TK2,

and TK3 as the candidate fuels are plotted in Figure 21 (Calculations

(Mass Flow Rate Ratio) and FIgure 22 (Volume Flow Rate Ratio).

for MKO, MXYA and MKII are not included, since they had been dropped by the time this nnnlysis wns performed.) As indicated on the figures, high

55

-- tC

.0

/m
C

o-

-~

*0 0

Ur

<2

EE
0

40 04j 3nl'VMdW3N

0)

56

I
zw
I

0
< zt
UiU
'f 0
0-

II

<~A

z S0z
A M-1

w
CIL
c

ca 0
tt -0%r

A~<

0 zo

040

Cor
m '
-r 0

NOuvInNfl'dOd NOkI09 'V*0S/3ivIGNV:)

'S31IVU MO-14 SSVV4 03kfint)3W :10 OIIVU

57

w
0

w
0 U

Y0

<:

U.L
________

LL
-

z.

ACA 0
Z

CL
07

:
o-

C;

IW
S3LV4

NOO9 oS/ivaawA NOI~fl~d~

I
58<

(nIi~1b~f1A0Il~~~

values of the abscissa are indicative of missions with low thrust coefficient
requirements whil.e Coefficient lying low values are indicative of missions with high thrust indicates that as indicated rank (in lower in

requirements.

Straightfoward

reasoning Thus,

curves represent

more favorable

fuels.

the figures, air/fuel ratio

for weight-limited

systems the candidates in the order

the low >

range of most interest) > [TK2] > [TK3], > [TK2] is

[MK]2]

> [TKI]

[MK2] > [MKI] rank is [TK1]

while for the volume-limited systems the > [TK3]

"

> [MK12]

[MK2]

> [MKI].

As might be

expected, this result

essentially in

line with the rankings in terms of at

mass and volumetric heating values, high air/fuel

the only exceptions disappearing

ratios (where heating values have more meaning) as curves As pointed out earlier, for all operating points

cross in each figure. to the left of the

intersection of the curve for a given formulation is theoreticallv

with the ordinate value of 1.0, the candidate formulation superior to the state-of-the-art boron

formulation used for comparison.

C.
1.

PROPELLANT FORMULATION AND CHARACTERIZATION

PreliminaryFormulation Work on Magnesium-Loaded Composite Propellants (Feasibility Assessment) Until recent years, the "work horse" (-1st.i1_e_ Fuelj-Rich

bindtrs for composite functionality which Unlike Al,

propeilants

have been butadiene polymers with carbrvI

have been crosslinked with epoxy and/or

tmine curing agents.

which has a protective oxide surface, the Mg surface MgO also crosslinks carboxyl or chemically polymer. alter groups. it

is very reactive.

It is therefore necessary to passivate can be incorporated in in a carboxyl development

Mg powder before

This problem was encountered

a previous propellant

effort under an AFRPL program involving development of high boron-loaded fuel-rich propellants. Mg was desired as an additive to improve the boron However, addition of Mg to the propellant

combustion in the afterburner. consistently resulted prior

in crosslinking (gelling) of the binder during mixing Passivation treatments were helpful, but in using a

to curing agent addition.

the problem was not solved. 65/35 Mg/Al

Success was attained, however,

alloy powder which was passivated.

Hydroxyl polymers were not

evaluated because of boron

incompatibility with the cure system.

. .9

AI W .... .

A series of small mixes was made compatibil itv point out p I vine rs . problem. 'rI. ,se are summarize-d in

to

assess the extent of TLse resl ts

the

Tab Ic e.

the serious

Inc,,mpattbII ity prohlem be tween Mg, powder and ca.rboxvl evaluate d were crosslInked On the other hand, wit ia ut curinrl alpent,

Both pol vners

and they gassed badly . this problem,

hvdroxyl

polviers did n( t show ormul.ltions

indicating that magnesium-loaded

castable

could be made with hvdroxvl-contain g polymers cured As part of this phase of the program,

with dilsorcyanate. aimed mainly at fuel-rich screen-

Assessing the feasibility tof making tastable

magnesium-loaded

propellants (and carried out before completion of the thermodynamic ing effort described earlier),

five processable castable magnesium-loaded

fuel-rich propellants were successfully formulated, with magnesium content up to 37.5Z. (See Figure 23 for compositions.) The highest magnesium-loaded

formulation, designated MFRIl8 has a theoretical heating Value of 8600 BTU/ib or 492 BTU/in 3 . Attempts to produce formulations with only it appears that the lower limit a 10:

binder were unsuccessful; is 12 or 13'/.

on binder content

An attempt was also made to formulate This effort was unsuccessful

15/35/50 ItTPB/ the propellant found that use ol

FIMX/Mg propellant.

in that

would not process and gassed badlv.

It was subsequentiv

a different cross-linking agent eliminated the gassing problem. spherical HNX in place of the rough crystall ine material of and control

Use of

used in this

first effort, and better selectIon sizes,

over ItNX and Mg particle

solved the processing problem, as evidenced by results described in In addition to the castabi]

the next section. mixes of 10/80/10,

formulation work, drz-packed

10/60//30,

and

:0/40/50 IITPB/tLMX/Mg compositions were

made for preliminary combustion studies. None of nitrogen at pressures. the above formulations although all 5 to 1M%All would sustain combustion under well at higher

one atmosphere, Substitution of

burned reasonably for HMX was found of these

to greatly The formubut they

Improve the combustion lations were very "dirty"

characteristics burners,

formulations.

producing heavy clouds

of soot,

did burn completely, leaving essentiliily no residue (an Important consideration as discussed earlier). The large quantity of soot ejected caused

60

z44

LU X LU

-~

LU

C) 0 0
0

4
000

U L) . 0r

< -J

00 4 J 0j
U LU

0 C.)
U. 0

0 0

0
LL 0

0. 0

LU

C) C)

LU

LL

LU

wLL z
<
.. 0 -

LU

x*>~>

z
4j

< -J

< -A

< -

I0O 0oC

0 0

0>

>

90
U

0C

9
LU

L LU

LU

LUJ

COtt

<C4 2> -U

< >
LU U U LU

< >

< >

40 z z z
LU 2

CCa
$n < >

C)

-It--

0000

00

C)

L&A

C '

Ln

f-

CC3

31VU

ONI.f.

062

AA W0 -~ ----

0-

attempts to obtain burning rate versus pressure curves in strand bomb to fail,

the standard Accordingly,

due to shorting of the timing trip wires.

burning rate data on the five castable formulations were obtained by visual observation of their combustion In a pressurized window bomb, with use of a stopwatch. These data are presented in Figure 23. The scatter was due,

to the use of the visual

timing technique and not to burning irregularities. ranging from

A;i may be seen, the burning rate exponents were reasonable,

0.2 to 0.6, and the burning rates were in the range of typical fuel-rich propellants of other types (0.13 to 0.38 inches per second at 1000 psia). This preliminary formulation work indicated of magnesium-loaded that production

fuel-rich propellants with 12 to 15% HTPB binder ,and

up to 50 to 55% magnesium would present no insurmountable problems and was used in conjunction with the castable propellant thermodynamic screening effort described earlier in selection of the candidate formulations listed in Section III-B-4. It also indicated that for burning rate control, use

of some ammonium perchlorate and/or an energetic binder would be necessary since varying WIX particle size has very little effect on burning rate. 2. Formulation of Candidate Castable Compositions Defined In Thermodynamic Screening Effort As discussed in Section III-B-4, designated MK1, MK2, five castable formulations

K6, MK11, and MK12 were selected for further eval'ation.

General compositional characteristics and heating values for these formulations are given in Table 7, though detailed compositions are not given Three of these formulations (MK1, MK2, and

for classification reasons.

MK12) utilize the R45M HTPB binder system, but differ in fuel and oxidizer. MK2 contains both AP and HMX as oxidizers, while MK1 has no AP. has no HMX and MKZ2

MK12 utilizes Mg/Al (2/1) alloy rather than the pure magnesium MKI and MK2. The other two formulations contain energetic binders with

metal used in

some nitrate ester content. for smokeless propellants.

MK6 utilizea a high energy binder developed It is composed of nitroglycerine, lacquer grade

nitrocellulose and polyester prepolymer (R-38) crosslinked with isocyanate, The MK11 binder is less complex consisting of BTTN-plasticited polyester

63

cured with diisocyante.

Ingredient incompatibility problems were encountered

with both of these propellants, as well as with the modified formulation, MY6A. The solution of these problems will require more time and funding than and further effort on these formulations

were available for this program, was eventually abandoned.

MKI, with AP oxidizer, This formulation

was selected for

Initial

formulation.

was used to optimize particle

size fit, to solve gassing Most

and curing problems, and to evaluato processing aids and additives. of the trial mixes were made with this propellant.

Once the problems were

solved, formulation of the other two HTPB propellants was relatively straightforward. Previous work had indicated that is is necessary to rigorously dry the Mg powder to eliminate gassing. vacuum at Freshly dried Mg powder (under

100C for a minimum of 6 hours) was therefore routinely used

Optimizing particle size distribution was made more difficult

throughout this program.

because the spherical Mg powder was available in only 3 average particle sizes: 18, 22, and 50 micron. Formulation with 3 to 5 micron powder was not it only under non-volatile

considered, because the manufacturer would ship liquid. (small). I

was available In only 2 sizes, Class A (large)and Class E AP was available in many sizes. The first series of MKI mixes to optimize particle size

is summarized in Table 9. micron Mg and 7 micron AP. gassing.

The best particle

size fit occurred with 50

All of the mixes made with IPDI exhlbit,,d some did not, it was subsequently used as the

Since those with DDI

curing agent.

The addition of 3% Velocite plasticizer to the binder did Tensile properties were

not improve processing so its use was discontinued.

determined on Mix 135, which had a NCO/OH ratio of 0.8/1.0 (74 psi maximum stress, 35% strain at maximum stress and 387 psi modulus). Subsequent mixes were made at 0.9/1.0 to obtain a harder propellant. Although the end-of-mix (EOM) viscosityofMix 135 was only

11 KP and it had good flow properties, it was observed to have a very short

64

'44

0) 4) 4) EAu) m 0 (4 .4

U -) 40 M C " Q00000 V ) cl to m ) (a V4C u ) 4 C-

0w
:10

0 0 z

0 0 z
00u

z-0 4~

ON

(' 4 Ul
I-

) co
-.

4 C
-

,4
-4 U

(1w4
U, A

.
CrN

444

(4
x -4u

040

4:

44

04 ,4

1-4

1--t4

UOeO

4N
C4 -4

-C,4

a~(N
(N

!
C-I

:1.~L~0
" C4

C) Cl)

0
C') U'

II

C>

1~ ' P4~~~~I:

L 00 1 0V4 N

0 0"

:'

Y1

65

pot life

(less than I hour).

This was attributed to the surface characterA series of mixes was therefore made The problem was

istics and

impurities of Mg powder.

to solve this problem. essentially solved bonding agents.

These are shown in Table 10.

-y changes in the R45M antioxidant in conjunction with

Mix 170 was chosen as the final candidate Formulation MKI.

It contains HX-752 bonding agent and the NCO/OH ratio was adjusted to 0.85/1 to obtain Improved mechanical properties. Because of the experience gained with formulation only four mixes were required for formulation of MK2. (No. 131) was processable, but had short pot life. in Table 11. of MKI,

The first mix made

The mixes are summarized

Mix 171 is the final MK2 candidate, and had properties very candidate.

similar to the MKU

As with MK2, relatively few mixes were required to develop MKI2, due to solution of major problems during formulation of MK1. differences in this formulation were noted. Two

The Mg/Al alloy has less The

interaction effect with the binder, resulting in a softer cure.

absence of AP also changes curing charazteristics requiring longer cure times. Table 12 presents a summary of the mixes made. Mix 172 became the

candidate MKl2 propellant. improved tensil

HX-752 bonding agent was required to get because of the large Class A HMX.

properties, probab],

It also exhibited yielding or dewetcing to a much greater extent than the other two propellants. The mechanical properties of the final candidate propellant

formulations were included in the preceding tables. of the three propellants is given in Table 13.

However, a comparison

These are microbone specimens, JANNAF specimen data

tested at ambient at a crosshead rate of I in/min. were not determined, nor were high

ind low temperature properties determined.

The safety and hazard characteristics of these formulations were also determined. A summary is given in Table 14. The data indicate

that no extreme hazards should be involved in handling these propellants. None of the samples ignited at 6 joules, 5000 volts on the electrostatic discharke apparatus, but all emitted sparks at low joule settings. This is

0 00

~~ ~ ~
. 0 00

-r

V) .

04

04

.4

0'

Si

00

C:

0en

04

-IIn

m CI en

04

v-

00

V454 Go 9-5

c5

I-

<

InS

0n

e4

o
1-4 0t 0

00
00

It

14

14

14

67

1
-44 u4

o3

I.O

F5
'.4

o
,.

a
-4

0
.4 -4 S c4 -4
I

C..4
I ;W". , r

0
S.,

W co
-4

~ -4

It

MiC-~

P4 0)&. 0
-

-4 4-4 '9 :U

-'.

-4000
W0-

C
(L(4 '-4

CV

0) 1~
4.,

Qn 0'
)
40C1

'.
('

Z 0
0) w4

0
U) n V)U4 9)

0d

0d

6)

4-4

.0

ul

..4

-4

.-4

.- 4

-,4

r.4

-4

r4

r4

68

En

-1

C1 4

C-4

m~

-~

04

04

en

4-

c*

000

I-4

-(3

41 r2

1 t

:IJ

9-4

'-

cA .
2 4 .4 co 0

"0 4)

U)

-z
-. '~~

a)
x-

4
06

"n
4.0

c~*~

~4

70

H.

. ........

MU

.14Du
0

fwu

U.

C.?

in

44

00cm0

z
le n
75 U w

NN

wl co ccr
0

z
Z 8A

wJ 00 i Z. Zw

2.
0 t 0O4

.0

c4

le

<

> <

cc 0

cc
cc LU LUc

c LU

b-

>

cc

cc

0. Z 0

L"u x-LU

0
-

zo

x~~ IzC-

~W

0wL

71

typical of Mg and Zr propellants, and is due to metal particles flying

off the propellant surface.

Whether or not these could result in an igniThe HMX piopellants are more sensitive

tion of a large sample is not known.

to impact, but are less sensitive than composite propellants containing. ultrafine AP. One positive test out of 10 (9 negatives) was observed with This may be

the cured AP propellant on the ESSO Screw with pyrex grit. a freak, since theother 9 are negative. before conclusions are drawn.

Further friction testing is indicated

All of the propellants have high rapid-rise

autoignition temperatures, indicating good thermal stability. The ejection characteristics were compared at ambient pressure in air. The MK1 propellant (Mg, AP) exhibits excellent ejection, burning The MK12 propellant (HMX, The MK2

like a stoichiometric aluminized propellant.

alloy) was difficult to ignite and exhibited poor ejection. candidate was intermediate in its ejection properties.

Burning rate data for MKI, MK2, and MK12 are presented in Figures 24 bomb.
-

26.

Initially, data were obtained in a conventional strand

Extreme difficulty was encountered in obtaining reproducible strand This was attributed to the very reactive Mg The Teflon washers and insulation were

burning rates in this device.

vapor from combustion in the bomb. severely attacked.

Premature clock starting and stopping, and clock chatterThis was attributed to arcing and shorting

ing were consistently observed.

out due to Mg vapor and condensed Mg., It was necessary to verify some of the data points in the window bomb using a stopwatch. not too accurate. This technique is

The burning rates of the three propellants, as obtained

by this technique, are shown in Figures 24 - 26 as the hexagonal points. The MKI and MK2 rates are quite similar in the upper pressure (700 to 2000 psi) range (0.40 vs 0.38 in/sec at 1000 psia). However, MK2 has a much This is a typical

lower exponent, with hiigher rates at lower pressures. effect of HMX.

The MK12 rates are considerably lower and this propellant

burned erratically nt lower presnures.

below 100 psia.)

(It would not sustain combustion

This is consistent with the ejection observation and

indicates marginal combustion.

72

<

IA

Ii
z
> LU 01'-'

--

-j_

< < z Z0

aLL

w
C) I 1

w
0
1.

0.

U___

a
Zu,/l)81
73--

p.-t

cc

oi

LA

<

0 0 L0
___ CA I

>A

A<
w
-_

_-

CO

__

-z
x

'Ao 0

wA

__

j -

<A

74

-~4

1F~i-

ti

LA

r-

z
>

z4:
--

0-

4:g

<

r
0

u-

0=

00

0~su)~

(3&/U)

Later, additional burning rate data were obtained, using an acoustic emissions technique to determine the beginning and end of burning of a known length strand. halocarbon grease. In these tests, the strands were coated with

Rates obtained in these tests are presented as the these bars indicate the variation

horizontal bars in Figures 24 - 26:

in pressure during the tests due to use of a closed, rather than vented, strand bomb to prevent noise interference from flow from the strand bomb to the surge tank. As may be seen, these tests indicate rates of about

60 to 70% oi those obtained in the conventional strand bomb tests. Motor tests with these propellants, described later in this report, were used to obtain motor burning rate versus pressure points which are also presented in Figures 24 - 26. These tests indicated that the

strand data obtained with the acoustic emissions technique were considerably more reliable than those obtained in the conventional strand bomb. 3. Formulation of Candidate Pressed Compositions Defined in Thermodynamic Screening Effort The three candidate pressed compositions identified in Section III-B-4 (TKI, TK2, and TK3) were formulated and evaluated for safety and ballistic properties. All of these compositions are highly magnesium loaded,

with TKl containing some aluminum metal as well (approximately 5/1 Mg/Al). TKl contains a proprietary fluorine-containing oxidizer, TK2 employs HMX as its oxidizer, and TK3 contains polytetrafluoroethylene (Teflon) as an oxidizer. All of these compositons are dry blends which are compacted under Compositions TKI and TK2 contain small amounts of isocyanate-

high pressure.

cured hydroxy-terminated polybutadiene (HTPB) binder as a composition aid, to impart adequate strength to pressed pieces for handling and use. ition TK3, made up with Teflon, did not require a binder. Compos-

Safety properties

measured were sensitivity to ignition by impact, friction, electrostatic discharge, and autoignition temperature, using BuMines tools. Results of

these tests, presented in Tables 15 through 11, indicate no particular hazards problems associated with these formulations. In terms of ICC

Shipping Classifications, TK2 will be a Class A propellant, while TKI and TK3 will be Class B. 76

TABLE 15.

SUMMARY OF TKI PROPERTIES

Safety Properties Impact Sensitivity, Eo Friction Sensitivity, Esso Screw Friction Sensitivity, ABL Block Electrostatic Discharge Autoignition Temperature Ballistics 0.514
In steel tubes r = 0.54 (P/100) r = 0.96
, ,

225 kg-cm Negative Negative at 500 psi/900 Negative at .09 joules/5000 volts > 390 0 C

P < 300 psia P > 300 psia

Free-Standing

r = 0.31 (P/100) r = 0.85

0.727 0.066
,

<

360 psia

(P/O000)

P > 360 psia

Burning Characteristics of 1/2" Diameter Strands in Steal Tubes Burns well with no clinker up to 1000 psia. Slight degree of clinker on tube walls for p > 1000 psia.

! I I I
77

TABLE 16.

SUMMARY OF TK2' PROPERTIES

Safety Properties Impact Sensitivity, Eo Friction Sensitivity, Esso Screw Friction Sensitivity, ABL Block Electrostatic Discharge Autoignition Temperature Ballistics In steel tubes
0 3 r = 0.395 (P/1000) .

100 kg-cm Negative

0 Negative at 200 psi/90

Negative at .09 joules/5000 volts 280C

Burning Characteristics of

" Diameter Strands in Steel Tubes

Small clinker formed at 100 psia. Extensive clinker formation at 200 psia. Clinker is very hard and soot ejection pour at 1000 psia.

I i

I I

78

TABLE 17.

SUMMARY OF TK3

PROPERTIES

Safety Properties Impact Sensitivity, -Eo Friction Sensitivity, Esso Screw Friction Sensitivity, ABL Block Electrostatic Discharge Autoignition Temperature Ballistics In steel tubes 0.62 r = 0.185 (P/100) r = 1.15 Burning Characteristics of No clinker at 100 psia. Small clinker at 200 psia. Extensive clinker and poor soot ejection at 500 and 100 psia. (P/1000) , P < 500 psia. , P > 500 psia. 150 kg-cm Negative
0 Negative at 500 psi/9 0

Negative at .09 joules/5000 volts > 390 0 C

1.20

" Diameter Strands in Steel Tubes

i i I I I

79

Burning rate versus pressure data were obtained over a pressure range of 100 to 1000 psia for TKI, TK2, and TK3, using cylindrical pellets approximately 1/2-inch in diameter and I-inch long, pressed at 30,000 psi in steel sleeves. Results of these measurements are given in Tables 15 through Both of the compositions containing a fluorithe proprie-

17 and Figures 27 through 29.

nating oxidizer (TKI and TK3) exhibit a pressure exponent break:

tary oxidizer reducing the exponent at P >300 psi, and teflon increasing the exponent at P >500 psi. Burning rate measurements were also made for Compo1" L 1 The use

sition TKI pressed as free standing cylindrical pellets - 0.5" dia. x inhibited with fluorocarbon grease. Results are shown in Figure 27.

of grease as an inhibitor in place of a steel sleeve causes a slight reduction in burning rate, and increase in exponent. The characteristic exponent break

is still present, but occurs at a higher pressure (400 psi vs 300 psi with steel). Burning characteristics (ejection) are summarized in Tables 15 through 17. The best of the compositions, in terms of burning without

formation of clinker is TKl; this composition ejects all combustion products when burned at pressures up to 1000 psia, the highest pressure tested. For motor firings (discussed later), two grains of composition TKl werE pressed up at 30,000 psi, free-standing, in seven increments. diameter of each grain was 2.05 inches. The

Grain I was 7.02 inches long and

weighed 696.5 grams. Grain 2 was 7.03 inches long and weighed 694.1 grams. D. PROPELLANT COMBUSTION PRODUCT CHARACTERIZATION Formulations MKI, MK2 and MK12 are thermodynamically predicted

to yield product streams containing approximately 1 gm-mole of magnesium vapor per 100 grams of propellant (about 30 to 40 percent by volume of the total gas-phase products of combustion). Since this magnesium vapor should be

extremely combustible, and in fact should promote afterburning of other fuel product species as discussed later, it is of interest to know whether or not these theoretical expectations are indeed realized, particularly at the high magnesium loadings involved. cient, If the primary motor combustion is not effi-

they may not be, due to lower than theoretical temperature preventing

4.0 (D Strands pressed at 30,000 psi in steel tubes, 1/2-inch diameter by 1-inch long. Burned in tubes. SStrands pressed free-standing at 30,000 psi, 1/2-inch diameter by 0.87-inch long. Inhibited with halocarbon grease.

2.0
0 5 14 0.54 (p/100) .

L~ I
r 0.96

0.31 (p/100) 0 .7 2 7

r =0.8 5 (p/ 1000)0.066

0.2-4

K
200 400 600 800 1,000 PRESSURE (psia) 2,000 4,000

100

Figure 27. Burning Rate Data for TK1.

81

2.0

0 Strands pressed at 30,000 psi in steel tubes, 1/2-inch diameter

by 1-inch long. Burned in tubes. 1.0 4-_1

0.8
0.6-6

t '

_......

0.4

0.2

0 r0.395 (p/1000)

0.1 100 200 400 600 800 1,000 2,000 PRESSURE (psia)

Figure 28. Burning Rate Data for TK2.

I
| 82

!I

4.0 Strands pressed at 30,000 psi in steel tubes, 1/2-in diameter by 1-inch long. Burned in tubes., 2.0
___

1.15 (p/ 1000)

20

_
0.6

-,_
_

/'A

i'

z z
0.4

0.2

r = 0.185 (p/i 00)0.62

0.1
100

200

400

600

800 1,000

2,000

PRESSURE (psia)

I
Figure 29. Burning Rate Data for TK3.

I I
| 83

!
I l l l ld a l lll I l l l lal d l lil lll

vaporization of the magnesium or resulting in a shift of magnesium vapor plus nitrogen to magnesium nitride (thermodynamically predicted at temperatures of 1700 K or less). In addition, while thermochemistry indicates significant if the magnesium reacts prefercntiil1- with the prolimitations, there may be c, iderably more

amounts of CO and Mg(g),

pellant oxygen, due to kinetic

YgO(s) and C(s) formed and less CO and Mg(g).

It is also i:,pcrtant that the

other products of combustion be characterized and compared to theoretical predictions. Two types of experiments have been devised ior characterization of the combustion products of Ml, MK2, and MI,12, tho f ir.;t boin7

an in situ optical absorption measurement of magnesium vapor density above

the burning surface of the propellant, and the second bei~ip- combustion of a propellant sample in a closed bomb with subsequent collectinn and analysis of the products. The r..oor advantage of the first technique lies in the fact

that results should not oe confused by events possibly occurring during cooling of the combustion products. For example, thermodynamic calculations indicate

that as the products are cooled two shift reactions shw ld cccur unless kinetally frozen: pr:,cedlure, g ( -. M:,) C, and 3 Mg + N2 tendinv Mg. N .Wi ti the seccad '1 ii ajor

sue'l- sIifts may occur,

to confuse tl.e is

re,:ILt S.

disadvantapt, of the first

technique is

that it

a very diffict] t experiment, (very fine carbon) in the

particularly with tie high level of particulates propellant combustion products.

The apparatus utilized for in situ Mg!(gl, density measurement is shown In chamber. igures 30 and il. Figure 30 shows schematically the reaction made out of Plexi ,las. TI'he prois

'Techamber is

cruciform and is

pellant sample,

a cylindrical strand 1.3 cm in diameter,

2 to 3 cm long,

placed In the center of the chamber,

the fue 1-rich, orously in magnesilim-contai xin the strands

just upstream of the side ar's.

propellant prodlcts bv an ig.,iito inert an as.

,ince

1,11rn vigArg on is

cxygen,

must he protected

therefore introduced

in three streams,

one from the extreme upstrem;i loration,

w
-J

z z
0

ziI
w

0 0LLL

nLLzL

za 0
0z
00 <o

w
03

-0
0

LnA
E'
U,,

z
<0 u

TU,0

z
co 0 :

00

I1 U

w05

WC

CL

x
CL V

LL Io
CL

and the other two through the side arms.

An optical beam from a tungsten-lamp

continuum source intersects the strand jet normal to the main axis of the experiment. Mg(g) densities are measured by the selective absorption of the Since this wavelength is in the ultraviolet, the entire

2852A resonance line.

optical system, including the side-arm windows, must be made of quartz. Since the Mg-loaded propellants sometimes do not burn well at atmospheric pressure (particularly under argon), the strand combustion was aided by laser radiation, normal to the burning surface. watt CW CO2 instrument, operating in the LWIR at 10.6 pm. The laser is a 100 This wavelength

is convenient, because it is not absorbed selectively by the propellant product gases to any appreciable extent. Some absorption by H2 0(g) and CO2 (g) We found it

must be anticipated, but this should not affect the experiment.

also very convenient to use the laser radiation for ignition of the strand: in each run, the laser was first turned on to a high power; after the strand ignited the power was turned down, or off, as necessary to give a manageable burning rate. The optical absorption apparatus is shown in Figure 31. A

detailed description of the apparatus and of the newly developed continuumabsorption method is given in Ref. 14. pellant strand products. Briefly, a beam from a tungsten-filaAfter the passage

ment lamp, collected by a set of slits and lenses, intersects the jet of proThe beam diameter is 0.9 cm.

through the jet, the beam is split into two parts, each entering an identical set of monochromators, preamplifiers, and phototubes. are 10A wide. The monochromator slits One is centered on the Mg-resonance line, 2852A, and the other

on an arbitrary nearby wavelength, 3200A. The photomultiplier signals from beth beams are displayed simultaneously on a dual-channel oscilloscope. The apparent absorption (A) by Mg(g) is obtained by subtracting the control channel absorption from that of the active channel (2852A):

87

I
-lo

p
A. I

S (VL)
P
p

Dr VL
D (V)

[
pS+F#

1)

(C)

DF . Power detected when light source is viewed through

DF
DP S. PD VL
-

plume
Power detected when plume is viewed alone

= Power detected when light source is viewed without plume = Atomic absorption line frequency = Neighboring frequency

The apparent ges absorption, A, is related to the gas volumetric absorption coefficient, KGA, bys

-exp

[K G

(S2' S)]lgiv' - vjh (v')dvdv'

(2)

.7
Iv

V -vlh

(v.')dvdv'
f (N)

KCA = Gas volumetric absorption coefficient path length S2 Sl= Plume optical gjv'-vj= Slit function h (v')= Exit slit response v - Frequency of incident radiation

V' = Position in exit slit of monochromator corresponding to frequency v for an infinitely narrow entrance slit, perfect monochromator optics, and no diffraction effects.

~88

'Ip'

The number density of magnesium vapor atoms is in turn related to the gas volumetric absorption coefficient, KGA , by:

K GA/SCv) ITe 2

I meCE g, exp (-E,/kT) g [exp (-E I/kT) 3 flu

S (v) = Normalized line shape factor me C = Electronic mass = Vacuum velocity of lighL

9, = Degeneracy of level i Ei = Energy of level i k = Boltzmann's constant T

=

Temperature

e = Electronic charge f u
=

Absorption oscillator strength

= Lower energy level Equations 2 and 3 have been evaluated through use of a computer program as described in Ref.]14 to develop theoretical curves of apparent gas absorption, A, versus magnesium atom number density, N, for various path lengths as shown in Figure 32. However, since there are considerable uncer-

tainties in this theoretical calculation procedure and since preliminary tests indicated very low (less than 1 percent of theoretical) magnesium vapor yields, it was decided that a series of calibration runs be made to develop corrections (if needed) for the theoretical curves. These tests were conducted with the

propellant sample shown in Figure 30 replaced by a magnesium vapor source capable of producing a stream of pure magnesium vapor. The vapor generator consisted of a specially constructed crucible (heated by resistance heating) filled with magnesium ribbon, inside a vessel with a one centimeter diameter orifice through which the vapor exited into the optical absorption test beam. The generator and test apparatus were placed in a vacuum tank and argon was bled in to the desired pressure level for a test. The resistance heater was

then activated, producing a jet of magnesium at the vacuum tank pressure and

r
U
. :[ ...
_ _ __ _ _ _ _

89

..

-,-

i' W
_ _

-...

.2 .

... ... .. .

...

. ...

100
MAGNIUWM 2852 Sso
A)ENS mAEXS

A LINE
10 A______
LE G H 5

25c

c Tm

000 0 KPATH
c m00C

z
II

0260

40

z
w

II
20 0 14
10

1015
1010

1016 NUMBER DENSITY N (atoms/cm

1o7

18

Figure 32.

Theoretical Magneisum Vapor Absorption Versus Number Density

90

a temperature approximately equal to the boiling temperature of magnesium at that pressure. From this pressure and temperature the actual magnesium vapor The apparent absorption A calculated from the

density could be calculated.

recorded optical beam power level oscilloscope traces by Equation 1 was then combined with Figure 32 (the 1 centimeter path length curve) for calculation of a magnesium vapor density based on atomic absorption measurements, for comparison with the actual density. presented in Table 18. Asmay Results of a series of four tests are be seen, use of the theoretical curves of A vs. N tends to underestimate the magnesium atom density by approximately a factor of 4. Accordingly, on subsequent propellant tests the atom densities calculated by use of these curves were multiplied by a factor of 4. Early in the program, strands of four different uncured mixes (hand-tamped into strands) were tested by this technique. These formulations, designated 118, 119, 120, and 122, contained 37.5, 30, 10, and 10% magnesium respectively. The first contained 15% HTPB binder, with the remainder of the formulation being HMX. The second and third were each 10% HTPB with the remainder JINX, and the final formulation was identical with the third, except for 5% AP replacing part of the HX. The 37.5% Mg formulation gave absorption results which were difficult to read, because the beams were very highly attenuated by product particles, probably C(s) and MgO(s). Both the 10% and the 30% Mg formulations gave readable apparent absorption measurements, fluctuating considerably over any given run, and varying from run to run. The absorpt-t-ns were 18.5, 11.6, 7.0, and 9.5% for the four formulations (in numerical order). Results are presented in Table 19, As indicated, in all cases, the number density of magnesium vapor atoms were considerably less than predicted by thermochemical equilibrium calculations. Approximately 10 to 15 optical absorption tests each were conducted with Formulations Ml, MK2, and MK12 burning at 1 atmosphere. Due to problems with particulate absorption tending to flood both data channels (giving absorption readings of 100 percent) and producing large amounts of noise even when the did not flood the channels, only about half of the tests

91

.....-. ..... . . .. ... . . . .... . . . . . .... . . .. .. . . ..... .. . .. . ..

_

uj
41~

00

C;

2 00

cogo
00 41 "U "4

0% 44
4

%C0

06~

al
C.)4 12

I
II

4U

ol

IA~I0

9.4

U4

41

U,

C!

O.4

-4

.44

0 j~
93

Ado

0
Z->C
wN
--

NN

44

0j
0s

.2C

ii~~~A Li9
I.-

F.

.2 00

E-4

0J

do

zlN

(0

1--

I-

cc M0G 2

0
_ _ _

44
_ _ _ _ _ __ _ _ _ _ _

were usable and even in these the uncertainty in the apparent absorption, A, was usually on the order of 10 to 20 percent. marized in Table 20. However, it is felt that the Results are sumdata obtained are at least semi-quantitatively correct.

As may be seen, for each formulation the percentage of

theoretical magnesium yield measured averages about 10 to 12, with the highest percentage being 19. Three possible explanations for these low yields of

magnesium vapor have been suggested: (I) In the propellant combustion, the magnesium is attacked preferentially by the available oxygen, resulting in MgO(s) + C(s) rather than the Mg(g) + CO(g). (2) Heat losses and/or combustion inefficiency results in sufficient reduction of the flame temperature that an appreciable amount of the magnesium appears as magnesium nitride (thermodynamically favored at temperatures below about 1500-K). (3) The temperature at the point of optical measurement is so far below theoretical due to inefficient combustion, heat losses, or both, that condensation of much of the magnesium occurs. Examination of the propellant stoichiometry enables us to quickly rule out the first explanation inasmuch as the three formulations are so fuel rich that even if all of the oxygen thermodynamically predicted to appear as CO instead was used to tie up magnesium, it would only account for 30 to 50 percent of the magnesium. Formation of magnesium nitride, on thermochemically predicted

the other hand could, in combination with the Mg + CO - MgO + C shift, tie tip all of the magnesium in the cases of Formulations MK2 and MK12: however, in the case of Formulation MKI, even an allowance for both shifts would still result

in approximately 50 percent of the theoretical mnanesium vapor still being present. Only as the temperature is decreased to approximately llO0-1200K.

K,_
I

SI

"ul

'.

for Formulation MK1 (compared to a 1 atmosphere combustion theoretical flame temperature of 1940'K) does the magnesium level fall to 10-15 percent of that predicted for adiabatic complete combustion, the final decrease being due to magnesium condensation. Results of a series of thermochemical equilibrium calculations for .NK1, MK2, and MK12, in which the enthalpy of the system is adjusted to bring the flame temperature down (corresponding to heat losses), demonstrating the possibilities discussed above are presented in Figure 33. In

this figure, the volume percent of the gas phase which is magnesium divided by the same parameter at theoretical adiabatic conditions is plotted against temperature. As may be seen, for the ratio to drop into the 10-15 percent

band indicated by the optical absorption tests, the flame temperature must drop from the theoretical value of approximately 1950K to 1400K for Formulations MK2 and MK12, or 1100*K for Formulation MUI. In all three cases, thko cor-

responds to approximately a 75 to 80 percent loss of theoretical heat release. At this point, we are unable to conclude whether or not the observed results of our one atmosphere propellant combustion tests can be explained in this manner. In addition to characterization of the combustion of the three highly-magnesium-loaded, fuel-rich propellants by optical absorption measurements of the products of samples burning at one atmosphere, we also burned samples in a closed bomb at 500 psia (argon) and analyzed the products. apparatus used is depicted in Figure 34. The procedure is as follows: The propellant sample to be burned (approximately 10 grams) is inhibited on the sides and placed in a 2700 cc capacity modified 1oke bomb. The bomb is evacuated and flushed several times with argon and finally pressurized to 500 psia with that gas. wire. The sample is then ignited with a hot The

The combustion causes the pressure in the bomb to rise to approximately

800 - 1000 psia, dropping back to opproxtuately 550 - 600 psia as the bomb

( I
/I

96

i-'

1.0 Ntow I1 ATMOSPHERE

-MKI

gC

.............. MKI 2N2MC .............

~ 0.8
0-

I.w CK 0.6
UL.

:)

_j a 0. 0.

~Mg

CONDENSATIONSHF
_ _ _

M3N
_ _

> 0.2 0.1 Wo s Fiur o10 140 20 TEPRTR aiaino hoeia Ga-Pas Prdcswih wit Syte Temeraur 97 10 rcino r ansu 80 20

z
33

<

z
LUU
LCL

>

CL

00

cc-

<

<0

LU

-J

CLC

LU;

.juj

11

0A0

-i~

LU

cc

-1

-0-

cools.

After the bomb has been permitted to cool thoroughly (and the com-

bustion products to mix well with the argon) a gas sample is taken for chromotographic analysis. The remaining pressure is then bled off and the bomb The is connected through argon-filled lines to a modified Orsat apparatus. bomb is then filled with a 4 percent solution of HCI In water.

Tests run

with blanks show that the IHC1 solution will dissolve MgO or Mg3 N 2 with no gas evolution while it will react with any nascent magnesium to yield one gram-mole of hydrogen gas per gram-mole of magnesium. Therefore, collection

and measurement of the volume of the gas evolved as the bomb is filled with the 11CI solution (with, of course, correction for the displacement of gas initially in the bomb by the solution) permits determination of the amount of magnesium produced by the propellant combustion.

The contents of the gas sample bottles (two per test) are
analyzed for volume percent hydrogen, oxygen, nitrogen, hydrocarbon, carbon monoxide, and carbon dioxide. With these numbers and the calculated initial

gram-moles of argon in the Hoke bomb (calculated from known pressure, temperature, and volume) the gram-molesof each of these species produced by the propellant combustion can be calculated. These are then compared with theoretical predictions (based on the known propellant sample weight). of a test with Formulation MK2 are shown in Table 21. Results

As may be seen, 0.070

gm-moles of magnesium are produced, compared to a theoretical yield of 0.111 gm-moles - the percentage yield is b5 percent. Now, it is most interesting to compare the shortage of magnesium with the shortage of CO and N2 . If we assume that the shortage of these two gases is due to reaction with magnesium, we find that the shortage of CO will account for a shortage of 0.010 gm-moles of magnesium (via Mg + CO - MgO + C) while the shortage of nitrogen will account for a shortage of 0.030 gm-moles of magnesium (via 3 Mg + N2 -.Mg 3 N2 ) for a total shortage of magnesium of 0.040 gm-moles, as compared to the observed shortage of 0.041 gm-moles.

99

I j

LU 0 E
U.

.
0-

0 0
06
0.9

U)

>

n
0
LU

w
0U
90 % g,,0000 9 cc

Vo

'U

0CL 0E
'U

4 >
U) 0 I-

go

a W

0 4 I x 6 s6 dc 02
).
Z

0
H I
-

0
zL

N~+ 2C0OO
CO-0

CAIL a

mg-

4-

0lc

tuIwLCC L.

0
0
>_ J

0L
.

9 0

0-

100

Results of two tests each with MKl, MK2, and MK12 are summarized in Table 22. As may be seen, the nascent magnesium yields with MKI and

MK2 are approximately 65 percent of theoretical while that with the mixed-metal propellant, MKI2,is about 25-30 percent of theoretical. It is instructive to examine the theoretical effects of cooling of the products of combustion of the formulations by mixing with argon. (See Figure 35) At 500 psia, the theoretical flame temperature of MK2 is

2353*K, with a total CO yield of 0.227 gm-moles per 100 grams of propellant, 1.191 gm-moles of magnesium vapor per 100 grams, and 0.378 gm-moles of N 2 per 100 grams. As the products cool to 2100*K, all of the CO is thermodynamically
-

predicted to disappear via Mg + CO a second shift, 3 Mg + N 2

-

MgO + C.

As soon as the CO Is exhausted,

Mg 3N 2 is predicted to begin, with all of the rem inIt is

ing magnesium being reacted by the time the temperature reaches 1700'K.

our contention that just such shifts are occurring though kinetic limitations prevent them from going to completion, permitting only about half of the theoretical CO shift and one-third of the theoretical N

shift.

Accordingly, we

feel that this test indicates near equilibrium production of magnesium vapor during propellant combustion at 500 psia with subsequent shift of magnesium to MgO and Mg 3 N 2 during cool-down.
'K12 may also be explained in a similar manner

some of

The results of tests with MIl and (See Figure 35.) In all eases,

the experimental nascent magnesium yields, shown by the squares on the theoretical curves, correspond to freezing of the thermodynamically predicted product composition shifts with temperature at about 1900-20000 K. Carbon particles produced by the strand combustion in one of

the MKI tests were collected from the final acid solution and analyzed for size distribution: the results are presented on a log-probability plot in

Figure 36. The straight line indicates that the particles produced have a log-normal size distribution. The mass median particle size is seen to be

1.9 microns, with 80 percent of the mass being contained in particles of 5 microns diameter or less, and 90 percent in particles of 8 microns diameter or less. These small sizes indicate that there should be little trouble en-

countered in burning this carbon in an air-augmented rocket afterburner,

particularly in the presence of burning magnesium vapor.

101

I
r
.. . . . . ., ,

". . .

r. ILu

z
0

LU-

0-4o
U.,U

zS
LuM-O. t

L10

rw
co0

1%
003c 0 0~
L)

N
0to
)
44

-4

0 0

zz

4.4

a-n

4: IL

z
u

%
C

z-J4 AC

-LJ

~..

-zw

00

I.LNVI13dOkJd suwi oo t/wnIS3ND9W4 IN3:)SVN S3T0YV4449

103

-90% OF WEIGHT-IN PARTICLES LESS -THAN 8 MICRON

-DIAMETER-

--

- -

'AIAMME

80% OF WEIGHT IN PARTICLES LESS k

E

TE

IE IU

i5OFWEIGHT

RETGETRTA

HTDAEE

1 04s

o9

89

WEGTPRETGETRTAIHTDAEE
Figure~~~~~~~~~. 36 atce Colete ieDsrbuino ro 50 s BmbCobstono abo atce

E. 1.

SINGLE-PARTICLE BURNING STUDIES Magnesium A very brief experimental study of single-particle magnesium

combustion was made using the Atlantic Research Corporation flat-flame burner technique. T g - 2200 + 200*K. Particles (d . 50 um) were injected into hot burner gases, As expceted, the burning times of even such large

particles are very short; indeed, so short (<0.5 msec) that they could not be measured quantitatively. The preignition delays, on the other hand,

were relatively long, several milliseconds. This may appear surprising, since magnesium has been reported to have a very low ignition temperature, below 1000K (Ref. 15). The explanation, no doubt lies' Itr the fact that,

in this technique, cold (room temperature) magnesium is injected into hot gases, and it must melt before ignition (Tm - 924"K). Thus, in addition

to the heat-up time from room temperature to ignition temperature, the preignition delay must also include the time, t particle:
,

necessary to melt the

TOTAL

273 - 924

S m

dpMg

(924 -273) + Hmel 12 k (Tgas - 924)

(4)

where p is the particle density and k is the thermal conductivity of the gas. Substitution of realistic numbers for the various properties in the

above equation indicates that approximately 2.5 msec is required to heat a 50p magnesium particle to its melting point and melt it. 2. Magnesium Nitride As indicated earlier, it appears likely that magnesium nitride particles are produced in the fuel-rich combustion environment of the primary motor with the formulations of interest; thus, some knowledge of the

manner in which magensium nitride burns in an oxygen-containing atmosphere (e.g., a ramburner) is of interest. Accordingly, a brief study of the

ignition and combustion of magnesium nitride single particles in the Atlantic Research Corporation flat-flame burner was carried out as part

105

--

~"--' -

1
l m NII

7
am N N

"
lN

mnnm

mn

of this program.

A sample of lump magnesium nitride was purchased from

a commercial supplier and ground to a fine powder with a mortar and pestle. A nominal 35 to 40 micron cut of this material was then sieved from the ground product for study. Some of this 35 to 40 micron material appeared to

be single crystals, but much of it consisted of a number of smaller crystals attached very tightly together (so tightly that vigorous mechanical scrubbing did not break them up). It should be pointed out that the particle size

quoted is nominal, since many of the particles were far from spherical or cubic, some being long rods with aspect ratios in excess of 4 to 1. Ignition and combustion of these particles was carried out at a flat-flame burner temperature of 2350 + 100*K at oxygen levels of 5.4 to 20.8%. Ignition delay time was calculated from the distance above

the point of injection of the particle into the flame to the appearance of a sudden brightness. This ignition delay time did not appear to be

particularly sensitive to the oxygen level, being approximately 3.5 + 1.0 msec. In light of the observation that the particles generally have fairly

large aspect ratios and thus are likely to have equivalent radii of 20 to 30 microns, we performed heatup calculations for such size particles, initially at 298*K placed in a 2300*K environment. These calculations,

which neglect any reaction contributions, indicate that for an ignition time of 3.5 msec, the particle ignition temperatures must be somewhere in the range of 1100 to 1600*K. Burn time data were obtained by observation

of burning particle streaks from initial brightness to burn out, through a high speed light chopper wheel. plotted in Figure 37. Results of these observations are

As would be expected, burn time decreases strongly

with increasing oxygen content, from 6 msec at y0 2 = 0.054 to 2.7 msec

Is
-

at Y0 2 F.

0.208.

THEORETICAL EFFECT OF MAGNESIUM VAPOR ON PLUME IGNITION A theoretical examination of the effects of the presence of

magnesium vapor on the ignition of a fuel-rich exhaust plume (containing CO and H 2 as other gaseous fuels) mixing with a co-flowing air stream was

1
i 106

U!_

w
I--

~ Surroundings Temperature m 23000 K

) Surroundings Temperature ft 24S0K

10 14 PERCENT 02 IN GAS STREAM

18

22

Figure 37. Magnesium Nitride (35-40 micron) Burning Times.

I
1
107

n'!7;n
Add

carried out with aid of a modified version of the Aerochem Computer Program
for Nonequilibrium Rocket Plume Predictions (Ref. 1%. This program treats parallel turbulent mixing between concentric chemically reacting streams, with both mixing and chemical kinetic processes being treated. Outputs of

the calculations include detailed temperature and composition profiles in the mixing/combusting plume, and percent of any given fuel burned at any given axial station down the plume. In the only case examined, the primary motor propelThe

lant is a 15/35/50 HTPB/HMX/Magnesium formulation burning at 200 psia.

product stream is expanded through a primary motor nozzle to 15 psia, at which point it enters a coflowing 500 0K air stream (also at 15 psia) in a six-inch diameter afterburner with a 5-inch diameter throat. is set. An air/fuel ratio of six

Under these conditions, the total mass flow of the air plus fuel

streams is (assuming

100% afterburning efficiency) 5.1 lb-m/sec, with a priThe air velocity is 575 feet/second,

mary motor exhaust flow of 0.73 lb-m/sec.

while the fuel jet velocity (and initial diameter) is dependent on assumptions regarding the primary motor exhaust temperature, which depends on primary motor combustion inefficiencies and heat losses. For adiabatic, 100% effi-

cient combustion in the primary motor, the fuel jet exhaust temperature is 1871*K, with a resultant jet velocity of 5400 ft/sec and jet radius of 0.058 ft (0.7 inches). For a fuel jet exhaust temperature of 1200*K (a second condi-

tion examined), the jet velocity is 4350 ft/sec and the jet radius is 0.052

jft

(0.62 inches). The plume calculations have been carried out under several sets of assumptions. In all cases, the effects of condensed phases are neg-

lected (decoupled) since the Aerochem program cannot treat condensed phases. The sets of assumptions examined are: Set I: 1871K primary jet assumed. Standard kinetics used

for H 2 and CO combustion reactions. (See Table 23.) Magnesium reactions neglected.

108

TABLE 2-#.

KINETIC RATE EXPRESSIONS USED FOR HYDROGEN

AND CARBON MONOXIDE COMBUSTION REACTIONS IN

PLUME IGNITION/COMBUSTION CALCULATIONS kforward Reaction OH + H2 0 + H H 20 + H OH + H CO2 + H

=

A exp [B/RT]/T

A* 3.46 (10) 1 1 3.72 (10) 2.88 (10

10 11

N 0 0

B -5150 -16800 -9450 -1080 -18000 -1800

0

H + 02 2OH + 0 2 OH + CO

-13 9.30 (10 3 ) 9.50 (10 1.39 (10-

0 0 0 0
1

OH + OH H20 + 0 = CO2 + M M CO + 0 + H + H + M Z H2 + M O 0 + + M 02 + m H 20 + M H + OH + M

) )

5.50 (1030)

1.00 (10"3) 2.76 (10-2) 1.00 (10 4.40 (10

-3 3

1 0 1 0

0 0 0 +1000

0 + H + M 4 OH + M H + 02 + M " HO2 + M

) )

CC-molecule-sec units

i I
~109

t ' I
---

4-.=-

Set II: Set III:

Same as Set I except 1200K primary jet assumed. Same as Set I included. + 02 except magnesium reaction kinetics are

The additional reactions included are Mg(gas)

MgO(gas) + 0 and Mg(gas) + 0 + M ' MgO(gas) + M.

The heat release obtained from condensation of MgO(gas) is assumed to be released immediately upon formation of

MgO(gas).

The rate constants for these two additional

32

reactions were estimated to be k = 10-10 exp [-32000/RT] and k = 10 , respectively (in cc-molecule-second units)

Set IV:

Same as Set III except data of Kashireninov (Ref. 17) are used to replace the rate constant for Mg(gas) + 02
MgO(g) + O(g) by k = 3.5 (10 " 10) exp [-160001RT]

Set V:

Same as Set IV except condensation of the MgO product is neglected (thereby removing the oxide condensation heat effect).

The results of the plume calculations, in terms of percent of potential H 2 0 formed (representative of degree of completion of burning of the plume) versus distance from the plume origin are plotted in Figure 38. Curve I, at 1871K primary jet conditions there is no As may be seen from

noticeable ignition When the

delay, with the plume combustion being totally mixing limited.

assumed initial jet temperature is dropped to 1200K, however, there is an appreciable plume ignition delay, approximately 0.5 - 0.6 feet in the absence of magnesium reactions (Curve II). Allowing magnesium reactions, as controlled

by the estimated kinetic parameters (Curve III) significantly decreases the ignition delay length to about 0.3 - 0.4 feet, while use of the kinetic data of Kashireninov (Ref. 17), essentially eliminates the ignition delay, as shown by Curves IV and V. (Note that Curves IV and V are not coincident with Curve

I since the mixing rates are different for the 1200K jet and the 1871*K jet.) Thus, it appears that the presence of magnesium vapor can be quite beneficial to ignition of hydrogen and carbon monoxide in a fuel-rich plume expanding

I
i110

IC_

CC

Ofa
0 c-

Ea

U, U4

w -Maew
W1U

A 400(

m~~

wp x

o.I aMa a
( m -

q
i

Za

1 .41

4 9

u p a~ a= a In 03V4U04
<

a a N d N~ ZV3Ld1JOJ a

CAH

111LItAC

----------c; qe o7

into coflowing air.

Examination of detailed concentration profiles indicates,

as expected, that this enhancement is due to the production of oxygen atoms by the first magnesium reaction, which is quite rapid compared to the consumption of oxygen atoms by the three-body second reaction. G. 1. COMBUSTION OF SELECTED CANDIDATES IN MOTORS Castable Composition Motor tests were conducted with formulations MKl, MK2, and MK12 for evaluation of their burning rate versus pressure characteristics in motors (as compared to strands), evaluation of the tendency of the exhaust streams to despoit in and plug nozzles, and evaluation of the efficiency of ejection of the combustion products (as measured in terms of percent residue left in the motors). The burning rate versus pressure data have

already been presented in Section Ill-C; reasonable agreement of the motor data with strand data taken in the acoustic emissions strand-burning apparatus was obtained. One motor case was employed for all motor tests despite a strong variation in grain sizes. are presented in Figure 39. Details of the primary motor assembly

The motor case, designed to withstand 3000

psi operating pressure, is 9.875 inches I.D. with a length of 20.75 inches from the inside of the head wall to the outer edge of the flange, where it mates with the integral blast tube case. A 2-inch diameter blowout

port and an igniter wire plug port were located near the aft end of this case, as shown. In addition, two static pressure taps (not shown) were A mixing aid was

included for measurement of primary motor pressure.

attached to the down stream end of the blast tube in some of the tests. Both end-burning and center perforated grains were evaluated in tests. Grain size was varied by using inert spacers and volume limiters In some of these tests (referred to as

of various internal dimensions.

primary motor tests), the product stream was simply exhausted to the atmosphere, while in others (referred to as connected-pipe tests), the products were fed to an afterburner for evaluation of their afterburning

112

I
i~!V -

Go
C 0

00C eg.

U
N

0 *1
-

N
-N
I

-m

.2'

Cc

. LL LiI

ED 6 CL I.. 0)

13

IC

characteristics, to be described in a later section.

A total of six motor Grain

tests were conducted with MKl, one with MK2, and three with MKI2. and nozzle sizes for these tests are given in Table 24.

In Test Number IP,

the graphite free-volume limiter was uncoated, providing a severe heat sink to the mateiial flowing through it; in all subsequent tests, it was coated with an insulator, to minimize this problem. Throat closure and residue

level data for all tests are presented in Table 25 and Figures 40 and 41. As may be seen, severe throat closure problems occurred for small graphite nozzles (0.25 to 0.40 inch diemater), probably due to plating out of magnesium/magnesium oxide/magnesium nitride eutectics on these nozzles. Such throat closure would obviously present a severe However, it should be pointed out, that for

problem to a motor designer.

practical systems, desired fuel flow rates will be such that for singlenozzle primary motors, considerably larger nozzles will be used. The two

data points on the right of Figure 40, which correspond to an initial throat diameter of approximately 0.8 inches indicate that the throat closure problem rapidly disappears with increasing nozzle size. If the designer

finds that he needs to use multiple nozzles to distribute fuel in a ramburner, and thus finds himself back in the less than 1/2-inch diameter nozzle throat range, substitution of asbestos-phenolic nozzles for graphite nozzles appears to provide another acceptable solution, as shown in Figure 40. Minimizing residue left in the primary motor is obviously important in providing a useful fuel generator system, since the sensible and possibly chemical enthalpy associated with this residue is not available for utilization in the ramburner, with effective degradation of the heating value of the fuel. Table 25 and Figure 41 indicate that unacceptably high

residue levels were obtained in some tests, while quite low levels (10% or less) were obtained in others. Two factors appear to be possible contributors First,

to the differences in residue levels obtained in different tests.

it may be seen that all tests with cylindrically perforated grains, burning in the central bore, had low residue levels. It appears reasonable that

114

.4

A~Aa A~

14

.44

.4

aa

a4

CU

~U~

008A0

*a
04

.4
A A0 0..0 0.0. 0 in .r6 .1 W
-4

.
00A 0:

0.

.00 OR 00 nCCC W !. N
i

a,~ .4

,4*

V4
LSO

1f4

ef

0SO

64 SO ..

16

61b

:10
6.4 A
o. " e

*'

4
o

A.
. 115*

on

110 40

a.

4'4.

1i V

@ ,

94 .4

1
4j

S"4:

di
I I16

111

'
t | ll_

Ia
4

wz uz

IL

C00

W4

3V

VkiNI

N&.13W

N3H

117

0
t4

00

co

-4 N

z
404 tom
wn
-

0
U IAJ I3
-

Nlie

cylindrically perforated grains should be less susceptible to residue formation than end-burning grains, since residues generally form as a growing ash structure on the burning surface which builds up by trapping carbon, magnesium oxide, and other such condensed-phase products as they are produced in the combustion zone: with cylindrically perforated grains, the sweeping action of gases moving through the grain port would be expected to tend to continuously break up this ash structure before it can form a hard clinker which cannot be dislodged. However, it is not unequivocally clear that the dominant factor in determining the residue level is grain configuration, inasmuch as the cylindrically-perforated grain tests were all run at comparatively low pressure (100 to 150 psia), while most of the end-burner tests were run at somewhat higher pressure (2 500 psia). Examination of Figure 41, In in which residue level is plotted against average pressure during the test indicates a fair degree of correlation of residue level with pressure. a residue level of only 5.6%. level of 36%.) fact, one end-burner test conducted with MK2 at low pressure resulted in

(end-burning

(However, another test with MKI2 in an

configuration at nearly the same pressure yielded a residue Thus, it is not clear whether grain configuration or

operating pressure level are dominant in determination of ejection efficiency, or whether (most likely) both play important roles in this area. A series of tests was conducted in which strands of MK1 were burned at various pressures from 200 to 2200 psia, to see whether there was a critical pressure level at which residue began to form (though it is recognized that this level may be quite different for strands than for motors). Up to 1200 psia, these strands (1-inch long, 1/2-inch diameter cylinders cast in epoxy tubes) ejected completely when burned. At the

next pressure level tested, 1500 psia, approximately 28% of the propellant mass was left in the epoxy tube, while at 2000 and 2200 psia, the residue levels were 40 and 47% respectively. Thus, there does appear to be a pressure effect on residue formation, at least for MK1.

A pressure effect on formation of residue, residue increasing with increasing pressure, for end-burning grains in particular, is not

119

unreasonable on theoretical grounds.

If we assume that the flow of gases

from a burning fuel-rich propellant past initiating residue clusters on the surface is such that the gas drag on the "particles" is in the Stokes Law flow regime, then the force tending to pull these clusters away from the surface is proportional to the gas velocity away from the surface, as indicated by Equations 5 through 7.

Drag Force

AclusterCD

(5) 24W

CD - drag

coefficient = 24/Reynolds Number -

Dcluster

(6)

Drag Force -

pV2A 2 cluster

24p Dcluster V

= kV

(7)

The velocity of gases away from the surface depends upon pressure through several factors: (1) (2) (3) Effect of pressure on gas density. Effect of pressure on propellant burning rate. Effect of pressure on moles of gas produced per gram of propellant. A series of calculations has been performed for the velocity of gas flow away from the surface of MKI, MK2, and MK12 as a function of pressure under assumptions of adiabatic combustion, 20% heat loss, and 40% heat loss: results are presented in Figure 42. These calculations were performed

using thermochemical equilibrium calculations to calculate the gas molar

yield under the three different assumptions regarding heat loss, and using measured burning rate versus pressure data. As may be seen, gas velocity

away from the surface decreases strongly with increasing pressure, particularly in the 100 to 500 psia range of interest, lending strong support to the idea that pressure is a critical facator in the formation of residue.

1
5120
I
. . ... d

*'--. .

..

40 V r (in/sec)] Gm-MOLES OF GAS PER [10.73 T (R) I 12J Prop) [GRAM OF PROPELLANT[ P(psia)

jf/e

f/

32
w

Z 24
MK2 MK2 0
IL

3:

ADIABATIC

- 0

200

400 600 PRESSURE (psi&)

g0o

1000

Figure 42.

Velocity of Gases Away from Surface of Burning Propellant Versus Pressure

121

An alternate possible explanation for increase in residue with increasing pressure is demonstrated in Figure 43. A series of thermo-

chemical calculations were run for MKl with the system enthalpy being fixed at various levels, representing different levels of heat loss (either due to combustion inefficiencies, or losses to the massive hardware used). These calculations permitted definition of the percent heat loss level which would result in the appearance of condensed-phase (liquid) magnesium in the combustion products. These calculations were carried out at various As may be seen, the perc'nt heat decreases with increasing

pressure levels from 100 to 1000 psia.

loss required for appearance of magnesium liquid pressure.

Since the appearance of liquid magnesium (in combination with

carbon and solid magnesium oxide) may be a key to producing a clinker structure, this mechanism may lead to increased residue with increasing pressure. Analyses of residue from one test with MK1 (Test IP) and one test with MKl2 (Test 5P) were carried out to determine what type of material (particularly in terms of potential heating value) was being left in the primary motor in tests yielding appreciable amounts of residue. With the

MKI propellant, the residue was found to be approximately 22% carbon and

(78%

magnesium oxide, with essentially no unreacted magnesium being present. Comparison with the amounts of carbon and magnesium oxide which should have been produced thermodynamically by the motor indicates that about 70% of the carbon product and 75% of the magnesium oxide product remained in the motor. The chemical enthalpy available in the carbon left behind

represents approximately 11% of the total theoretical heating value of MK1. In addition, sensible heat, which would have contributed to heating of air in a ramburner, was lost. With the MK12 propellant, the residue was approximately 27% carbon, 12% MgO, 18% AI 2 0 3 (with perhaps some combination of the MgO and A1 2 0 3 to MgAl 2 0 4 , spinel), 32% aluminum, and 11% miscellaneous (probably nitrides and carbides). Thus, approximately half the carbon

and nearly all of the aluminum remained in the residue, with only a small

I
i 122

!)

---m mm -mmmmmm mmwmmlmm mmmm mmm

I
9
100

80

I
I

z u

60 I
I

40

! w
* w U
IL

20

200

400

600

800

1000

I
Figure 43.

PRESSURE (psia)

I
1

Alternate Possible Explanation for Increase in Residue Level with Increasing Pressure

123

I
l~a

fraction of the magnesium being left behind.

The carbon and magnesium in

the slag represent approximately 30% of the potential heating value of MKl2. 2. Pressed Compositions Two TKI propellant grains were fired in motor tests during this program, one in a test in which the exhaust was simply dumped into the atmosphere and one in which the exhaust was burned in a ramburner in a connected-pipe air facility test. The grains employed were end-burners,

2.M' inches in diameter and approximatley 7 inches long, weighing approximately 1.5 pounds each. Both were designed to operate at about 200 to 300

psia, with a planned burn time of approximately 10 seconds, and a fuelflow rate of 0.15 pounds/second, using an asbestos-phenolic nozzle of 0.25 inch diameter. Unfortunately, it appears that in both tests, the propellant either sideburned (liner separation from the propellant), or cracked upon ignition, leading to much higher than planned surface area early in the test (with consequent attainment of much higher pressure and mass expulsion

rate early in the test) with subsequent rapid decay of pressure and mass expulsion rate as the propellant consumed itself. In the first test, the

residue level was approximately 13%, while in the second (connected--pipe air facility) test, the residue was 10%. Pressure-time traces for the first The propellant continued

8 seconds of each test are shown in Figure 44.

to eject at a very low level for several seconds beyond this time period. Depletion of funds prevented further examination of the problem and attempts to avoid it. However, it should be pointed out that a formulation of

similar composition (slightly lower heating value) was successfully fired in two tests on an Internal Research and Development Program: accordingly,

it is felt that further effort would result in successful firing of motors using the TK1 formulation. H. 1. CONNECTED-PIPE AFTERBURNER TESTS Castable Compositions A total of five connected-pipe air-augmented rocket tests

124

I
#r

00

40 v
-

40C'

IL(olv) 3wSS3W

125

were carried out with the castable Mg-loaded fuel-ri,, prcpellants formulated during this program. These tests were conducted at test conditions highly

adverse to efficient afterburning (low pressure, high air/fuel ratio), conditions under which state-of-the-art boron formulations do not burn well. Four of these test were conducted with MKl grains (Tests 401, 401A, 402, 451) and one with an MK12 grain (Test 403). Grain sizes for these tests

were presented earlier in Section III-G, as was a descriotion of the primary motor hardware. The fuel-rich solid propellant motors of MKl and MK12 were fired into a secondary duct where the primary rocket exhuast mixed and burned with air injected coaxially from a high-pressure biowdown facility. (Desired

air temperature was achieved by passage of part of the air through a pebblebed heat exchanger.) Static pressures in the secondary duct just upstream

of the afterburner nozzle were measured and used in combination with measured air flow rate and total temperature and propellant flow rate (determinable from metering chamber measurements of pressure as described later) to determine the efficiency of mixing and combustion in the afterburner. In addition,

force balance measurements were made and used with the air and propellant mass flow rates for an independent determination of afterburning efficiency. Afterburner tube outside wall temperature measurements were also made to permit correction of efficiency data for heat loss to the uninsulated nickel afterburner tube used in this study. A schematic drawing of the test hardware employed for the connected pipe tests is presented as Figure 45. As shown, the entire test

hardware assembly was suspended in pendulum fashion from a knuckle joint near the top end of the hot air downpipe, the knuckle joint permitting the entire test assembly to swing in one dimension through up to 1/2 degree of arc. Fastened to the end of this downpipe was a plenum, open at both ends.

The fully assembled primary motor (including metering chamber and mixing/ flameholding aid) bolted to the upstream side of the plenum. The boss on

the front end of this motor was sized such that a load cell fastened to it just contacted a stationary thrust backstop when the downpipe was exactly

1
~126

0.

06

'

aG ma

-e-

3c

.u

Ga

0-0

LL'' .0
wc

00
0 o

.0.a
C-J

C127

vertical.

This thrust backstop included a thrust plate mounted on a roller

bearing to permit vertical movement of the entire assembly resulting from thermal growth of the downpipe as it heated up. The system was so balanced

that in the absence of operation of the motor or ejector/diffuser, the above mentioned load cell received a positive compressive load. However, with

the ejector/diffuser operating and the primary motor not operating (the ejector was activated several seconds before ignition), the ramburner base pressure reduction was sufficient to result in a net force away from the stationary mount; consequently, in tests where the ejector/diffuser was employed, a "preload" load cell operating in tension was added to the system. As may be seen, fuel generator (primary motor) products exhausted from the fuel generator through a blast tube which carried them to a throat where they passed through the mixing/flameholding aid into a coaxially flowing air stream. (Air from the downpipe was passed through an air guide providing The two streams were then

a uniform coaxial flow around the blast tube.)

mixed and burned in an afterburner tube and exhausted through a sonic nozzle. Since the afterburner operating pressure in these tests was sufficiently low to cause unchoking of the afterburner for an ambient pressure downstream of the nozzle of 14.7 psia, the ejector/diffuser was used in these tests to lower the base pressure sufficiently to insure choked afterburner operation. ThL jector, which was mounted on the ground, was separated from this

assembly by a small gap. The outside diameter of the blast tube, extending through the plenum into the ramburner, was 2.0 inches. This tube was constructed A silica-phenolic

of stainless steel with a wall thickness of 3/16 inches.

tube of 1/4-inch wall thickness was slipped inside this tube to thermally protect it from the hot primary motor exhuast products. The length of this

blast tube from the downstream side of the primary motor nozzle to the upstream edge of the metering chamber nozzle was 8.75 inches. The metering

chamber nozzle, located at the end of the blast tube was machined from a 1.625 inch O.D., 1.125-inch long graphite cylinder. This nozzle insert

was machined to taper smoothly from the silica-phenolic tube inside diameter

1
~128

1.125 inches) to the desired throat diameter with no further expansion (sonic throat). In all tests except Test 451, a mixing/flameholding aid,

described later in more detail was attached to the end of the blast tube. As mentioned earlier, air of the desired temperature (nominally 900'R) was brought into the system by means of the downpipe from which the entire test system was suspended. This pipe was precisely

at right angles to the ramburner axis and the system pivot point was on the centerline of this pipe, with the result that thrust generated by the system did not have to be corrected for air inlet momentum in calculation of system performance. The air leaving the downpipe entered an annular section From this

around a tube with the same inside diameter as the ramburner.

annulus it passed through a high resistance (large pressure drop) perforated grid, turning into the annular section between the above-mentioned tube and the blast tube. This annular section of constant cross-sectional area

extended 8 inches to the end of the blast tube and an additional 2-1/2 inches to the spoiler disk on the mixing/flameholding aid. A picture of this

aid, which screwed onto the end of the blast tube, just downstream of the metering chamber nozzle, is shown in Figure 46. This aid was a copy of

the one developed by Marquardt Corporation in slurry ramjet testing and used previously by Atlantic Research with beneficial results. The barrel and

wedge converted the primary exhaust to a subsonic stream and diverted it away from the centerline into the coaxially flowing air. The serrated

disk spoiled the airstream near the center of the ramburner, causing it to recirculate (entraining at least part of the primary exhaust) and thereby serving as a piloting flameholder. The outer diameter of the serrated disk of approximately 25% of the air flow

was 3.75 inches, resulting in blockage passage around the blast tube.

In the aforementioned previous Atlantic

Research experience with this aid, difficulty in avoiding burning up of the aid was encountered. In that work a stainless steel barrel was used, with Accordingly, for the current

failuregenerally occurring in this region.

program, the barrel was made of molybdenum and the wedge of silver-impregnated tungsten.

129

Figure 46.

Mixing/Flazneholding Aid used in this Studv

130

The ramburner chamber itself, which was attached to the dowrstream side of the plenum by means of Marmon flanges and a Marmon clamp, was a 43.5-inch long, 5.89-inch I.D., 0.33-inch thick nickel pipe. The

nozzle housing for the ramburner nozzle was constructed of stainless steel, with the nozzle insert being made of graphite. The overall length of the

nozzle package (which fastened to the ramburner tube with a Marmon clamp arrangement) was 3.4 inches. inches. The nominal nozzle throat diameter was 4.62

The graphite insert tapered from the ramburner tube inside diameter

to this throat diameter in approximately 2 inches, the remainder of the nozzle being constant area (sonic nozzle). The end of the mixing/flameholding aid extended approximately 3.6 inches into the ramburner tube. Thus, the effective length of the ram-

burner was approximately 40 inches and the L* was 65 inches (nominal residence time of 5 milliseconds). During each test, up to 48 channels of data were continuously measured and recorded on three Honeywell Model 1508 Visicorders with M-100-350 Galvanometers providing flat frequency response up to 60 cycles per second. These data were read from the visicorder traces at intervals selected as needed throughout each test by means of a Data Scaler Model 400. types of measuring devices were employed: Two basic

(1) strain gauges (load cells and Four-step calibrations were

pressure transducers); and (2) thermocouples.

used to convert strain-gauge measurements to engineering unit data, while manufacturers calibrations and/or standard thermocouple EMF tables were used to obtain engineering unit data from the thermocouple traces. The

quantities measured continuously during each test are listed in Table 26. In addition, the following pretest and post-test measurements were made: (I) (2) (3) Initial propellant weight. Residue weight. Initial metering chamber throat diameter.

(4) Final metering chamber throat diameter. (5) Initial ramburner throat diameter.

(6) Final ramburner throat diameter.

131

TABLE 26.

PARAMETERS MEASURED CONTINUOUSLY DURING EACH CONNECTED-PIPE RAMBURNER TEST.

A. B. C. D. E. F. G. H. I. J. K. L. M. N. 0. P. Q. R. S. T.

Static Pressure Upstream of Air Mass Flow Venturi Total Temperature Upstream of Air Mass Fio., Venturi Air Total Temperature At Plenum Primary Motor Static Pressure Primary Motor Total Temperature Metering Chamber Static Pressure Metering Chamber Total Temperature Compressive force on Primary Load Cell Tension Force on Preload Load Cell Pressure on Base of Ramburner Nozzle Static Wall Pressure in Ramburner 2" Upstream of Nozzle Entrance Static Wall Pressure in Ramburner 10" Upstream of Nozzle Entrance Static Wall Pressure in Ramburner 18" Upstream of Nozzle Entrance Static Wall Pressure in Ramburner 26" Upstream of Nozzle Entrance Static Wall Pressure in Ramburner 34" Upstream of Nozzle Entrance Outside Ramburner Wall Temperature 4" Upstream of Nozzle Entrance Outside Ramburner Wall Temperature 12" Upstream of Nozzle Entrance Outside Ramburner Wall Temperature 20" Upstream of Nozzle Entrance Outside Ramburner Wall Temperature 28" Upstream of Nozzle Entrance Outside Ramburner Wall Temperature 36" Upstream of Nozzle Entrance

2 channels 2 channels 2 channels 2 channels 0 to 2 channels 2 channels 0 to 2 channels 2 channels 2 channels 4 channels 4 channels 1 or 2 channels 2 channels 2 channels 2 channels Up to 4 channels Up to 2 channels Up to 2 channels Up to 2 channels Up to 4 channels

132

(7) (8)

Ambient pressure. Air total temperature at the aft end of the ramburner (prior to ignition).

Measurements of pressure and temperature upstream of the air mass flow venturi were used straightforwardly for calculation of instantaneous air mass flow rate throughout each test. Initial and final propellant

weight, initial and final metering chamber throat diameter and primary motor pressure versus time data were used to calculate instantaneous fuel flow rate, assuming a primary motor C* independent of time and assuming a linear variation of throat area with time from initial area to final area. Once the fuel flow rate and air flow rate were determined, they were used in combination with the measured air total temperature at the plenum, the measured ramburner static pressure, and a thermodynamic equilibrium computer program (or tabular lookup) for determination of theoretical ramburner flame temperature, polytropic constant for expansion of the mixed, burned stream from the ramburner to the ramburner nozzle throat, and throat molecular weight. The fuel and air mass flow rates were

then used in combination with the static wall pressure just upstream of the ramburner nozzle, the nozzle throat area, an assumed or calibrated nozzle discharge coefficient (representing the deviation of mass flow rate from that to be expected in the absence of boundary layer effects) and the aforementioned polytropic constant and throat molecular weight to determine actual ramburner product total temperature. Alternatively, a force balance,

using the force data from the two load cells, the measured pressure on the ramburner nozzle base, and ambient pressure, could be used to calculate the vacuum sonic thrust generated. From this quantity, along with the nozzle

area, discharge coefficient, polytropic constant, and throat molecular weight, an alternate calculation of actual total temperature in the ramburner could be made. In general it was found that use of static wall pressure

data at the end of the ramburner was more reliable than the force balance procedure.

133

The ramburner outside wall temperature versus time data was used with the known thermal capacitance properties of the ramburner to estimate the rate of heat loss from the stream flowing through the ramburner at any given time. This heat loss rate was then used in combination with the total mass flow rate through the ramburner and the thermodynamically calculated heat capacity of the burned air/fuel mixture to calculate a stream temperature loss due to heat transfer to the ramburner walls, and this latter quantity was then added to the calculated actual total temperature to yield a "delivered adiabatic ramburner stream total temperature." Finally, this temperature was combined with the theoretical ramburner stream total temperature and the-measured air temperature to yield a temperaturerise efficiency. The equations used in this data reduction procedure were:
WDOTA = Kventuri Pventuri /iT/ venturi (8)

WDOTF

Pmc Amc (Propellant Weight - Residue Weight) IP A de

mc mc

TT4T - f (WDOTF/WDOTA, P4, TTAIR)

f

(10) (11) (12)

2 2 y

Y MW
y(MW)g TT4P

f2 (WDOTF/WDOTA, P4, TTAIR) f3 (WDOTF/WDOTA, P4, TTAIR)

+)/(Y-1) P4 A 2e2(P4)2 WDOTF)2 T+ -1

-Y

j(M4)

(13)

M4

1 +

(M4)21

j(y+l)/2(y

-1)

(14)

Aramburner

12

134

'I7

de

dTramburner backside avg (Ramburner Mass) (C p,ramburner) dO

(15)

(TT4P)

(TT4P~corrected

TT4P +

dQ/de (WDOTA + WDOTF)C p, product stream

(16)

(TT4P)corrected - TTAIR TT4T - TTair where: WDOTA - Air Mass Flow Rate WDOTF - Fuel Mass Flow Rate Kventuri - Venturi Constant PVenturi Pressure upstream of flow venturi

(17)

Tventuri - Temperature at flow venturi Pmc - Metering chamber throat area

TT4T

= -

Theoretical total temperature for 100% efficient

afterburning Theoretical specific heat ratio for 100% efficient afterburning

MW = Theoretical product molecular weight for 100% efficient afterburning P4 - Static pressure at entrance to afterburner nozzle contraction section TTAIR - Air total pressure
= Functional relationships calculated using equilibrium

thermochemistry program for the propellant being tested

TT4P - Total temperature at end of ramburner based on pressure measured there

W Gravitational constant R - Universal gas law constant At - Ramburner throat area

Ke - Ramburner discharge coefficient

135

',

M4 = Mach number at entrance to afterburner nozzle contraction section A Ramburner cross-sectional area dQ/dO = rate of heat loss to hardware (TT4P) corrected - TT4P corrected for heat loss to hardware TAT - Temperature-rise efficiency. In all tests, the air total temperature was approximately 900*R, the afterburner residence time was 5 to 6 milliseconds, and the afterburner length/diameter ratio was approximately 8. Afterburner operating

pressures ranged from about 20 psia at the higher air/fuel ratios to 30 to 40 psia at the lower air/fuel ratios tested. Efficl .:y results for three

test with MKl employing the PLAAR mixing/flameholding aid are presented in Figure 47, while efficiency results for an MK1 holding aid appear in Figure 48. test with no mixing/flame-

As may be seen, the afterburning efficiencies

in general were quite good, generally in excess of 90% with use of the PLAAR aid, and in the 82 to 95% range without it. It is of interest to compare the three tests with MKl with a PLAAR aid to the one test without it, since this is the only formulation which was tested both with and without the aid. It appears that the aid is

worth approximately 10 percentage points in terms of temperature-rise efficiency, though obviously at the expense of increased combustor drag (not measured in these tests). It appears likely that the relatively low

efficiency without the aid Is due to mixing limitations rather than flameholding considerations. The primary motor nozzle exit diameter for this thus, there This

test was 0.833 inches, while the ramburner was 45 inches long;

were only about 54 primary nozzle diameters available for mixing.

is probably not sufficient mixing length, since the relatively low overall air/fuel ratio prevailing through most of this test requires nearly total mixing to get sufficient air to all of the fuel. During the test, a very

bright flame was observed in the gap between ramburner and ejector, indicating the mixing/combustion was still taking place. In tests with the PLAAR

136

0
aI

-2

('4 C44

wp

00Q

a
@10-

00
I
0

IAAn

El

0L. ZU

E3

<

980'1~~~~=

.IJPLJ0D-()4L4PJ~~TI

4OH .n2d

El

C137

<IA
,-.-----.

IL In wN IL 0 00 + 4

oc II

0I CL-

.-

cn

881 01Pl~lOO 8H

DOTT93aS1 anjid0j

138

Ada~-

aid, no such flame was seen, indicating, in combination with high efficiency numbers, that mixing/combustion was complete. If the efficiency increase observed in going from no aid

to the PLAAR aid is indeed due simply to better mixing, rather than to recirculatory flameholding, this suggests that such performance increase might also be obtained simply through the use of multiple nozzles in cases where a single primary nozzle would lead to relatively low values of afterburner length/primary nozzle diameter ratio. rather than Such use of multiple nozzles

the PLAAR aid should be considerably less detrimental in terms

of primary stream momentum degradation and combuqtor drag. Results of a single test with a second castable composition, MKl2, are presented in Figure 49. Test conditions employed in this test

were particularly severe (afterburner pressure of 16 to 18 psia, air/fuel ratio of 28/1 to 39/1) conditions under which state-of-the-art boron formuAs may be

lations yield temperature rise efficiences of 40 to 60 percent.

seen, this formulation, which has considerably higher theoretical performance than MKI due to use of 2/1 magnesium/aluminum, rather than all magnesium as the metal additive, performed quite well, yielding temperature rise efficiencies in excess of 90 percent at air/fuel ratios below 30/1, although this efficiency did drop off to about 80 percent at an air/fuel ratio of 37/1. The main conclusion from this test series is that the castable magnesium propellants formulated in this program generally afterburn well under adverse conditions (low pressure, high air/fuel ratio) even without a high drag mixing/flameholding aid. It appears that mLxing

limitations can cause some performance degradation at near stoichiometric conuitions for relatively low values of afterburner length/primary motor nozzle diameter: a possible low-drag solution to this problem is the use

of multiple nozzles under these conditions. The probable reason for the excellent afterburning characteristics of these formulations, even under highly adverse conditions, has been alluded to earlier. These formulations have been shown through

139

0)

0 o

0 0

<I

IL'

0 CM

00

SscYI lvaH

I0J

Pal11OI0

M%

J(~u9TD3JJ

asTH ainipjduia

140

atomic absorption experiments to produce large amounts of magnesium vapor in the exhaust streams of the primary motors. This magnesium vapor, as

shown in the theoretical analysis presented in Section Ill-F, reacts extremely molecules to produce 0 atoms, which are violently reactive, 2 promoting vigorous ignition and combustion with all fuel species which they rapidly with 0 ,ontact. Thus, as rapidly as the fuel stream can be mixed with air, the

Mg vapor reacts leading to vigorous combustion of the fuel stream and the system is essentially mixing limited. The high afterburning efficiency of the MK12 formulation is particularly gratifying in that it has, by a considerable margin, the highest theoretical performance of the castable composttions examined (heating value of nearly 10,000 BTU/lb or 600 BTU/in 3 ) as shown in Table 7. 2. Pressed Compositions As indicated earlier, one connected-pipe air-augmented rocket test (Test 460) was performed with the pressed composition, TKI. The same equipment and data reduction procedures as described above were used in analysis of this test. However, as a result of the unusual burning

behavior of the primary motor (See Figure 44.), the afterburning performance data obtained with this test is somewhat suspect. described previously was employed in this test. The PLAAR mixing aid Air tota7 temperature was

again 900*R and the afterburner operating pressure was approximately 15 to 25 psia, again representing highly adverse condit'ions as regards efficient afterburning. Observation of the pressure-time trace of Figure 44 indicates

that only efficiency data obtained in the first 0.5 to 0.6 seconds of the test are likely to have any validity, and that even these data points are open to question in that it is difficult to accurately calculate instantaneous fuel-flow rates from such a pressure-time curve, and also in that the heat losses to the hardware are likely to be larger than calculated by the data reduction procedure described above during this initial hardware heatsoak period. Nevertheless, efficiency data for the first 0.6 seconds of the test are presented in Figure 50. As may be seen, temperature rise efficien-

cies of 60 to 80% are calculated (even with the conservative assumptions

141

01O IVHoia-3Nl:
____ J~d ___1131 3SIN _ Y4.

Inja

142

regarding heat losses to the uninsulated ramburner). In addition, one connected-pipe ramburner test was run with a formulation quite similar to TKl (same ingredients, slightly different

proportions, approximately 2% lower theoretical heating value) on an Internal Research and Development program. This test did not encount-: he

same primary motor interior ballistics problems and yielded meaningful data over an air/fuel ratio range of approximately 3/1 to 7/1. Air total

temperature was approximately 900OR once again, but afterburner opt-rating pressure was slightly higher (30 to 40 psia) and no mixlng/flameho]ding aid was used (two factors which should tend to compensate). In this test,

temperature rise efficiency, corrected for heat losses to the hardware, ringed from approximately 80% at air/fuel = 7/1 to 90 to 100% at air/fuel
=

31.

Thus, it appears that pressed compositions in the TKI class 'ffer considerable potential for air-augmented rocket applications, having theoretical heating values in the range of 10,000 BTU/lb (750 BTU/in 3 ) and offering the promise of very high afterburning efficiencies under adverse ,onditions (relatively low air temperature, low afterburner pressure) typical of high altitude cruise applications where boron-loaded forumlatLions encounter afterburning difficulties.

I I I I I
I 143

I
... .. . . .. -

REFERENCES

1.

King, M.K., and Ma~ek, A., "Studies of the Ignition and Combustion of Boron Particles for Air-Augmented Rocket Applications," Interim Report AFOSR-TR-75-0043, October, 1974. Friedman, R., and Matek, A., "Ignition and Combustion of Aluminum Particles in Hot Ambient Gases," Combustion and Flame, 6, 1, pp. 9-19 (1961). Friedman, R., and Matek, A., "Combustion Studies of Single Aluminum Particles," 9th Symposium (International) on Combustion, Academic Press, N.Y., pp. 703-712 (1963). Friedman, R., Ma~ek, A., and Semple, J., Heterogeneous Combustion, Academic Press, N.Y., p. 3 (1964). Ma~ek, A., "Fundamentals of Combustion of Single Aluminum and Beryllium Particles," lth Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, Pa., pp. 203-217, (1967) Macek, A. and Semple, J., "Experimental Burning Rates and Combustion Mechanisms of Single Beryllium Particles," 12th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, Pa. pp. 71-81 (1969) Ma~ek, A., and Semple, J., "Combustion of Boron Particles at Atmospheric Pressure," Combustion Science and Technology, 1, pp. 181-191 (1969). Mavek, A., "Combustion of Boron Particles: Experiment and Theory," 14th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, Pa., pp. 1401-1411 (1973). Prentice, J.L., et al., Metal Particle Combustion Progress Report, Naval Weapons Center Report TP4435, Appendix A, p. 106 (1968). Mohan, G., and Williams, F., "Ignition and Combustion of Boron in O2!nert Atmospheres," AIAAJ, 10, 6, pp. 776-83, (1972). King, M. K., "Boron Ignition and Combustion in Air-Augmented Rocket AfterBurners," Comb. Science and Technology, 5, pp. 155-64 (1972). King, M. K., "Boron Particle Ignition in Hot Gas Streams," Comb. Science and Technology, 8, 6, p. 255 (1974). King, M. K., "Predictions of Laminar Flame Speeds in Boron-Oxygen-Nitrogen Dust Clouds," 15th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, Pa., 1975. Macek, A., Goede, P., Scheffee, R.S., and Zukovich, P.P., Final Report, Contract No. F30u02-73-C-0313 (ARPA through RADC), March, 1974.

2.

3.

4.

5.

6.

7.

B.

9.

10.

11.

12.

13.

14.

144

-,

.,.

15.

Cassel, H., and Liebmann, 1., "The Cooperative Mechanism in the Ignition of Dust Dispersion," Combustion and Flame, 3, 4, December 1959, pp 467-75. Mikatarian, R.R, Kau, C.J., and Pergament, H.S., A Fast Computer Program for Nonequilibrium Rocket Plume Predictions, AFRPL-TR-72-94, August 1972. Kashireninov, 0., Oral presentation at AIAA 13th Aerospace Sciences Meeting, Pasadena, CA, January 1975.

16.

17.

145

-JI

CHRONOLOGICAL BIBLIOGRAPHY OF PAPERS AND REPORTS RESULTING FROM CONTRACT F44620-71-C-0124

Macek, A., and Semple, J.M., "Composition and Combustion Characteristics of Condensed Exhaust From Boron-Containing Fuel-Rich Rocket Motors," 9th JANNAF Combustion Meeting, Monterey, Calif., Sept., 1972, CPIA Publication 231, Volume I, p. 359, Dec., 1972. King, M.K., "Boron Particle Ignition in Hot Gas Streams," Combustion Science and Technology, 8, pp. 255-273, 1974. King, M.K., Combustion Studies of Fuel-Rich Propellants, Progress Report (Jan. 1, 1974 - June 30, 1974), AFOSR Contract F44620-71-C-0124, August 21, 1974. King, M.K., and Maek, A., "Studies of the Ignition and Combustion of Boron Particles for Air-Augmented Rocket Applications," Interim Report, AFOSR-TR-75-0043, October, 1974. King, M.K., and Ma~ek, A., "Combustion Studies of Magnesium-Loaded FuelRich Propellants," l1th JANNAF Combustion Meeting, Pasadena, CA, Sept., 1974, CPIA Publication 261, Vol. III, p. 65, Dec., 1974. King, M.K., "Predictions of Laminar Flame Speeds in Boron-Oxygen-Nitrogen Dust Clouds," 15th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, PA, 1975. King, M.K., Combustion Studies of Fuel-Rich Propellants, Progress Report (July 1, 1974 - June 30, 1975). AFOSR Contract F44620-71-C-0124, August 8,

2.

3.

4.

5.

6.

7.

1975.
8. King, M.K., and Childs, L.B., "Development of Highly-Magnesium-Loaded Fuel-Rich Propellants," 12th JANNAF Combustion Meeting, Newport, Rhode Island, Aug., 1975, CPIA Publication 273, Vol. III, p. 201, December, 1975. King, M.K., and Fields, J.L., "Combustion of Highly-Magnesium-Loaded Fuel-Rich Propellants," 13th JANNAF Combustion Meeting, Monterey, CA, Sept., 1976. To be published in CPIA Publication.

9.

146