INet - DTIC

INet

Distribution unlimited 98

Department of the NavyNaval Air Systems CommandCode AIR-52032C 83 09n 09 0Washington, D.C. 20361 7 UJ O6r

OTIC FILE GORY

SLCURITY CLASSIFICATION OF THIS PAGE (non Data Entered),

PAGE READ INSTRUCTIONSaEPORT DOCUMENTAON BEFORE COMPLETING FORM1. REPORT NUMBER 2. GOVT ACCESSION NO. 3, RECIPIENT'S CATALOG NUMBER

NOl0019-81 -C-0184 , 84. TITLE (and Subtitle) 5. TYPE OF REPORT & PERIOD COVERED

Processing Science for AS/3501-6 Carbon/Epoxy Final Report for PeriodComposites 1 Aug 81 - 30 Apr 83

6. PERFOPMING OG. REPORT NUMBER

7. AUTHOR(S) S. CONTRACT OR GRANT NUMBER(s)

J. F. Carpenter NO0019-81-C-01849. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT. PROJECT, TASK

AREA & WORK UNIT NUMBERS

McDonnell Aircraft CompanyMcDonnell Douglas CorporationPO Box 516, St. Louis, MO 63166

II. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE

Dear ;,ment of the Navy 8 September 1983Naval Air Systems Command 13. NUMBER OF PAGES

Washington, DC 20361 Code: AIR 5304 4414. MONITORING AGENCY NAME & ADDRESS(If difftrent from Controlling Office) IS. SECURITY CLASS. (of this report)

See Block 11 Unclassified15s. DECLASSIFICATION/DOWNGRADING

SCHEDULE

16. DISTRIBUTION STATEMENT (of this Report)

Approved for public release, distribution unlimited.

17. DISTRIBUTION STATEMENT (of the abstract entered In Block 20. it different from Report)

1$. SUPPLEMENTARY NOTES

19. KEY WORDS (Continue on revers, side it necessary and identify by block number)

Composite, Carbon/EpoxyPhysiochemical CharacterizationModels for Resin Curing ProcessesRheology of ResinsThermosetting Resins

.2q. ABSTRACT (Continue en reverse side ti necessary and Identify by block number)

S- 'The objectives of this program were to investigate the curing behavior ofthermosetting matrix resins under conditions relatible to the processing para-meters of the companion carbon/epoxy prepreg material and to investigate theinterrelationships of volatiles transport, physiochemical and mechanicalproperties.-

, ~~~~FORM 17 DTO F'NV SOSLT

DD I JA 1473 EDITIONOFINOV6SISOBSOLETE Unclassifiediii SECURITY CLASSIF TION OF THIS PAGE ("h.. Dat. Entered)

3 ZL

-. ~ 7

N00019-81-C-0184

MCDOaN.VLL AIRCRWAFT COMPRANY

FORWARD

This f inal report covers the work performed from 1 August1981 to 30 April 1983 under Contract N00019-81-C-0184 by theMaterial and Process Development Department of the McDonnell Air-craft Company, McDonnell Douglas Corporation, St. Louis, Missouri.The program was administered under the direction of the Naval AirSystems Command by Mr. Maxwell Stander and previously by Mr.Richard L. Dempsey.

The program was managed at McDonnell by Mr. R. J. Juergens,with Dr. J. F. Carpenter as Principal Investigator. Majorcontributors were Messrs. T. T. Bartels, G. T. Stillwell and C. E.Wilson of the McDonnell Materials Laboratory.

iAccession ForVTIS GRA&T

JustieatiO

D ilbility *odescAvsiailit ana/O

Iist Specia

1 46

PACDOP4NELL DOUGLS-4 CONAVORAT5OPJ

v

N00019-81-C-0184

This Page Left Blank

9RCDON'WAELL OI.DLAs COMPWAVW

vi

IL7~

N00019 -81-C-O 184

AqCDOPINWELL- AIRCRMjAFTy cOMPAON

TABLE OF CONTENTS

SECTION PAUE

1 INTRODUCTION AND SUMMARY...................1

2 PROGRAM4 PLAN...............................2

3 TECHNICAL RESULTS....................5

3.1 Phase I -Develop a Model for Resin Rheologyand Flow..........................5

3.2 Phase II -Correlate*Physiochemicai Propertieswith Resin Rheology .. ................ 2

3.3 Phase III - Determine Resin Volatile Effects 283.4 Phase IV - Relate Resin and Laminate

Properties......................31

4 SIGNIFICANT RESULTS............. . . . . 33

5 RECOM4MENDATIONS........................34

R.EFER.ENCES...........................35

DISTRIBUTION LIST..............................37

MJCOONN4ELL [email protected] COMOMSATIOM

le- vii

:1 z ___ ____ ____

N00019-81-C-0184

A2DOWNV&LL AIRCRVAFT OMAN


AfCAMOMMLL DOVOIL.4 COWwpORATjojW

viii

- -- i4 . - Z!

N00019-81-C-0184

*UCDONVEZLJL AIRCRAAFT COMPANY

LIST OF ILLUSTRATIONS

FIGURE PAGE

1 ProgrPam.la. .. .. .. ............ 2

2 Isothermal Rheograms in Replicate for 3 501-6 Resin,

4--.Batch B-i @ 135C . .................. 7

3 Computer Drawn Isothermal Rheogram for 3501-6Resin, Batch B-i @ 135C . ............... 10

4 Computer Drawn Family of Isothermal Rheograms for3501-6 Resin, Batch B-i @ 125, 130, 135, and 140*C .11

5 Rheograms for 3501-6, Batch B-i, for DifferentHeating Rates ..................... 12

6 Computer Drawn Dynamic Heating Rheogram for 3501-6

Batch B-i @ 10 C/MIN...................15

7 Computer Drawn Cure Cycle Rheogram for 3501-6,Batch B-i . .. .. ... :. .. .. ... .. .. .. 16

8 Computer Drawn Isothermal Rheograis for 3501-6Resin Comparing Batches B-1, B-2, and B-3 @ 135*C . 17

9 Computer Drawn Dynamic Heating Rheograms for 3501-6Resin Comparing Batches B-1, B-2, and B-3 @1 0 C/MIN........................18

10 Computer Drawn Cure Cycle Rheograms for 3 501-6Resin Comparing Batches B-1. B-2, and B-3 ....... 18

~.11 DSC Thermogramn fcr 3501-6 Resin, Batch B-i @V10

0 C/MIN. ....................... 21

12 DSC Thermogram for 3501-6, Batch B-i @ l.25*C/MIN . 25

13 DSC Thermogram for 3501-6, Batch B-i @ 2-.5C/MIN .. 26

14 DSC Thermogram for 3501-6, Batch B-1 @ 5*C/MIN . . . 26

&UCDOPIP*ELL 0OUGL COORPORATION

M- ix

L-

.4 _______

N00019-81-C-0184

i MCDONNELL AIRCRAFT COMPAIV

LIST OF TABLES

TABLE PAGE

1 CONSTANTS FOR ISOTHERMAL VISCOSITY EQUATION,BATCH B-I ........... ..................... 8

2 RHEOMETRICS DATA FOR VARIED LINEAR HEATING RATES . 13

3 CORRELATION OF LINEAR HEATING RATE DATA ...... . 13

4 BATCH PARAMETERS FOR MATHEMATICAL MODEL ...... . 16

5 FLOW NUMBERS FOR 3501-6 NEAT RESIN BATCHES .... .. 19

6 COMPARISON OF EVENT TEMPERATURES FOR VISCOSITY(RDS) AND THERMAL (DSC) DATA FOR LINEAR HEATINGRATES ......... ...................... 22

7 DSC-2 KINETIC FACTORS FOR 3501-6 RESIN, BATCH B-i . 24

8 THERMAL DATA FOR COMPLETE CURE OF 3501-6,BATCH B-1 ....... ................... .27

9 PHYSIOCHEMICAL PROPERTIES ... ............. 28

10 VARIATION OF RESIN BUBBLE SIZE WITH DISTANCE FROMTHE SURFACE ........ .................... 29

11 DSC EVALUATION OF CHEMICAL BLOWING AGENTS,@ 50C/MIN ..................... 30

12 LAMINATE MECHANICAL PROPERTIES ... .......... . 31

13 LAMINATE PHYSICAL PROPERTIES ... ............ 32

" MACDONNELL DlOUGLASl CORPORATIONI

~xI"

N00019-81-C-0184

_ MCD@NNELL AIRCRAFTr COMUPANIV

LIST OF ABBREVIATIONS

RDS Rheometrics Dynamic Spectrometer

CPS Centipoise

HPLC High Performance Liquid Chromatography

KCPS Thousand Centipoise

GPC Ge] Permeation Chromatography

DSC Differential Scanning Calorimetry

TADS Thermal Analysis Data Station

phr Parts per Hundred Resin

T Temperature

Tgel Temperature at Gel

Tos Reaction Onset Temperature

Texo Reaction Exotherm Temperature

TCR Complete Reaction Temperature

ri Viscosity

"no Zero Time Viscosity

it Time

T Tgel Time to Gel

G' Storage Modulus

Got Loss Modulus

AHH Enthalpy

.(Dynamic Frequency

M Torque

* Linear Heating Rate

k Kinetic Factor

HPCDONNELL IDOJUGLAS CORPORATION

xi

RN'___~- -

N00019-81-C-0184


LeoJWF 2008-fe4ojofwDxiI

N00019-81-C-0184

1.0 INTRODUCTION AND SUMMARY

The processing parameters currently used in the fabricationof carbon/epoxy composite structures have essentially been derivedfrom experience, as a result of extensive iterative testing. Thepurpose of this program was to develop a generic, and in-depth,scientific understanding of the effect of processing parameters onthe behavior of prepreg resin.

Specifically the objectives of the program were to investi-gate the curing behavior of the neat resin under conditionsrelatable to the processing parameters of the companion carbon/epoxy prepreg material and to investigate the inter-relationshipsof volatiles transport, physiochemical and mechanical properties.

These objectives wece accomplished by the development of a* mathematical model for the rheclogy and complete cure of the

3501-6 resin system. The first part of the model is an expressionwhich describes 3501-6 resin viscosity as a function of the timeand temperature of cure. The model was used to compare the totalflow behavior of the resin under varied conditions of cure and thebehavior of three batches of resin in wbich tne major epoxide com-ponent had different starting viscosities.

A computer program was written which uses the parameters ofthe mathematical model to plot viscosity profiles and calculatethe total flow of 3501-6 resin for any desired combination of timeand temperature of cure.

The mathematical model was expanded to include an expressionfor the time to complete cure as a function of cure temperature.

The model is applicable to other thermosetting resins and can' be used to establish cure cycle conditions fromn a minimum of

- experimental determinations. It is also envisioned as being a

useful methodology for the real-time control of automaticprocessing equipment (i.e. autoclave control).

Variations in the resin rheology of three special batches ofresin were compared, as were differences in chemical compositionof the prepreg and mechanical properties of the laminates. Theeffects of volatiles on resin processing were also investigated.

WADCOM ELL 0400a COR4P4MJWAT

C 1

- -- t.r.

N00019-81-C-0184

2.0 PROGRAM PLAN

The technical approach is given in the Program Plan, Figure1. The rheology of 3 501-6 resin was determined over a broad rangeof varied heating conditions. A model for the curing behavior ofthe neat resin was developed through the derivation of a mathemati-cal expression which relates the resin viscosity to the time andtemperature of resin cure. Physiochemical properties were deter-

z mined arnd the thermal analysis data was used to expand the curingbehavior model to include a mathematical expression which relatesthe time required for complete cure to temperature. A computerprogram was written which uses the parameters of the mathematicalmodel to plot visco~sity profiles for any desired cure cycle.

Phase~Deiv Mode Testle Vaidt ofe ei ielg

0 3 Resin Batches e Phsomet ic: Dynamic 0 Viscosity ProfilesSpectrometer 0 Total Flow

Phae N. Curreat Physladwsodcl Prepds 1 el el

Thifermtal SroannigCeia roete rwest ei k

CaOnM"t 0 Infrared0 Atomic Ahsorption

Phmnveluoda Resin D.d. fetrieMtoso

0 Types0 Transport

Pe VPopert Daae-d reeesirpete

= UI" 1. P IPIWQ*WWOWWEL m -OW CPm 3@

2

Z- V

N00019-81-C-0184

WCDONWWEL-L AIRWCRAA coiPANVy

j Three batches of 3501-6 resin were compared. These batcheshad been formulated to contain tetraglycidyl methylenedianiline(TGMDA) having three different starting viscosities. The types ofresin volatiles and their transport mechanisms were investigated.As a last step, the resin properties determined in this programwere compared with previously published data for their compositelaminate properties.

PHASE I - DEVELOP A MODEL FOR RESIN RHEOLOGY AND FLOW

The objective of this phase was to develop a rheologicalmodel which represents the behavior of 3501-6 resin under condi-tions relatable to the processing parameters of the companioncarbon/epoxide prepreg. This objective was accomplished bydetermining the rheological characteristics of the resin using theRDS-7700 rheometrics dynamic spectrometer. During resin heatingthe viscosity was measured for four different isothermal con-ditions, three dynamic heating conditions, and a typical curecycle. A mathematical expression which relates the viscosity tothe time of heating was written for each of the four isothermalconditions. The four equations w9..- 'hen used to derive a generalexpression which relates viscos.- , the time and temperature ofcure for all isothermal conditions.

It was found that the general expression for isothermal vis-cosity change could be converted to an expression applicable todynamic heating (varied linear heating rates' by integrating thetime dependent term. Then, together, the equations for isothermaland dynamic heating constitute a set of equations which predictsthe viscosity for any time/temperature sequence of resin cure.

This set of equations is proposed as a mathematical model forresin rheology during cure. A computer/plotter was programmed toinclude the equations of the model, and viscosity profiles areautomatically produced for any combination of cure cycle

parameters.

The rheologies of three batches of 3501-6 resins werecompared. These batches were made by standard processing and withnormal component concentrations, but contained TGMDA with variedstarting viscosities.

P ASE II - CORRELATE PHYSIOCHEMICAL PROPERTIES WITH RESIN RHEOLOGY

The objective of this phase was to determine relationshipsbetween physiochemical properties and resin rheology and to deter-mine the applicability of physiochemical data to an expanded

V curing bc.iavior model. Differential scanning calorimetry (DSC)was used for thermal analysis of the resin system and the chemicalcharacteristics w re determined using infrared, liquid chromatogra-phy and atomic absorption. The physiochemical properties werecompared with the variations in rheology for the three resin

I batches tested.

S ACDONNELL DOUGLAS CORPOR.ATION

3

;4 .

N00019-81-C-0184

MCDONNELL AIRCHt.FT COMPANY

The DSC data was used to expand the model for resin curingbehavior to include a mathematical expression which relatesCemperature to the time required for complete cure.

PHASE III - DETERMINE RESIN VOLATILES EFFECTS

The objective of this phase was to characterize resinvolatiles and explore techniques useful to modeling theirformation and migration during cure. As-received resins wereinitially compared under ambient pressure cure and vacuum cureconditions for the types and amounts of volatiles. Bubble sizeand distribution was compared using a Bausch and Lomb imageanalyzer. Later, prepreg materials were compared for outgassingcharacteristics by submerging the specimens in silicone oil andobserving bubble formation during resin cure.

A means was developed to introduce known amounts of bothcondensible and non-condensible volatiles using a method that isapplicable to the study of resin volatiles transport in neat resinand the characterization of voids in cured composites.

PHASE IV - RELATE RESIN AND LAMINATE PROPERTIES

The objective of this phase was to determine whether the neatresin properties determined in this program exhibited any corre-lation with previously determined laminate properties. The objec-tives were achieved by comparing laminate physical and mechanicalproperties, determined earlier in Reference (1), with the rheologi-cal and physiochemical properties determined in this program. Theresin batches compared differed only in the viscosity of the baseepoxide (TGMDA) used in their formulation.

V

MCDONNELL DO.GLAS CORPORA'TION

4

-_ _ -

N00019-P'-C-0184

MCDONNIELL AIRCRAFr COMPANY

M 3.0 TECHNICAL RESULTS AND DISCUSSION

Three batches of 3501-6 resin were s'..ected for the program.Thae differences among the three batche.. of resin is in thestarting viscosity of the major epoxide component, tetraglycidylmethylenedianiline (TGMDA):

BATCH TGMDA VISCOSITY (CPS)B-I 10,700B-2 14,000B-3 18,000

All batches were formulated by tne supplier's standard pro-cessing procedure and all were formulated to contain the followingstandard concentrations of starting components:

TOTALCOMPONENT PERCENT (WT)

Tegraglycidyl Methylenedianiline 56.5Alicyclic Diepoxy Carboxylate 9.0Epoxy Cresol Novalac 8.54,4' Diaminodiphenyl Sulfone 25.0Boron Trifluoride Amine Complex 1.1

In earlier work, these batches gave different behavioralresponse to processing, Reference (1).

3.1 PHASE I - DEVELOP A MODEL FOR RESIN RHEOLOGY AND FLOW - Inthis phase resin flow measurements were made under varied curemodes and a mathematical model was developed which describes3501-6 resin viscosity and flow as a function of the time andtemperature of processing. The effects of varied cure cycles onresin flow was demonstrated. The model was used to compare the

j rheologies of three batches of resin having varied startingviscosities for the major epoxide component. The flow propertiesof the three batches were shown to be quite similar.

Instrumental Technique - The Rheometrics RDS-7700 instrumentwas used for this program. It is a dynamic, oscillatory rheometercapable of measurements over a wide range of temperatures. It can

be programmed for any combination of linear heating rates,isothermal holds, and simulated cure cycle conditions. Itprovides continuous printout and data plots for selectedproperties, including dynamic viscosity (n), loss modulus (G") andstorage modulus (G'), together with time (t), temperature (T), andtorque (M).

The rheological properties for the neat resins were measured= under the following conditions:

Starting Temperature = 500CStrain = 50%Shear Rate = 10 rads/secPlate Gap = 0.5mm + 0.01mm

F. MCDON*VELL DOUGLAS CORPORATION

5

N00019-81-C-0184MFCIDONNELL AIRCRAFT COMIPANY

All samples were preconditioned at room temperature by pres-sing about 1 gram of resin between two pieces of non-porous Tefloncloth (CHR-6TB) to produce a resin wafer about 5mm thick. Thewafers were exposed in a vacuum desiccator for 24 hr and removedjust prior to running the rheograms. The sample preconditioningwas used to remove moisture and any other vclatiles that mightcause erratic variations in rheology.

The RDS-7700 instrument was used to gener, e plots of viscos-ity vs. time for the temperature conditions o. the tests. Theseviscosity profile data were used to calculate a measure of thetotal flow achieved for the varied time/temperature parameters ofthe tests.

The reciprocal of viscosity is termed "fluidity". Fluidityintegrated as a function of time was taken as a measure of total

Rflow under the pressure conditions of the rheometrics instrument.

Isothermal Heating Conditions - The change in viscosity as afunction of time and temperature was first studied for the iso-thermal case.

Under isothermal conditions the change in viscosity with timecan be expressed by the equation:

logn = k(t) + logno (1)

where: n = Time Dependent Viscosity (poise)o = Zero Time Viscosity (Constant for Isothermal Runs)k = Kinetiv Factor

Et = Time (minutes)

This model for the isothermal case has been proposed byseveral investigators, References (2-6). Figure 2 shows a plot offour isothermal runs for viscosity as a function of time at 135 C.A linear regression line for logrn vs. time is also shown. Theminimum viscosity was reached in an average time of 6 minutes.

,

MCDONNELL OOVULAS CORPORATION

6

4.- _77-

N00019-81-C-0184

1i03

it0

2P 0120 4 so 0 16Tim -mi O33/W

Er/Fiur 2 sohema htgrm I Rplctefo 30-6Rei, ath . @130

MCOWOVLLC OOLAW ONAON~~a

7/

/V

N00019-81-C-0184

M~CD)OVNELL AIRCRAFT" COM'PAWV

Replicate runs were made for Batch B-I at four differenttemperatures. The values calculated for the constants k and logno at each temperature are given in Table 1. The kinetic factor(k) decreases with decreasing temperature and zero-time viscosity(no) increases with decreasing temperature.

Table 1. Constants for Isothermal Viscosity Equation, Batch B-1log il= k(t)+ log 101

Temperature k log u/o Corral(OC) Coef tr2)

140 0.0554 - 0.0517 0.982

135 0.0401 0.1193 0.985

130 0.0318 0.2856 0.988

125 0.0247 0.4001 0.988

OP33aMU.SI: The kinetic factor (k) can be expressed as a function oftemperature (T, *K) by an expression similar to that derived inReference (7).

c 10Al/Tk T2 (2)

* 'For Batch B-i, log (kT2 ) vs. l/T is solved by linear regression togive:

2log 4140.2 + 13.9885 (3)lg TT

r= 0.994 (correlation coef.)

Thus:

k = -74 i01 [• 0 - 4 1 4 0 / T ]

kT9.74x 10 [2 (4)

MCDONLL DOUGLAe CORPORATION

8

N00019-81-C-0184

SMCDONNELL AIfCRA~r COMPAWV

In a like manner, the isothermal values for log no (Equation1) for Batch B-i was reduced by assuming:

710 C2 10 A 2/T (5.T 2

The linear regression of log (no T2 ) vs. l/T gave the values of

the constants as:

C2 = 8.81 x 10 - 7 (6)

A2 = 4645.8

with a correlation coefficient of r2 = 0.989.

Thus, a mathematical model for isothermal resin flow is givenby the following expression, which relates viscosity to the timeand temperature of processing:

logn = cl [l lT (t) + log 1OA2/ (7)

For Batch B-1 the expression is:

log-7 13[1f 4 14 0 .2/T (t) + g8.81 x 10 - 7 1 4 (8)

A desktop computer, interfaced with a digital X-Y plotter,was programed to solve Equation (7) and automatically plotviscosity as a function of time/temperature.

A computer generated plot for the isothermal viscosity pro-file of Batch B-l, at 135*C, is shown in Figure 3. The computerprogram provides for the superimposition of experimental data.Each point shown on the rheogram in Figure 3 represents an averageviscosity value for replicate determinations. Figure 4 shows acomputer drawn family of curves for isothermal determinations at125, 130, 135, and 140*C.

k

*UCDON WELL DOUILAS COPOMtRAT

9

I,4

N00OO19-81-C-0184

4 11003 Experimental

3

LogVismoety

Poise

0100 04 0 2 ;

Tim- min OaMuse

Figure 3.Computer Drawn Isothermal Rheogramn for 3501-6 Resin, Batch B-1 @1350C

*WCDWNKLWUMLAS OC@AWP@ATU@N

10

M--t5 -7

N00019-81-C-0184

MCDONNELL AIRCRAFTr" COMPANY

1400 13135*C

4 8100

3

* LogViscosity

Poles

Time - in P3IS

Figure 4. Computer Drawn Family of Isothermal Rheogramsfor 3501-4 Resin, Batch B-1 @ 125, 130, 135 and 1400C

Dynamic Heating Conditions - Dynamic runs were made for thethree test batches of 3501-6 at linear heating rates of 1, 2, and4*C/min. Typical rheograms for Batch B-I at the three differentheating rates are shown in Figure 5. Data for critical points(minimum viscosity and gel) of the curves for Batches B-1, B-2,and B-3 are given in Table 2. As expected, the faster the heat-uprate, the lower the minimum viscosity and the shorter the time togel.

MdUCDO4PIELL OOUtLAS CORIPORATION

81

N00019-81-C-0184

1341 CImln 20CImin 10C/min

SViscesity 12Flo w=5.6 Flow =6.4 Flow 1 72Poise

I1 _.00

10

0 -04 6 o10 2

Time - minGP33001S

Figure 5. Rheograms for 350146 Batch, 8.1 for Different Heating Rates

11

El N00019-81-C-0184M1 1 tCDONNELL AIRCRAk-F COMPANY

Table 2. Rheometric's Data for Varied Linear Heating Rates

Minimum Temperature atBatch Heating Rate, 0 Viscosity, 1 Temperature, T (C) Time to Gel', tool Gel Point,

Number (OCimin) (Poise) at Minimum ui (min)•T,,,(°C)

1 7.2 110 111 167

B-1 2 4.7 120 63 184

4 2.7 130 36 194

1 6.8 112 115 171

B-2 2 5.7 126 63 186

4 3.2 136 36 202

1 8.6 114 112 168

B-3 2 5.8 122 64 186

4 3.2 140 36 206

1.0o0 poise GP334M17

The temperatures for minimum viscosity and ge- can bepredicted for other linear heating rates by the equations given inTable 3. Calculation of these temperatures provides usefulinformation; however, what is needed is a means of using the dataobtained from a minimum number of isothermal runs to predict theviscosity profiles obtained under the dynamic conditions of variedlinear heating rates.

Table 3. Correlation of Linear Heating Rats Data

Batch Equation Relating Temperature Equation Relating Temperature

Number and Heating Rate for Minimum and Heating Rate for ResinViscosity Gel"

4,535 log 4,451

-1 log € T(OK +11.8465 T(*K) +10.4,45

r2 0.999 r2 =0.982

390024,0959B-2 log = 3 +10.1149 log 0= +9.2256

T('K) T('K)

r2 = 0.959 r2 =1.000

0. Heating rate (Kftini

° WCDONNELL DOULAS CORPORATVION

13

N00019-81-C-0184

MCO0d)4ELL AIRCRAFr CO"WAP

Equation (1) can be converted into a form applicable to thedynamic heating case by replacing the first (isothermal) term withan integral expression to give:

tgel

logn = f k dt + log (o (9)

For linear heating rates, time can be expressed in terms oftemperature by:

T -T O dTt and dt-=-- (10)

Equation (7) is then modified to give the followingmathematical model for dynamic heating:

log n = CI f {l ! IdT + log 1 A2 1T (11)-"OTO

The computer/plotter was programmed to include Equation (11).A computer plot for dynamic heating at °C/min, together withexperimental points, is shown in Figure 6.

Equation (7) for isothermal heating, together with EquationR(11) for the dynamic heating case, constitute a mathematical model

which provides the capability to calculate viscosity for anytime/temperature sequence (i.e., the resin rheology portion of aprocessing model).

L

I

*CDONPIELL OduLMAS CORaRArsTO

14

N00019-81-C-0184

0 Ex~ ont

3 1

LOUVWsosty \0

Poise 2N

I

00

Time - min oP3Mega

Figure 6L Computer Drawn Dynamic Heating Rheogramn for 350146, Batch 8.1 @ IlCImIn

To verify the suitability of the model, a typical cure cyclerun was made on the Rheometrics instrumrent. The cycle included aramp at 2*C/min to 116*C, a hold at 116*C for 60 minutes, a rampat 3*C/min to 117*C, followed by a hold at 177*C. Figure 7 showsexperimental data together with a computer drawn curve using therheology model.

15

r-J~

2MR

N00019-81-C-0184

a2) HoWd 116C for 60 mmn

4 () Rw*~ 3'Cftfin to 1WC05',

LogViscosity

2

1E

0E0 20 40 6 80 00 10 14

D ~ ne-mtE cl

Figure~ 1 7.CmutrDaw ue yl Regamfr351,Bac

fromli ths7dt. The~ praeer fa yWor the model foBach of-t1

thre w Batch is1 tive ins able . ei elctea 2,10

Ta"l 4. atch pa~erweter for Mathematical Model

10g 17 = Ci [loAl~frI (t) + log [C21OA2ff1T2

Batch C, As C2 A2

B-3 2-16 x 1012 -3.467.i 1.45 x l~y 4,140.8

Where: 7) = Viscosity (Poise)

T=Temperature (K)

16

I N00019-81-C-0184

MCDONNELIL AIRCRAFT COMPMANY

The rheology models for the three batches were used togenerate computer curves depicting the viscosity profiles.

The rheological properties of Batches B-I, B-2, B-3 werecompared using computer drawn curves. The isothermal viscosityprofiles are shown in Figure 8. Figure 9 shows batch comparisonat a heating rate of 1C/min and Figure 10 shows the modelrheograms under the conditions of a typical cure cycle. Theviscosity profiles for the three resin batches are quite similarand show no significant trends relatable to the viscosity of theTGMDA used in their formulation.

-B-

41B-1 \B133

LogViscosity

Poise

A -

0 20 40 60 80 100 120 140

Time. min

Figure 8. Computer Drawn Isothermal Rheograms for 3501.6 Resin Comparing Batches B-i,B-2, and B-3 @ 1350C

MCDONNELL DOULAS CORPORATION

17

N00019 -81-C-0184

MCDONN~E&L AIRCAFT~ COM&ANV

B-=83/

3

LogViscosity

Poise

2

11

0 20 40 080100 120 140Time- mini OP33a40o

Figure 9. Computer Drawn Dynamic Heating Rheograms for 3501.6 Resin Comparing BatchesB-1, B-2, and B-3 I l0 CImin

5

3Log

,COAEL DOGACOPNTO

0 2040 6 80 00 10 18OPS0M~

N00019-81-C-0184

M*CDONNELL AIRCRAFT COMWPANYV

Total Resin Flow - The total flow at constant pressure isdescribed as the value obtained by integrating the reciprocal ofviscosity (fluidity) from zero-time (to) tc the point where theflow ceases, i.e., time to gel (tgel):

f tgelFLOW NUMBER =j dt/- (12)

to

Flow numbers for Batches B-l, B-2, and B-3 are shown in Table5. Individual values obtained in replicate runs are shown inparentheses. These values were obtained by machine-integration ofthe viscosity profiles from the RDS experimental runs and not fromthe rheology model. The flow-numbers are a measure of total flowfor each condition of time and Lemperature under an equivalentpressure.

Table 5. Flow Numbers for 3501-6 Neat Resin Batches

Flow Numbers (Minutes/Poise)BatchI

B1 B.2 83

1. Dynamic (°C/min)

1 7.24 7.51 6.16(6.89, 7.59) (8.43, 6.58) (5.93, 6.40)

2 6.43 5.52 5.55(6.42, 6.44) (5.52, 5.53) (5.26, 5.85)

4 5.61 5.05 4.19(5.36, 5.86) (4.91, 5.19) (4.61,5.76)

2. Isothermal (°C)

125 5.71 5.38 6.04(5.63, 5.66, 5.85) (5.13, 5.63) (5.69, 6.39)

130 5.55 4.53 5.30(5.46, 5.63) (4.46, 4.59) (4.67, 5.93)

135 5.08 4.25 4.46(4.86, 5.00, 5.02, 5.40) (4.25, -) (4.22, 4.71)

140 4.91 4.29 4.33(4.36, 5.45) (4.35, 4.22) (3.98, 4.09, 4.91)

3. Cure Cycle 5.96(6.14, 5.77)

WCONNELL OUOLAS CORPORATIONI' 19

N00019-81-C-0184

MCDONNELL AIRCRAFT COM*PANV

For the three batches tested, the following observations weremade:

o Dynamic Cure Mode - Increasing the heating rate giveslower total flow. The total flow of a resin batch is inline with the viscosity of the starting YGMDA.

o Isothermal Cure Mode - Increasing the isothermal tempera-ture also gives lower total flow. The total flow of BatchB-3 is higher than expected from the relative viscosity ofthe starting material.

o Cure Cycle - The cure cycle tested (Ramp at 20C/min to116°C - hold 116 0C for 60 min - ramp 3*C/min to gel) did

not maximize flow.

o The data indicates that it is possible to achieveequivalent flow for all batches by tailoring cureconditions. These batches represent variation in TGMDAviscosity over the full range of commercially availablematerial.

Flow numbers can be calculated directly from the viscositycurves generated by the rheology model. The calculated values arein general agreement with those obtained by direct measurement ofthe individual experimental rheograms; however, they are notconsidered to be as accurate. To improve the accuracy of thecalculated flow numbers it will be necessary to assume a higherorder regression fit than the linear regression assumed forequation 7.

3.2 PHASE II - CORRELATE PHYSICOHEMICAL PROPERTIES WITH RESINREHOLOGY - The Objective of this phase was to determine inter-relationships for physiochemical properties and the rheology ofthe resin batches studied and to determine whether physiochemicaldata can be used to expand the processing model. The thermalanalysis data was shown to be applicable to the development of a

,l mathematical expression for calculating the time to complete curefrom the cure temperature. This thermokinetic model, togetherwith the rheology model consitute a viscokinetic model for resincuring behavior.

Determine Resin Thermal Properties - Thermal analysis of the3501-6 resin system was performed by differential scanningcalorimetry using the Perkin Elmer DSC-2 instrument.

A typical DSC thermogram is shown in Figure 11. The initialminor exotherm is associated with the BF3 catalyst and may be dueto the dissociation of the BF3 - amine complex together with thereaction of the novalac epoxide with the sulfone (DDS) curingagent. The major exotherm is mainly due to the TGMDA and carbonylepoxides reacting with the remaining curing agent. The DSC instru-ment pr~rits out the exotherm peak temperatures. The points forminimum viscosity and gel, as determined by RDS, are shown nn thethermogram to locate these events relative to the exotherma! curereaction.

i i MUCDONNELL D1OU.GLAS CORPORATION

20

I -

N00019-81-C-0184

: °MCDONNELL AIRCRAFT COMPANY

5.00

3.75

MCAL/sec 2.50

1.25 --

SGel (216"C1 TFxo = 229°C (Major)o I ,,_ I I I I I, I I 1

30 60 90 120 150 180 210 240 270 300 330 360Temperature- "C

Figure 11. DSC Thermogram for 3501-6 Resin, Batch B-1 @ 1O0 Clmin

The DSC exotherm peak temperatures are compared with the RDSviscosity temperatures (minimum viscosity and gel point) in Table6. There appears to be no well defined quantitative correlationsbetween these data. A very general observation for the variedlinear heating rates is that minimum viscosity is near the minorDSC exotherm peak and the gel point is near the major DSC exothermpeak. The reaction associated with the minor exotherm causes theviscosity to start increasing and the flow is decreased comparedwith some other epoxide composite matrix systems, which do notcontain BF3, such as Narmco 5208: Fiberite 976 and Hercules 3502.

PCOOPUNNLL DPOUGLAS CORPORATION

21

II

N00019-81-C-0184,: MCDONNELL AIRCRAFT COMPANY

Table 6. Comparison of Event Temperatures for Viscosity (RDS) andThermal (DSC) Data for Linear Heating Rates

Temperature (C)

Heating RDS DSC RDDSCBacRate N h (Minimum (Minor (MajorNumber Viscosity Peak (0 Peak(OCImln) Point) Exotherm) Point) Exotherm)

B-1 123 120 186 1962.5 B-2 128 116 191 196

B-3 129 120 192 196

B-1 133 148 200 215

5.0 B-2 141 145 207 213

B-3 144 145 213 211

B-1 145 159 216 229

10.0 B-2 155 155 225 228

B-3 159 158 235 229

B-1 157 172 233 249

20.0 B-2 170 172 244 248

B-3 175 172 259 248

OP33-OOMe20

4'To study the kinetics of the 3501-6 cure reaction the Perkin

Elmer DSC-2 was interfaced with a thermal analysis data station(TADS) for acquisition and manipulation of data.

Kinetics of 3501-6 Cure Using DSC-2 Data and the TADS Pro-gram - The exothermal cure reaction for transformation of theB-staged starting resin to the cured matrix polymer may beexpressed as:

[RESIN] K [POLYMER] + AH (13)

-MCDONNELL DOULAS CORPORATION

22

p',l---. I

N00019-81-C-0184• MCDONVELL AIRCRAFT COMIPANYV

Based on the decrease of resin concentration with time, therate equation can be written as:

dad = K (1 - ap (14)

Where:a = degree of polymerizationt = timeK = rate constantn = parameter for data fit

The reaction expressed in Equation (13) is assumed to bediffusion controlled, in which case the order of the reaction ismeaningless. Therefore, the term n is a parameter for mathe-matical data fit and is not applicable to defining the truereaction mechanism.

The classical Arrhenius equation can be used to express therate constant (K) as a function of the temperature (T):

K = KO exp (-E/RT) (15)

Where:R = gas constantT = temperature (*K)E = activation energyKO = constant

Combining equations (14) and (15) gives the kineticexpression:

da (16_),)

- K0 exp (-E/RT) (l)n6)

R. Rush, Reference (8) and R. Fyans, Reference (9) havedescribed the use of Model DSC-2 equipped with the thermal analy-sis data station (TADS) and the TADS kinetic software package fordata manipulation in solving equation (16).

* - It is assumed that E and n are independent of (1- a ) and aredependent of the DSC scan rate. These independent relationshipswere shown for a solid diffusion controlled reaction by Crane,Dynes and Kaelble, Reference (10). The degree of polymerizationa, can be expressed in many ways. The simple approach is usedhere.

In order to calculate a, the partial areas of the exothermmust be obtained. This is done by a subroutine in TADS whichdetermines the total AH and partial areas by using Simpson's Rulefor the area under the DSC curve.

AH partialx a (17)

AH total

1 - a is then calculated and becomes a known for the solving of4 Equation (16).

MCDONELL DMOVOLAS CORPORATION

23

N00019-81-C-0184i i MC DNNELL AIRCRAFTlr COMltPANY"

Equation (16) can be reduced to a linear form by taking the

natural logarithm to obtain:do'

in d-f = in K. - (E/RT) + n in (I - a) (18)da

Equation (18) has DSC knowns of in -, l/T and ( - a) as

variables to solve for Ko , E and n. This is done by multi-linearregression in TADS.

SThe kinetic factors calculated for 3501-6 neat resin Batch9-I are given in Table 7. The data indicate a pseudo first orderreaction. The magnitude of the activation energy indicates adiffusion controlled network extension of epoxide/hardener system.

Table 7. DSC-2 Kinetic Factors for 35.1-6 Resin Batch B-1

1.25 2.5 5.0(0CImin) (Crin) (°CInin)

i - Kinetic Factor~Constant,

in K. (sec"1) 15.4 13.4 10.5

Activation Energy,Ea (callg) 19.8 18.2 15.6

Psuedo ReactionOrder (n) 1.2 1.5 0.9

GPJ3ISe- 2i

The 3501-6 resin has a bimodal cure exotherm. The set ofkinetic factors reported here is for the major exotherm. Asimilar set of data was determined for the preliminary minorexotherm that is due to the BF3 catalyst. The minor exotherm datawas not consistant and reflects a limitation that has not beenresolved.

DSC Data and a Thermokinetic Model for Complete Cure - Thekinetic factors calculated using the TADS sofware program gavedata that may be used for reference and in comparing 3501-6 withother thermosetting resins. However, the form of the kineticsequation provided by TADS is not immediately applicable to aprocessing model.

An approach that will allow the use of DSC to predict thetime/temperature combinations required for complete cure has beenreported by Carpenter, Reference (7).

MCDONNELL D)OU LAS CORPORATIO

I 24

__ _ _ _ _ __ _ _ _

N00019-81-C-0184

MCDONNELL AIRCRAFT COMOPANV

This technique uses the temperature for complete reaction(TCR), shown in Figures 12-14. A plot of log 0 vs l/TCR for threedifferent heat up rates gives the following equation:

log 4875 + 9.821 (19)TCR

0.8

TCR=

o 0 .6_ -2 8

MCAUsec 0.4

0.2 Min: 184.49

0 I III20 50 80 110 140 170 200 230 260 290 320

Temperature - °C

Figure 12. DSC Thermogram for 3501-6, Batch B-1, @ 1.250Clmin

M~ACDONNELL IDOIJLAS CORPORA'lION

25

-. --- -S

NOOC 19-81-C-0184

2.0

1.5

MOAL/sec 1.0

0.5

20 50 80 110 140 170 20)0 230 260 290 320

Temperature -Figure 13. DSC Thermogram for 35014., Batch 8-1, @ 2.5 CImin P3U1

2.0

1.5

MCAL~sec 1.0

0.5

Min: 215.94

020 50 80 110 140 10 20 30 260 290 32

ii Temperature - *CFigure 14. DSC Thermogram for 3501-4, Batch B-1, @ S*ClmlnAVCDONNVELL SPOMLAS COMPACOARATUOP

26

- -. l

N00019-81-C-0184

MUCDONiNELL AIRCRAFT COPANV'M

The data is given in Table 8.

Table 8. Thermal Data for Complete Cure of 3501.6, Batch B.i

Critical Heating Rate, 4 Temperature EquationPoint (*CImln) (OC)

1.25 22 og = 4875 + 9.821TTemperature forComplete Reaction (TCR) 2.50 245 r2 = 0.9987

5.00 261

GP33O0423

As was shown in Reference (7), the isothermal time to com-plete reaction (tCR) can be expressed as a function of cure temper-ature by the following equation:

tCR = c T2 104875/T (20)

The constant, (c), can then be determined from:

c =JTCR 10-4875/T dT (21)( OT2

TO

This gives the final expression for time-to-complete reaction

as:

tCR = 1.345 x 10-14 T2 10 4875/T (22)

The isothermal time to complete reaction (tCR) at a cure temp-erature of 350 0 F, by Equation (22), is 3 hr:8 min. A calculationthat includes the contribution of the cure cycle steps prior toreaching 350*F, decreases the calculated time needed at 350F byonly a few minutes. Only final cure and postcure temperatures arerequired.

The mathematical model, equations (7) and (11) predicts therheology and flow of the resin system under conditions of cureencountered in processing. Equations (19) and (22) constitute athermokinetic model which uses DSC data to predict conditions ofcomplete cure. Together, these equations are important parts ofan overall model which describes the behavior of the resin underconditions of composite processing.

Determine Resin Chemical Properties - The three batches of3501-6 resin, made from TGMDA base resins with varied viscosities,were tested for chemical properties. The data, shown in Table 9,fall within limits established in earlier work, Reference (1).

Mt*CDONNELL OULAS CONPORAVION

t~ 27

* -

N00019-81-C-0184MACDONWELL AIRCRAF1T COM4PAIV

Table 9. Physiochemical Properties

_ _ I -Viscosity. DDS Carbonyl Epoxy BF3 DOS TGMDA Epoxy

Batch of TGMDA (Total) (Total) (Total) (Unreicted) (Unreacted) (Unreacted(CPS) x 103 (%) (%) (%) (%) (%) (%)

B-1 11 23.8 8.4 1.1 17.8 53.4 7.8

B-2 14 23.5 8.3 1.3 22.2 53.8 8.9

8-3 18 23.7 8.4 1.3 21.6 50.1 8.5

GP33*M24

The variations in base resin viscosities is due to differ-ences in the type of quantity of oligomers produced in the manufac-ture of TGMDA. In general, the greater the oligomer content, thehigher the viscosity. It was thought that the oligomer contentmight affect the kinetics of the cure reaction. If this were thecase, these three batches, which were all B-staged the same,should have different amounts of unreacted components. However,no significant trends in chemical composition were found.

3.3 PHASE III - DETERMINE RESIN VOLATILES EFFECTS - The objectiveof this phase was to characterize resin volatiles and their trans-port and the relationships t voids in the cured system.Exploratory work was conducted relevant to modeling the behaviorof volatiles during resin cure.

As a preliminary investigation, samples of as-received 3501-6neat resin were melted and held at 950C. At this temperature theresin h-.s an approximate viscosity of 10 poise. Even at this lowviscosity no bubbles were formed at ambient pressure. A vacuum of29" Hg was applied and bubbles quickly formed throughout theresin. The bubbles slowly rose to the surface. After about 45minutes most of the bubbles had risen to the top one-fourth of theresin. The vacuum was vented and the sample was quickly returnedto room temperature. The bubbles in the upper layer of the room-temperature solidified resin were thus "frozen" in place.

The bubble size was measured as a function of resin depthusing a Bausch and Lomb Image Analysis System. The bubbles in agiven layer were quite uniform in size. The average bubblediameters at nine different depths are given in Table 10. Thebubble diameters decrease as the bubbles approach the surface.This would suggest that the bubbles are not coalescing to formlarger bubbles as they rise to the surface. The decreasing sizeof the bubbles as they near the surface is also contrary to whatwould be expected from hydrostatic pressure effects in a fluid ofuniform viscosity.

M~ACDONNIELL 1DOUG LAS CORPDORATION

28

I-

N00019-81-C-0184

MCDPONNtlELL AIRCA&r COMPPANW

Table 10. Variation of Resin Bubble Size with Distance from the Surface

Surface, Inner MostLayer Layer

Distance 1 2 3 4 5 6 7 8

Bubble 43 66 77 83 81 98 89 95 95Dia (p)

*Arbtray distance with depth of focusGF33,00IS M

These preliminary experiments indicated the following:

o V7olatiles present in the as-received 3501-6 resin werecapable of producing large quantities of bubbles duringresin cure.

o Of the two suspect volatiles, water and air, absorbedmoisture is the most likely candidate, since copiuousbubble forwation resulted only after reducing the pressureon the melted resin to below the vapor pressure of water.

Fabrication of low-void composites thus requires that prepregproduction and composite processing conditions be controlled toeither remove the resin volatiles or retain them in solidsolution.

Direct Observation of Outgassing - A second method used forquantitative assessment of off-gassing characteristic: of theresin was a technique reported by Brown and McKaque, Reference(11). This method consists of submerging the speciamens indegassed silicone oil and observing the bubble formation duringthe resin cure cycle. A vacuum oven with a large viewport allowedthe use of video equipment to record the out-gassing experiments.The video recording system used provided a continuous readout oftime and temperature superimposed above the sample.

Observations and results obtained using this techniqueincluded the following:

o Careful control of tl-e vacuum level above the silicone oilwas crucial for reproducibility from run to run. Only pre-preg specimens could be compared, since weighted .pecimenshad a tendency to break away from anchors and fl at to thetop of the silicone oil. Making close comparisonsrequired that specimens be observed side by side, duringthe same run.

OWCI CDOWELL fl6L COROARA 01 P4

29

N00019-81-C-0!84MACDONNELL AIRCRAft COAFP4NV

o As-received AS/3501-6 prepreg was compared with preprcgdried in a vacuum desiccator overnight. The two prepregspecimens were heated at 4-5F/min to 350*F under vacuum.The as-received material off-gassed vigorously throughoutthe run, while the dried prepreg showed significantly lessoff-gassing. Moisture is again indicated as the majorvolatile. Complete removal of moisture, 'however, appearsdifficult and would be a costly step.

o As-received T300/5208 prepreg was compared with theas-received AS/3501-6 prepreg for conditions of 4-5* /minheating to hold at 375*F under vacuum. Both specimensbubbled vigorously during most of the run. The 5208prepreg bubbled more than 3501-6 at 370"F and beyond.This may have been due to the shorter gel time of 3501-6.Otherwise, this technique did not indicate mucn differencebetween the two specimens.

Quantitative Introduction of Volatiles - To evaluate vola-

tiles transport mechanisms a method is needed for introducingcontrolled amounts of the representative volatiles into the matrixresin. Six chemical blowing agents were selected for this purposeand evaluated by DSC. The blowing agents investigated and theirexotherm (or endotherm) peak temperatures are given in Table 11.

Table 11. DSC Evaluation of Chemical Blowing Agents, @ 5"Chnln

Blowing Agen Temp., C

Celogen, TSH (Uniroyal) 168

I ]Celogen, OT (Uniroyal) 166

Kempore (Olin) 209

Opex, 93 (Olin) 203

Nitropore, OBSH (Olin) 155

Expandex (Olin) (206)

)Endotherr. gP 34

The maximum rate of gas release should occur at the peaK temp-eratures. On this basis the Nitropore OBSH was selected since itdissocaiates between the point of minimum viscosity and the on-setof the major reaction exotherm.

"*ICDONELL DIGLAS CORMPON lOR3

N00019-81-C-0184

MCDONNELL AIRCRAFT COMPA IV

Nitropore OBSH chemically is 4, 4'-oxybis (benzenesulfon-hydrazide), which decomposes to give 15% water and 15% nitrogen byweight. This compound was shown to dissolve readily in 3501-6resin at 90C and to dissociate at a suitable temperature (155*C)for volatile formation. A known quantity can be introduced ofboth condensible (H20) and noncondensible (N2 ) volatiles.

Experimental determinations were made using varied concentra.-tions of Nitropore, i.e., 2, 0.2, and 0.02g Nitropore/lO0g resin.The mixtures were subjected to cure cycles with and withoutvacuum. Concentrations in the range of 0.02g Nitropore/lOOgmatrix resin appear to strike a practical balance for futureinvestigations involving volatiles transport mechanisms.

3.4 PHASE IV - RELATE RESIN CHEMORHEOLOGY AND LAMINATE PROPER-TIES -The objective of this phase was to determine whether corre-lation existed between the chemorheo]ogy properties determined inthe first phases of this program and laminate propertiespreviously reported, Reference (1).

The laminate property data available to this program waslimited to the vendor certification and receiving physical andmechancial properties, shown in Tables 12 and 13.

Table 12. Laminate Mechanical Properties

0* Tensile Proper'!es Interlaminar Shear

Strength (R.T.) Modulus (R.T.) Elongation (R.T.) R.T. 250OF 250OFBatch (ksi) (msl) pln./ln. x 103 (ksl) . (ksl) 24 hr Boll

B-1 224, 242, 227 21.2, 20.0, 19.6 11.0, 12.3, 11.8 21.2, 20.3, 20.8 15.0, 15.0, 15.1 10.4, 10.8, 10.6(231) (20.3) (11.7) (20.8) (15.0) (10.6)

B-2 265, 296, 277 20.1, 21.1, 20.4 13.4, 14.2, 43.8 19.5, 19.6, 19') 14.6, 13.7, 15.1 11.4, 11.8, 11.5(279) (20.5) (13.8) (19.4) (14.5) (11.6)

B-3" 257, 274, 269 19.4, 20.2.21.0 13.4, 14.0, 13.3 19.8, 18.5. 18.4 14.6, 14.1, 14.0 12.0. 11.5, 11.4(266) (20.0) (13.6) (18.9) (14.2) (11.6)

Minimum Acceptable

Requirements 200 18.0 10.0 15.0 9.0 7.5

Average"Data from equrvalent batch CIP3340C -24

MCDONNELL 0OUGLAS CORPORATION

31

N00019-81-C-0184

MCDONNELL AIRC~RAFT COwMPANy

Table 13. Laminate F., ical Properties

B -in Flow Volatles PIy 1I ickness ! -lisin Content Fiber Areal WtBach (%) M% ,n) Tc, % gi2

8-1 - - 0.0051 Conforms 41, 40, 40 153, 152, 152(40) (152)

B-2 21.4 0.77 0.0051 Conforms 42, 42, 39 149, 148, 150(41) (149)

B-3* 25.9 1.00 0.0050 Conforms 42, 43, 43 148. 147, 147(43) (147)

Requirements 10 - 30 1.5 Max 0.0052 Adhere to 42 _ 3.0 145 - 155+ / - Itself for

0.0003 10 mm withLess Than

- 10% Damage

) Averagefata Irom equivalent batch OP33000-25

All batches exceeded the minimum property requirements. Notrends were evidenced. Thus, as with the chemorheolgy, thephysical and mechanical tests traditionally used for certificationand receiving inspection were insensitive to variations in theviscosity of the TGMDA component within the suppliersspecifications.

0 MCDONNIELL DOUGLA. CORPORATION

32

Lb

N00019-81-C-0184

MCDONNELL AIRCRAFT COMPANY

4.0 SIGNIFICANT RESULTS

o A viscokinetic model for the curing behavior of 3501-6resin was derived from rheometrics and thermal analysisdata.

- The rheological part of the model describes the flowbehavior of the resin for any cure cycle using amathematical expression relating viscosity to the timeand temperature of cure.

- The thermokinetic part of the model describes theconditions required for complete cure using amathematical expression relating the time for completecure to the curing temperature.

o The viscokinetic model, now available, is a pre-requisitefor the future development of a composite processing modelwhich includes the effect of the reinforcing fibers.

o Total resin flow during cure was found to decrease withincreased dwell temperature and with increased heatingrates.

o The form of the viscokinetic model for curing behavior of3501-6 is applicable to other thermosetting resin systemsand is proposed as an approach to rapid optimization ofcuring parameters for future resin systems.

o The curing behavior was quite similar for three batches of3501-6 resin in which the base resin (TGMDA) viscosity hadbeen varied across the extreme of currently manufacturedMY720 epoxide.

- A comparison of the rheology of the three batches withtheir chemical and laminate mechanical properties didnot evidence any correlateable inter-relationships.

MCDONNELL DOUGLAS CORPORATION

42N__ 33Jeff.

SIN00019-81-C-0184t 9CDONELL AIRCRA 1 COM4PANYV

1 5.0 RECOMMENDATIONS

o Refine the resin rheology model for more exact predictionof experimental viscosity profiles and flow numbers

o Verify generic applicability of the viscokinetic (rheologyand thermokinetic) model to other thermosetting systems.

o Determine the effects of varied autoclave and vacuum bagpressures on rheology and volatiles transport and offgassing.

o Determine the effects of the reinforcing fibers on resinrheology and cure.

o Use a chemical blowing agent for future modeling of thetransport properties of condensable and noncondensablevolatiles.

o Volatile matter requires further study because of thedeleterious effect in composite properties.

t34

NOOU19-81-C-01841MACDONNELL AIRCRAFT7 COAE, PANY)

RFENCES

1. Sewell, T. A., "Chemical Composition and Processing Specifica-

tions for Air Force/Navy Advanced Composite NatrixMaterials", Contract No. F33bl5-78-C-5177, Final Report forPeriod 18 September 1978 to 18 January 1982.

2. Roller, M. B., Polymer Engineering and Science, Vol. 15,

No. b, pp. 40b-412, June 1975.

3. Hollands, K. M., and Kalnin, I. L.., "Epoxy Resins", Advanced

Chem. Ser., -No. 92, bO, Am. Chem. Soc., Washington, D.C.(1970).

4. Kamal, M. R., Sourour, S., and Ryan, M., Technical Papers,31st Annual Technical Conference, Society of Plastics Engi-neers, 187, Montreal (May 1973); Kamal, M.R., Polym. Eng.

Science, 14, 231 (1974).

5. White, R. P., Jr., Polym, Eng. Science, 14, 231 (1974).

b. Mussatti, F. G. and Macosoko, C. W., Polym. Eng. Science, 13,236 (1973).

7. Carpenter, J. F., "Instrumental Techniques for DevelopingEpoxy Cure Cycles", 21st National Symposium and Exhibition at

Los Angeles, p. 783, April 197b.

8. Rush, R. M., "Monitoring the Cure of Therraosets via Thermal

Analysis", Perkin-Elmer TAAS (Draft Copy, dated 5 March

1982).

9. Fyans, R. L., "Computerized Thermal Analysis", AmericanLaboratory, 101, January 1981.

10. Crane, L. W., Dynes, P. J., and Kaelble, D. H., Polymer

Letters Edition 11, 533 (1973).

11. Brown, G. G. and McKaque, E. L., "Processing Science of EpoxyResin Composites", Air Force Contract F33bl5-80-C-5021.

MCDONNELL DOUOLAS CORPORATION35

~~ra


AUCROONELL OQIJOL.A COROPOARATION

36

OIL________________________A,

N00019 -81-C-0184sMCDONNELL AIRCRAFT COMwANV

DISTRIBUTION LIST

(One (1) copy unless otherwise specified)

Address Number of Copies

Commander bNaval Air Systems CommandAttn: AIR-5304 (5 copies)

AIR-311 (1 copy)Washington, DC 20361

Commander 2Naval Air Development CenterAttn: Code 60bWarminister, PA 18974

Commanding OfficerNaval Air Rework FacilityAttn: Code 340Naval Air StationAlameda, CA 94501

Commanding OfficerNaval Air Rework FacilityAttn: Code 340Marine Corps Air StationCherry Point, NC 28533

Commanding Officer 1Naval Air Rework FacilityAttn: Code 340Naval Air StationJacksonville, FL 32212

- I

Commanding Officer 1Naval Air Rework FacilityAttn: Code 340Naval Air StationNorfolk, VA 23511

Commanding Officer 1Naval Air Rework FacilityAttn: Code 340Naval Air Station, North IslandSan Diego, CA 92135

Commanding Officer 1Naval Air Rework FacilityAttn: Code 340Naval Air Station, Bldg b04Pensacola, FL 32508

-MCDONNWELL ZPOUOLA& CORPORMATAO

37

N00019-81-C-0184

AUCDO~tdPEELL AIRCRAFT' COMPAIV

Distribution List (Cont'd)

Number of Copies

Director 3Naval Research LaboratoryAttn: Codes 6383

66546120

Washington, DC 20375

Commander 2Naval Sea Systems CommandAttn: Codes 051

05D23Washington, DC 203b0

Director 2Naval Ship R&D CenterAttn: Mr. M. Krenske & Mr. A. MacanderWashington, ;C 20084

DirectorNaval Surface Weapons CenterAttn: Dr. J. Auyl (R-31)White OakSilver Spring, MD 20910

Office of Naval Research 2Attn: Codes 431

413I i Washington, DC 20350Air Force Materials Laboratory 4Attn: Codes LC

. [LNLTFLAE

Wright-Patterson AFB, OH 45433

Air Force Materials LaboratoryAttn: Dr. J. M. Whitney/MBMWright-Patterson AFB, OH 45433

Army Materials & Mechanics Research CtrDept of the ArmyAttn: LibraryWatertown, MA 02172

NASA HeadquartersAttn: Mr. C. F. Bersch600 Independence Ave., S.W.Washington, DC 2040b

ACDONELL D G, aLAS C ORP TOoI3

N00019-81-C-U184I M~ACDlONNELL AIRCRAF1' COMIPAIIV


Number of Copies

NASA 1Langley Research CenterAttn: LibraryHampton, VA 236b5

NASA 1Lewis Research CenterAttn: LibraryCleveland, OH 44185

Defense Technical Information Center 5Cameron Station, Bldg #5Alexandria, VA 22314

Director 2Plastics Technical Evaluation CenterPicatinny ArsenalDover, NJ 07801

U.S. Applied Technology Laboratory 1AVRADCOMAttn: DAVDL-ATL-ATSFort Eustis, VA 23b04

Brunswick Corporation 1Technical Products Division325 Brunswick LaneMarion, VA 24354

Celanese Research Company 1, SOS 1000

- Attn: Mr.- R. J. LealSunwait, NJ 07901E. I. DuPont de Nemours & Co. 1Textile Fibers Dept

~Wilmington, DE 19898

- Fiber Materials, Inc. 1Attn: Mr. J. Herrick

- Biddeford Industrial Park- Biddeford, MIE 04005

General Dynamics 1Covair Aerospace DivisionAttn: Tech LibraryP.O. Box 748Fort Worth, TX 76101

CDO lELL DOUGLAS COR PORIBO

N00019-81-C-0184l M'CDONNELL AIRCRAFT COMPANY


Number of Copies

General Dynamics 1Covair DivisionAttn: Mr. W. Scheck

Dept. 572-10P.O. Box 1128San Diego, CA 92138

General ElectricR&D CenterAttn: Mr. W. HilligBox 8Schnectady, NY 12301

General Electric CompanyValley Forge Space CenterPhiladelphia, PA 19101

B. F. Goodrich Aerospace & Defense

Products500 South Main St.Akron, OH 44318

Graftex DivisionEXXON Industries2917 Highwoods Blvd.Raleigh, NC 27b04

Great Lakes Research CorporationP.O. Box 1031Elizabethton, TN 37b43

Gruman Aerospace Corp. 1Attn: Mr. L. PoveromoBethpage LI, NY 11714

Hercules Incorporated 1Attn: Mr. E. G. Crossland

Magua, UT 84044

HITCO 11600 W. 135th St.Gardena, VA 90406

Illinois Institute of TechnologyResearch Center

10 West 35th St.Chicago, IL bObl6

MUCDOFNNLL [email protected]* CO MPORATOVl

40

24- M__

Eg - -

N00019-81-C-0184

afVMCDONNELL AIRCR4A1r COI#PWANV


Number of Copies

Lear Fan Corp. 1P.O. Box 60,000Reno, NV 89!Ob

Lockheed California Co. 1Attn: Mr. J. H. WooleyBox 551Burbank, CA 91520

Lockheed-Georgia Co. 1Attn: M. L. E. MeadeMarietta, GA 30063

Lockheed Missiles & Space Co. 1Attn: Mr. H. H. Armstrong

Dept. 62-60Sunnyvale, CA 94088

Material Sciences Corporation1777 Walton RoadBlue Bell, PA 19422

McDonnell Douglas Corp. 1Douglas Aircraft Co.Attn: Mr. R. J. Palmer3855 Lakewood Blvd.Long Beach, CA 90801

North American AviationColumbus Division4300 E. Fifth Ave.Columbus, OH 43216

Northrop Corp.3901 W. BroadwayAttn: Mr. G. Grimes

Mail Code 3852-82Hawthorne, CA 90250

Philco-Ford Corp.Aeronutronic DivisionFord RoadNewport Beach, CA 92663

Rockwell International Corp.Attn: Mr. C. R. Rousseau12214 Lakewood Blvd.Downey, CA 90241

SCDONNELL IDO.QLAS CO1 PORATION

( m o • • 41

LiNO0019 -81-C-0 184

MCONELL AIRCRAFT" COaPANV


Number of Copies

TRW, Inc.Systems GroupOne Space ParkBldg. 01, km. 2171Redondo Beach, CA 90278

TRW, Inc.23555 Euclid Ave.Cleveland, OH 44117

Union Carbide CorporationChemicals & PlasticsOne River RoadBound Brook, NJ 08805

Union Carbide CorporationCarbon Products DivisionP.O. Box 6116Cleveland, OH 44101

United Aircraft CorporationUnited Aircraft Research LaboratoriesE. Hartford, CT 06108

United Aircraft CorporationHamilton-Standard DivisionAttn: Mr. T. ZajacWindsor Locks, CT 06096

United Aircraft CorporationSikorsky Aircraft DivisionAttn: Mr. J. RayStratford, CT 0bb02

University of CaliforniaLawrence Livermore LaboratoryAttn: Mr. T. t. ChiaoP.O. Box 808Livermore, CA 94550

University of Maryland IAttn: Dr. W. J. BaileyCollege Park, MD 20742

University of WyomingMechanical Engineering Dept.Attn: Dr. D. F. AdamsLaramee, WY 82071

*tWCDOWELSL V L CON OAVW4~42

NOUU19-81-C-O184

ACOOWIAELL AI CR4AFT COMAWPA


Number of Copies

Westinghouse R&D CenterAttn: Mr. Z. Sanjana1310 Beulah RoadChurchill BoroPittsburgh, PA 15235

II

43

a;- MUM 511 - . . .. . .. .. •--,l+-

Documents

INet - DTIC