AD-AO96 273 NAVAL AIR DEVELOPMENT CENTER WARMINSTER PA AIRCRAFT --ETC F/G 11/4EPOXY RESIN DEVELOPMENT FOR COMPOSITE FIELD REPAIR.(U}OCT 80 L J BUCKLEY, R E TRABOCCO
UNCLASSIFIED NADC-N0128-60 NL
"llilllll"ln~nnnnunnnnnnnm
11111~* ~ 1 .5
''l1.151*
F-l
MICROCOPY RESOLUTION TEST CHART?ATIONAL BUREAU OF STANDAROS-1963-A
~Imo arm RISIN DEELPIENT FOR COMPOSITE WIED REPAIR
L. J. BUCKLEY
I. E. TRADOCCO
Aircraft and Crew System Technology DirectorateNAVAL AIR DEVELOPMENT CENTER
Warminster, Pennsylvania 18974
8 OCTOBER 1980
FINAL REPORT
AIRTASK 30, ZF61-542-OMl
vw UNIT 0. Em 501
APPIM FO PUILIC RELEE; DISTRI ION Nl ALITED
DTICftepagled for MR1Imueinme efs wav SELECWAA AIR Scoma, Y 12 1A.
hslmfteC 20M'
NOTICES
wW DJMEH- IM in. mrig o 'VI i -~ inpots hind bY Urn bd Ak QPmi40
VW b uII 2*1 rnib ado o r Cuui 01. do Urn fts A&oto ft i~ WAU
qu"L r =ot MWu Nm NAOC7U1521 icntu Urn O~n& Caowr ipu for d. oo 117 IM mol ui1
bydoS o umM i Mw am lm mi
Il cm idr. N.m Ok Oswdopunh CatoSI Talm"o Diref Nod Air &hdp" Censor
is Slow ais I
40a l iiod do.um I Fwojmftasm
IS Msk.. CI 31d T@&Auiq uwa
A5 Fl, dg Asnuu bus..w
FRODI nn EIOsMFr IM. i..udm or haiom u.M rei - cmmr ~dd t d immbb6s na A- b mun imtmdo da Sovammornor do km o anl dght a mmd ado.
MVIVM By: -OAM &eaIMru
UNCLASSIFIED
__________I. TYPE OF REPORT & PgEOD cowSKE
POXRU E RI N tAIATION MEN AOR ftPS I
Warminster,~~~~IMOR PenMlaum19maWR"NI O Z 0
partnt of th OGNIAvyNA9AD DS1
11.~ie CONTROLLNG OFFIC NAMEbAN ADDRES
Nava AiTrIST STtemN Command 8 epmtj
Dar PROeED ofO PUhLe RELAvE DITRUI UOLIITE
~7.DSTISUIO STTEMNT .1~ ~.e ~mEt.S~sl N Stdt~in ken7
Wasintn SUPLEENAR NTE
1. REY WORDS(N hhueemmm. .5*AM G megeemp01 WfMI *f ip WgeI Oftle) I EUIYCAS &W
17. DSRACUTI STATUEMN &1m 8110~ Me5* Itim in 8" 0 NE 4110WOR *4 10u1000
AOTwn" evala tn ow cofducOtd a go, etrmn the* opimm oymratmEroal tobuedsfor compsit eeam Thematria mut aenytempate stReand adeuatxsreha, lvtdtmeaue.Mcaa
CifsAetDifferential canning aloriietryaddnmcmcaia nlsswr
pefreaonten cadcduei system Sh er shearu endlfmerari
strengths were examined for each system. The most promising resin systems
AN,',173EIIWCt ~UU.Y UNCLASS I FIED
T ""MTV CLAIWOCA.Wn OF Tull PgUi
Do* besnu
proved to be stable at room temperature In the B-stage condition. Thesesystems Incorporated a sterically hindered amine hardener that providedlatency. The effects of cloth style on void content and shear propertieswere also examined.
UNCLASSIFIEDUswUYY GL*AMX*~ OF IMl P;;&=. 3m b 0
------- --
TABLE OF CONTENTS
le.b.
LIST OF TAMLSS 2
LIST OF FIGURES 3 -34
INTR.ODUCTOI PRCDR . . . -
DISCUSSION . .- * o o 9 -11
COMCLUSIOUS * . . . . 12
R3~gNATIONS . . . . . . . 0 0 0 12
A~N~L3G~US 0 . 0 . . . . & 13
..........................................
LIST 0F TABLES
Table No. Titlepo N.
I Resin System Evaluated 15
II Short Doe Shear Strength (PSI) of Initially 1Screened System
III Short Bea= Strength (PSI) at Room Temperature,22091 17(104 0 C), and 22007 + Moisture of Candidate System
IV Flexure Strength (PSI) at Room Temperature, 22001 1(104 00, and 22001 + Moisture of Candidate System
V Arnox 3110 Lap Shear Strength 19
VI Percent Change In Short Beam Shear Strength 20
VII Percent Change in Flexure Strength 21
VIII Moisture Absorption Data 22
ix Resin Formulations Evaluated by Differentia 23Scanning Caloribetry
X Resin Formulations Evaluated by Dynamic 24Mechanical Analysis
MADC-80128-60
LIST OF FIGURES
Figure No. Title Page No.
1 Photomicrographa of 332/DUEDA, S-Harnss end 255-Harness Cloth Laminates, 70X
2 Photomicrographs of 332/DENA aud 332/2991=HHDA, 268-Harness Cloth Lainatee, 70Z
3 Photomicrographs of 332I1299/iNHA Showing Void 27Content Decrease In 5-Harness Cloth Laminates, 701
4. Photomicrographs of Arnox 3110 and 332/1299/SEUA, 288-Harnees Cloth Laminates,* 701
5 DuPont Differential Scanning Calorimeter with 291090 Therml Analyzer
6 DSC Plot of 332/DHHD In the Freshly Mixed Condition 30
7 DSC Plot of 332/MMHA in the 5-Staged Condition 31
8 DSC Plot of 332/299DNIIA in the Freshly Mixed 32Condition
9 DSC Plot of 332/1299IDHHD In the 5-Staged 33Condition
10 DSC Plot of RDG%/MD in the Freshly Mixed Condition 34
11 DSC Plot of RDG2K/MD in the 5-Staged Condition 35
12 DSC Plots of RDGI/HD In the 5-Staged Condition 36Showing Mdvancaent with Tim
13 DSC Plot of RDGE/MNHD In the Freshly Mixed 37Condition
14 DSC Plot of RDGEIDNUA In the 5-Staged Condition 38
15 DSC Plot of 332 1'D In the Freshly Mixed Condition 39
16 DSC Plot of 332/ND In the 5-Stage Condition 4
17 DSC Plot of Arnox 3110 4
18 DuPont Dymmidt Mehamical Analyzer with 109 41Thezmul Analyser
LIST OF FIGURES -
Fiiu@ so. D-0*l ?go No.
19 DN& Plot of 332113UDA, 8-lam...s Cloth~ Laminate 43
20 DNA Plot of 3321D26M, 5 Harness. Cloth Liate 4
21 BMA Plot of 332/1299mUD&, 5-Narnes. Cloth Laminate 45
22 DNA Plot of RDGZlUSIEA, 5-BNexuss Cloth Laminate 46,
23 DNA Plot of Arox 3110, 8-8sivass Cloth Laminate 47
4
NADO-0012 860
INTRODUCTION
The use of graphite/epoxy as an aero-structural material has Increaedgreatly over the last few years. The Navy, with the 7-18 and AV-85 aircraft,uses graphite/epoxy to such an extent that it has become necessary to considereffective repair of these structures. Graphite/epoxy structures are usuallyfabricated by the application of heat and pressure in an autoclave. Repairsmade with the use of an autoclave, where such procedures are possible, do notpresent any serious technical problems. Repairs made where an autoclave isnot readily available, such as on an aircraft carrier, require materials thatcan be cured with vacuum pressure only to form high Integrity laminates.
Storage of repair materials may also be a problem. For exmple, mostresin systems when mixed must be used imediately or else, when B-staged,must be scored at low temperature (OF, -18oC).
An ideal resin system for repair of graphite/epoxy structures should havethe following characteristics:
(1) Ambient or low temperature cure;
(2) Low viscosity;
(3) Room temperature storage in the B-stage condition;
(4) Short time cure (I hour or less);
(5) Elevated temperature and moisture resistance; and
(6) Vacuum pressure during cure.
Item (1) eliminates the problems associated with conventional heat sourcessuch as heating blankets and hot air guns. Current technology, though, doesnot permit elevated temperature performance of room tempertule cured epoxyresins. Therefore, a moderate temperature cure (2000F to 3000F; 930 C to 1490C)with glass transitions at or above the cure temperature is the next best approach.
Low viscosity is preferred to enable the resin to flow in between thefibers. Room temperature storage,either unmixed or mixed (in the B-stage),isa necessity for a repair system on the basis of supply and storage at the fieldlevel. A short time cure of 1 hour or less decreases the total repair timwhich decreases the down-time of the aircraft.
Elevated temperature and moisture resistance is best achieved by havinga higher glass transistion temperature.
Current composite repair techniques include Gruman's F-14 boron/epoxyrepair kit (Ref. D). This kit is intended for repair of the horizontalstabilizer which is constructed of titanium and boron/epoxy face sheets overan aluminum hdn4yeomb core. Titanium foil is used with a film adhesive whichmust be stored at OF, -18oC.
Current techniques with graphite/epoxy Includes precured patches that areadhesively bonded to the composite structure. The patches are procured andtherefore limited to flat areas of the structure. One part of the soaiivesystem has a room temperature shelf life of only 3 months. This is clearly
5* k4 t
MADC-M0128-60
unacceptable for field level repair where the distribution of materials canbe very time-consuming.
Bolted or mechanically fastened repairs have been successfully demon-strated by McDonnell Douglas for graphite/epoxy lminates 3/16 and 1/2 inchthick with through-the&-thickness hole damage (Ref. 3). These repairs weredeveloped for application on fuel cll composite wing surfaces using titaniumalloy patches and backing plates. The problems with this type of repair isthe damage to the composite caused by the bolt hole drilling and the aero-dynamic Incompatibility of the titanium patch. Also, a bolted repair systemcannot be used on honeycomb structure.
The optimum situation for the field repair of composite structures isto use both methods of repair, bonded or bolted, depending on the locationof the damage. Honeycomb, for example, requires bonded repair methods whilethick, flat, non-honeycomb structures are suited to bolted repair.
These considerations led to the use of a resin system that is stable inthe B-stage at room temperature. It consists of a difunctional epoxide resinwith a tetrafunctional amine hardener. This system was developed by J. Rindeat the Lawrence Livermore Labs. (Ref. A). The resin is a diglycidyl ether ofbisphenol-A; Its advantages consist of high purity and low viscosity. Otherresins, such as polyglycidyl ether of orthocresol-foraldehyde novolac andresorcinol diglycidyl ether were used either as an additive or a substitute.
The hardeners used were 2, 5-dimethyl 2, 5-hexane diamine and menthanediamine. Both of these tetrafunctional hardeners are sterically hindered toprevent or delay the secondary amine hydrogen reaction.
Other systems and modifications are presently being explored. Attemptsare being made to decrease the time and temperature of the cure and extendthe storage life.
DER 332 - Diglycidyl ether of bisphenol - A
RDGE - Resorcinol diglycidyl ether
DMHDA - 2, 5-dluethyl 2, 5-hexane diamine
ECN 1299 - Polyglycidyl ether of orthocresol-foraldehyde novolac
NC 513 - Aromatic ether (epoxy reactive diluent)
MD - Menthane dismine
EXPERIMENTAL PROCEDURE
This investigation was conducted in two phases. Phase I includedepoxy resin selection and evaluation by mechanical properties of a curedgraphite cloth laminate. Phase I consisted of evaluation of various epoxyresin systems through thermal analysis techniques, specifically DSC (dif-ferential scanning calorimetry) and WNA (dynamic mechanical analysis). Alsoincluded in Phase I was a check on laminate quality using optical microscopy.Initially, all evaluation was based on mechanical properties. The aquisit ionof thermal analysis equipment added a second dimension to the resin eVeluatis.
6
HADC-9012--60
Phase I
An industry and literature search led to the selection of candidateresin systems. Laminates were fabricated using these resin systems in
conjunction with Hercules A-370-O1 and A-370-5H graphite cloth. Their layupwas a 9" x 9", quasi-isotropic type using (. 45, 0, 90)s . The candidateresin systems were mixed, when necessary, and applied to the cloth. The resinand the curing agent was reacted at a stoichiometric ratio which was calcuatedas follows:
(Au) (100)
phr - REQ ka t Amine Hydrogen Equivalent Weight
phr - parts by weight of REQv Epoxide Equivalent Weight of Resinmine per 100 partsresin
The resin was distributed throughout the laminate by the use of a rolling pin.Excess resin was removed by vacuum bleeding before cure. Room temperaturecures were run using a mylar bag with a 10 oz. fiberglass tooling cloth asa bleeder material. Higher temperature cures were run using vacuum pressureonly. All laminates, with one ex~eption, which employed either UHHDA or MDas the hardener were cured at 260 F for one hour. There were no elevatedtemperature post-cures.
Test specimens were cut from the laminates to evaluate the followingmechanical properties:
(1) Short beam shear strength at room temperature, 2200F (104°C), and220oF + moisture.
(2) Flexure strength at room temperature, 2200F, and 2200F + moisture.
All mechanical testing was done on an Instron test machine. Elevated temp-
erature short bean shear tests were usually performed first to determine if
further testing was justified. Data was collected and reduced for eachlaminate. The moisture content was obtained by placing test specimens in a
humidity chamber set at 100% relative humidity, 1400F (600C), for a minimum
of 96 hours.
Phase II
Thermal analyses were conducted using DSC and M& on cured and uncuredresin systems. The cured laminates were sectioned, mounted, and polishedto examine their cross-sections. A Buehler "Minimet" Automatic Polisherwas used to complete the polishing.
Photomicrographs were taken of the near surface and center cross-sections
to determine void content and resin distribution.
7
NADC-8128-60
RESULTS
A list of all the resin systems evaluated is given in Table I. Shear andflexure data are presented in Table II for the system that were initiallyscreened. Tables III and IV show shear and flexure data for the more promisingsystems. The testing was done at room temperature, 2200F (1040 C), and 220oF withmoisture.
The mechanical properties improved as the number of voids in the lminateedecreased. Figures 1 through 4 show photomicrographs (at 70X) of the laminates.Laminate quality improved with fabrication experience as shown by the decrease inthe size and number of voids.
The tensile lap shear data for Arnox 3110 Is presented in Table V. Arnox 3110
was evaluated in two ways, as an adhesive and as a laminating resin.
The percent decreases in shear and flexure strength from room temperaturevalues are presented in Tables VI and VII. Table VIII shows the moisture absorp-tion data obtained for the laminates. It can be seen that the specific type ofweave, 8-harness or 5-harness, affected the moisture content.
Thermal analyses were conducted using a DuPont 1090 thermal analyser withDMA and DSC modules. Figures 6 and 7 show DSC plots of 332/DMHDA in the freshlymixed and B-staged conditions. Figures 8 and 9 show similar DSC plots for the samesystem with the ECN-1299 additive. Figures 10, 11, and 12 show the DSC plots ofRDE/MD with respect to time. Figure 10 depicts material in the freshly mixedcondition; figure 11, in the B-stage (4 days later); and figure 12, an additionalday at room temperature. Figures 13 and 14 show DSC plots of RDGE/DKHDA inthe freshly mixed and B-staged conditions. Figures 15 and 16 show the same typeof DSC plots for 332/ND. Figure 17 shows a DSC plot for Arnox 3110. The Arnox3110 resin system is premixed and stable at room temperature in this condition.
The peak exotherms and melt endotherms are summarized in Table IX for theDSC plots.
Dynamic mechanical analyses (Fig. 18) were performed on cured laminatesections. Due to the dimensional restrictions on the DMA sample, the laminatethickness became the sample width by taking a very thin slice of the laminate.This meant there were no continuous fibers between supports. Thus, the matrixwas evaluated with as little contribution from the fibers. as possible. The .
DNA plots of 332/DMHDA 8-harness and 5-harness cloth laminates are shown inFigures 19 and 20. There were no significant differences between the 8-harnessand 5-harness cloths. Figure 21 shows the IMA plot of 332/DMHDA with theECN-1299 additive. The frequency and damping plots (modulus and tan .) areshifted to the right which shows an increase in the glass transition temperature.Figures 22 and 23 show DNA plots of RDGE/DNHDA and Arnox 3110. The RGE/DNHmhas a relatively low Tg. All of the Tg's and tan 8 peaks are summarized inTable X.
DISCUSSION
Of the resin systems investigated the room temperature curing systems likeDow's D.E.R. 332/D.E.H.20 and CIBA's 6005/0500/HY956 flowed when tested at ele-vated temperature (22001) and thus were not further evaluated (see Tables I ad11). Shell's Spon 828/Z system, which was cured at 3000F, yielded acceptable
j8
HADC-80128-60
mechanical properties but was difficult to work with due to the danger of theZ catalyst. This would be especially true in a field or Intermediate-levelmaintenance type of environment. The Apco resin system was much too viscous tobe used as a laminating resin. Epoxylite 8130 did not have adequate strengtheven at room temperature. The Iaidazoles were examined very briefly and showpotential as adhesive-type curing agents. Because they were in the form of apowder, mixing was a problem. The Arnox 3110 resin did not have the mchanicalstrength either as a laminating resin or as an adhesive. (See Tables III, IV,and V). This material was fabricated early on in the program and evidencedvoids in the cured resin. It is planned to obtain a second batch of thismaterial to substantiate initial mechanical property results.
The 332/WIDA system with and without modification showed potential.These resin systems were more thoroughly investigated with hot-wet mechanicalproperties and thermal analyses. The RDGE/MD, 332/MD, and RDGE/IMADA systemswere also examined with thermal analysis (specifically DSC).
Important considerations for any potential repair system are the matrixdominated mechanical properties. One such property used for material character-ization is short beam shear strength. Short bean shear properties are anexcellent measure of overall laminate quality. They are a resin-dependentproperty and therefore very sensitive to temperature. Adequate elevated temper-ature shear properties require good fiber - resin adhesion and higher glass-transition temperature than the test temperature desired.
Good fiber-resin adhesion can be achieved or improved by decreasing thevoid content in the laminate. This was attempted in two ways. First, a 5-harnesssatin graphite cloth was used which is a more open weave of cloth than the 8-harness satin. (See Figure 1). Second, a reactive diluent, Cardolite NC-513was used. This low viscosity fluid has one epoxide group per molecule and,therefore, combines chemically in any cured epoxy resin formulation. As littleas 10% by weight reduces the viscosity of a liquid epoxy by 70Z. Lower viscosityimproved wetting and penetration into the laminate and also reduced air entrap-ment. Photomicrographs showed that the number and size of voids n the cross-section decreased with the addition of Cardolite NC-513. (See Figures 1 and 2).This addition, however, has also decreased the elevated tempeinature strength ofthe resin. This reduction in strength is due to the plasticizing nature of theadditive. The failure mode changed at this point due to the influence of theadditive. Instead of a typical shearing type of failure, the sample plasticallydeformed. This decreases the overall load bearing capability of the laminateby making it easier for the resin to flow.
The mechanical properties that were evaluated for the 332/DEDA system seemto depend on the processing conditions. Because the graphite cloth was impreg-nated manually, penetration of the resin into the cross-weave of the cloth hasnot been optimized. It is believed that machine prepregging of this systemwould improve the mechanical properties. Efforts are presently underway tohave this attempted.
The photomicrographs show the 5-harness cloth to have better resin penetra-tion than the 8-harness (see Figure 1). The addition of NC-513 to 332/UEDAimproved penetration and flow of the resin but also decreased the umchanicalstrength of the composite. Both shear and flexure values decreased much morewith temperature than the other resin systems (see Tables VI and VII). The
9
MADC-80128-60
1CN 1299 additive did not show much of an effect with respect to the voidcontent. The processing (hand laminating procedure) greatly Imroved withexperience (see Figure 3). Consequently, the laminates fabricated towardthe latter part of this investigation contained less voids and were of a higherquality than the laminates fabricated early in the study. Laminates usingArnox 3110 contained a higher void content than laminates of the other resinsystems.
The addition of CIBA-eigy's ECN-1299 to the 332/M NDA system was made toimprove the elevated temperature performance. PM-1299 is a polyglycidyl etherof orthocresol-formaldehyde novolac with an epoxide equivalent weight of 225.It requires mintmm chain extension and crosslinking reactions to reach a highmolecular weight and crosslink density necessary for elevated temperature per-formance. Short beam shear and flexure data do not show my significantimprovement of elevated temperature (See Tables III and IV). Laminate qualityof 332/1299/D MDA, as shown in Figure 3, Improved to the extent that voids wereno longer a major problem. As shown in Table VIII, the decrease in short beanshear strength from room temperature to 2200F (1040 C) and 2200F plus moisturewas about the same or less without the 10N-1299 additive. The same behavior isnoted in flexure strength in Table IX.
The mechanical properties of the Arnox system did not decrease as much asthe other resin systems with elevated temperatures and moisture (see Tables IVand VII). This system has relatively low strength to begin with (at roomtemperature) as is shown in Tables III and IV. Adhesive lap shear data usinggraphite/epoxy coupons also showed that the system has low strength (seeTable V).
Moisture absorption data in shown in Table VIII. The 8-harness laminatesabsorbed twice as much moisture as the 5-harness laminates with the same resinsystem. In general, the 5-harness laminates had less voids (due to better
* 'resin penetration) than the tighter weave 8-harness. A greater void contentprovides a greater volue for absorbed moisture. The NC-513 additive seemed
, Ito cause an increase in moisture content over the base system. The Arnox 3110laminate contained the highest void content of all (see Figure 4) and conse-quently absorbed the greatest amount of moisture.
A higher temperature cure of 3000F was attempted with the 332/1299/MDNsystem. No significant increase in mechanical properties was noted. Therewas no improvement even at elevated temperature. This laminate contained voidswhich could have adversely affected the mechanical properties.
The differential scanning calorimetry apparatus is shown in Figure 5. The332/DDA system shows two ezotherms when examined in the freshly mixed condition(see Figure 6). These exotherms correspond to the primary and secondary hydrogenson the amine groups in the hardener (2, 5-dimethyl, 2,5-hexante dimined). Asshown in Figure 7 with the B-staged system, only one broad exotherm remains alongwith an endotherm., The single, broad exotherm represents the reaction of thesecondary amine hydrogen and the endotherm represents the melt of the uncrosslinkedpolymer material. This sam type of behavior is noted when CW 1299 Is added tothe 332/MD A system. (Sea Figures"8 and 9). In this case, though, the melt
r endotherm is not quite as deep.
The RDGf/ND seemed to cure at room temperature with time (see Figures 10 and11). The curing agent, enthane dismine or 1, 8-dimino-p-suthane did not pro-
10
NADC-80128-60
vide the latency that was expected. As shown in Figure 12, the exotherm diminishescompletely with an additional day at room temperature. No melt endotherm wasobserved with RDGE/MD after storing at room temperature in the B-stage. TheRDGE/DNNDA did show a slight melt endotherm and latency (see Figures 13 and 14).With additional time at room temperature the melt endotherm decreased and theresin system approached the fully cured condition. The 332/D system did notcure at room temperature. It remained stable with a melt endotherm and a curingexotherm (see Figures 15 and 16). Arnox 3110, with the advantages of room tempera-ture stability in the mixed condition and fast cure, has a sharp exotherm atabout 1000C (2120F) (see Figure 17).
With the exception of Arnox 3110 all of the systems using the resins, DER332, RDGE, and ECN-1299 and the hardeners DKHDA and ND showed two DSC exothermswhen examined in the freshly mixed condition.
The RDGE/ND resin system cured at room temperature as is shown by the absenceof a curing exotherm after B-stating. The 332/DHMDA, 332/1299/DHIDA, and 332/1Dresin systems did not advance to cure at room temperature. These systems melted,as shown by the endotherms (see Table IX), and cured as shown by the exotherms.They did this after being stored at room temperature in the B-stage. During thistest all of the resin systems were left exposed to a laboratory environment. Noefforts were taken to prevent moisture contamination which, to the extent itoccurs, advances the cure. Considering these points, the above resin systemsshow promise as future composite repair materials. The dynamic mechanicalanalysis apparatus is shown in Figure 18. The dynamic mechanical analyzeryielded measurements of the glass transitions and tan 6 peaks. The MIA plotsshow the resonant frequency and damping (energy dissipation) as a function ofthe sample temperature. The resonant frequency is directly proportional tothe elastic modulus. The damping or energy dissipation corresponds to tan 8values. The DNA plot for 332/DMDA, 8-harness fabric is shown in Figure 19.The dynamic glass transition and tan 6 peak can be seen as 1330C and l500Crespectively. The same Tg and tan 6 peak was found for the 5-harness fabric(332/MHDA). (See Figure 20). Therefore, the type of weave of reinforcingcloth has no effect (as was expected) on the frequency and damping responseof the sample. The addition of ENC 1299 was shown to increase the Tg and tan 8peak (see Figure 21). The drop-off In the frequency (modulus) curve s moresudden. This could mean that the modulus is less affected by temperature untilthe Tg is reached. Once the Tg is reached, a sharp transition occurs. Theadvantage to this type of behavior (as opposed to a more gradual change inslope of the curve) is a higher use-temperature.
RDGE/DMADA has shown stability in the B-stage at room temperature. TheTg and tan 6 peak for this material occur at a lower temperature than the othersystems (see Figure 22). The plot for Arnox 3110, 8-harness fabric, Figure 23,gradually changes slope with a lesser decrease in frequency than the othersystems. Table X lists all of the glass transition and tan 8 peak taupertures.The Tg was taken as the intersection of the two tangent lines drain to fit thestraight line portions of the curve. This glass transition temperature Is oftentaken as the use or service temperature. The tan 6 peak, which Is 200 to 300 Chigher, is referred to as the dynamic Tg. Because crosslinking of the polymerrestricts the mobility of the polymer segments, higher dynmic Tles areassociated with more efficient cross-linking.
11
OWI N I I.. . . . . . .-- ,' "
ADC-80128-60
CONCLUSIONS
1. The most promising resin systems based on the ground rules of short time,low temperature ,vacuuIm pressure cure and long shelf life at room taeraturein the B-stage were:
(1) 332/DDA
(2) 332/1299/D IDA
(3) RDGE/DMHDA
(4) 332/MD
2. The ECN-1299 additive increased the glass transition temperature of the332ADHDRA resin system.
3. There were less voids in the laminates with the 5-harness graphite cloththan with the 8-harness cloth.
4. The NC-513 additive lowered the viscosity of the 332/IMIA resin system.This addition also caused a marked decrease in the elevated temperaturemechanical properties.
5. The Arnox 3110 resin system evidenced low mechanical properties.
6. Dynamic mechanical analysis indicated glass transition temperatures nearor above the cure temperatures.
RECOMMENDATIONS
Additional efforts in this area should be directed towards increasing thestorage time of the B-staged resin. The advancement of the system should bestudied as a function of storage time sd emvri t 1 oisture Is thoughtto cause crosslinking in the system by allowing the secondary smine hydrogensto react. This effect along with how such flow occurs after time in theB-stage should be examined. During the initial part of this work, all resinformulations were mixed stoichiomtrically. Varyin the amounts of hardenerabove and below this point could Improve the latency and mechanical properties,especially at elevated temperature.
Msuthane diamine with more steric hindrance could Increase the stabilityof the resin system and thus offer a longer storage time in the B-stage.
Small-scale repair simulations using the above mentioned materials withadhesives, If needed, are a natural follow-up to this work. A compatibleadhesive must be found in this case with similar storage capabilities
12
NADC-0128-60
ACKNOWLIODGSIN TS
The authors would Igo to thauk Nomrs. K. Crlark, T. n l,,R. McCartney, D. McCauley and Mm. P. 1oUsW for their asIptem An fabucatitand testing laminates with the a repair system.,
1
t~L 13
Il
NADC-80128-60
(a) Rinds, J. A., et al. "2,5-Dlaethyl 2, 5-Hexmn Diamine:A Promising New Curing Agent For Epoxy Resins," 11th NationalSampe Technical Conference, Nov. -1979, Boston, MA.
(b) Mahon, J., "The Development of Large-Area Damage Rep airProcedures For The F-14 Boron/Epoxy Horizontal Sbabilizer,"U. S. Navy Contract No. N00019-77-C-0032, Grman Aero-space Coep.
(c) Watson, J.*, et al, "Bolted Field Repair of Composite Structures ,"U. S. Navy Contract No. N62269-77-C-0366, McDonnell Douglas Corp.
(d) Carpenter, J., "Teat Program Evaluat ion of Hercules 350 1-6 Resin,"McDonnell Douglas Corp., St. Louis, NO., May 1978.
(e) Crabtree, D., "Room Temperature Curing Resin System forGraphite_/Epoxy Compsite Repair," U. -5. Navy Contract -No.562269-79-C-0224, Northrop Corp., Huuthorne, CA., Dec. 1979
14
N.A-002S0
TABLE I
RESIN SYSTUS lYMUATID
RESIN ADDITIVES HARDENu RESIN NWtlhu
Spon 828 -- z SellEpon 815 - Z SbellEpoxylite 8130 - 8130 PoZylte Inc.Apco 5393 - 5393 Applied Plastics Inc.Der 324 DER20 Dow
Araldite 6005 Araldite 0500 K956 CIDA-GetyDer 332 -- DMA DoM
Der 332 ITA/ Dow
Der 332 NI-I DowDer 332 Cardolite NC-513 DMDA DowDer 332 EWN 1299 1NHDA DowDer 332 -- 2PZ-OK Dow
Der 332 -- 2PZ-4M Dow
Der 332 2P4MZ DowArnox 3110 .... General Electric
DGE -- Menthane Diamine DuPont
Der 332 -- Menthane Dimine Dow
RDGE -- DMHD DuPont
NipS
': r,
TAUSI 11
SHORT um su Room VT3Uil orSI OF fgx KNRTILL SCEIMST
RESIN SYSTEM ADDITIVE ROMK TUPUk~f3 22007 (104C
Epon 828/2 -6531-
Epon 815/z- 63 1117
Der 324/D31 20 -- 3516 794
CIBA 6005/H"Y56 CISA 0500 5247 1717
Der 332/DNHDA Se"86 4264
Der 332/NI-I - 4323 1167
Der 332/DMHD MC-513 6358 3587
Der 332/~DMA 10U1299 6185 4199
Arnox 3110 -- 4352 3854
77
MADC-0128-EO
laa
0 an do %0
~4
i"f0~~s0o in r%4"0 a i
SAD-O12-60
C4-
In C4 00 fet %D0 0 -%0 -+a C14a a
s ~ C' -Z m t
0-4
M inoI
N C*4
I. Pil 11. f N
RAD-80128-60
TABLE V
ARNOX 3110 LAP SIMhZ STREOMT
TYPE OF COUPONS TENSILE LAP SHEAR STRENGTH (PSI)
Graphite/Epoxy 939
763
882
961
643
AVERAGE -837.6
L9
1ADC-8o128-60
TALS VI
nCr CHANCE IN SHORT UM SHEAR STRNGTH
RESIN SYSTSM CLOT TYPE R.T.-*220°F R.T.-.220°1 + o
332/DI MA 81 -29.7% -40.22
332/DNMDA 51 -35.52 -37.32
3321513/1HHDA 81 -43.52 -76.7%
332/1299111DA 81 -32.12 -50.3Z
332/1299/--DHDA 5H -31.72 -46.0%
Arnox 3110 8H -11z -21.32
20
IIII I I
IhMC-00129-W0
TABLE VII
PERCENT CHANICE IN FLEXURE STRENGTH
RESIN SYSTEM CLOTH TYPE R.T.-122 0 R.T.f220 0 F+ H 0
332/DMDA 8H -22.62 -30.01
332IDNIIA 5H -20.01 -43.6Z
332/513/DHDA BR -60.11 -82.21
332/1299/Dt4HA 81 432.82 -60.52
332/1299/1MIHA 5H -40.21 -43.61
Arnox 3110 8H -28.32 -41.82
21
UMD-801 28-60
TABLE VIII
MOISTURE ABSORPTION DATA
RESIN FORMULATION CLOTH TYPE 2 INCREASE TIM (DAYS)
332/DNHDA SR .83 7
332/UNUDA 5H .31 7
332/1299/DMHA 8H 1.06 7
332/1299/DM9A 5H .48 7
332/513/DMHDA 8H 1.54 8
Arnox 3110 A 8R 1.64 8
.22
TAKLEI
RESIN FORMULATIONS EVALUAT~ED BY
DIFFERENTIAL SCANNING CAtLORDIETRY
FOSMAMflN PEAK EMXOER TEPERATURE( 0 C) MEL !!TIM~eC
332/UWUDA (Freshly mixed) 104,147
332/1299/DWHDA (Freshly mixed) 100,144
RDGE/MD (Freshly mixed) 87,134
RDGE/T*INDA (Freshly mixed) 83,123
332/MD (Freshly mixed) 109,161
332IDMHDA (B-stage) 133 50
332/1299/HHDA (B-stage) 138 52
RDCGE/MD (B-stage)
RDGE/TJ4HDA (B-stage) 114 60
332/MD (B-stage) 154 42
LI 23
XADC"601 2640
RESIN FORNUATIONS EVALUATI BY
P~jKC M3AMIC&L ANALYSIS
332/DNDHA, 85 133 ISO
3321DHHD, 511 133 150
332/1299/RIHA, 511 144 160
Arnox 3110, 8H 140 171
RDGE/DHHA, 5H1 106 124
24
NADC-601 28.60
j 332/DMHDA, S-HARNESS SATIN WEAVE
FIGURE 1. A VOID CONTENT COMPARISON OF THE $-HARNESS AND5-HARNESS GRAPHITE CLOTH USING THE 332/DMHDA RESIN SYSTEM.
25
f NMADC-SO0128.60
332/513/DMAHDA, S-HARNESS
332/1299/DMHDA, S-HARNESS
FIGURE 2. A RESIN FLOW COMPARISON OF THE NC..S13 AND ECN1299 ADDITIVES TO THE 332/DMHDA RESIN SYSTEM.
26
HADC-801 28.60
il
332/1299/ DMHDA, 5-HARNESS
FIGURE 3. A COMPARISON OF THE LAMINATE QUALITY(RESIN DISTRIBUTION)IMPROVEMENTS WITH EXPERIENCE.
27
NAIDC-801 28.60
332/ DMHDA, 5-HARNESS,
ARNOX 3110, S-HARNESS
FIGURE 4. A VOID CONTENT COMPARISON OF 332/DMAHDAAND ARNOX 3110.
28
NADC- 80128-80
Iz
J
z
II-
0
z
j~l IL
0
290
?4ALC-80128-80I.
:a ** Oil S 111100 361
,10U
0mlx
3 2
-IIma
I
Uw Ii. I0 *
L3 0
I-
0
3 0S0
* I
_________ * di
F .3 d d d 'p 'p 'aCa., u ~in~g
30
NADC-6OI26-~
I
3**II
I
S4I-
N UU
I-
I
U*0 RI n
N.-L
* 0
3 I-* 0
a
I.3PU.
d d 'p'Ii: ii~
~AU) -u ~H
31
ImenC-SO 2S-60
0
E
a
IL*0
IL
dm dCAN) &lid V
32 -
NADC-80 12860
I75
I0
16
MIA V"
33(
77-
NA- C-8O1 28-60
0
z
0
+C
I 3II
NADC-SI2S-~O
I'U.
I* z0U
0* MI
(I-S* I.
.ini
zU a~1J 3.3
0 MII
I-0
* -J
* US* a
1~I-
0h.2S
CU
.3 ~3 ~ gigA I.3 d d d .3 .1
£
55
s..-,- ~
NADC-80128-60
- I-
0z
° z(N Z
(A.,(
0
0
0 1a
36
M m
KADC81260
I0
lbld d dan im 106
37I
NADC-S0i 2840
U
I...
Iz* 0
(I0
I 4
I aUzI-
KL~) (
uJ 0N
0 aUSI.-
0
I I-0
3 1I It
I.
b3
* PS.
4 d d d 4 41' £ I
(As, -U ~W
3.
NADC-SSi2~~O
asses as asia as:
I* (3
II. N
Imaz
Li I-
NI-L0I-0
U a
* SI.
3 b.
U
d .3 .3 d d
qAu) -~ ~
3,
-~ .L~.,,
r NADC-601 26-60
sgml sq I 5~g~*l g g*~g 5
IN
I2
* 5-S
Ia
9 DCo IU)
0 :1!I
Ma
U 0b
I
(P)U1U~H
40 -'4
NALC-h12S-~@
U
I* K
0zg
(U p
I-U) 0
o iiiIs-
f.m.
* 2
a
U
I. U
U&-'~i 1r4-: 4~ '4 S J* g ~p *
41
I-
VIADC-801 28-60
an@ J
N44
(
z0
424
- - I
WAPC-aoIzS.49
I (A~ Ssq~m 0 C-2
I I I'. I,;aIU
I
0 *1I
II I"pII0 15
4 ~
4
SU
'I
~f. -~
.12
.7' . *
*t'~fi" ~ - I.
-
NAIC-SOlIS 40
I. (A) ESdIdinQ 1-2
I I I~ I I * U___________________ Up g~rp I *S I6SOS~
SU
** a
U UHg
p
II0
0
I-0
UI
I aU
m (-2
4
A.. tI1
' .
~c-So1 2S-~O
U 'ml0 00 63 30060 30 30 -
4
US 0
-aI
I6(
RI3-
* 0a
* U5SU
31
(q. - ~4
,1~.
NADC-801 aS-a
(A) SMIdUSa c-3
3 3 1 3 3S I I S I I S I I S a p p
UI-4
* N4-I *
zI..
* 0~1U
6Smlz
S
4 40Ix0
0 SISg
0a.
0-0-l
3 0. 4-
4No
Uaa03SC
IS, Sm IU *'~'k'~'~A U
NAD-801 28-60
333I 1113 330
IP
*W& ~womso POMONS*rp $ Ii-- 0-9"0"
47'..... "Mo
DISTRIBUTION LIST (Coutlamed)
Kingg o. UI0Co 2S-60
AIRTASK NO. ZP6l-542-0lI=KUNIT NO. ,( 501
no. of copes
NAPTC, Trenton,(J. Gla) . . .. .. . ... .. .. . . .. 1Cwu~u~wCN, Cina Lake,. CA........ .......... 1
Celanese Research Co., Summit, NJ (R. J1. Leal) . . .
NAVSIA, Washington, DC (Code6101x) ........ ..... 1Vought Corp., Systems Div., Dallas, TZ (G._ ourland) . . . o . 1NASA Headquarters, Qahington, DC (Code V-2 N. Meyer) . o . . . . 1The Bosing Company, Aerospace Div., Seattle, VA. . 0 . .. . . 1Boeing-Vertol Co., Phila., PA Dept. 1951 ........ *... 1NAVSBIP R&D rnta , Bethesda, ND (4. renske, Code 1720) . . . . . 1NASA, Langley Research Center, Hampton, VA (W. Novell) , . . 1United Aircraft Corp., United Aircraft Research Labs,
last Hartford, CT 06108 .. . .............United Aircraft Corp., Pratt & hitney Aircraft Div., 1
last Hartford, CT 06108 . . . . . . . . . . . . . . . . . .United Aircraft Corp., Hamilto-Standard Div., 1
WIndsor Locks, CT 06096 (T. Zajec). . . . . . . . . . ..United Aircraft Corp., Sikorsky Aircraft Div., 1
Stratford, CT06497 (J. Ray)..... * .... ....Union Carbide Corp., Carbon Products Div., Cleveland, OR. . . . . 1Philco-Ford Corp., Aeronautic Div., Newport Beach, CA. . . . . . . 1University of Maryland, College Park, MD (Dr. W. J. Bailey). 1University of Wyoming, Mechanical Engineering Dept.,
Laramie, WY 62070 (Dr.. . Adms). . . . . . . . . . . .. 13M Compamy, Medical Products Div., St. Paul, WI 55144 (J. G. Blair) 1Amy Materials and Mechanics Research Center,
atertown, MA 02172 (G. L. 8agnauer, R. 1. Saeir) ..... 23. 1. DuPont Co., Instruent Products Div.,
Wilaington, DR 19805 (t. L. Blaine) .... ..... . 1onR, Beasten Central Regional Office, Boston, NA 02210
(Me Nelen voloch). . . . ... . . . . . . ...... . 1nTrC . . 12
-,add
