UNCLASSIFIED

AD NUMBER

LIMITATION CHANGESTO:

FROM:

AUTHORITY

THIS PAGE IS UNCLASSIFIED

AD880416

Approved for public release; distribution isunlimited.

Distribution authorized to U.S. Gov't. agenciesand their contractors; Critical Technology; NOV1970. Other requests shall be referred to ArmyAviation Materiel Laboratory, Fort Eustis, VA.This document contains export-controlledtechnical data.

usaamrdl ltr, 23 jun 1971

© <» USAAVLABS TECHNICAL REPORT 70-65

< STUDY OF THE PERFORMANCE OF ADHESIVE BONDED JOINTS

i

o

Li_

('.Li

FINAL REPORT

Harold K. Shen

lohn L. Rutherford

November 1970

U. S. ARMY AVIATION MATERIEL LABORATORIES FORT EUSTIS, VIRGINIA

CONTRACT DAAJ02-69-C-0094

KEARFOTT DIVISION

SINGER-GENERAL PRECISION, INC.

LITTLE FALLS, NEW JERSEY

I hll .1», uturnl II ■. b,r, I I FNpurl • untn.U, »nri «-«i h lu lotrl|n i^.rrnn.rnl. ,,r

pnot •p|irnv«l ol I'. S. An M.t.- nrl 1 «Wr»tnni ■ 1 ii

ly A i,.ti

;■ ■> :'"■

n

, i

i i

trr^

/Y7

DISCLAIMERS

The findings in this reporl are not to be construed as an official Department of the Army position unless so designated by other authorized documents.

When Government drawings, specifications, or other data are used for any purpose other than in connection with a definitely related Government procurement operation, the United States Government thereby incurs no responsibility nor any obliqaticn whatsoever; and the fact that the Government may have formulated, furnished, or in any way supplied the saio drawings, specifications, or other data is net to be regarded by implication or otherwise as in any manner licensino the holder or any other person or corporation, or conveying any rights or permission, to manufacture, use, or sell any patented invention that may in any way be related thereto.

DISPOSITION INSTRUCTIONS

Destroy this report when no longer needed. Do not return it to the origi nator.

.(iFin Whin !«T1« D ';„,1G iJUfF SE6TIW

; ^ •..Ttun

I ;,.) hW HMUIM-rtV WO]

mi~ AVAR. m<» «*«*

DEPARTMENT OF THE ARMY HEADQUARTERS US ARMY AVIATION MATERIEL LABORATORIES

FORT EUSTIS VIRGINIA 23604

This program was carried out under Contract DAAJ02-69-C-0094 with Singer- General Precision, Inc.

The information contained in this report is a result of research conducted to study the behavior and performance of adhesive bonded joints when sub- jected to tensile and shear loading conditions and service variables. Through selected experimental Investigation and using high-sensitivity capacitance-type extensometers, tensile modulus, shear modulus, elastic limit, yield stress, fracture stress, end viscoelastic strain were de- termined for various combinations of adheslves, adherend materials, and service conditions.

This report has been reviewed by the U.S. Array Aviation Materiel Labora- tories and is considered to be technically sound. It is published for the exchange of information and the stimulation of future research.

Task 1F162204A17001 Contract DAAJ02-69-C-0094

USAAVLABS Technical Report 70-65 November 1970

STUDY OF THE PERFORMANCE Of ADHESIVE BONDED JOINTS

Final Report RC 70-3

By

Harold K. Shen John L. Rutherford

Prepared by

Research Center Kearfott Division

Singer-General Precision, Inc. Little Falls, New Jersey

for

U.S. ARMY AVIATION MATERIEL LABORATORIES FORT EUSTIS, VIRGINIA

This document Is subject to s pec al export controls, and eac h transrr, ittal to foreign governmeni "s or fore gn nationals may be made only with prior approval of U.S. Army Aviation Materiel Laboratories, rort Eustls, VI rglr ila 23604

ABSTRACT

The purpose of this work was to observe the effects of selected experimental and material variables on the performance of metal/metal adhesive bonded joints. Three adhesives were used: FM1000, EC2214 and Metlbond 329 with aluminum and titanium adherends. The major experimental variables were test temperature and strain rate. Material variables included composition of adhesive, adherend material, bond-line thickness and tapered bond-lines. The following properties were determined for tensile and shear loading: modulus, microyield stress, precision elastic limit, viscoelastic flow, and fracture behavior. Lap shear properties were also measured. The strains were determined using a high sensitivity capacitance- type extensometer.

It was found that the tapered bond-lines did not degrade the properties; in fact, the viscoelastic flow was considerably reduced. The adhesives at 720F, and higher, were elastic only up to about 5 to 15 percent of the fracture stress. The elastic limits were lower and the viscoelastic flow was higher for shear loading than for tension. Raising the test temperature and lowering the strain rate produce similar effects. Temperature cycles below the curing temperature did not influence the room-temperature properties. Poor bonding procedures and out-of-date adhesives provided the most damaging effects on the properties.

... in

FOREWORD

The work reported here was performed in the Research Center, Kearfott Division, Singer-General Precision, Inc., at Little Falls, New Jersey, under Contract DAAJ02-69-C-0094, Task 1F162204A17001. It was sponsored by the Structures Division, U.S. Army Aviation Materiel Laboratories, Fort Eustis, Virginia, with Mr. Thomas Mazza serving as project monitor. This final report covers work performed from 12 June 1969 through 12 June 1970.

The project was conducted under the supervision of Dr. John L. Rutherford, Program Manager, with Dr. Harold K. Shen serving as Principal Investigator. The following people also contributed to the program: Messrs. Laverne Dunham, Thomas Magnini, and Andrew Skurna. All machining and adherend refinishing was the responsibility of Mr. Charles Bing, Manager, Model Shop.

TABLE OF CONTENTS

Page

ABSTRACT i i i

FOREWORD v

LIST OF ILLUSTRATIONS viil

LIST OF TABLES x

INTRODUCT ION 1

EXPERIMENTAL PROCEDURES 2

Test Materials 2 High Sensitivity Extensometer 2 Adherend Preparation 5 Specimen Preparation 9 Shear Tests 13 Tensile Tests 18 Lap Shear Tests 18 Data To Be Reported 19

Test Program 22

RESULTS AND DISCUSSION 28

Phase I, Base-Line Data 28 Phase II, Adherend Material 44 Phase III, Cryogenic 55 Phase IV, Strain Rate 67 Phase V, Cryogenic/Strain Rate 80 Phase VI, Bond-Line Thickness 90 Phases VII and VIII, High Temperature/Strain Rate 105 Phase IX, Ageing/Environment 118 Summary of Results 127

CONCLUSIONS 133

LITERATURE CITED 134

DISTRIBUTION 135

VM

LIST OF ILLUSTRATIONS

Figure Page

1 Capacitance-Type Extensometer 4

2 Tensile Adherend Gripping Assembly 6

3 Talysurf Profile of Tensile Adherend After Lapping 10

4 Talysurf Profile of Tensile Adherend After Several Bonding and Cleaning Cycles 11

5 Shear Adherends and Bonding Fixture 12

6 Torsion Testing Apparatus 16

7 Viscoelastic Strain Parameters 21

8 Nonuniform Bond-Line Configurations for Shear T«sts 24

9 Nonuniform Bond-Line Configurations for Tensile Tests 26

10 Tensile Tests, FM1000, 720F, Cross-Head Speed - 0.02 in./min 33

11 Shear Tests, FM1000, Aluminum, Cross-Head Speed- 0.5in./min, 0o/0o 41

12 Shear Tests, FM1000, Titanium, Cross-Head Speed- 0.5in./min, 0o/0O 54

13 Tensile Tests, FM1000, -650F, Cross-Head Speed - 0.02 in./min, 070° 62

14 Tensile Tests, EC2214, Aluminum, 72 F, Cross-Head Speed - 0.02 in./min, 0O/0O 96

VIII

LIST OF ILLUSTRAriONS (cont'd)

Figure Pag^

15 Shear Test, EC2214, Aluminum, 72 F, Cross-Head Speed-0.5 in./min, 0O/0O 103

16 Tensile Tests, Metlbond 329, Aluminum, Cross-Head Speed-0.02 in./min, 0O/0O Ill

17 Shear Tests, Metlbond 329, Aluminum, Cross-Head Speed - 0.5 in./min, 0O/0O 1 «7

IX

LIST OF TABLES

Table Pog9

1 Cure Cycles for Aluminum and Titanium Aü>>er«nd Sp*ci*r*n» 1*

I! Summary of Test Program *•*

III Tensile, FM1000, Aluminum, Phase I ^

IV Tensile Summary, FMI000, Aluminum, Phase I 32

V Shear, FM1000, Aluminum, Phase I 37

VI Shear SUP lary, FM1000, Aluminum, Phase I 4<-

VII Lap Shear, FM1000, Aluminum, Phase I 4i

VIII Lap Shear Summary, FMlOOO, Aluminum, Phase I ^

IX Tensile, FMlOOO, Titanium, Phase II *7

X Tensile Summary, FMlOOO, Titanium, Phase II '?

XI Shear, FMlOOO, Titanium, Phot« II 51

XII Shear Summary, FMlOOO, Titanium, Phase II S3

XIII Lap Shear, FMlOOO, Titanium, Phase II b6

XIV Tensile, FMlOOO, Titanium and Aluminum, Phase III 57

XV Tensile Summary, FMlOOO, Titanium and Aluminum, Pho»« III 61

XVI Shear, FMlOOO, Titanium and Aluminum, Phas« III ^

XVII Shear Summary, FMlOOO, Titanium and Aluminum, Phase III <>ö

XVIII Lap Shear, FMlOOO, Titanium and Aluminum, Phot« III 67

XIX Tensile, FMlOOO, Aluminum, Phase IV 69

XX Tensile Summary, FMlOOO, Aluminum, Phase IV 73

Toble

XXI

XXII

XXIII

XXIV

XXV

XXVI

XXVII

xxvm

XXIX

XXX

XXXI

XXXII

XXXIII

XXXIV

XXXV

XXXVI

XXXVII

XXXVIII

XXXIX

LIST OF TABLES (cont'd)

Page

Shear, FM1000, Alominum, Phase IV 75

Shear Summary, FM1000, Aluminum, Phase IV 79

Lap Shea-, FM1000, Aluminum, Phase IV 81

Tensile, FM1000, Aluminum, Phase V 82

Tensile Summary, FM10OO, Aluminum, Phase V 84

Shear, FM1000, Aluminum, Phase V 8C

Shear Summary, FM1000, Aluminum, Phase V 88

lap Shear, FMIOOO, Aluminum, Phase V 89

Tensile, EC2214, Aluminum, Phase VI 91

Tensile Summary, EC2214, Aluminum, Phase VI 95

Shear, EC2214, Aluminum, Phase VI 98

Shear Summary, EC2214, Aluminum, Phase VI 102

Lap Shear, EC2214, Aluminum, Phase V| 104

Tensile, Metlbond 329, Aluminum, Phases VII and VIII '06

Tensile Summary, Metlbond 329, Aluminum, Phases VII and VIII HO

Shear, Metlbond 329, Aluminum, Phases VII and VIII 112

Shear Summary, Metlbond 329, Aluminum, Phases VII and VIII 116

Lap Shear, Metlbond 329, Aluminum, Phases VII and VIII 119

Tensile, FMIOOO, Titanium, Phase IX 121

LIST OF TABLES (cont'd)

Table Page

XXXX Tensile Summary, FMI000, Titanium, Phase IX 125

XXXXI Lap Shear Summary, FMI000, Titanium, Phase IX 126

XXXXII Effects of Nonuniform Bond Lines 132

xii

INTRODUCTION

The objectives of this research program were to: provide a better understanding of adhesive bonded joints, obtain information needed to improve present technology in the H»sign of bonded structures, and evaluate the effects and interaction of adhesiv material variables and service test variables on bonded joints. To accomplish these, microstrain techniques were used to determine the mechanical properties of metal/metal bonded joints as a function of a number of experimental and material variables. Microstrain analysis employs high sensitivity capacitance- type extensometers capable of resolving extensions of the order of 5 x 10"' inches and angles of twist of 6 x 10"° radians.

EXPERIMENTAL PROCEDURES

TEST MATERIALS

In this program, two adherend materials were used: 7075 aluminum alloy in the T 6 condition and titanium - 6 aluminum - 4 vanadium in the ful I hard condition. Three adhesives were studied:

1. Metlbond 329, Narmco Materials Division, Whittaker Corporation. This is a 100-percent-sol ids, modified epoxy adhesive film supported on a synthetic fabric carrier. Since it is symmetric, it is not necessary to orient the film with respect to the adherends. Metlbond 329 provides for load-bearing metal-to-metal and sandwich applications capable of long-time operation at temperatures ranging from -423° to 450oF.

2, EC2214, 3M Company. This is a one-part, 100-percent-so I ids, thermosetting liquid adhesive which gives off no volatile by-products during cure. It has high impact, peel and bend strengths with a service temperature range of -70° to 200oF.

3. FM1000, Bloomingdale Department, American Cyanamid Company. This adhesive is a white, elastomeric unsupported film designed for structural bonding of all metals, sandwich construction, wood and plastics. A primer is not required except where processing requires "tacking" of the film in place.

HIGH SENSITIVITY EXTENSOMETER

In order to measure strains of a few percent in a specimen whose gage length may be as small as 0.005 inch (the adhesive in a bonded |oint, for example), it is necessary to have an extensometer that can resolve deformations as small as 10~° or 10"' inches. Such an instrument has been developed and is in daily use in the Research Center '#2/3< j^e basic instrumentation was originally conceived for deformation studies of metals. A description of that extensometer is presented here.

-8 -7 • It has the highest theoretical (10 ) and practical (10 ) extension

sensitivity of all available gages.

• Since the extensometer is mounted on the specimen, it does not record any deformations in the loading assembly.

• This extensometer measures bulk behavior, and does not modify or alter the surface structure of the specimens.

• Sensitivities of 2 x 10 inches are routinely obtained, even after large prestrains (25 percent strain with a 1-inch gage length).

• The present extensometer has been used at temperatures between 4.2° and 4730K. A design has been worked out for an extensometer capable of operating up to about 1000oK.

• By matching the coefficients of expansion of the specimen and the extensometer material, temperature fluctuations as large as iO.^K do not influence the results.

• The capacitance bridge output is easily recorded.

• Present extensometers can be used with specimen diameters ranging from 0.060 to 1.250 inches and lengths from 0.002 to 3 inches.

The extensometer consists of two parallel copper plates, in close proximity, which are mounted concentrically on the tensile specimen (a schematic diagram is shown in Figure 1). A suitable circuit determines the capacitance of the system, which Is a measure of the separation of the plates according to the formula

D

where

C = capacitance

tt = a constant

K = dielectric constant of medium between plates

A = area of smaller plate

D = distance between plates

Thus a change In the distance between the plales means a change in the capacitance, so that any changes In the length of the test section can be recorded. The extensometer forms one leg of a capacitance bridge which is balanced at the beginning of the test. As the specimen is extended, the capacitance of that leg decreases, thereby unbalancing the bridge. The bridge out-of-balance voltage is amplified and fed to the X-axis of an X-Y recorder.

TO LOAD CELL

COAXIAL CABLE

HIGH POTENTIAL PLATE

MICROMETER THREAD

LOWER BODY

SOCKET

SPLIT COLLAR

UPPER YOKE

UPPER BODY

LUCKING SCREW

INSULATION

GROUND PLATE

W\ —- SPECIMEN if MICROMETER DIVISIONS

LOCKING SCREW

IT"11

-LOWER YOKE

TO CROSS HEAD

i Figure I . Copaclfance-Type Exfensometer.

To calibrate the extensometer, one plate Is mounted on a precision micrometer thread so that it can be moved accurately a small distance. The response of the pen on the X-axis gives the sensitivity, which may be as high as 5 x 10"' inches per division (20 divisions per inch of graph paper). The sensitivity is controlled througli three variables; initial plate separation (usually abouf 0.025 inch), voltage range of X-axis on recorder, and amplification of bridge out-of-balance voltage (Robertshaw Fulton Proximity Meter). Other things being equal, the smaller the gap, the greater the sensitivity, since a given displacement represents a larger percentage of change in a small gap than in a large gap.

During the development of the capacitance extensometer and the evaluation of alternate methods, consideration was given to the amount of bending introduced during tensile testing. Misalignment was established with the use of both bonded resistance strain gages and a Huggenberger mechanical strain gage placed at different points around the circumference of a tensile specimen. The amount of bending was reported as the variation in readint of the gages for different stress levels. Combinations of chains and universal joints were first used and were found to introduce strain differences of 10 percent and higher. The minimum amount of bending (about 5 percent) resulted when a ball-and-socket joint was mounted on each end of the specimen. It was found that a small (bias) load had to be maintained at all times to preserve a' Mty.

ADHEREND PREPARATION

All the adherends and bonding fixtures used In this program were machined from 7075 aluminum in the T-6 condition and Ti-6AI-4V as heat traated. The tensile specimens were in the form of a butt joint made with adherends having a circular cross section and a diameter at the epoxy/adherend interface of 0.500 inch. A schematic view of an adherend and the gripping assembly is shown in Figure 2. The machining specifications for the adherends are as follows: (1) the adhesive face to be flat within one lighf band of helium (5876 A); (2) the adhesive face to be perpendicular to the surface of the 0.500-inch diameter within 0.00005 inch Total Indicated Reading (T.I.R.); (3) the 0.500-inch diameter to be round within 0.00005 inch anJ to match the mating adherend diameter within 0.00005 inch T. I.R.; (4) the adhesive face surface *o have a roughness of 4 to 5 microinches (arithmetic average, using the standard cutoff width jf 0.030 inch) wirh the stylus traverse mode in several direction«; (5) the adhesive face circumference to have a sharp edge with a maximum radius of 0.002 Inch. The several machining processes have been organized so as to meet fhese specification».

HALF-BALL

TJIRLAÜS FOR ATTACHMENT TO UNIVERSAL JOINT

HEMISPHERICAL SEAT

ADHEREND

ADHESIVE FACE

(not fo scale)

F'gure 2. Tensile Adherend Gri PPi'ng Assembly,

For shear tests, napkin-ring type adherends were used. At the adhesive face, these adherends are right circular hollow cylinders, having a 3-inch outside diameter and a wall thickness of 0.125 inch. The opposite ends of the adherends have wall thicknesses of 1 inch for attaching to the torsion test apparatus. The machining specification for the shear adherends was the same as for the tensile specimens with these additional requirements: the adhesive face to be perpendicular to the surface of the inside diameter within a T.I.R, of 50 x 10"6 inches and the inside and outside diameters of the adhesive face to be concentric to within 10"4 inch

T.I.R.

There were three major reasons for selecting this type of shear test:

1 . With tests such as these, the loading mode is pure shear; that is, there are no tensile forces acting on the joint. With more commonly used lap shear tests, cleavage loading is always present due to elastic or plastic deformation of the metal adherends. Thus the lap shear test data are difficult to interpret in terms of shear characteristics.

2. Pure shear tests do not involve dimensional changes in the specimen. This is in contrast to tensile tests where the material is elongated in one direction and tends to contract laterally to maintain constant volume (Poisson's ratio). In tensile tests and lap shear tests using metallic adherends bonded together, the difference in Poisson's ratio between the metal and the adhesive material influences the deformation characteristics of the adhesive. That is, when the joint is strained in tension, the adhesive material is not free to contract laterally according to its own Poisson's ratio because it is constrained to follow the Poisson contraction of the adherend. This constraint on the adhesive material is present in both tensile tests and lap shear tests. In napkin-ring tests, where the only deformation is pure shear, the absence of dimensional changes eliminates any effects resulting from differences in the Poisson's ratio between the adherend and the adhesive material.

3. In all torsion tests, the strain is a function of the radial distance from the axis of rotation. The strain is zero at the axis of rotation and increases to a maximum at the perimeter of the specimen. With this variable strain, it is difficult to calculate the shear modulus. Thus, a thin-walled tube is used where the wall thickness is small compared to the diameter of the tube. The strain changes so little across this wall that it can be considered uniform.

The lap shear adherends were carefully machined to 1 inch by 6 inches in the

various thicknesses.

Two adherend cleaning methods were used prior to bonding: ora for aluminum and the other for titanium. These tv/o were selected after evaluating other

procedures. The cleaning method for aluminum was developed by the Forest Products Laboratory, University of Wisconsin:

1. Degrease.

2. Dip in chromic acid solution at 150 ± oT for 10 minutes. sodium dichromate - 1 part/weight distilled water - 20 parts/weight concentrated sulfuric acid - 10 parts/weight

3. Rinse the metal thoroughly in cold, running distilled or deionized water.

4. Oven dry at 145 ± 5 F for about 10 minutes or air dry.

For titanium adherends, a procedure developed by the American Cyanamid Company was used:

1. Wipe with methyl ethyl ketone.

2. Degrease with trichloroethylene vapor.

3. Pickle in the following water solution at room temperature for 4 minutes:

nitric acid - 15% by weight hydrofluoric acid - 3% by weight

4. Rinse in tap water at room temperature.

5. Immerse in the following water solution at room temperature for 2 minutes:

trisodium phosphate - 50 grams/liter of solution potassium fluoride - 20 grams/liter of solution

hydrofluoric acid (50% solution) 26 milliliters/liter of solution

6. Rinse in tap water at room temperature.

7. Soak In 150 F tap water for 15 minutes.

8. Spray with distilled water and air dry.

One of the methods for observing the surface finish of the tensile and shear adherends was a Talysurf profile. In this method, a stylus is drawn across the surface and the vertical displacement is recorded on a strip chart. A diamond stylus with a tip radius of 10"^ inch was employed; the stroke length was 0.5 Inch. The strip chart moved at a rate 20 times faster than the stylus, giving an effective linear

magnification of 20. The sensifivity of the vertical travel was 10 x 10 inches per division of chart paper. On al! adherends Talysurf profiles were made in two directions, at right angles to each other. A typical trace on an as-lapped adherend is shown In Figure 3. To ensure that the cleaning procedures used prior to bonding

did net cause the flatness and roughness to exceed the above tolerances, Talysurf measurements were made on each adherend at regular Intervals. If the Talysurf profiles indicated that the flatness or roughness tolerances were exceeded, the adherend was reflnished by lapping. A typical profile of a specimen damaged by the cleaning solutions is shown in Figure 4.

SPFCIMEN PREPARATION

The test specimens were bonded Immediately after cleaning. Very careful fixfuring was used to guarantee concentricity and axiality of the three types of specimens. In each case, the fixtures were made of the same material as the adherends so as to minimize differential strains due to different coefficients of thermal expansion.

The tensile specimens were prepared in a V-block assembly. Two V-blocks were mounted on a base with a removable end plate; one V-block has two spring clips for holding the adherend. To obtain a constant folnt thickness using this fixture, the following procedure Is employed: a precision shim (having a thickness equal to the desired bond-line thickness) is placed on the end plate, and one of the adherends is placed In the V-block In firm contact with the shim and held in place with the spring clips. The other adherend Is then placed In the second V-block and pressed against the first adherend. A U-type clamp Is used to hold this adherend in position. The base plate and shim are then removed as was the adherend held In position with the spring clips. The adherend still clamped in the V-block Is then coated with the adhesive. The uncoated adherend is replaced In the spring clamp V-block and positioned about 1/16 inch away from the coated surface. The end plate is replaced In position (without the shim) and slowly tightened down. This procedure results in a joint having the proper bond-line thickness. Excess adhesive Is wiped away from the joint Immediately after assembly. The adherends are held this way until the curing cycle Is completed.

The shear specimens were bonded using a specially designed central arbor. Figure 5 Is a photograph of two napkin-ring type adherends and the central bonding arbor used to fixture them during the cure cycle. The arbor consists of a metal shaft surrounded by a Teflon ring (a) at the position of the adhesive line. Three steel pins (b) located at 120° Intervals are press-fitted into the shaft through the Teflon ring to provide the arbor with a dimensionally stable diameter. During cure, these pins seat against the inner surface of the adherends at the bond line

T

u c

o m e IM

4. »—

O

1

I ■••

rj

10

^

j-;: 2-\.

z m

«<o

C o

c

s >

u 0)

< -o c V 1»

-o <

c 0)

1..

4» .Si ~ 'u 9u

11 ,20

^s;

11

12

and maintain concentricity of the adherends with respect to each other. The Teflon ring serves to prevent bonding of the arbor to the adherends, and also keeps the inner diameter of the adherend free of adhesive. Close-fitting slip rings (c) are located at both ends of the arbor, thus providing axiality of the adhesive- bonded adherends. Since both axial ity and concentricity of the adherends is insured, the adherend faces (which have been carefully machined perpendicular to the axis of the adherends) remain parallel to each other and thus provide a uniform bond thickness. The bond-line thickness obtained Is constant to within ±0.0005 inch for the majority of joints prepared in this manner. For cast adhesives, joint thickness is established by holding the adherends apart with gage blocks located between pins (not shown) which have been press-fitted into the adherends. A leaf spring (d) bolted to the end of the arbor maintains pressure during the cure, preventing movement of the gage blocks. Gage blocks are not used for sheet adhesives, the bond thickness being determined by both the initial sheet thickness and the pressure applied through the leaf spring. The pressure is continuously adjustable to 25 psi by controlling the position of the bolt (e). Higher pressure can be attained by using a stiffer spring.

After completion of the cure, the adherends are removed from the fixture in an arbor press. Since adhesion between the arbor and adherends occurs only at the steel pins, the adherends are easily dismounted.

The bond-line thickness for both tensile and shear specimens is determined by measuring the distance between pairs of diamond-shaped indentations located on the adherends. These indentations are formed with a micro-hardness tester and spaced at 90° intervals on the adherends behind the grooves. The ■separation of the marks is measured first with the clean adherends butted together, and again after bonding. The precision of this measurement is of the order of 0.5 x 10""* inches.

The lap shear specimens were made in a Carver Hydraulic Press using a fixture designed to produce aligned specimens with a uniform bond-line thickness. Each lap shear specimen was prepared individually. Finger-type specimens that must be cut apart after bonding v/ere not used.

The cure cycles for the three adhesives are shown in Table 1 and are identical for both aluminum and titanium adherends.

SHEAR TESTS

2 3 The shear tests ' use napkin-ring type adherends as described above. The shear specimen consists of two relatively thin-walled tubes bonded together end to end; they are loaded by rotating one adherend relative tc the other about an axis

13

TABLE 1 . CURE CYCLES FOR ALUMINUM AND TITANIUM ADHEREND SPECIMENS

FM1000 EC2214 Metlbond

329

Temperafure 350 (0F)

250 350

Trme 2 (hr)

2/3 2

Pre-Heaf Temperafure 160 (0F)

160 160

Rafe of Heafing ^6 (0F/mln)

^4 ^6

Shear Bonding Pressure 26

(psi)

0 26

Tensile Bonding Pressure «^150 (psl)

|_

0 =«150

14

coincident with the longitudinal axis of the specimen. The torsion testing apparatus is shown schematically in Figure 6. A cenlral axle supports the bonded shear

specimen by means of close-fitting bushings. One adherend is bolted to an end plate; the other adherend has a torsion sprocket fastened to it. Torque Is applied to this latter adherend through two chains on opposite points of the torsion sprocket. Two chains eliminate any bending moments on the specimen. The apparatus is attached to the cross-head of an Instron Testing Machine, and the loading chains are connected to the load cell in the top of the Instron. The capacitance-type extensometer is mounted straddling the adhesive bonded joint so that one capacitor plate Is stationary along with the fixed adherend while the other plate moves as the second adherend rotates. The resultant change in plate separation introduces a new capacitance value which is calibrated in terms of angle of twist. A second extensometer, not shown in Figure 6, is mounted on one of the adherends for use as a load cell, It records the deformation in the adherend as a shear stress. Load-extension curves are drawn on an X-Y recorder, with the output from the extensometer that measures the angle of rwist fed to the X-axis while the "load cell" extensometer signal is put Into the Y-axis.

For measuring the angle of twist, the extensometer is attached to split rings which fit into grooves in the adherends near the adhesive bonded joint. The grooves serve to locate the split rings. In a reproducible manner, ut a distance of 0.030 inch from each adherend face. Thus the amount of adherend deformation contributing to the observed extension is limited to a gage length of 0.060 Inch, representing a significant Improvement over multiple joint designs in which the large adherend deformation obscured the results. Here, for a typical choice of adherend material and adhesive system, the adherend deformcSon represents less than 15% of the total deformation and may be accurately accounted for through calibration procedures.

The standard test procedure is to record load-extension cu-ves for load-unload cycles starting at very low maximum stress values. Load cycles are repeated at Increasing maximum stress values until the specimen fractures. Such procedures are necessary to reveal the precision elastic limit, microyield stress, viscoelastic behavior, etc. When necessary, the procedures were modified to accommodate the test objectives.

The standard cross-head speed for shear tests was 0.5 in./min. To observe strain rate effects, two other speeds were used: 2.0 and 0.05 in./min. For the standard bond-line thickness of 0.008 in., these cross-head speeds represent strain rates of 2.9 x 10"', 7.2x 10~2 and 7.2 x 10"^ per minute, in order of decreasing speeds.

15

c

I

c o

s

1/1

16

Tht rnfirc t -"aior opparofut fill intid* on cnvironmcnfol chomber io fHaf f«?»fi nyay

b« made a» fcmpcroture» ror>gif>g from -100° fo 3750F. High-frmpcrafurr »eiting

uV0% on oir otmoipKrre, whil*« fhe low-»empcro»ufe work ii June in carbon dioxide

901. Temperature it confrolled fo * t F, and a fhijrmocouple it offaclicd to fbc

tpecimen fo record ift femperafure. Thit tome environfT>cnfol cbambrr wot uted

for fbe fentile and lap ifiear fetft.

Tbe tbear modulut calculafion iforf^ wifh fh«» forouo T applied fo f^e bondrd joint

T PD

where P it half the oofpuf of fhe odhcrend load cell (fhc load in each chain) and

D it fh(? for»ion tprockef diam^fcr. The iheor tfrett applied fo fhe adhetlve joinf ii

T TC/J

where C it the radial ditfonce to fhe center line of fhe joint and J ii the polar

moment of inertia given by

J »/J2(d4 - d4) O I

where d. and d ore the ir.ier and outer dlometert of the ocfrerendi. The width of 1 o

the odherend face it 0. PS inch, tofficiently tmoll compared to the 1.5-inch

rodiut of fhe o<^>erendi f!-il the iheor ifrett acrovt it may be contidered contfant.

Subttituting for T and J in the equation for f ,

PDC

- 32 (d4 - d4) o 1

The thear tfrain > It relofed to the angle of fwitf, Ab, by

> C/L AS

where L it the gage length of the tpecimen (bond-line thicknett). For tmall

ditplocementt, A9 is approximated by At r (the ditplocement meotured with fhe

exfentometer divided by the radial ditfance tc the center of the exfentomeferK

The thear modulus G it given by

G T/>

Subitifuting, fhii becomej

PDLr

rr/32(d4 - d4) A* o I

Puffing In the numerical value» from the equipmenf,

G = 8.O6PL/A1

TENSILE TESTS

AJ described above, the tensile jpecimen consists of two J-inch-diometcr metal rods

adhesive bonded end to end. Each orfierend is gripped in o ball-and-socket joint as

shown in Figure 2. Previous studies hove shown that such an arrangement minimizes

any bending in the specimen during application of the load. A imoll bias load,

about 10 to 20 pounds, is always maintained on the specimen to preserve alignment

in the tensile assembly. Similar to the shear tests, the tensile specimen is subjected

to load-unload cycles at increasing maximum stress values until fracture occurs.

With these tensile tests '»*', the applied load is measured by the Inst.on load cell

whose signal output is fed to the Y-oxis of the recorder. The tensile deformation is

measured using a capacitance-type exfensometer that straddles the adhesive bonded

joint. This also requires that the recorded data be corrected for the adherend

deformation included in the extensometer output (this correction is of the order cf

15 percent for a typical specirr.en).

For standard tensile tests, a cross-head speed of 0.02 in./min was used; with a

bond-line thickness of 0,008 inch, this was a strain rote of 7.5 x 10 /min. For

strain rate effects, two other cross-head speeds were used: 0. 1 and 0.002 in./min,

corresponding to strain rotes of 3, 75 x 10"* and 7. 5 x 10""/min, respectively.

LAP SHEAR TESTS

All lop shear tests were made in the environmental chamber mounted in the Instron

Testing Machine. Wedge-type grips were used with spacers so that the test

alignment was axial. The stando'd cross-head speed was 0.5 m./min for all the

tests; additional speeds of 0.05 and 2.0 In./min were also to be used to observe

any strain rate effects. All specimens were taken directly to fracture.

18

DATA TO BE REPORTED

The following dolo, where opplicoble, were recorded for each ffci»;

1 . Spoclrnen Number - Specimen» are identified by o code ining a

combination of letter» and number», where the fir»t three letten identify

the adhe»ivc, the adhcrend, and the type of te»t (F for FMlO'X),

E for EC2214, M for Metlbond 329, T for titanium, A for ilummum,

S for »hear, T for ten» le and L for lap »hear). The next element i»

either a pair of angle» »eporoted by o »lo»h or one or two digit». The

angle» are the taper» of each acfierend bonding face (»ee Te»f Program)

u»ed for nonuniform bond-line thickne»»e», while the digit» identify a

particular pair of adherend» u»cd for bord 'ine» of uniform thicknet».

The number following the do»h i» the te»t »equence.

2. Bond-Line Thickne»» - When one value i» given, it refer» to the uniformly

thick bond. If two value» ore reported, they ore the minimum and

maximum thickne»»e» of the tapered bond line.

3. Effective Ten»ile Modulu» - Thi» term i» u»ed when di»cu»»ir>g the terwile

modulw» of o<^e»ive» in the thin-film form between two a<#>erend». Thi»

it in recognition of the fact that the obterved modulu» i» for a material

under »evere con»traint. When the epoxy i» bonded to a metal a<#>eiend,

it i» not free to deform according to it» own Poaton'» ratio, but mmt

conform to that of the adherend which generally ho» a lower value for the

Poiwor«'» (atio. By retfricfing the lateral contraction of the epoxy, the

tentile modulu» i» effectively increo»ed. Thu», the meotored value in

the»e thin-film te»t» i» an effective modulu» rather tbon the modulu» found

in the material when it i» free to deform according to if» own dictate».

4. Shear Modulu» - The »bear modulo» i» the ratio of the »treu to »train,

within the elastic limit, when the »train i» measured at tf e displacement per

unit length caused by »hear stress per unit area. In napkin-ring tests, the

only deformation i» pure shear (as opposed to tensile or lop shear test»),

and thus any effects resultiitg from difference» in the Poitson". ratio between

the acfierend and the ciiesive material are eliminated.

5. Precision Elastic Limit - The precision elastic limit is the maximum stress

to which a specimen may be loaded and still behove elo»ticolly, with the

deformation remaining a single-valued function of the lood. The word

precision is u»ed »ince very »moll deviation» from elottic behavior may be

observed. G«r>erally, the elastic limi» depend» on the sensitivity with

which the stress-strain relationships are observed.

19

6. Microyicld Sfre»» - TM» i*. >he lowe»f itrcsi level, during a ioad-unload

cycle, which re»uU» in the finf pcrrrxincnf ihain at fhc bioi :freii.

Only the ifroin fhot pcniifi leverol minute» after unloading i»

comidered permanent.

7 Mode uf Fracture - There are two po»»ible mode» of failure in thin-film

epoxy aHhe»ive joint»: cohe»ive and oc^ejive. In cohe»ive failure,

the cpoxy itKlf fracture». Thi» type of failure can be recognized by

the pretence of a layer of a(#ie»ive on each of the adierends. Another

term for thi» fracture ir " center-of-bonrf' failure. Acfrejive failure

occur» when the fracture i» at the interface between the ac^e»ive and

the o^erend, i.r., the adhe»ive/ot#ierend bond foil». Thi» i»

generally characterized by what oppeor» to be a clean metallic »urface

after fracture. However, clo»e examination may reveal a

microtcopically thin layer of epoxy remaining on the adherend »urface.

The revolt» arc pretented a» percentage of cohe»ive failure.

8. Fracture Sfre»? - The fracture »tre»» it fhat »treu at which the joint fail».

It i» calculated by dividing the load at fracture by the crou-»ectionol

area of the ac^eiive bond.

9. Vi»coela»tic S'roin (r. ) - Upon completion of o load-unload cycle from a

»tret» level above the mlcroyield point, there i» a »et remaining In

»he tp«clm«n. In fact, thi» I» the major method for determining that

; te yield point ha» been exceeded. Very often the »et dltoppeari with

rime; that I», the tpeclmen »lowly return» to it» original length. The

value reported here i» the »et in tne bonded joint immediately upon

unloading to the bio» »tre»» (the minimum load maintained to preserve

axiallty In the loadlr>g aitembly). When di»cui»ing vl»coelo»tic,

recovered or permanent »train, it I» necenary to report the maximum

»tret» le el uted In that load-unload cycle.

10. Recovered Strain (fj) - Thi» I» the amount of »train that I» recovered

within two minute» after unloading; it I» vltcoelo»tic recovery.

II Permanent Strain {<*) - The »train remaining in the »peclmen after the 2-

mi'-ute recovery ('j i]'fj)- The»e 3 vl»coela»t I c »train parameter»

are »nown in Figure 7. For »near »train», the »ymbol» y., >- and y. are

u»ed.

20

> i 1

A i

1/ /

/ y /

•'3* -'7 •

Strain

Figure 7, Viscoelostic Strain Parameters.

21

TEST PROGR/M

The test program was divided into nine phases. It is summori/ou in Toblo II

and ''".cribed in more detail below.

Phase I - Base-Line Data

The objectives of this phase were to: establish fh« base-line values of the ttandbrd

tests for comparison with subsequent tests hn Wig different test condition», emomir*

six different bond-line tapers so as to lolecf hree for use in the balance of the

program; select the best adherend thickness hr lap shear tests.

The standard tests were made with FMlOOO on aluminum ac^erends at the

intermediate cross-head speeds (0.5 in./min for shear and 0.02 in. min for tension)

at room temperature with a 0. 008-inch-thick bond line.

Six different tapered bond-line thicknesses were investigcted for shear festt

(Figure 8):

1 . an included angle of 22 minutes vifh the bond center plane normal

to the longitudinal axis of the specimen (designated a* 11'/I I')

2. an included angle of 44 minutes with the bond center plane normal

to the longitudinal axis of the specimen (designated as 22'/ 22')

3. an included angle of 22 minutes with the bond center plane tilted

1) minute: from normal lo the loncjitudinal axis of the specimen

(designated as 2270)

4. an included angle of 11 minutes with the bond center plane lilted

5.5 minutes from normal to the longitudinal axis of the specimen

(designated as D'/O)

5. an included angle of 0 with the bond center plane tilted I I mi.iute«

from normal to the longitudinal axis of the specimen (designated

6. an Included angle of 0 with the bond center plane tilted 22 minutes

from normal to the longitudinal axis of the specimen (designored cs

22'A/22'C)

22

t 1 ►■ 1 UJ 1 H* 1 U.

0 >

1 ^

1 Ui | ai

1 *" I

i z 1

! U

i o or

1 3 | z

J ** 0

ao t S S at s s 00 fN "" 1

a.

!

-i ll J 'S

*

»8 1* -0 fN CM rsi (N (N (N rst fN (N fN <N

c 5 ! •

i-O 1 ^ 1 i-J r^ (N* (N Ol ^ 1 CN [N fN fN

^ 1 ^ c^ r* <N rsi (N [ CN fN

• ■f 1 ^ rsi fN (N fS «^ fN (N fN »^ r^

u ^ rs* «N r^ r>* rs rsi fN CN rsi fN fN fN4

fN

i X

• c* rw rv* IN IN (N rs# (N fN <N CM

• 1 ^ r^rv ^~ fN fN CN CM (N4 (N

• ^ ~ fN f>* (N (N <N fN fN (N fN fS

CM fN fN* CM <N fN (N rj <N <N fN rsi

UJ

<

>

i UJ a. X UJ

O

0 X

See

text

(Exp

erim

enta

l P

roce

du

re«)

i I

X

X X X X X X X X X X X

S s X X

■0 X X X

X X X X X X X

Yi X X

■p »^ X X X X X X X X x

X X

■5 c

■6 »- X X X

< X X X X X X X X X X

* >

<

n X X X

X X

i U- |

X X X X X X X X

II a. uj

1

ii Is SS 11

'c « 01 0 ^

»1

1 .y 0 c «

ll 8 a

II t !

1 1- c

c ? 8

a. x 0 wt «1 { = = O J3 > 0 -0 > > 0 -0 > 0 -Q | > 1 - 1

23

\

u CM

CN

U o ♦- c

r in 0)

O 4)

-C

i-

e .?

u 4) C

c

w CN CM

it 'c i o Z

oo

ä

24

Six different tapered bond-line thicknesses were investigated for tensile tests (Figure 9):

1. an included angle of 4 with the bond center, .e normal to the longitudinal axis of the specimen (designated a* io/20)

2. an included angle of 2 with the bond center plane normal to the longitudinal axis of the specimen (designated as 10/10)

3. an included angle of 2 with the bond center plane tilted 1 from the normal to the longitudinal axis of the specimen (designated as 2 /0 )

4. an included angle of 1 with the bond center plane tilted £ from the normal to the longitudinal axis of the specimen (designated as i7o0^

5. an included angle of 0 with the bond center plane tilted at an angle of 2° from the normal to the longitudinal axis of the specimen (designated as 2° /2 )

6. an included angle of 0 with the bond center plane tilted at an angle of 1 from the normal to the longitudinal axis of rhe specimen (designated as 10A/1 r)

From each group of 6 tapered bond lines, 3 were to be selected for use in Phases II through VIII.

For the lap shear tests, aluminum adherends were used in 4 thicknesses: 0.088, 0.062, 0.048 and 0.032 inch. Following evaluation of the results, one thickness was to be used for all subsequent lap shear tests.

Phase II - Adherend Material

To determine the effect of an adherend having a higher modulus of elasticity, the tests made in Phase I were repeated using TI-6AI-4V adherends. Only 8 shear, 8 tensile, and 2 lap shear tests were made using the thickness values selected in Phase I,

Phase III - Cryogenic

The low-temperature characteristics of FM1000 were measured at -65 F using both aluminum and titanium adherends. All other experimental variables were the same as in Phase II.

25

0^

u

in 0)

C

o

u 0CV.

0cs

8

c

I t c

c o

E Z) O)

kC c o U

c '_j -o c .2 E

.0

c 3 C 0 Z

£ o^

26

Phase IV - Strain Rate

Tests were made with FM1000 on both aluminum and titanium adherends at high and low cross-head speeds (2.0 and 0.05 in./min for shear; 0.1 and 0.002 in./min for tension).

Phase V - Cryogenic/Strain Rate

Since low test temperatures and high strain rates both are known to increase modulus values, they were combined in this phase. Tests were made with FM1000 on aluminum adherends at -65 F using cross-head speeds of 2.0 in./min for shear and 0.1 in./min for tension.

Phase VI - Bond-Line Thickness

Earlier work had shown that some mechanical properties depended on the thickness of the bond line. To evaluate this effect, shear, tensile and lap shear tests were made using thicknesses of 0.004 and 0.020 inch. Since FM1000 has an optimum bond- line thickness (0.008 inch), an alternate adhesive was selected, EC2214. This is a paste-type adhesive having no carrier cloth so that a thickness-dependence study could be made.

Phase VII - High Temperature

Adhesive properties were measured at 160 F using a high-temperature adhesive, Metlbond 329, with aluminum adherends. For comparison, another set of specimens was tested at room temperature; all other variables were kept constant.

Phase VIII - High Temperature/Strain Rate

To observe the combined influence of two factors that tend to lower the modulus values, tests were made on Metlbond 329, with aluminum adherends, at 160oF with the lowest strain rates (cross-head speeds of 0.05 in./min for shear and 0.002 in./min for tension).

Phase IX - Ageing/Environment

At the beginning of the program, tensile tests were made with fresh FM1000 on aluminum adherends. That batch of adhesive was stored until the end of the program, and new specimens were prepared and tested using the old adhesive.

27

RESULTS AND DISCUSSION

The test results and discussion are presented by phases. Each phase is introduced by a brief summary of the experimental variables, followed by three sections for the tensile, shear and lap shear tests. The results are given in tabular form and typical stress-strain curves are shown (the fracture point is Indicated by an X at the end of the curve). It should be noted that the results presented here are for those specimens considered to be acceptable in terms of bond-line thickness and uniformity and having reasonable values for the fracture stress. Many specimens were rejected out-of-hand because the bond-line thickness was not correct or because it varied in random fashion. The test results for specimens having very low fracture stresses were discarded, for premature fracture was considered strong

evidence of an improperly bonded specimen.

The test results for each phase are summarized in separate tables. To report experimental scatter values, the maximum and minimum variations from the average

value are averaged. For the viscoelastic strain measurements {€,, Cji ^i/ ^2^' only approximate values are given because these stress level-dependent measurements were not ail made at the same stress value. In some cases only one value is included in the summary table.

PHASE I, BASE-LINE DATA

The basic data, to be used for comparison with the other phases, were accumulated for FM1000 adhesive on aluminum adherends at room temperature with the standard cross-head speeds: tensile - 0.02 in ./min, shear - 0.5 in./min, and lap shear - 0. 5 In./min.

Tensile Tests

The results from the tensile tests of Phase I are presented in Table III, Including the 6 tapered bond lines (see Figure 9), Table IV is a summary of the results with the experimental scatter. A typical tensile stress-strain curve is shown in Figure 10 (curve A). The elastic effective tensile modulus (E*) Is calculated from the load- extension X-Y traces using the bond-line thickness as the gage length for determining the strain and the cross-sectional area of the adherend for the stress. For specimens with a tapered bond-line configi ration, the gage length was taken to be the average thickness.

28

^ ! .^ 1 ^ g J

^-8 8 S3 32 >0 Q

00 in Ö 00 3S C30 CO 0 00

10

! 4,

e - U.

0 ^

§1 O CN

0 0 0 0 88

(N O §i 0^ CN

0 0 CN

8 00

CN O — CM O- — — 0 6- 0- O j U. 1/1 t) — '■— '~ ' "

*o ! oj

*>» UJ 1 O M v-i k- >rt ,

8 ^ 0 o

CN fN

O O 0 0 * O 0 0 0 0

CN

0

< Q.

• — 1- u>

5^Ä r>« 00 — CN r- CN

IN, CO CN "i K S

CO IV in CM 22 •0

^ 1 i ^ t

C

i ^ ■^ .y z j a! LU Hi 3

0 CJ

O —

0 0 0 "Jl ^ 0 — O

0 m 0 -0

55 O CM ° 1

1 5 •0 t^ >o 00 00 tv >o fv O m

D

-< < 1 ^ ^"

I ^ 1

8 i U- 1

•- ^ "5 S- J2 £ 0° UJ »^ ^ ^^

C» CN • • lO IT)

0 0 • • 0 in • • • •

■<t IT)

>o 0 co* ro

O • CO

m

1 B

1 ^ _ v 1 Ht UJ

— i ^

CM - - - : = : - = = = = = =

z Ol 1— TJ

* 1 8 c Ms-^

1 UJ $ ^ > 2 S. c

CN O

1 < U Ä" • O = - - : r = = = - = : - =

! ,_ lO 0 • •

— in • •

CN in

CN r—

in -c 00 CN • •

? -r1 c '" '«t ^l- CN CO CN C^

C ^

OK

-o 0 • • ■Tf 00 rs 00

55 CO CO

• • O 00

• 00

CN • 1

00

— CN 1— CN

— CN 1 1

— CN r- CN 1 t

<— CN t 1 o'^'u 1 c O O 0 0 0 0 •^Q^ CN CN r— r— 1

0) u ^— CN CM p p 0 0 \\ \ \ .iü 1 O«

O 1 oV- °:°: ^^ o_ 0^ ^ o<o<

• 1 ÄZ

r- CO CN CN CN CN CN CN t^ »—■ I

►— 1— y— »— 1— h- 1- 1- *- ^z ^ ^ 1— (- < < U. Li.

< < UL. U_

< < U_ U.

< < LI- u.

< < U_ U-

< < U. U-

< U-

< 1 u. j

29

T3 0> D C

c o U

en <

2 »> b

oo

CO i

«2

CO

CN42

CO

o

0) c

U «>i w

(U c

E ^ 0) r > < £

c 0) i-

E J5 o F 0» o # Z

_^co

° 2

CO oo

CO "

CM

CN O

O O

I O

I O

O

— I o

o in o o rv oo r^. »>

o o o o CN K o rv o^ >o

o o o —

o o o o

TJ- in o o io o co >o • • • • o o o —

u->

§ ^ o K A .A • o

• •*t

• to

^ r—

1 CN

3 <> A CN

0CN

0CN

o o o o o

in co —

o d o o o

m in TJ- o m co o o 00 CN

r- in in

oo co

i^ m o o «o cs rx o^

o TJ- lO CN O •n in ■<■ o o ^ oo S2 o m o >— — CO K -o 00 — in m o co O CO o o

r^. o in m o CN

■ • • •

o — o o

•n co o o — in

O«— OOO r— lOOO

CN •—

CN CM

m CO 8 in

CN S «O o o r^ >o

^ CM CO — I I I I

o o o o

o' o^ o^ o° ^- — ^- CM

00

CM I

o o

CM

30

lie o o in •O O- rv

rv p rs o •o & o o- o p o ^ tv CK r^ c>

CO 1

r)2 ^ Q CN — iO CO

K O Q »o — o^ <7) m ^00 CO W -w • • •

o — o ■ • • •

o — o o • • *

o — o ^-

n S?S8 00 o o — 00 tx ?§?s

/v. o • • • • • • • • ■ o o o o o o o o o o —

_ CO

1 ^^8 lO O Q W) co rx io CM

O O Q Q oo <o -^ «O

-Q _ o • • • • • • • ■ • • t

V w ^ O CM — O (N O ^ O CN O CN

c 4- c 0 ■

U a. CM

1 ^^S. rv = = = = =

UJ _l ■a m o --. <

8 O cs o • O : = = - -

2 s "^ OO CN

0)_| c — * •

2 JL _^C0 3 5 — o ■*t CN • • 00 00 O 00

• • 00 00

— CN — CN — CN 1 1 1 1

c 1 1 oU 0u QU o^

se

cim

e

lum

be

r

a a 0 < 0 < o"< o"< CN CN (— r^

h^ P h- t- I- t— üiZ < < < < < <

LL. U- LL U_ LL. LI-

31

— o o ul lO U~l o O

i > f Os CO K S8

K CO oo

^ü£ 1

o

V 11

IT) rj CM — o CN

n i' H H w ■+' i' i o UO R a-

1^- 8 w -s o d d d O

"

n CO § $ K IX

m o

CO -H ■H -H -H -ti M -ii 00 hN CO s o o m

o (N o^ N o in

^'— -—- o O CM CN

UJ g § g 8 8

ro

in oo

to < £ - -fl 7 57 Sf tf M

3 „, ^ Q O ^^ o o +1 m X a. 5? o

o § o 1 8 o o o r— o o ^n o

2 13

u- t/i b "" 1— '~ o~

Z ■D UO o o

5 0 s IT) 00 $ CO lO o

3 0 t/» -H ■II -H CN -H -H -H _J u w

o

IT) o ■H Q in

2 < 2^Ä

CM K CN ^ in ^ CN — ■-" CN — ,— CN

§ O •— c Ul CN m 2 u.

0 o <N $ -tj" ro CN CN

-H •— 'T ••- -ti ♦1 w ■H -H o

> or

P o E ^ cT m ^ -£:

-tl

o 03 CM

CN IT)

n

tv CN CO

CN O

m

B

2

UJ _j

M ^ | 1 'S V3 T;IO

-s 5 .2 2 LU i- <: ^

-H

n -H IT)

u->

o

-H O

-H CN

+1 CN

-H 00

d 00 z a. UJ t— 2|o5 = = r : i = >

■o UJ —; —1 ö c 00 < T "o ^-

p o •

i; 2

o o = = = r

oU oU —■ o O 0 0 0 CN

^ „^ O^ o^ O

0^ > °> Jj c.U o CM IN CN /

32

12,000

10,000

8000

I 6000 to

4000

2000

/7 . ^^ «»«»«

A

c :%■"*

/""

B -"K

/// i - Aluminum

- Aluminum 't

. - Titanium 0

o7o0

,o /2o "A7 C D/0O f

Curve ß

Curve B

Curve C

f f

25 50

Strain (10' )

75 100 125

Figure 10. Tensile Tests ^\ '.IGOJ, 720F/ Cross-Head Speed - 0.02 in./min.

33

Due fo the large experimental scatter, it wa» difficult to determine fKr mflupncc of

nonuniform bond-line configuration on the elastic effective tontilo mociuli/i, wlrhin

the experimental scatter there was no effect due to nonuniformify. Similar

observations were made for the precision elastic limit, the mi .royield ttrew, »ht

fracture stress, and the fracture mode. However, the strain u» unloading, <. ,

was diminished when nonuniform bond lines were used. Comparing the uniform bond-

line value of 1.13 x lO"3 for fj to that for the F/tf value of 0.58 « 10'3, •» con

be seen that there is a 49 percent decrease. For the other twe selected c<jnfigurotiom,

the decrease was somewhat smaller.

One striking feature of the results is the deviation fi^m elastic lehovior at very low

stress levels. The average precision elastic limit (P.E. L.) for all bon'Mme

configurations was 870 psi, which is only 8 percent of the average ffocfure fret»

(10,500 psi). The P.E.L. is the first deviation from elastic loouing, in thif work it

was in the form of a hysteresis loop. When the speclme.i was unlocded, tK re wo» no

change in the length, but energy had been dissipated during ff-c lood-unlood cycle.

The P.E.L. was largely independent of the tapered bond-lin? configuration. The

first permanent set (microyield stress) was found at on average value of I 560 pti.

Thus, at only 15 percent of the average fracture »tre«, the bonded joint undergoet

permanent damage. During load-unload cycles at higher maximum itrcv» level»,

proportionately greater damage was incurred. Abovt the microyield »•rei»,

viscoelastic flow was usually observed. The tapered joint» had only o negligible

influence on the microyield stress. For a quantitative description of vi»coelo»tic

behavior, three strain parameters are used as described in the section " Data To

Be Reported' (also see Figure 7). The total strain remaining in the specimen at tKe

instant unloading is finished is designated as ( .. In all case» '| wo» greater, the

higher the maximum stress used in the load-unload cycle. For example, in Table III,

specimen number FAT10-1 had a value of 0.35 x 10"^ for *• wben the »tret» wo»

70 percent of the fracture stress. When a cycle was mode at 90 percent of the

fracture stress, f, was 1.35 x 10" . Thus, the strain upon unloading wo» greater

by a factor of 4 for an increase of only 29 percent (7700 to 9900 p»i) in tbe maximum

slress. Sometimes the strain (f ] ) in the specimen at the instant of unloading woJd

persist for long times (of the order of tens of minutes); in other case», tbc »pecimcr

would recover some of that strain almost immediately.

For each load-unload cycle, the length of »ht specimen wo» monitored for 2

minutes after unloading. Experience has shown that the major port of tbe recovery

after unloading occurs within 2 minutes. The column» of Table» III and IV labeled

f 2 show the strain recovered within that 2-minute waiting period. <•» i» the difference

between f| and f ^ and is called permanent »train; the word permanent mean» only

that it exists afte, the 2-minute recovery time. Whether «-j would be foond after

several hours or days was not determined. For oil »pecimen», the higher the mo«imum

stress, the greater the amount of recovery and olio the larger the amount of total

34

strain at unloading. It appears that for stresses below about 65 to 70 percent of the fracture stress, there is no recovery after unloading. Recovery occurs during the unloading process; most of the load-extension recordings do not have straight- l ine unloading that is typical of elastic behavior.

It appears that the viscoelastic strain, c^, was greatest for the two configurations (2°./2PQ and w h ° s e normal to the bond center plane is not parallel to the longitudinal axis of the specimen. These are the two bond-l ine tapers that introduce the greatest amount of shear during the tensile loading. The values shown for columns €1 and a n c ' stress level in Table IV are averages for a l l the data reported in Table III; it should be remembered that and ^ depend on the stress level (percent CTp) to which the load-unload cycle was made.

The fracture stresses averaged 10^500 psi for a l l specimen configurations. The highest values were for the 1° /1 tapered jo in t , whi le the lowest were for the 2 ° . / 2 ° £ configuration. In general, the tapered bond lines had only a negligible influence on the values of the fracture stress. For the 2°^ /2°c and configurations (see Figure 9), the bond center plane is parallel to the two adherend faces, but a l l three of these planes are t i l t ed . Thus, there is introduced a shearing act ion in addit ion to the tensile force. This may account for the observation that the lowest fracture stress was found for the 2°^ /2°£ taper in which the center plane t i l t is 2° , the largest tested. A typical stress-strain curve is shown in Figure 10 (curve B) for a 2°^/2°Q configuration. The fracture mode was characterized as percent cohesive (center of bond) using a visual estimating procedure. On the average, the fractures were about 80 percent cohesive in nature.

As discussed at the beginning of this section, many specimens were discarded because of fracture at very low stress levels. Excmination of these specimens showed that the standard cleaning procedures had not been fol lowed or that the adhesive had spoiled or that there were many voids in the jo in t . Thus, it appears that the bonding procedures are much more important in control l ing the fracture stresses than is the introduction of nonuniform bond-l ine thickness.

Based on the results shown in Tables III and IV, three taper configurations were selected for use in the balance of the program:

1. an included angle of 4° wi th the bond center plane normal to the longitudinal axis of the specimen (designated as 2 ° / 2 ° )

2. an included angle of 2° wi th the bond center plane normal to the longitudinal axis of the specimen (designated as 1° /1 )

3. an included angle of 2° wi th the bond center plane t i l ted 1° from the normal to the longitudinal axis of the specimen (designated as 2°/0°)

35

The first two of these are symmetrical about a plane whose normal is paral lel to the longitudinal axis of the specimen. The center plane of the third configuration is t i l ted 1 and the taper is not symmetrical as are the first two taper configurations (see top three sketches of Figure 9).

Shear Tests

At the beginning of the program it was extremely d i f f i cu l t to obtain satisfactory results wi th the tapered bond-l ine shear specimens. In order to obtain good results for at least 14 tests, over 42 specimens were made and evaluated. Results for 16 of the satisfactory tests are presented in Table V . The shear modulus (G) was calculated as described in the section "Shear Tests," using the average bond-l ine thickness as the gage length. A summary of these results, along with the experimental scatter, is presented in Table V I . A typical shear stress-strain curve is shown in Figure 11.

Deviations from elastic behavior occurred at very low stress levels. Taking the average for a l l specimens, the precision elastic l imit was about 320 psi, which is only 3 .4 percent of the average fracture stress (T^ = 9500 psi). The average value for the stress necessary for the first permanent strain (microyield stress, T ) was 620 psi, or 6 .5 percent of the average fracture stress. These percen tages"^ lower than those observed for the tensile tests (P.E.L. = 0.08 Op and cr = 0 . 1 5 Op). Wi th shear tests, the fracture mode was less cohesive (center of Isond) in nature than was the case for the tensile tests. In these latter specimens the typical fai lure was 80 percent cohesive, whi le for the shear tests the average was only 33 percent cohesive. There was no strong correlation between mode of fai lure and fai lure stress. There were too few specimens having 100 percent cohesive failures to establish a l imit ing value in the total absence of any adhesive fai lure. Thus, it is not possible to say whether the fracture stress is established by local ized adhesive fa i lure, wi th the path of the propagating crack determining the specific mode of failure (cohesive vs adhesive).

In association with the low stress values for the onset of nonelastic behavior, the viscoelastic strain values were very high. These parameters, yj, >2 ar>d ^3 / a r e

the shear equivalent of those shown in Figure 7. In comparison to tensile tests, the viscoelastic flow occurred at lower stress levels and in greater quantity for comparable fractions of the fracture stress. Viscoelastic flow in shear occurred at stresses as low as 15 percent of the fracture stress. As the maximum stress for a load-unload cycle was raised, the viscoelastic strain increased rapidly. For specimen number FAS22'/ 0 -5 , the strain at unloading, yy increased from 0 .3 x 10"^ at 0 .15 Tp to 148.0 x 10~3 at 0.85 T^. To a great extent, the strain y^ was not permanent in the shear specimens. Averaging a i l the results for columns yj and y^ ' n Table V shows

36

LU to < X

D Z

D

o o o

tt: < LU X

CO

<

> 0)

o o

u 5 a

a. to K1-

ir>

c o

£ i J « • u

8 H o o

o ^

U

^ o u |

c

o E

22

o o o — >o o

88^ i 8 5^8 ON O •- - o>

•<■ o

22 S?5? 5^

o o

o o o o

p O .x O Q Q O^ r- Q 00 CO ^ O OO ^ O lO NO

r^ c«4 o o o o rx o lo a« — cs Sgf^ co rv Tf CO rv o K oo

O CO 00 '^■ IN Tj- O 00 u-> o o cs o o K cs o o* 00 o 92: "- ö O^ K m o CV4 CM — — io ^r CNJ CM u-) CM

5? !/> CO

vO CO §8 88 CM

g 00 CO tv CO

Co CO o — s

m g

5 I CO >o

O"— ^J-co >oco inio coo oco t^m — cs r-^-^r- .— .— ^-r- r— — ^-^ COCN — i—

<N tx= = : == :: =r :: == : =

IT)

O r = : := :z :: :: := ::

Ov O O- O ON CO O (> CO _ CO 00

NöNO fv.'O COCO O^Ch ON,-J IN.'«

11 II II II '" *ri o r^oo i^»— "«rco

O »■" 'mm CN •• •• •• •• ... OOtV OO'O -»t'* ONO

rvrxixiv — (vjco— CN -— CM ^ CM I ' ' '

CMCM II 11 tNJ^ Z Z. en* mo ^ \^ ^^ ^^ ^K, \\

I I I I - • - - - - - - - <- < - < - < — — .— .— CMCM — — CMCM — — CMCM .— — — — '—'— CMCM — — CMCM — — CMCN — — l/1tOI/>V/1 lOcO COCO «/)bO oOCO tOl/) oOOO <<<< << << << << << <<

37

«■s;?- hi «O CM CM "O lO CO •^■ ■^F to Tf vo co "t io

Q CO Q O io »O lO 00

^■oopocoiorNCM u->>o'?)ü5>ooou-)rN

o» ■* «- o oo rN • • • — >0 CM CM 00 lO -«t • • • • • • • ^3 o rs — O r- CM CM o «o o — CM CM — m — 1 O CO

co" L u->

Oj o •n «o rs o oo o O O K ■^ CM O ^ w • t • • • •

CM CM O O O O <> • • •

CO — o o o --* o ^ o too

^^ CO

■o V

1 ^- o ■^ o oo o oo u-> o O O ^ 00 — O CM "t 00 O lO CO

"«^g-*^^ CO* NO* O i—' CM '«r —* vo .— m o V

D C ^ • c 0 u

1

Test

Te

mp

(0F)

R = = = = z : : : =

> -D

UJ -j CD <

Cro

ss-H

eo^

Spe

ed

(in

./m

in)

• O = : =

0» o o o CS CO '•""^ • ■ • • • • s» .1 2 c ^ ^ ix uo co co

CM CM CM CM 0)-l c *"

Ilia O — — CM

N! IN! IS! IN!

'-1

8

.5-

'-2

7.6

-

'-1

8

.7-

'-2

5.8

-

5 4.7

-

6 4

.1-

c CM CM — >— 1 J

IJ CO "«f IO o> 1 1 1 1

C. CM — — o o \ \ \ <

y ^— ^— ^^ ^^ M CM ^- ^- fj fj ^— ^— ^^ ^— CM CM — — CM CM »^ oo t/> m l/l l^ «/) OO OO 00 ÄZ < < < < < < < < < <

38

gf^ ^i^ O CO ^Q tX

Tf <o >o CO O IT) r- IT) CO CN —

CO 1 o^ -* o o n "O CO CO

n ö • • ■ • • • • • ^ 3 O CO CN «O CN O

ro CM b

• CO • O O O CN o o O O

i — o cs •* o o ,— u-) CO CO

■ • • • • • • 4)

O 00 CN O CN r— CN O

D ^C 4- • C . 0. o t» E fT U CM

IN : = = = =

> -o

LU o ^- —i 0) c 00

< O 4) J IT) I- n C • u ^^: o =

o o o* o*

S? -i c 4 S CO ^_ CO CO

Ave

rc

Bon

d-

Thic

k • • • • • o o rv. vo

o^ o

CM CO — CN 1 1

c 4) L.

— CN &

1

CN -U ^U

p p CN <N \ -<

\ \ lü X \ ^ .< .<

Spec

N

um

r— w~ fM ^™ w~ ^- r~* CN CN »— •— U0 1/1 OO (/) UO LO

< < U_ LU

< U.

< U.

< <

39

*—^ V) — U.

« 5 h 4* s « 8. h «

t l-g LO un IT) o -^r co CN 01 _J S^

^^ CO

^1 • o

■

o o o 00

d O

Uü l/V < X CO

1 o o _ 00 CN - 00

a. _ o • • • • • • • o CO rg F—' o CN CN

■k rC~ C- (N

5 3 z

_i U_ 00 K-

CO

-H

O o

fs.

CN ■O

IT) r) CN H

IO

8

8

O

00

O

■H

8 o

< ■o o O •* o CN 2 CO 00 — '—

8 o 8

« CN 00

uo •o IN 00

s u a> J-^ H H H ^ 5

H H u. • •• L* »/>

5 ^Ä IT) m CN

IN

CO

in

0£. c o >o es

1 !5 •- ö "ft •? rr

Ou LU _J C>

-H -H H

CN 00 CN

<

o C30 ■"t 00 p-M CN 00 00

CO % 00 CN ? s

3 <rt CN CN C-i CN K ^_ LU • • o • • • • X +1 H -+l -H H ti H LO CN in IT) in in — CN

. uj oo $ 3- —* —' r-' —" ^-' ^J —'

> #

UJ —i ca CN < r^ : S : ; - s i—

2 c

iO C

U l>) w d S I i - - z E

II 3

.U .u 1 CN * CN

CN

_<

C o

11 p CN

V. O O

,< «

<£ 0 CN PM P^ CN t*m

a CN ' rsj '— CN ~~

AC

12,000

10,000

1000

Strain (10 )

Figure 11 . Sheo rT«t»,FMlCX)0, Aluminum, Ooss-Heod Speed 0.5 in,/min,

41

that the recovered strain ^ i s about 69 percent of y j . For the comparable case in tensile tests, the recovered strain *2 w a s o n ' v 31 percent of the total strain at unloading ( f j ) . These comparisons cannot be rigorously made because of the different types of loading and also because they were made from results at different stress levels. The tensile data were for load-unload cycles at about 80 percent of the fracture stress, whi le for the shear tests the numbers were for stresses averaging 47 percent of the fracture stress. However, this difference in stress level merely emphasizes the observations; for i f f j , ^ a"1 ' e3 c 'a^a were shown for tensile tests at 0.47 Op, there would be pract ical ly no viscoelastic f low. With tensile tests, the stress threshold for viscoelasticity recovery, w a s o r <^e r of 0 .6 cy .

Specimens made with nonuniform bond-l ine configurations had shear modulus values 25 percent higher than those of the uniform bond-l ine thickness. The nonuniformity sl ightly increased the precision elastic l imit and had only a negligible effect or: the microyield stress and the fracture behavior (stress and mode). There was a strong effect of tapered joints on the viscoelastic strain, 7 j . At a stress level of 50 percent Tp, y , was 20 x 10"3. Nonuniformioints at the same stress level had values for yj of only 0 .9 x 10~3 to 3 .0 x 10 , a decrease of 90 percent.

Based on the test results shown in Tables V and V I , three taper configurations were selected for use in the balance of the program:

1. an included angle of 22 minutes with the bond center plane normal to the longitudinal axis of the specimen (designated as 11 ' / l l 1 )

2. an included angle of 44 minute;, wi th the bond center plane normal to the longitudinal axis of the spe-imen (designated as 22 ' /22 ' )

3. an included angle of 22 minutes with the bond center plane t i l ted 11 minutes from normal to the longitudinal axis of the specimen (designated as 22 ' /0)

The first two of these are symmetrical about a plane whose normal is parallel to the longitudinal axis of the specimen. The center plane of the th i rd configuration is t i l ted 11 minutes and the taper is not symmetrical as are the first two taper configurations (see top three sketches of Figure 8).

Based on these observations of tensile and shear properties and on those from succeeding phases, it appears that the deviations from elastici ty (precision elastic l imit and microyield stress) are associated with the tendency for viscoelastic f low. If for any reason the viscoelasticity can be suppressed, then the elastic range of the material can be extended. Lowering the test temperature (see Phases III and V) eliminates the viscoelastic f low so that the adhesive is elastic up to the point of fracture. With in the strain rate range used in Phase IV, a higher cross-head speed diminishes the

42

viscoelastic f low so that elastici ty continues to a higher stress level . Raising the temperature, Phases VII and V I I I , promotes viscoelasticity wi th an attendant decrease in the elastic behavior. The use of tapered bond-l ine configurations markedly reduced the viscoelastic flow propensity (except for the cryogenic tests). This was probably due to the introduction of more complex stress states that interfered with the molecular viscoelastic flow mechanisms; for example, creating a compressive stress in addit ion to that for pure shear, or adding a shear component to a pure tensile stress.

As w i l l be shown in later phases, the use of tapered bond-l ine configurations had the effect of lowering the values of the effect ive tensile and shear modulus. In homogeneous materials, measured in bulk, the modulus values are uniform throughout the sample. However, in thin f i lm form, or when the adherends modify the properties, adhesives may not have constant modulus values. Consider a tensile specimen being loaded at a uniform rate. The stress is the same throughout the sample and is given by the load divided by cross-sectional area. However, the strain in the adhesive bond is not uniform; it varies inversely as the thickness. For the thinnest part of a tapered jo in t , the strain w i l l be the greatest because the elongation is the same at a l l points. Since the modulus is defined as the ratio of stress to strain, increasing the denominator results in a lower modulus. Another factor that depresses modulus values is viscoelast ici ty. This contributes strain for a given stress that is addit ional to that of the elastic strain. It is not clear whether the onset of viscoelasticity is a strain or a stress threshold. However, in a tensile test wi th a tapered bond l ine, deviations from elast ici ty would be expected to occur at the thinnest portions f irst. For a l l calculations of modulus, an average bond-l ine thickness was used which raised the calculated modulus for the thin sections and lowered it for the thick sections.

O f course, another factor that influences the modulus of a material is a change in response due to alteration of the method of loading. In a shear test, the molecular chains uncoi l , stretch or move in a certain fashion. If a compressive stress is superimposed (as wi th tapered bond lines), the chains w i l l react d i f ferent ly . What-ever the mechanisms of that reaction are, they have the net effect of providing more strain (for a given stress) and producing a lower modulus than is the case for a pure shear test.

l ap Shear Tests

To determine the optimum adherend thickness for lap shear specimens, 12 sets of 7075-T6 aluminum adherends were bonded wi th FM1000. Adherend dimensions were 6 by 1 inches with four thicknesses (0.032 , 0.048 , 0.062 , 0.088 inch). The specimens v»ere taken direct ly to fracture at room temperature using a cross-head

43

speed of 0 .5 i n . /m in , wi th results as shown in Table V I I . A summary of the results is presented in Table V I I I . The adherend thickness of 0.062 inch was selected for a l l addit ional lap shear tests in the other phases of the program. It was chosen because it produced the highest fracture stress and because it had the best fracture characteristics.

PHASE I I , ADHEREND MATERIAL

The tests made in this phase differed from those of Phase I in three respects: a l l adherends were made of a heat-treated t i tanium al loy (T i -6AI -4V) , only three tapered bond-l ine configurations were used for tensile tests and three for shear tests, and only one adherend thickness was used for lap shear tests. The curing cycle was the same for the titanium adherends as for aluminum adherends.

Tensile Tests

More than 30 specimens were prepared to obtain the 13 successfully tested and reported in Table IX. The results are summarized in Table X , which also shows the experimental scatter. A typical stress-strain curve is shown in Figure 10 (curve C). it was expected that the effect ive tensile modulus of an adhesive would increase when used with a higher modulus adherend material. In this comparison, the adherends used in Phase 1 (aluminum) had a Young's modulus of 10 x 10^ psi, whi le in Phase II the adherends (titanium) had a modulus of 17 x 10^ psi. Referring to Table IV, the average adhesive modulus for the uniform and selected tapered bond lines was 5 .0 x 10^ psi. From Table X , for t i tanium adherends, the average adhesive modulus was 5 .9 x 10^ psi, an increase of 18 percent. A stiffer adherend material seemed to have a greater effect on the uniform bond-l ine properties than on the tapered joints. From Tables III and IX, it can be seen that the increase for uniform specimens is about 11 percent.

The precision elastic l imit and the microyield stress were about the same for the Phase II tests as for Phase I . The fracture stress values were not appreciably altered by using t i tanium adherends, and the fracture mode was also about 80 percent cohesive in nature The viscoelastic behavior was essentially the same in Phase II as in Phase I. The recovered ^ w a s about 50 percent of € ^ on the average, compared to 31 percent of for the Phase I tests.

Introducing the tapered bond lines does not influence the effect ive tensile modulus, w i th in the experimental scatter. Tapered bond lines do appear to lower the precision elastic l imit and the microyield stress values. Wi th the exception of the 1° /1 ° configuration, the strain at unloading, , is diminished. There was no effect on the fracture stress or mode due to tapered joints.

44

V > to J) <D

-> - C -2 <u o u -O <J 2 o u_ >

t o < X a.

•» s Z> z D _ i <

o 5 C* < LL) X t / l Q-<

CCJ <

n < U - -C i - ifl i .t: o-

c K '"T 31?

"O V) • C i -a, « .5 0) _*co

_C u | - D • " O < £ C,

c 0 u. E ® . = JQ u E £ § Q- - 7 i/> t -

S> S 5) o o o

"O "O ~D _a> "o "5 "5 u - u -~o - Q "O c C C a> a> a)

Urn u. i— <D a> <1)

_C - C - C -O "O "D o D O

o rv o o i \ i \

§ 8 § 5? 8 8 < < < 8 m o >o >o

o o m o o — >o m >o

o o ^ o o o o ^ o o o o o r x ^ - N O C O O O C O ^ - O - O C N C O C N

C O C O C O C N C N C N C N C N C N C O C O C O S o o o o > o > o > o c o c o c o ^ F ^ ^

r— CN CO I I I

"5t I

ir> o I I

00 I

o —• CN4

i 7 T I < < < < < < < < < < < < U» LL. LL- UL Li- LL. LL. U_ LL. LL. LL. LL.

45

to < X Q_

s z 2 3

<u "O o

<D u. 3 u E

0) a> > > *</» </> a> a> _c _c o ' O o \ a) o a> > > > *«/> w 'w <D a> <D

_C - C _C "O o o o o o

o o o "O <D

>

2 ID to c* < L U X tO

u to

*1 E I

>o 41

S i n o

8 U-> -H

m

-H 5? 8

-8 C P

~U O

CO < £.£ -o<? is <U

* <

00 00 CN >o 3 CN

00

46

LU to < X

5 z <

o i LL.

LU -J

z

CO

<

0) > V»

0) 9> i- -C

<u o U -o u E 1 ^

EIS '-■*-. Ll_ u- to b

5ä1 c o ? .^ S!-2 .§ a

Q. UJ _i A

o J> ^ S.

.2 £ oo LU |— ^ v^

S (Uo"-

s ^-o 'i s « \ o <u * 4 Q. #c U t>i "

a> U) • a> .£ 0) C

5>_i C

Ave

ro

Bon

d-

U 1

c 4) i_

.§J{

Spe

c N

um

iv. N. K

S S1 o 00 o t^ 00 00 o

O O o ^o >o ^ a- — o CM (N ^

O -"t w. • • . O -O IT)

IN1

O

O =

lO O 00 00

o o Q O ■^ CN

si >- CM

<> «O

in in

-t ^>

o o

CM o

S 9 0 00 ^T 00

O O ^t ^t — 00 o o

CM O K lO CM CO ^

r- CO

CM O

-2 • • lO in CM CM

IT) "^ • • CN

4^ • •

r^ co • NO O r^ i-v

— CM — CM

CO ^R. o o r— ~- \\

1 1 &■ —

CM 0^^ OJO

i- ^- tu H- 1-

o o 00 o

o o O O o o m o 00 CM O 00 O — CS O

o o ^ ^o

rv o Q CM

LO »O

't oo • • CM CM

I I O^ CO

co ■^■

o o o o

^ 0c.

47

i > ^ 00 _J 2_ ■»t vO >o o

CO 00 00 tx K oo m 00 K 00

CO ,- — <> 00 00

K CM o CO CO CO

CO i o o o

(NO O

do id

o o o CO CM tX • • • o o o 0

.55

1.1

0

0.6

0 o o -<*

O O "O

d d d

CO

w . •

0.8

0

0.5

0

0.0

5

0.6

0

0.1

0

0.7

0 moo

CO CM 00 « • • o — c 1

.00

0

.72

0

.48

a> D

#C 4- C o U

CO 1 _ o

w 3-

o o o m •o — o o

O C Q CN CO ^

o o o a- co ^ O CM CM

O CO O

r- ^. ^ O O O r- O CM —

i

X UJ

co <

Ave

rage

Cro

ss-H

ead

Tes

t B

ond-

Line

Spe

ed

T

emp.

T

hick

ness

(in

./m

in)

(

0F)

(I

D*3

in

.)

CM r^ _ - -

CM O • (—1__. ..- ..- _..

io -^- ^- o»

ts, co' CO K 6.4

-41

.4

10

.5-4

5.5

6.1

-41.1

7.7

-25.2

7.5

-25.0

7.4

-24.9

9.9

-27

.4

6.9

-24

.4

7.3

-24

.8

Spe

cim

en

Num

ber

i— CM CO r* CM CO CM CO "^

. 1

:TT

9-1

:TT

9-2

:T

T12-3

O O c <N CM CM

CM <M CM t- *- t- h- H- 1-

0 o o oSFo~

0Ö0Ö0Ö CM CM CM t- h~ i- t- t- h- „

48

i CO CM f—•

! i 5» "^T tju-

1 n CO

/I oo 00

1 ^ ^^ I — ^ lO ■ ~~ CM CM CM ! i u-1 f * • **" 1 | 1 ^

< 1 x 1 ^ iii %

2

CO 1 IT)

! * ■ o _

CM

H O

•

-H" 1 o o

H

in IT) m LO

1 D i ?"> "t

i < i 1

^ 0 ^ •

■H -H' | w '— o UO N 1

CO 00 r—• I

i t CN • • i

i •- <U d — — J 1 1 •* 1 »-

P S ^

O g 8 0 1 CO 1 o 1 0 CM ?r g Tf 1 ^ i: LL +1 o m I

1 i o o Q r^ o CO •O CO 1 u. | ! t» '^f o o 0 1 I ' — o 1— ^— 1— ! |

i > .- in o o o : i ^ X CO CM F— CO

< o CO ^t 1— ■ ! i ^

u o) r- +) ■w +t +1 ; 1 •— L. m

5^3 o o CN

CO IN

o CO

3 CO tv O LO

S LU 1 0

CNJ 1— '~~ '— J 1 o

1 —' 1 "^ .U tx iO o 0

1 """ 1 1 *■* 7" <*- IT) oo m MD 1 (/I I

1 z H J .§ i o. tu _i -3

-H O

CO -H ■ft

CN

■H i Lu CO IT) o o i i— CM O u-) o

•— 00 rf ^9 ! |

1 ^ U 4) 3 in X - "5 Q- iO o

-* o 1

3 UJ 1 ! —' 1 1 co 1 1 <

-2 £ o o +1* o •

+1

•

-H

•

-H 1 <N

* 1 1

1 '"" •

^ ^ 2-

UO uo >o

h* - m. — 1 1 i "O

o «u 'c ^ -^

1 ^^ i 2 « ,-• CM

ull | O • o z z - 1

c 1 .2 ! *-

c E • - D —ID)

1 =1 ^ o O 0 <N o

<2<3 ^ o^ o^ o^ o CM CN 1

49

Shear Tesfs

The fesf resulfs for the shear specimens are shown in Table XI and summarized in

Table XII, A typical shear stress-strain curve for these experimental conditions is shown in Figure 12. There was no significant change in the shear modulus when titanium adherends were used as compared to aluminum adherends Using the average values of Table VI (including the selected tapers) and Table XII, the comparative modulus values are 1.43 and 1.48 x 10 psi. Also, comparing the results configuration-by-configuration does not reveal any large differences.

Deviations from elastic behavior occurreJ at very low stresses as In Phase I. For these Phase II tests, the average precision elastic limit was 3.4 percent of the fracture stress and the microyield stress was 6.5 percent. Fracture stresses were somewhat lower than with the aluminum adherends, with the 22,/22' specimens having significantly lower values for Tp. The fracture modes averaged about 75 percent cohesive with the exception of the same 22,/22l specimens. The oth'" mechanical property values for these two specimens were in close agreement with the other specimens in this phase. Viscoelastic flow was similarly prevalent her? as in the specimens of Phase I, although the recovered strain y« was slightly lower with titanium adherends than with aluminum adherends.

As with the Phase I shear results. Introducing nonuniform bond-line configurations increases the effective shear modulus by about 25 percent; likewise, the strain at unloading, y,, was markedly reduced (about 90 percent for similar stress levels). In contrast to the Phase I shear results, the tapered bond lines resulted in a lowering of the precision elastic limit, the microyield stress and the fracture stress.

It is not surprising that the shear modulus is the same for [oints made with aluminum adherends as with titanium adherends. With this method of pure shear loading, there are no dimensional changes in the adherends when they are loaded. In the absence of size changes, there is no Poisson contraction and the modulus of elasticity is not invoked. In tensile tests, the adherends deform at different rates depending upon their modulus values. Since the adhesive is forced to deform along with the adherend, it has c foreign modulus imposed upon it. Thus, the tensile modulus of an adhesive is affected by the adherend properties. In these shear tests the adherend modulus values do not participate, so the adhesive modulus in shear is the same for titanium adherends as it is for aluminum adherends.

50

1/1 < X D.

_J

o o o

< LLI X

<

-8 o

0) > 4)

o

"- ^ S^

■2 a '5

ja

o « u iß

c o

IT, «> "^ T^" JU o £ « a. rr; "^ o.

8 Si 3 in

o o

-D O Ü r

U 0) * OL.E

I 8 0) c

_yco

ft» •-

•i-l

o o o sO o o

■o >o IN

T)

o o — -o o o 00 o

CM -«t o CN

$> o O IN o o «o ^

o o N IN

?8^ 5^8 S o 5?^ O- IN CN CO o — ^^ 00 o>

00 CN IN 00

CN ■^:

lO IT)

O CO CN O ^ o n O CN IN ^ IN rN CO CO CO o CO CN CN

^ CN CN CO CN

CS — CO u-)-^- -o^ -o^o

CN K = : = = - - - -

CJ ~ r s = z ; z z

^ tN >0 CN CN CN • • • • • • a u-> 00 o TJ- o ■"t ■»t CN CN

4^ CN CN

o CO lO O •o OJ CN •

• ■ « r— rN o* o IN r^ IN r—

CN — CN —

i 1 1 CN CO CN CN r— i— 1 1 CN CN fm ^ o o

CN CO r~ \ •\ w W | | I CN CN f— f^ CN CN

CO 00 ^— CN CN p» ^-* CN CN 00 1/0 oo OO OO to 00 LO 00 1— u.

1— Li-

i- u. u.

51

« « ^- SS ^^ S in ^o

co oo in <) IT) K ^ f^

CO

-o

c

CO

co

5.5

10

.7 r^ o . •

o —

t^ — CO

o — o

o < o o d r^ -* CM

O CO • •

l-x CO

o O CM

CO CO CN . • •

,- O CM

in «y

o CM* o o

m m c CO

O "^ o CM >— O

c o o ^> • *

o o — o*

52

M — U- X; «> ^ 0 ITl r— 't

1 t* IT) CN IO m

oo _i 2- $ li N 'O

CO •< O

«'s in » S

00 0

r» ^^• 't d ^_^ 10 CO _

CO a O- • . 1 CO • -H 4)

LU

^. o » ')

t^ O

< 0) 0

•0 O x S: ^^ LO K O o. t 2l

C *. LL.

"^ -H y $ ^

S s CM U0 CM 5?

U_ tS) t-T" a. ^ K 0 D 0^ "* 00 00

z 2 < fl) t

2 K^ o- cs _

h- CN 0 m 0 U 4) J- +1 -H -H -H

O o o ü.^^ CO 0 CO

CO IT)

f-a*

^ C o

u_ « .2 •~ *^ .*• £ 00 CM CO

>

3.

£ JO .5 » Q- UJ Hi 3»

,, 3 M

■A IT» 00 <N

V % cs

-H

CO CM

-H

CM

Z> .iJ u. 'S O. -H

• -H -H -H

< LU 00 2: ^^

CM • IT)

• • •

us X • to

_J K = = s

>< -0 LU 0 -^ _J « .£ < I— 1 ?^ 0 o> _•

L. R C 10 • 0 = = :

c 0 ^B

g E

la ^ 0

0 \ > \ \ JN r— CM cy CM •— CN

J

53

14,000

12,00C

10,000

8000

!8

6000

4000

2000

-K

/ -650F

,/

\f 720F

S

1

\

\

200 400 600 800 1000

Strain (10' )

Figure 12. Shear Tests, FM1OO0, Titanium, Cross-Head Speed - 0,5 ir./min,

o7o0.

54

Lap Shear Tests

The results for the lap shear tests using titanium adherends 0.051 Inch thick are shown in Table XIII. The average fracture stress wns 7075 t 550 psi, which is sliphtly

higher fhan the average value of 6750 i 550 psi for aluminum adher' nds. With the best Phase 1 specimens the failure was 100 percent cohesive; with these tests It averaqed 75 percent cohesive and 25 percent adhesive.

PHASE Ml, CRYOGENIC

The major experimental variable change in this phase was that of the test temperature. Tests were made at -650F wi'h both titanium (Part a) and aluminum (Part b) adherends, with all oth^r conditions as in Phases I and II.

Tensile Tests

The tjst results are presented in Table XIV and summarized n Table XV for both titanium and aluminum specimens. A typicc' stress-strain curve for low-temperature tensile tests of FM1000 on titanium adherends is shown in Figure 13.

Referring to the Part a (Table II) results for titanium adherends, the elastic tensile modulus was Increased about 45 percent, from average values 5.0 x 10^ psi at room temperature ro 7.3 x 10^ psi at -650F. The fracture sfress was raised to an average value of 15,000 psi, an increase of 39 percent over the room-temperature value. In general, the fractures were slightly more cohesive in nature at the lowest test tempenture of -650F - an average of 90 percent versus 75 percent for room temperature. The moit significant property changes as a •'esult of testing at low temperatures were associated with viscoelasticity. In contrast to the large instantaneous strains (f ,) at room veniperature (Table X), averaging 1.23 x 10"^ at about 80 percent of the fracture stress, these specimens showed no viscoelastic flow even at stresses very close to the fracture stress. Also related to the lack of viscoelasticity was the 200 percent increase in the precision elastic limit from an average of 825 psi to 2500 psi at -650F. The average microyield stress increased from 1.750 psi to 5460 psi at the lower temperature, <"n increase of about 210 percent.

With the aluminum adherends, a greater cryogenic effect was observed. The average tensile modulus was raised fron 5.0 x lO"5 psi to 7.15 x 10^ psi, an increase of 43 percent caused b/ lowering the test temperature to -650F. An increase of 330 percent was seen in the precision elastic limit, going from 870 psi to an average of 3750 psi. In this case the P.E.L. occurred at 22 percent of >he fracture stress, while at room temperature it was seen at 8 percent of OV. the microyield stress showed a similar effect - about a 400 percent increase m absolute value - and if was u\ 47 percent of the fracture stress. At the low test temperature, the fracture stress

55

56

'm >

1 in 4) 0) t -c ■2 «> S 0-0°

lie m o * o o in in o o o o o o

CO o o> o O IX tV O» OO I

~

1 LU l 1 n -~- CO

< ^8 g g §1 ^9 co Co O O Q O o Ql 88 1 0 u —'

C *■ Li. u. to b

CO Q IN. ^

c3 co CM 00 CM CO 00 1 C1- 5 2 <N CO ts. >0 "^ CM O« IO IS j

1 ^ 2 73 ^ | Q) z >> ^ ^

9 * • - U W) g| S^ s% lO 00 888 5S

t —' ' 5Ä 3 1^ — hs -- ix — «o — < tV ^ In ^ t^ -«»■ «o o «o rv K 0 C

'""

z 0

< !E •- 5 | 8S § 8 88 Q O

O 00 ^88 8§ 3 CL UJ _l Ä ^ ^ ^ D; O to

CM r- ^ ^ § iÖ S IO (O

z i 1 < 1 !« ir- j t; i u j* — a. i ,— i Hi J! TJiO ' *• i ^ s ^ o

tu ►- i w ^> «o O^ CM CM O» O — CM O CO CO <— 1

o | • • • • • • • • ■ • •

O tN hs ■ • :

o I 1 •

^ 2fc u- 1 *" 1

i LU ?. . _ = s _ . - = . ~ - 1 _l

co i z 73 LU | S ;? •— 1

s X =

CN

> X

o d s z s = = = = s = : s :

LU oo <o «o ^o »o »o 00 CO 1 —' 1 -*—*. • • • • « • * * f

! co • .1 2 c oo rs CM CX ^— •—' >— CM

< CO CO CM r^ CM CM "* *■

h- CO cv • • &• CO

00 »o • • CO CM

i i

»O »O

J.6 • * «o -o o • • • oo o- o

cio!, -o rs

c -* if) ■* «o iO "G rx oo I « 1- 1 1 1 1 ■ i i i

co ■^■ 0CN?4

tt

0 0

CM CM 1- 1-

tt

■* — ^

1 1 1 — CM CO t- t- h- <<< U. U. U.

0 0

0«MRI

u. u. ]

57

■o 0) 3 C

i o u

> X

CD

<

0) >

-C

■Ele

«I

I gl E ^ Tu- u- to b

• — U Wl

5Ä3

OL UJ -I 3

is Is

ll€

X

oÄi O) .1 2 i • c ü U

u E

«no Q o

o o

o o CM

• • *o oo

•n >o i =

CM o

00 Ot

<N CM

O» 00

O —

0%^

o CM

88 §8 ^ S CO S

■*• CM • • K 00

« •

CM CM

c«,4 46 o r^

oio o q — r- WO* »- >- I- I- << < <

58

g €) b

s

CO

o2 o

rT (N2 •

o

CO

• 7^ O C : = = = r = = = = = =

4)

C

0

U

•

1 s = : s = = = = = = = = |

> j X i LU

—< CD

< 1—

Cra

ss-H

ead

Spe

ed

(in

./m

in)

CM O • o = : z r = z z = = : = = 1

Ave

rage

B

oncH

-ine

Thic

knes

s (l

O-3 i

n.)

CO CM • • O» 00 3

.8-3

8.8

2.6

-37.6

5.1

-22.6

5

.1-2

2.6

4.1

-21

.6

4.0

-21.5

lO o o • • • oo a«, o 6

.8-4

1.8

7.8

-42.8

Spe

cim

en

Num

ber

FT

T9-

3 F

TT

9-4

■ i 0 o

0CM0CM

U- UL

o o

i=E

m -o i i o o o o

Li. LL. FAT

1-4

FA

T2-1

FA

T3-1

f>. 00

o'o'

CM CM

Ll_ U- 1

59

2 «» ö

CO

CMS w , ,

-o 0 D

tC

C o U

CO

_ o *c = : :

i

> X

- i~ = s

UJ —i CO <

Cro

ss-H

ead

Spe

ed

(in

./m

in)

CM O • o =

Ave

rage

B

ond-

Line

Th

ickn

ess

(lO

-3in

.)

9.3

-26

.8

8.4

-25

.9

10

.4-2

7.9

7

.0-2

4.5

Spe

cim

en

Num

ber

o — i •

a o

U- u_

CM CM

u. u.

60

UJ

< X Ck.

D z

D <

z <

z <

>■

00

00 z

>

<

i CM2

CO I

— O

III «^ b

8 o ÜÄ S;

c o

U «I 3 « Z - 3 O-

- S ^2 ^ H- i 3

2 = uT

2 X c-

c o

.ll -I O)

11

61

14,000

12,000

10,000

8000

a.

I 6000

4000

2000

y /

/

/

Titanium Adherends

y / /

Aluminum Adherends

/

/ 1 ) 0 1 5 ; 10 25

Strain (10"3)

Figure 13. Tensile Tesfs, FM1000, -65*^, Cross-Head Speed - 0.02 in./mm, o7o0

62

a wi

veraged 17,120 psi, compared to a room-temperature value of 10,800 psi. As ./1th the titanium specimens, there was no viscoelastlc flow at any stress level for the specimens made with aluminum adherends. A typical -650F stress- strain curve for FM1000 on aluminum adherends is shown in Figure 13.

Generally speaking, the most significant effect of low temperature was in the viscoelastlc behavior. No viscoelastlc flow was observed at any stress levels, and the precision elastic limit and micoyleld stress values were increased 200 to 400 percent. The increases in the modulus values and fracture values were more modest in comparison.

For the specimens made with titanium adherends, there was little effect due to the tapered bond-line configurations. In the case of aluminum adherends, there was a i'ipht increase in the elastic tensile modulus, the precision elastic limit, and the m'croyield stress.

Shear Tests

Thirty-two specimens were prepared and tested for the shear portion of this phase In order to obtain the 16 satisfactory rests reported in Tables XVI and XVII. A typical stress-strain curve for a -ö^F shear test using titanium adherends is shown In Figure 12.

With regard to the specimens made with titanium adherends (Part a. Table II), the elastic shear modulus Increased 34 percent at the low test temperature. The fracture stress Increased from an average value of 7900 psi at 72 F to 13,400 psi at -65^, an Increase of almost 70 percent. However, the fracture mode changed from principally cohesive In nature to completely adhesive failure. The effect of the low temperature was more pronounced for those properties associated with viscoelastlc behavior. As with the tensile tests, there was no viscoelastlc flow at any stress level, even very close to the fracture stress. The precision elastic limit showed a massive increase (650 percent), going from an average of 264 psi at 720F to 1860 psi at -65°^. The microyield stress was raised from 530 ps» to 3120 psi, an Increase of 500 percent. These low-temperature values of the P.E.L. and the lrmys were, respectively, 14 and 23 percent of the fracture stress Tp. At room temperature, the P.E.L. and T were 3.1 and 6.7 percent of Tr, respectively.

Specimens made with aluminum adherends showed similar low-temperature behavior, but the changes were not so pronounced. The elastic shear modulus Increased 38 percent to an average value of 13,260 psi. The precision elastic limit was raised from 320 psi to 525 psi, an Increase of only 64 percent; this value was 4 percent of the fracture stress. The microyield stress, at 1470 psi, was 11 percent of Tp. At room temperature, these values were 3.4 and 6.5 percent of the fracture stress, respectively. Fracture behavior changed, on the average, from 33 percent

63

^^ at >

■Z a, O

£it lO lO Os 00 O O O O o o o o o o o o OO

V

2 « "5 u S a p S -^ U- 1/1 »~

O O O O 88 88 88 88 §§ 88 88 53 C^j <N

CM Q CO ^ ^ 5 ?S 1? — ix

CM CM §? = ^O

LU 4) CO X < o •« X o 4) vC" t a.

<3 '^ 88 88 8g 88 82 ?? 00 ^ ^ 5 Sz o — CM rv vO ^ K CO K •? O — rx tx

CM CO CO CM CO CM — CM

D c o

Z !i .y 5 -J <

e JI ;i Q- UJ _. S

o o O- CO O In m o-

88 CM »O

^ rx CO oo

5? OS ■t 00

IO m m in O- CO CO CO ^8 — — — — CM — CM —

O «» cC' z < ^ f o o

"-■ ui 5 3 5 D

O IX O OD

CS —

O O • • — CN

— CM • • CM CM

o >o • ■

CM —

»n oo • • CM —

o ^ • • CM —

00 00 • •

Z < _ a. ^ >-

8 o ■o

£ ~ : s

0 . . i 1 .£ u.

^ UO 3

a£ iO < LU

• O : . „ _ . . • m — — . ,

X ir>

CM • >

LU O CN

co ir>

CM O

24

00 in • • CM CM

1 1 •O CO "O o

• •

in o

o o • • CM CO

1 ^ OP '<V>

CO CO • • 00 CO CM CM

1 1

_l • • • • . • • • • CM 4 k • • ao oo rx o> o o m in rs ■<»• CM •o — o -^f < ^— FMW

t— C 4) u CN

o E CO M-

1 1 co ■»»■

1 1 -^ m >0 X

1 1 co -^r

■ i — CO

Oi ^ CM ?J r— ^— i i CN CM r^ —— 1 1

lz — CO

4-1

(N CM \\ ?J CM

\S M3 — 1 1

— CM

CMCM >>

oo

CM CN ^_ ^_ CM CM ^™ ^— CM CM f-» —— CM CM 1/1 CO (/) */» »/) l/) l/> 1/1 uo oo m i/) m in co m

U. U- u. u. U_ Li. U. U- < < UL U.

< < U. UL.

< < << u. u.

64

13 a» 3 C

■4- C o

> X

<

gut,"- ist

i co2

CO I

^2

i . o

2 Jo-

O ^ 4» C

e Xc

4>

c V > <

8 c

O I '22

t* iJ8

&7L

c i

i

O :

O CM

oo rs

T" 4-L

o in

o" o

CM «O

<M CM

00 «O • • <N CM

I

o^ ■«»■ O O CO CO

o •<•• ^ co 9-~ • • •

CO I ■t «o

\\ \< ^- ^- ?M ?J «— ^- CM CM UO «/I lO »/>

m o

r- CM

I/) »•> < <

»A >o • CM

CM

>o — CM CO

I I GO 00

O* —*

s^

CO I

CM o x

i i CM CM — .;- CM CM — — W >> CM CM ^- ^- CM CM — — OO I/) «/»(/) < < < <

CM

o -t

— CO

^^ CM CM

< < a

65

I/O < X a.

Z) z

—i < Q z <

2 Z <

Of

5

< X

> X LU _i CO <

i

CO

f gl JO

.2

§

e if .§ s a. uj -j 1

^c

c =

8 8

■< CO CO CM

5? IC

»n o

" 5? R 8 « CM

o o o o -M -H -M -H 00 (> O CM

• • c • — >— CM CM

1»

o

C. CM

§ CM — CM

i o^

«o o o 2 2$

? 8 O CM

S So CM IT) CM •— ■H -«

Jfl_Jt

-H -H

— en -o CO

■^ CM

o o o ■H -H -H

00 CM 00 GO • • I •

— CM — —

CM

9

< < * ~ Z ~ s < N — v 3 CM — O 0 \ \ \ ^ fj =- CM O CM ^ CM

66

cohesive at room temperature to 100 percent adhesive at -65 F. Lowering the temperature raised the fracture stress from 9500 psi to 13,260 psi, an increase of 40 percent. No viscoelastic flow was seen at any stress levels at -650F as compared to an average strain at unloading (y,) of 5.2 x 10" at 72 F.

Introducing tapered bond-line configurations in the titanium specimens changed some of the properties. The elastic shear modulus was increased as was the P.E.L., while the T appeared to decrease. Although the fracture stress did not change much, the fracture mode changed from 90 percent cohesive with a uniform bond line to total adhesive failure in the tapered joints. There was very little effect on the aluminum bonded specimens due to the tapered adhesive joints.

Lap Shear Tests

Lap shear specimens made of titanium and aluminum were tested at -65 F; the results are shown in Table XVIII. The titanium specimens fractured at an average stress of 8430 psi, which is 19 perr^nt higher than the room-temperature value. A similar change was observed with the aluminum specimens: the average fracture stress at -650F (7950 psi) was 18 percent higher than tSe value at room temperature (6750 psi).

PHASE :V, STRAIN RATE

The tests performed in this phase were similar to those of Phase I except that different cross-head speeds were used: for tension, 0.1 and 0.002 in place of 0.02 in./min; for shear, 2.0 and 0.05 in place of 0.5 in./min; for lap shear, 2.0, 0.1, 0.05 and 0.002 in place of 0.5 in./min.

Tensile Tests

The tensile test results for the low and high cross-head speeds are presented in Table XIX and summarized in Table XX. These data are to be compared with those of Tables III and IV. The results for the tests at the low cross-head speed of 0.002 in./min we-e very close to those of the Phase I tests. There was no appreciable change in the tensile modulus, the precision elastic limit, the microyield stress, the fracture mode or the viscoelastic flow parameters. Only the fracture stress was different: it decreased from 10,700 psi (average) to 9100 psi at the lower cross-head speed. Thus, lowering the strain rate by a factor of 10 did not significantly change the tensile properties. However, raising the strain rate by a factor of 5 did have an effect on the mechanical properties. At the higher cross-head speed (0.1 in./min), the effective tensile modulus was raised, on the average, 22 percent from 5.0 x 105 psi to 6.1 x 10^ psi. All the parameters associated with viscoelasticity

67

ist < X Q.

z

D < Q Z <

z <

o o o

< X

Q. <

> X

00 <

0) >

'in 0)

u-S 8 £ D

(J <o

"D 0 Ü 'c x '^ J. T) .E « o \ 2 2 c in •

o

_J c -6 -^^ c .z, O I

c2

% 5? o o o o

s o 00 — CD N

O O

K 00

IT)

o

CNJ o

o o K in

o o • • 00

o

c 4) ■_

.5 _Q

o § oo Z

< < rs oo CN CN -I _l < < { 12 tz

— CM

U- U-

oo

So

-g o E

< *

68

tu to < X a.

z D

to Z

X X UJ —i CO <

lit

f 81 es •— u. U) 0

4)

ÜÄ3 c

a. uj Li -Ä

■ft Ö TJ»0 -X 4) o o S3

I «o"-

• £

SJ-i c

5 e 0 U

9) u.

ÄZ

o p oo rv

00 tv

o o CO oo

o «o

oo oo

Ks

8 o =

00 cs • •

^2 < <

o o >o >o

o o tv o CM 00

8S? 7 =

O» tv

CO

CO i , CO

CN CO

o o rv ^

CN 8

in in o — tv 00

co -^r • •

in to

CO o* • •

<N CO

co -4- • •

p O K »O

o 00 8i 8 S

CM ^ *o c*^ o» Ov ON O ON

o

gl

rv «o

co ^- • •

CM IN. CN CN

00 ON

■^- ON

kO «O N© 00 II II

0 0 GO

^ o_o_ o^o^

SS !<!< !<!<

o o o m O O ON ON

O ON <0 IV ^ o o o

co o u^ io

— "«t CO CO

8 5? oo Q CN O I-N. ^•

O s

^J- CN • *

oo r«s

IV ON CO "«t • • • « m -^ to NQ

CN tv

ON IT)

9 CN^ ON lO

• • IN IT)

69

-a 0) 3 C

i

x x

oo <

4)

s I t-S 8 lit

f|l TO

.i

•r; »- '«I 5Ä 3 c o

a. UJ _i ^9»

UJ h- 5 3

lo-

-o

g §

o o as — o

O IT)

9 o II

o 8 CN O CO CO

o o

0° «O

88 00 <N CO CO

>— CM

o co • • m oo

Rs

c -JiCO

c .r U I

^£2 CN CM

00 if)

h 00 IS *- _

< < < < 2^ j

70

g «) b lO CO fN 00

CS O O 00

CM »O oo oo 00 O»

co~" o o^ O CN

• • — o

in oo CM o • • o —

CS

• • CO —

>q co • • — o

rs co — o • * o —

CM If)

o d

CO

^2 o >o • • o o o o

«o

o d CM CO

d o CO

d o ^- o CM CM

• • O O

CO^

■o « D c

_ o O lO CM CO • * — o

^ CM • • o —

in o ^ • • CO —

»O CO o» rv • • — o

"5 co in o • • o —

^ CM CM IV • • o o

i 0 U i

•

R = CM

X X LU -1 CO <

Cro

ss-H

ead

Spe

ed

(in

./m

in)

CM

8 • O =

_ m ^ . * - • O = _ ^

Ave

rag

e B

ond-

Line

Th

ickn

ess

(lO-3

in.

)

CO CM ■ •

CO

CM CO

00 Ov • • O CM CM CO

CO "V • •

co -^ • • CM IV CM CM

00 O • • ^ o»

■^ CM • .

00 K

•n o^ • •

• ♦

ci7

«o >o o'o1

•O 00

oV O O

Ll_ U_

CO o

CM CM

lO

^ci

55 u. u.

TT CO

o^S- CM CM

u. u.

71

i 2 « t)

lie §§ o -o o« oo |

9 CO

o d o d i|

• • o o o o |t

4)

1 i rP

• • o o

•o ^ i <n es n d d H

0

1 u 1 * 1 x

X

- i-- .5 IS: Kz • » II

UJ

<

Cro

ss-H

ead

Spe

ed

(in

./m

in)

• O : = = 1

Ave

rage

B

ond-

Line

Th

ickn

ess

(ICT3

in

.)

8.1

-25.6

7

.2-2

4.7

11

.3-2

8.8

11-2

8.5

Spe

cim

en

Num

ber ■ i 0 0

oS^

u. u. << u. u. 11

72

If?" «^ -I £.

o o to o •* CO p CO * CO oo CO CO 00 00 00

« .1 S » 2 8

00 8 <o o- CM CO ^^ o O — CM

CO 1 -H" -H -H i -H -H

Mi CO CM CO

o CM 3. s •o s o- CM

(SI • w o- -O O o o O O O

00 5? UO m ■* 5 io rv 5 CM CM o 12 CO -H -H -H 4

00

-H ■H -H -H* I

_ O s n 00 CO K 9 $ ?

W w -* o CM -■ d • o d d

§ 8 o 5 5?

-H -H § -

> ■H -H -H -H 8 s 8 S tu i-r- s rt ^ S i o o < "- «^ O CO 0^ o> r X o.

^ s ^O >o

z 3

•

| 2 7 1 s 7

o m o

-H 5? « CO 8 lO s 12 8

j sll K <o ^ >o a «I mm in

< ^ •^' r~ •" CM CO CO CO

i u.

Pre

cisi

on

Ela

stic

Li

mit

(psi

)

CO -H

00 tv

JQ CO

CM

8 +1

g io

p m

> M ^^ IS 5 0 •* — '5 CM ^— ^- ■<»• •* >o CO

s 1 i -^ J2 g oo ^ H! 5 "

-H -H* ■«' -)l" -H* ■« -H -U

3 to

00

IT) m' o« IT) o' >o'

UJ 3 s IF Z

UJ • « o i- t- iL- R s s z R s S 5

t-

• ■o

s 1 1 tu _J so <

ine

C

ros»

- ur

atio

n

Spe

ed

(in

./IT

s • o : s r d = = =

-1 O) 0 0 0 o O 0 CM O CM — O

1 § p \ \ < p \ \ *s. \ 0 0_ o \ 0 o 0 o CM CM o CM — CM

73

to operate. Accordingly, the P.E.L. and the a strains, f, and ^z were decreased. The fracture !

lO"3

were changed to reflect the shorter times available for the viscoelastic mechanisms were increased while the

stress remained about the same, but the fracture mode became more cohesive at the highest strain rate. The precision elastic limit nearly doubled in value so thot at 0.1 in./min the P.E.L. was 14 percent of CTp instead of 8 percent of (V at the standard strain rate. Likewise, the microyield stress was increased to 30 percent of the fracture stress compared to 15 percent from the Phase I tests. The strain at unloading, c, , averaged 1.48 x at the standard strain rate; at the higher cross-head speed, this value dropped to 0.52 x 10"3. There was a similar decrease in C«. At the high strain rate, the recovered strain (fo) was about 50 percent of C,; at the standard strain rate, €2 was about 0.31 e^

The effect of bond-line configuration was indeterminate for some properties because any changes were within the experimental scatter. However, at the low strain rate, it appears that: the modulus was increased slightly; there was little effect on the P.E.L.. the a ,, and the fracture behavior; the viscoelastic flow, € i, was raised. This latter obseivation may have been somewhat in error because of the possibility of instrument drift during the long loading-unloading times encountered at a cross- head speed of 0.002 in./min. The same generalizations can be made for the results at 0.1 in./min with one exception: the strain at unloading, C., was diminished about 45 percent when tapered bond-line configurations were used.

Shear Tests

The effects due to changing the strain rate were minimal for the shear tests. The most significant change was in the elastic shear modulus at the highest cross-head speed; this was an increase of 26 percent. All the test data are reported in Table XXI end summarized in Table XXII. In the Phase I work, it was shown that the strain at unloading, ?j, was about 15 to 30 times greater in the uniform bond line than in the tapered |oints. Similar observations were made here for the two alternate cross- head speeds. The variations in €, due to the nonuniform configurations ranged from a factor of 7 for the low strain rate to about 19 for the high strain rate. Changing the strain rate did not alter this marked effect caused by introducing a tapered bond line.

The tapered bond-line configurations raised the elastic shear modulus values above that of the uniform bond-line case, with the effect being more pronounced at the higher strain rate. The other strength values and the fracture behavior were not appreciably altered by the tapered bond lines. For both strain rates there was a mar'red decrease in the viscoelastic flow. The strain at unloading, y,, decreased about 85 percent at the high strain rate when tapered bond lines were used.

74

< X o.

Z

D _J <

o o o

< Ui I to

X X LL) _J CO

<

41 o

^5

0) >

o o

'x 2 " o $ (^

• — C- v>

c o

ß ^ J

« li-

o co o

o 5? co — o oo

.SJloo i" to 5 w

g J, J<:CO

u E

m o o =

o o —

o o O CO

o o oo —

o o o o r- 00 O K

O — O

O 8 00 O« co ■*

00 o o lO >o <> rv rx CO -* m m

oo »o o io 00 Ö 00 00 co '«t tv. '■r

«8 O CM

CO CM

CM CM oo ^l- §9

CN CO CM CM CN CM

CM O

4$ 1 1

— CM 00 CM • • • • o r-s o. o •— '"

co ■^■ 1 1

CM CM

CM 00 c c 1 1

CM — CM CM f—m tm— CM CM to to to t« << <<

»n CM • • o^ — CM CO

I I CO CM

«O N. I I

CM • •*

If) CM CM I

I 00 >0^

^2 PP ^ci

co to <<

CM CM CM CM to to <<

o O CM O

O O Q O O ^ h> CM m O^ CM OO

8 ^

o co .— CO CO CM

o CM =

«O O

O CM*

IX CO f— I I I

— CM O

to to to << <

O "O

o o o o CM 00 o» oo

o CM "«t CO CM o ■^- O CO o co in >o — K

00 K 00 ^ CO ""T

^t try CM«O r^is Tfoo oo-tco oo — — r-' — — — r-' r-' — ■—' ■—' ~ F-' CM*

CM CM K = == = s = s N. = = c s

"t -*

- «•>

75

(0 o c c o U

X X

CO

<

>

£ a, O y -fl o

He

•2 a ^

IS

'x 0 v. 5 8^

c O

ü .2 »

.5 2 o o

hi

«5 c ü L

c

u £

o —

88 CN O rv. o

00 00

co o •— CM

<o in »- o»

X X i i

o co o

o o O Q o ^ 00 NO

^o oo

00 00 co

o o CM CS

o CM = : s

30.7

28

.7

27

.4

34.5

CM CO

oo *o

■^ W)

P= £<*>

< <

2M JM CM CM co to < <

76

-a <o D C +- c o U

X X

CO <

«> «) i- " > o

10 _J w

CO 1

CN O

1 — O

13 O ^- 4) C

8 8^ c ft C

rs. o

o o« • •

"t CM CM ■—

O O • • N- CO

O CN « • 00 CO CM —

CM

IT) O

o =

CO 00

CM »O

00 CM

»O CM

00 o CO 00

Tf <0 CM CM CM CO 0 a « • • o

O — -* CM O '-

00 — o 00

Tf >o •«0 CM CM 00 • • m » mm

O — lO CM O —

CM

o> .£ 0) J

2 • %<$

• ♦

CM CO u-J -^l- CM CM

4» c _yco u 1 r- CM 00 CM • •

1 1 CO CM M3 00 •

<£ 0 0 t^ O O O ^ 1— v^ ' 1

c CO -*!■

1 1 >0 N,

1 1 ^2 4) L. CM CM 1— 1— 1 1 tj{ CM 00

CM CM \\. O O

Spec

N

um

1 1 «^ ••' » <■ CM — CM CM 1—• r"* CM CM

CM CM 1— 1—• CM CM CO to t/) CO to to to to << << < < <<

-«t V) >Q LO *o «O

CM O O • • •

co IO rv CM ■*

-0010 • « •

"«t CM (> CM lO

CM

CN

«O O • • • O CM f-

to to to < < <

•o 00

o >o

-«t o o^ o» « • • «

.— r>. •- 00

O T)

r-' CM*

• • co rx

.- .A CM

CN

77

0) D C

i o U

X X

<

2 « K""

CO

n O

CO

CM O

CO

X c

u «re

« c u I ^2

c

v» E

z 1

O» CO

o —

co -o • • o o

CO t>»

CM

o

• • o co CO CM

— -«r X X

I I

to m <<

o oo

o o« • • — o

o o

o o • • — o

• • CM CO

CM CO • • oo >o

"* m

~~ £<*> CM <N CM CM U0 CO <<

78

>

< I

i ^ Z

5

>-

5

Of < UJ I

X X

<

« « ►-

1/1 _J ü_

n i o

u g a

S .-S a-

S J .§ " CL LU _J -S-

-2i 52 uj in 2 ^

0 Ac. c

? I •= D —I O)

<£ c o u

in

o

CM

o

u-) -o

o o o

K ■^ t >o O (N •" «( « H H

CM n (N n 5 N i— ro u-( ■t o

o

5 R — o <*) CM

O

CM <N \ CN

CN

H

■O

— P CM CN

IT) -o

in ■*t ■*

IT) tN CN +^

CN

H *l *■ m H >o K Tf *t in

CO o d o d d

■* ro r^ o r-N oo

>i 00 CN H O ** ^i -H ro H H

O o cs o

^ oo m

CNI — rö — ro — —*

o § CN

O ro 00 g 8

CO '-~ CM *— ■— +* CN ■H n t< n t ** *< in u-) o in Q a CM s § K Q CSl oa o ?5 m o- r^ 00 o o 00

♦l u")

R

O

CN

I CN CN

CN

-O

o o CN

1/1 — ~N 2? ro ro Q O — ti ro ■— ti o ^ **

« o 5 ^* •* — OO tv

s s ^ ^ -H 4< « -tl ■♦ oo is ■* ■o — •« — (N ^t <H ^

CN CN O o

■H « fl o rv o

— oi —* CN

CM CN

79

Lap Shear Tests

Although the test program called for only two alternate cross-head speeds in this phase, four speeds were used - 2.0, 0.1, 0.05 and 0.002 in./min. The results, shown in Table XXMI, are to be compared to those of Phase I (Table Vlll}. The highest cress-head speed produced the largest fracture stresses. As the strain rate decreased, the fracture stress also decreased. Under the standard test condition (Phase I), fhe average fracture stress was 6750 psi, somewhat less than the comparable value of 7300 psi for the 2.0 in./min cross-head speed.

PHASE V, CRYOGENIC/STRAIN RATE

Tests made in this phase differed from those of Phases I, ill and IV in that they were made at the highest strain rate and the lowest temperature. The purpose wa> to evaluate strain rate effects at -650F. Each of these experimental variations was studied separately in Phase III (low test temperature) and Phase IV (high strain rate). Thus, the results presented here must be compared to those of Tables III through Vlll and XIV through XXIII.

Tensile Tests

The test results for FM1000 on aluminum adherends at -65 F using a cross-hend speed of 0.1 in,/min are shown in Table XXIV. A summary of the results is presented in Table XXV. This combination of high strain rate and low test temperature produced the highest average values for the effective tensile modulus in the entire FM1000 program. Recapitulating, the average effective tensile modulus for the standard cross-head speed at room temperature was 5.0 x 10^ psi; lowering the test temperature to -650F raised this value to 7.15 x 10 psi (Table XV); increasing the strain rate brought the value up to 6.1 x 10^ psi (Table XX). When the two chonges were made simultaneously (Tables XXIV and XXV), the average effective tensile modulus was 8 .6 x 10 psi. The fracture rtress in this phase averaqed 17,900 psi, a value only slightly greater than that of Phase III (cryogenic). The P,E,L. -aloe was comparable to that of the cryogenic tests (3750 psi vs 3420 psi here). However, the microyield stress values in this series of tests averaged 10,100 psi, by far the highest seen in the program. This a was 56 percent of the fracture stress. The mode of failure was not much differenf from those of Phases I, III and IV. As in the earlier cryogenic tests, there was no viscoelastic flow at any stress level.

The major effect of the nonuniform bond lints was to increase the elastic tensile modulus by about 12 percent. There was also a slight decrease in the precision elosti'; limit, while the other properties did not change very much.

80

t/1 < X a.

z

8 o

< LU X

5

X X UJ

eo <

4)

S o

0)

9)

o u

0)

ü 9» :^ O v- irt

-o O ^^ 0) C

8 S^ u a c U ^ C-

C «n _

c 0) u E *» •= JQ u E

co S o o o CM O

8 8 8 IN

o o (N CN O O

IO IT) o o • • o o

o o

88

c» CM

88 • • o o

o o O O O «1 o o • • • • a • • • O^ K ■^r co — "O co es

— es co ■^, co -«t m -o <s es es es

i i 1 1 i i i i -i _» _l _J _i _J _i _J < < u. u.

<< < < < < u. u.

81

Ul

<

5 z 3

o i

z

X X

<

?!

lie O u

0) i- 3 u 8 S. 2iz Tu-

8 *"

5Ä 3

c o — o <n .—

•~ *. ■*-

£ ^ .§ «• Q. uj 13 3

<■» jü — 's. t; « -So o S o o

UJ »- i ^^

IJo"-

I fc"S

c

u i>f c

?-» c • « c u •

4> i-

JJ! o E ÄZ

«OO OO OOW) PQ

CM OO

8« 5 CO

O s

88 88 CN

88 8 CM K

" 8

8 8 — CN >o 00

CM CM

■o o

o o co go ^ S "O O 00 CO

00 8 CM t

CO uo »o K co o oo io a« • • • • • • • • • r^ oo <> K ON Ch 00 o >o

^r o»

^s? ■«U- >o ON NO CM CM CM CM

00 CM 6J> «nJ-J» ^4 • • • « • ■ • • • O O^ »o oo 00 K 00 — ON

82

■o 0) 3

JC *• c o O

X

<

2 • to 00

CO I

«3

CO

~'2 w

CO I o

0) g i «=' o> —1 c -

c 1 J^CO

c 4> 1-

E J; u E

3 z

I

o =

>o

00 CM

»o co I I — CO

•o o Tf (^

< <

^9 Tt >o <> vO CN CN <N CN

ci>i tO-iu^ dL4 • • • • • • • O 00 oo rv oo •- Os

o <N co ^- o o -- ^" w^ r^ 00 — 1 1 1 1 1 i i

0cv,^ o o o O 0 o o

oV- o o o O 0 <N <N CN CN 1- h- 1- 1- 1- h- 1- < < < < < <<

83

> LU

< X a.

$

z

8 o

>- <

1 D CO

z

> X

cQ <

■i V D I"* CO

I CN2

I — o

E |^

'>. p <* O 4) r^

— um

C

111 ?

z - 3 IL g «a T>«O

-o

2 Si c

c o 4» p .1 i i g

'c

8 eg

o o oo Cs

8 CO

in

io

9 0^

I

§ a 2: -fl -H -H

00 00 O

§8 5? r«i '— '—

-H

00 £4 C«4

-H

8

8

s?

-H

oo

o o oo —

M -H

00 9-

CM CO

ts

s o o

00

o — -H -H o r«s • • O* 00

o CM

84

Shear Tests

These shear tests combined low temperature and high strain rate with results as given in Tables XXVI and XXVII. The average base-line shear modulus (Table VI) for this system was 1.43 x ICP psi. Increasing the cross-head speed raised the modulus to 1.8 x 10 psi; dropping the test temperature raised the modulus to 1.9 x 10^ psi. Making both these changes together (high strain rate, low test ,. temperature) produced the highest shear modulus seen - an average of 2.15 x 10 psi (an increase of 50 percent over the Phase I value). The P.E.L. showed a 100 percent increase over the Phase I value. The fracture stress was most greatly influenced by the test temperature and not very strongly by a higher strain rate. The fracture stresses of this Phase V work were essentially the same value as for the cryogenic phase (III). As with the tensile tests, the property showing the largest effect was the micro/ield stress. The average base-line T was 620 psi; lowering the test temperature increased that to 1470 psi. Then combining the low temperature with the high strain rate, the microyield stress was 2370 psi (18 percent of the fracture stress). In Phase I, T was 6.5 percent of T-. Viscoelastic flow was not observed at any stress level during the experiments of this phase.

The shear modulus was increased about 16 percent when tapered bond-line configurations were used. Another effect caused by the nonuniform joints was an apparent lowering of the microyield stress and the fracture stress.

Lap Shear Tests

These lap shear tests were made at three cross-head speeds - 2.0, 0.1 and 0.02 in./min at -650F. The results are shown in Table XXVIII and should be compared with those of Tables VII and VIII from Phase I. The average base-line fracture stress value was 6750 psi. Lowering the test temperature increased the fracture stresses at all cross-head speeds. Strain rate had a much smaller influence on the fracture stress. Referring to Table XXVIII, the greatest average fracture stress was 8200 psi. At -6o F, all the specimens showed complete adhesive failure, listed as zero percent cohesive.

That the low test temperature had a stronger influence on the properties than did the strain rate is consistent with the observations of the tensile and shear parts of this phase.

85

0»

V

o o

"-5 o o o o

o o — .- o o

< X a.

z D

o o o

< LU X

> X X

CO

<

4i t/> u a. E *■ IJL u. lO K

TJ <D X o t/>

o s • — u V>

5 4» 3 c o </> u o

0

4-

^ E « OL LU Ü 3

1) 3 M •*- (A o

1-

s 5 Q-

o o "J Ä 5 w

^5 E LT

-o o <D X

i ^ co ^_^

0)

O) .

u I •- o u^s

c

u E a> 3

o o o p -sF ro

<N :

rs «o N is

I I

«/> I/» <<

o o

CS CN

O O

'"t (N

"t Tf IN

• 00

^ 9

12-3

1 9

.5-3

r^ u1)

CO X (s

i i ?^ eg CM CN

CN CN CN CN t/) to < <

CN m X

i i

LO CO

o o o «o O. CN CN CN

o o o o o io 5?S o o o io

O» is. r^ o«. hs ITS •o oo CN CO CN — •— •— — CN

o o o o IO o io o CO 00 00 CN 'O 00 co o O N IT) O- Co >o

IS.- CNIS CNCN 0OCN

i—' CM* CN CN CN CN -^ CN

•O I = = = = = = =

CM CN • ■

'«t CN CN CO

I ■ IO CO

IS 00 I I

!^ CN CN t>0 OO

< < < <

86

3 C

c ^3

> X X LU -J CO

<

CO

CO 2

CO I

CM O

CO I — o

CO 00

CN

C s

lO

C>4 C =

r -o o 4)

8 r-' w n C U ^w

a> .E $ c

c a> u

0 E 11

CM =

GO "t ^J" (M CO CM CM

44 JiA • «

■<»• CM IV 'O CO • CM CO • •

r^u«, 1 o» (M ~- lA CO •~ r—

CO CM X tv m X

i i i i rs oo m co CM CM

CM CM ^ ^ c^.^ i i w "^v^v

CM <N *-• f—• CM CM #— <— CM CN i—• »— CM CN to to to to to to to to << Uu U.

< < u_ u_

<< < < U_ Li_

87

^ > in — U.

lO -J w

^^ CO

XE ^^ V

CO __ E _ > ^b *c _ S 'c LU ?s~ C- to < X a. 2 ^

5? 8 8 lO CM '* CM rv CO

C 4. 11 g

M -H O

D U. U1 K 5 CM CO CM z ^~ """ '~ •"*

2 5 .2 o m lO

O ut § P5 lO 3? < 2 JJ ^. ■H -H -H -H

o 8

U 4> •— 8 CO

m CM CO

if) CM CO

CM CO CM i— CM

c i u.

> Pre

cisi

o

Ela

stic

L

imit

(p

si)

-H

g -H

D=5 -H CO CM

00

+1 00

5 <« ä r^ rv -o R

CM • CO • o CM • o

tt UJ io ^ C- -H -H -H M < O« «O CM O UJ • • • • X

• '— CM <N CM to ■a ?"c:

« «ft «o «. »- »-Hw- ^o > ■o

1 = = s

X 0 —v X V C

1H UJ —i CQ <

c^ = S = c 0 ^

« E •- 3 -1 0> Ö *~

^1 p s ^

16 \ CM tmm CM o^ CM f— CM

88

0) >

£ | 5 -S *: a» o o -S Ü

* ^1^ o o O O o o > UJ (/) < X a.

(U V i.

5 3 *- in

D Z 88 85? «8

oo >o f— »—• IT) C>

D

u- uo v3 00 K >o 00 t^ rv

-I <

8 o r— (U c

5 l"S^ u.

CN CM % «

o o » • a i o o • •

< U »/> ^ CM <N o o o o UJ X »/l

Q. < _l

• _j g — M -xi J«:« > c " i O O o o m U-) u. o • ■

t-x 00 • ♦

^ CO • •

CO CO IT)

UJ •#>» -1 0 GO <

c

1 M *-

IT) >0 Os o IN. 00 a>

tN CS — CN M

« i _l _l 1 1

—1 —i _J —1 <

£z < < u_ u.

< < U- u.

< < U_ U- «

89

PHASE VI, BOND-LINE THICKNESS

To study the effect of bond-line thickness, it was necessary to use a paste-type adhesive in place of the sheet adhesive having n carrier cloth (FM1000). WI*h FM1000 there is one optimum thickness (0.008 in.) that must be used. EC2214 was selected as the paste-type adhesive for fhe thickness study because it is widely used in the aircraft industry. Standard tests were performed as in Phase I except that the adhesive/adherend system was EC2214/aluminum.

Tensile Tests

A large number of specimens were prepared and rejected because of unacceptable

bond-line thicknesses. Finally, 9 specimens with thicknesses ranging from 0.0019 to 0.0054 in. were accepted for the thin bond-line configuration (0.004 in.), and 10 specimens with thicknesses of 0.0183 to 0.023 in. were accepted for the thick bond-line joints (0.020 in.). The results are shown in Table XXIX and summarized in Table XXX. A typical tensile stress-strain curve from these tests is shown in Figure 14.

Considering average values from Table XXX and the experimental scatter, there were no appreciable differences in properties between the thin bond lines and the thick bond lines for EC2214 adhesive on aluminum adherends. The average effective tensile modulus for all tests was 0.91 x lO^psi. In contrast, FM1000 on aluminum adherends had an average effective tensile modulus of 5.5 x 10^ psi. The values for the P.E.L. and the cr for EC2214 were nearly double those for FM1000. The fracture stresses and modes or failure were comparable for EC2214 and FM1000. Viscoelastic flow was readil/ observed in all specimens.

Using tapered bond-line configurations with EC2214 caused a change in the viscoelastic parameters. At similar stress levels, f,, for the thin bond lines, averaged 1.8 x 10"3 for uniform specimens. The average €, for the thin tapered joints was 0.27 x 10~3/ a reduction of 85 percent. A similar effect, but not so pronounced, was seen in the thick bond-line specimens. Introduction of tapered bond lines had a mixed effect on the other mechanical properties, with consistent trends in evidence.

Shear Tests

Because of difficulties in preparing suitable specimens, a number were prepared and rejected. Finally, 8 were accepted having bond-line thicknesses between 0.0019 and 0.0056 in.; 9 were used with the thicknesses ranging from 0.021 to 0.0288 in.

90

00 < X a.

s

z D

U

to Z LU

X X X UJ _1 oo <

a»

^-8 8 £ 5 t

4) 3 *- U

LL. CO b"-

2

c o

D. UJ Jj 3

a.

UJ . c - -

a> i-

ÄZ

o o o o o o

«O '«t Os — o o

^ in

CM O — CM

•ft « TJVO

UJ K- 5 w

CN O

O s =

O K «o • • • — CO CO

«"> ro -* I I I

CO «N CM I— I— t- < < <

Q8

88 00 «Q CM lO

CO ^ in o — CM

s^ O K 8 öi iO • • • » « — o -- o o

CM K = =

CM — • • 00 CN CO CO

I I CM — • • CO "t

o o o o

Q5? o 2 o ~ — "«t CM a« <>> o« o o

o _ CN

<MS

o o

^ CM

8 0 00 • •

— o

"t J^

CM CN

O* CN

(O -^

88

^ o

o o

95?

? o o 00 c CN CM

0O K rv CN • • o o

• • CN Ch CN r-

Ar ON

if) —

< < < < < <

o o o o Ov o» o o

o o o o 5? 5? CN CM ^- — ^ O

8cN öS? 88 5? S

O Q O O O -^ «O C^l co io in in CM CM CO F-

s • in 0^ • o • o

00 • o

R

CM O

o = : =

o o «"> »n • « • • co o^ o» oo CM — — ^-

o TJ- o. tx —

CO — CO — h- I- H- h- << < <

91

a> D C *- c o U

X X X

<

4) #>

0) 4)

l-S 8

ti S Or P w — U.. CO t)

P «> u a» i^

•— I- w>

5Ä 3 c o

9) O E w c — •— o o. uj _J c?

.H ^ | H

-i Q> o o

8 Wo""

a>

c B 4, ^co «j 1= y • > o < Ä h- C

c

Ml u E

o 5?

o

o oo

Q^

Q O o o CM CM

CNJ 00 00 O • • O '-

CN

CM O

CO 00 • » CO -Q

i i CO 00 • • 00 '— — CN

o o o o

§8 o o

o p

o o • • o —

<<

• * SO K r> ro 4-i — o CN CM

CN If)

I t o o

<<

o o o o

05 o Os o o r^

S8 ■* CO

o o rx K ^ o CN i— O CO -- CN CO CN

uo o> O Ü0 • • o o

o — • • CO -o co n

• • o oo CN —

IV — — CN

o'o'

CN CN I- I— < <

92

0)

C *. c o u

>< X X

CO

<

</» «I

£ > 00 _J #

CO I

r>2 w

V

CO I

CN2

CO I

. o

«> 4» O h- »- 2^

0)

O). c —

Ave

ro

Bon

d-

JiCO U 1

22

c 4> L.

JJI

Spec

N

um

^ 0 »o oo

• • o o

— o

CM O

CM

cs o

o o o co o rx >o oo oo oo oo oo »o 00

• • • • « • « o o o o o o o

o m

o o o o o o o

— ^ CN — ^ ^ CO • • • • I • •

o o o o o o o

O K >o • • • — CO CO

CN4 —

oo o> CO CO

I I CN —

m •

co ■*»■

m co -»i- i i i

CO CN CM I— I- H- < <<

^ rv

CN CM I I

O CM ■ •

CO ^

K •— 00 «O »o oo "O »o

IT) O 00 lO CM CO CM >- * • • • o o o o

CO O "«t W) r- CM ^- —

■ « • • o o o o

00 Q CM O CO Co TT CO

• • • • o o o o

CM O» CM —

4c^ O • o in *o • ■ • • • CO O* O 00

IT) •— CM F— »^ F—

o -«t o^ rx — i i i i

CO >— CO — \— \- i- y- <<<<

93

I L

CO I

w

CO

CN2

■D V S—Si

3 CO ^C 1 jl „ o c u/ r* o U i

■

X X X

ZZl LU —i CO •o < o ^,

1 I

«lie

« r^ co oo oo is

~ 8Rj gg o o o O O O

82 = g o o o o o d

8$ 88 8^ o o do d d

R

CM o O =

«00 O^ •<) o —

00 — — eg

00 N. CO CO

4^ j. ss «o

— o CM CM 82

^< << .<<

94

S £ b

i

CO i O

4)

3, c *. u.

U- oo b

T3 J) X 0 *, o o C^" 2^ ̂Q_

c 0 •^ ." "u t; •— •—. (U O £ *«

^.9: O. tu

- Q- -2 5 UJ 1—

0 O 5^

QL 15 E li_ w o0

-o 0 ^, 01 c

o o • '— o u

it en sj

O O CN t^ 00 00

+1 00

000

00

s -H* 00 o

o IX 00

8 9 8

CM O

-H 10

o +1 10

-H -H o 00 00 CN

—' o

n in o o -H

iO CO

10 CN IT) -H

o

10 10 lO fs CN

^ ?? .7 +1 10 10 iP K rv K io in CN o o ^ — ^

in in 00 in CO CN +j -H in in in 00

s o

— — CO CN

-H o

CO

«0 K — o ■M -H-

IP co ■H 00 00

o o

CN

CN O

IE t- o o ""'

CN — O

1- «kT ox ^ O CN — <N

-H

O m -o +1

8

in

-H

CN O

-H' CN CN

S3 O

-H

O

CN O

-H" CN CO

O 00

CO o -H CO o

CO o +1* CO CO

in in in CN CN <N ? SI H m m m rv CN CN <N CN CO 0 0 0

o CO

-H o IV o-

m oo CO

-H m

R -H O CN 00 CO

Ö ^ 5? -H -H -H O 8

O O

CN CN

o 00 CO +1 o

CN

rv o -H -H co m

m co o o -H -H" CN CN

OO —

CN IV

CN O

U

O o u

& i£ t0CN °-

O CN — 0

CN

95

12,000

10,000

8000

S. 6000

on

4000

2000

Strain (10

Figure 14. Tensile Test, EC2214, Aluminum, it*, Cross-Head Speed - 0.02 in./min.

o7o0.

96

The nominal bond-line thicknesses were similar for both the tensile and shear specimens. The shear test results are listed in Table XXXI and summarized in Table XXXII. A typical stress-strain curve for these shear tests is shown in Figure 15.

Comparing the uniform bond-line properties for thin and thick bond lines, it was seen that the thicker specimens had somewhat lower values. The average shear modulus decreased about 6 percent with the thick bond lines. The thicker bond lines also caused the following decrease in average values: precision elastic limit, 28 percent; microyield stress, 28 percent; fracture stress, 8 percent. There was about a 23 percent increase in strain at unloading (>i) for the thick bond-line samples compared to that of the thin samples.

The effects due to introduction of the nonjniform bond-line thickness were not very strong; no trends could be observed. The one exception was the viscoelastic behavior, especially 7,, the strain at unloading. For the thin bond-line samples, the tapered joints had an average value for y. 65 percent lower than for the uniform bond-line specimens. For the thick bond-line tests, the reduction was 77 percent.

The tensile modulus for EC2214 was shown to be only 17 percent of the value for FM1000. However, the shear modulus for EC2214 was 2-^ times greater than that for the FM1000 under similar test conditions.

Lap Shear Tests

The two thicknesses chosen for testing were nominally 0.006 and 0.014 in. Table XXXIII lists the results for these lap shear tests and includes one test for a specimen with a bond-line thickness of 0.022 in. The average fracture stress for the 0.006 in. bond-line joints was 5950 psi. Increasing the thickness to a nominal value of 0.014 in. did not materially affect that value. The specimen having a bond line of 0.022 in. had a lower fracture stress (4900 psi).

97

V ,>

2 fl, 3 "O

O O O o »o 05 K ss oo 8 532 2°o CM O

£ _ ■2 a'"» U Ä CL 5?8 ^8 Q O

Q CN is o ^ 5? 8^ CO IS in TT O O- O. CM «O O CM

if ^ K"- 00 o oo oo — 00 — ^ o^ oo oo 00 00

^ .2 "x

> 2 r ^ 58 co o S8 ^ o o o o CO "O 00

o rx ■- k. in — 00 O K 5» »o oo o UJ 5 Ä 3 CM —

< I C

o is .y

z

"Ö *• .^ ^-

Q- S _i Ä

Tf 00

CN c? K CN

5 CN oo p CO CN

O CM CM in -^ co CO CM CM

o io

5 S ^

ALU

u .2 »

^ J Q o ^ in 5 C

m "- ■^ CO .- IT) Os CN CM 00 — o- o • • • • • • • • • • • • • v

■* en ro CO CM ■* CO CM ^ CO CM CO CO CO

CM , <N *. a. U UJ 5 £ "^ cv»

rs = • s . _ - - - •

^ UJ X

1 T •

X X X

• O :

CO

< .12^ • rv. co • • • • o rv • • CM CN

0) _j c "* 55 O CM — CO -o >o h- CM CM CM CN 1 1

'O O- u«,^ J>^ CN «O • • NO CO • • <«8^£ • • • • • • • • IO 00 CO co ■*

CM ^• CO m — CO CM ■**• CM CM CM CN CN

•o o- u-) •— 00

•N >- X

1 1

?M<N CM CN

oo X i i O CN

53 >-X

i i

CN CM

*z ■ i X\ \\ w ■^ CM ^

— CM ^^ a «

CJ f< £66 CM ^ r~ r— CM CN ^—" r—' CN CN CM CM vn in I/) (/» V> CO v> un co co to v> uo < < < < < < < < << < < < UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ

98

-o

c ■I- c Ö

X X X tu _i oo <

'(ir M>

« s 3 -C

g-« 8

£le o ^ ß^

0) u. ^^^

•2 a« C«J CO «5? 00 00

CO IT) 00 00

^o 4) 'x 2 a u 8 c^ O CM

00 ■— §3 5 ä -S» c

«o o> "t CO

Ü o E „

UJ iS 5: 3

-o

^ CO w

O) i c

Ilia

z 1

o o — CN) CO CN

— -"t eg CO

eg

IT)

o

CM »n i

CM I

co «o cs eg

co o» X X

i i

to l/) < <

eg eg

-^ co

o

° 00 I ■

— 00 eg eg

35 cs eg IN eg co co <<

99

<ü

C •^ *- C o u

X X X

<

8 'S a) w C >

CO »J

CO

co »

CO I

CO I

— o

■ft

a>

-o

I - U «>i w

4> «1 ^

£ c 0 i. < «8 K- C-

c <U i.

JJi " E

1/) ^.

o o» «O IT) m »o

^- CO o >o

tO IT) >0 lO K 00

o o p 00 00 (^

m -«i- O CN

«o o O» IN

o —'

>o m m ^? a a • 9 •

o — o o o o

O Q O O 00 -^ (NU-)

lO ^O — CN

M O O —

.—' CM

0 P CO IN

CM —

O 00

o o

8 8 «N CN

O Q O io 0O

O O 00 no

CM -sf «O ^- —

Tf lO IN • • • O — CM

^ o »o CM O 0O

IN CO

d d

5? CM —

O 00

CM IN s

lO • o =

O IN o IN CO M- IN • •

• • • • • CM CM

55 O CM — CO •o <o CM CM CM CM

>0 CO • • o o ^^ J>± 1 CM IA • •

• * • m • • • • m oo co co ■^■ CM ^ co m — CO

iO

CM Tf CM CM CM CM CM

«- oo ^X i i

oo X o CM >- X

1 1 m m r^ ^^

CJ ?^ i

^ CM CM

^ IN CM CM \\ ^ •^r CM -^

A66 CM CM

1 1 — CM fj r>i

w» m CM C-t ^ Ö

CM CM *— *—• CM CM CM CM 00 (/) oo oo oo oo 00 to 00 00 00 CO 00 < < < < << << < < < < < UJ LU UJ LU UJ UJ UJ UJ UJ UJ UJ UJ UJ

100

^ *ö aw IX» V 4) K

1 Ä »^ O "t co io i *" _J ?1^ CO io >o "«O I

CO

too o m o >o ^ 3 iO o O CN 1

■ • « « II <N — r— ■— II

^ CO

CNO O CO IT) O 1 N "" • • • • ! 1

— o o o

^^ en

— o OS co «s >» •— • • • • 11

CO — fm r— II '° 4) 3 C •

1 0

U

1 '

"g ^G: s s 11

>< ■o

8 -c X X ^ -o J UJ 2p 1 <B m < u iJr^ «

O s * * II 1 »—

Ave

rage

B

ond-

line

Thic

knes

s (1

0-3

in.)

2.2

-45

.4

40.2

8-

48.0

■t CM

sis 1 * ? i — « Oi <M II

c

.11 s* — O ü E 1 .' ^~ 9mm t *l ll/.l

•ll/l

^4 ^^4

p>w «— <N <M 1 it 00 (/I to |l < < < < UJ UJ UJ UJ 11

101

I/) < X a.

D Z

Z

X Ui

<

_ o

2 »*5 S ^

is s »^ •—UM

c 5

u sill

llli 8 -c

o o • i- Q. C

H u ,2

— O O CN IV ts >o >o

» 8 $

eg CM

00 • ■» . Q -H oo >o % 9 o o o o

oo ro rs S ^ • • • . +1 -H -H 41 — o o ■* <0 ^ (N CM

8 S I

15 ^ 8 ^ ^ <M ro -H -H -H "> p <n o -H

I S f5 ^

00 o tv <0 oo ■« -H -W -M

In S rt «

<N «o n

•H ■« +( p» o» OB <o ro ts n co

CM ^ K -«r N. *0 ^0 ^O

o CM CO

• CM

m io o in --' d d d

o o "". co _• -. 00 ■ -H -H -H +1 CO vQ O CM ■* -O

O« CM CM —

IO IO IO CM CM ■* 00 >— •— ■H -H -H fin io

00 00 00

8 +7

IO 'M 4> <o io rt S - M

— K i-v o>

o

s? = IO tv CM

IO ^ IO

A i

CM ■* <0 IO

H M -H -H O IO 00 «

CO CO CM CO

IO ■O

o 1 1 « O : R = 'c T -8 -8 _

5 -e t ^E -8 J£ ^

b» f— ^ ?M z. f fM ^

1 ^ ~ ^ i 5 x. ^ ft CM

CM ft — &

102

10,000

8000

6000

t o Q_

in

<D u.

4000

2000

500 400 300 100 200

Strain (10

Figure 15. Shear Test, EC2214, Aluminum, 72°F, Cross-Head Speed - 0 .5 i n . /m in 0°/0°.

103

9) >

£ S 3 -S ■h « Q o -g u

> u_ 2g o^ o CN

o CM

o CN S o o o

UJ in < X a.

S

5 3 z £ 3 « 5 D

<

a z u» u. oo -i?

g 8 o g 8 00

8 8 8 >o -0 lO >o m ■*

H -<*■ r» CN CN U LU

«. ( oc a, «; . < c S c

j c - X T3 O 1

c •- O o o o o o o o a. 5

&£~ ^ • •

CN «

ID •o •

CM

" '" " CM

—*

X X X c LU m i_

.iü ,- 'N "t CO <o •o r^ <

oo Z

1 -J

1 i -J —1

1 —1

• «j

l- < < < < < < < LU UJ LU UJ UJ UJ LU

104

PHASES VII AND VIII, HIGH TEMPERATURE/STRAIN RATE

The specimens for these phases were Metlbond 329 on aluminum adherends. Phase VII was devoted to establishing the base-line properties for this system. Tests were made at room temperature using the standard cross-head speeds and bond-line thickness. In Phase VIII, similar specimens were used for tests at löCPF using the standard cross-head speeds as well as lower rates. The same nonuniform bond-line configurations were used here as were selected in Phase I

Tensile Tests

The tensile test results for Phases VII and VIII are presented in Table XXXIV and summarized in Table XXXV. Typical stress-strain curves for tests at 72° and 160 F are shown in Figure 16. In general it can be said that raising the test temperature lowers the modulus and the strength levels and increases the viscoelastic flow. If the strain rate is lowered at 160^, the same trends continue. Due to the relatively large experimental scatter, it \i difficult to delineate the effects due to the introduction of nonuniform bond-line thicknesses. However, it can be said qualitatively that tapered bond lines raise the modulus; lower the P.E.L., the a ,, and the fracture stress; and diminish the viscoelastic flow slightly, mys '

Considering average values for the various properties, the effects of higher test temperature and lower strain rate can be discussed. The average base-line modulus under standard conditions was 8.3 x KPpsi. Raising the test temperature to 160oF lowered this value to 6.5 x 10^ psi. A further decrease was noted when the cross-head sperd was reduced to 0.002 in./min - 5.95 x 10^ psi. The absolute values for the precision elastic limit decreased as the temperature was raised and as the strain rate was lowered. In addition, for each of those two factors, the P.E.L. as fractions of the fracture stress also decreased in the order 39, 25 and 17 percent of Or. The same phenomenon was noted for the microyield stress. Average absolute values of a and as a fraction of C- were 2700 psi (58 percent), 2000 psi (40 percent) and 1450 psi (33 percent) for the test conditions, respectively. The fracture stress was not changed much by test temperature or strain rate, and the fractures were nearly all 100 percent cohesive in nature. The viscoelastic flow (f. ) was increased about 47 percent due to raising the test temperature. Lowering the strain rate did not change this characteristic very much.

Shear Tests

Results for the shear tests of Phases VII and VIII are given in Tables XXXVI and XXXVII. Typical shear stress-strain curves for tests at 72° and 160 F are Kown in Figure 17.

105

> Q z <

< x o.

z

—i <

CO

Q Z o CO

oo z

X X

<

> M

0» «

5 a» o

2 S -z iZ2e

2 « 's P w ^—

u. to b"-

0)

u 4 r^-

5Ä

c o l/t

u u

4- ^_,

a. 0 h w»

LU —i A

o c o 0 — 0) ^ _ LU »— i ^

« in""

s X

5 2 O

o>

« c 0 •

c

ÄZ

8888

00 Ö CN § 9^«

O O Q O >o CO O 00 CN CO CO CO

(N «O lf> 00 — CN CN CN

00 CN O^ O» • . . . «o oo h^ oo

CN

CN o

rs o ^- hs « • • •

IX O tv 00

88 88 88

< < <

O Q si

85? CN CN

8 82 «o o*

IT) 00 • •

00 00

p o rx co §9

O Q O Q 00 CO CN CN

S CN r^ o»

CN lO

CO K

II II co rx o CN •*»• o

— lO X X

i i o o

CO CN -^ i— CN CN

T T T T oio^ CO — CO CO CN CN

CN —

CO O

I I 0 o

o o

< 5

< 5

o p oo oo co ^F

o o CO CN m ■^■ CN CN

00 8 rx —

rx CN

4<i

— CN

I I o o o o

CN CN

< 5

88

o CO

^ CN o co CN CO

00 tX

o NO

CO CO

oo o

"7 i CO — h- h- < < 55

88

lO 5 ^ ^

o o IX CN

O O

R8 O CO CN ON

CN CO O 00 i— CN

s = = = : o = : =

29.5

28

.7

22.9

23

.5

>o

4 I 00

ON

NQ K X X

o o CN CN

0%^ CN CN I- h- < < 55

106

-o 0)

c *- c o U

X X X

<

P 0)

itS

v

V« 0)

o u

9)

u- tn b

TO

O in

C o

CL LkJ 13 vS»

0) 3^ o o 53

8 «o"-

8 ^

^^ 8 X c U i>v^

0)

c 4) u

0 E

o o

8^ 5^

o o o —

o

cs o

o 8 «8

8 rx o to io

O CN CM —

o »o CN h>

O O o o

o o 00 00

o o 8^

95

o o — CN W) O

CN

8

88

5?«

§13 CM N

88

o o

CO IO

o o 5 lO «O O ss rv IN. O CN CN 00

o o in io

CO IO co >o ^ — ■«t lO CN >o — ^ •o »o 00 io ^t 'f io rv vO K io rv

o §

o = - £ o = = = = = = =

CO CO CN 00 IO CN rx io io io • • • • • • • • • • >o ^" CN "«l- — oo IO Os o o <N CS CN CN ■* CO CN CN CO CN

00 CO rl co o CN lA CN 1 1

CN O 66 • • • • • • • « • • • • 00 ^ ■<*■ K K 00 O CO 00 CN CO Os

o — CN CO CN CO 00 Os M CN "t >o F— 1"^ CN CN

1 1 f— r— 1 1 1 1 1 1 O 0 66 IO >o 0CN0CN

0 0 Oo0o 0%^ 0^0^ T ^—

0%^ O^ o^ CN CN CO CO CN CN CN CN

h- h- h; ►- 1- 1— 1— 1- »- l- H- h- < < < < < '< 55 55 5 5 55 2> 5 5 5

107

T3 9) 3 C

•4- C o u

X X X

oo <

2 » b oo _i 2^

CO i

o2

CO I

CN2

CO I . o

4) « O

U Ji

0)

8.S J^CO U I

^J^s c a> i.

« I

CO »O CN CM 00 OC 00 00

CM

CN o

tN. O — N. • • • • rv o tx oo

CO CM "^ »- F— »» r»» P^

I I I I CO »— CO CO t- t- t— h- <<<<

CN O K 00 00 00

IT) ^ S O IT) O CO CM

W) Q co Sri

• • • « • • • • o o o o o o o o

CO fv.

CO tx • • ■* o

X X

0 o CM CM

< 5

Ov' oo CM CM

O CM • • CM --

00 O

I I o o

< 2

o to

CM CO CM CM

•— CM

o o o o o^^ o"^ o

CM CM

< 5

o co CO 00

3

o a

o =

CO CO

00 o

«"> CO

I I CO —

< 3.

K oö

O» O O O CO >^ 00 CO « • . . o o o o

CO

d o 0.0

5

0.50

0.3

0.

13

0.90

0

.60

0.30

0.

40

d d o o o d CO

d o 0.4

0.

17

CO

o d o o

^8 §§ 8^ o o o o o o

3 3 as «o

oo

M3 hs X X

I I o o

CM CM

< 5

108

■o 4> C 4- c 0 u

X X X UJ —i CO <

t > i^ -j 2^

o2

CO I

CO I

. o

S E ^ «o

O ^.

2 0) \

Q-.S U1 w

a> to ^ , 0) c v>

in c

E > <

c

c _^co U 1

c 0) k.

E J! o F S -» Ä Z

00 »O r- CM r^ to ix r^

CN o tx cs

• • o —

o o

o

o

o =

CO 00

00 "t

CN CO o> CO

o o

o o d

CNJ 00

O — CN '^■ CM CN Csl <M II II

oo co rv co

^ r-s

CM lO O» 00

00 IX IX 00

o o

<N O O» CM S? 8S O — o o — o

o

CN o o

O :

I I 00 CO

^5

o o

o o co o«

• • — o

lO CM

CM ""t CO IX

Ö 00 CM O O -Q CO >o

— o — o o o

o o

o s? CO "O

• • o o

rx ID

CM co

i i o o

o o CM CM »— I— < < 22

CM CO CM CM

I I 0 o

< < 22

oo m

"t IX ix -«t

• • o o

00

d o

CM IX oo ^

• • o o

«O if) • • •— oo

• • • • o >o

Tf CO CM CN CO <N

O CN »O CM C^d, 66 • • • • • • • • rv oo >o co CO CM CO 0^

00 o

I I

^^ o o

CN CN

109

> o Z <

to < X

Z) Z

Q z o CD

a:

Z) 10

in z

> X X

CO <

g • t) tn _l i^

i

ro o

w^" C-

41

-*- O s a «) — y i . u. isi o

T3 4) X n M

o »^

c o

4) o E in

UJ t— i ^^

0) «o

8 -c X

-a .E tu o 0) • n o i/>

c

?■ E ^ i r c

5 o U

CO -O *0 «"> oo rs. oo oo

CM CO K 5 fv 00 <N 00 —

00 00 K h.

'« '? >> ? » » » >i I 11 «

K U") — o — — ^

u-) CO _ o

in m 0.0. ^

-H -H +1 -H „o in o- {St— — CN

+1 ID

* -H — o

-H -H mm « o o t

o o o o

-. S. S. ^ -H +( +1 +1 o oo n Q vo cs ■* m o o o o

op m ^ K o co CM o> oo m -H -H +1 +1 •o fx m n tx ■<t — CO ■* m ->*• ■*

o o « m

m -H m

(O CM CM CM

CM +1 +{ -H

CM m m

o 0 o K " -o CO -H -H -H +1 o o o o K oo m r^ c-' rx oo •^■

-H -H -H -H o K o m oo oo rv co

CM

CM O

O = z :

Ä2-Oo O CM — CM

O O O O

5°9S ^o ■— "> CO -ff +1 -H +(

oo co n m >o m — ^o CM — CM —

m in <N -* m

-* t5 CM ^O -H -H

m -H m oo oo K — — K m co o — o> — —

•H +1 +1 -H oo o- ■* o

m vo ^o t^

o o

CM o O :

o o o CM ^ O

o o o o

52^ o 2 2 o • • • • CM • • -H -H -W -H -H • +1 +1 o m -o >o m +i co m o> co o -o o. — ■<* ^0

o o — o o — o o

>o CM CM m — — ■«t CO S28§ -H +1 41 -H +1 -H +1 +1

5965

4545

4420

5350

o m o o co O p oo — CM is l?? ■«t ■* -«t ■«)■

o o in co o oo o> o

N. ^ >0 O

>n3 CM 'O •- ^CM ^ O +1 -H m 4H co m — o o >o o — rv K — in

CM

o — • — -H +1 -H -H — m o< co xf so >0 "O

o o

CM o o

p o o 0_ CM r- O

O CM — CM

110

5000

4000

3000

3

Si

2000

1000

/

r720F

/"

1 ^

/

/

/

/IöO0^

/

yx /

/

V 1

/ 10 12 14

Strain (10"3)

Figure 16. Tensile Tests, Metlbond 329, Aluminum, Cross-Head Speed - 0.02 in./min, o7o0.

in

> o z <

CO < X Q.

Z

ON

CO

O z o

< LU X

> X X X

CO <

a> >

4- d) o o -ö o E o o £ 5 £

u £ a. p C -^

0)

2 ^ •— i- 1/»

c o '« o

jf ^ E « ^ UJ Hi 3

^ * o o

O «> X

s 0) ox*

g)_i c '" E J, -JCCO « e ü U ^l£2

c 4) u

I | o E

28

88

CO CN • • CO CO

CN

O :

CN CO

I I

to lO << 55

o 5?

s?«» w ^ in o

o o o» co -^ >o

co «o • • •^- CO

CO >o

^'^ I I

O* <N • • 00 00

ov X I I

CM CM

CN CN CN CN «/) in < < 55

o o

o o «n CM

o o o o ^ — io ri co o co .— ^. ^- o« 00 ^^

p <N CN CN 0* .-r 'S ^

CN

CO CO

o r^ • • o — CO CO

00 U1)

O CN

o —

to oo < < 55

o o 8

CO Q

5» «n 2^

CO '«J-

o o • • — o« CO CM

"4 00 CN O^

o CN

I o

CN to

CN

o m CN CN IS)

lO o o o o «n

88 8S CN h«. ts. «O

CN — NO "O

CO If) oo co •n ^

co co co rv »o CO

CN o « •

CO CO

o «o

io oo « •

CO CO

in o

CN CN

o in

in — CN —

to «/I < 5

o» o

«A NJ • ■

CN tN.

O 00 —

I I CN £l CN CN

CN ^1 CN CN to to

08 >o o

8 o o

CO —

lO o co fx

CN o 00 O O CN O CN W» ^^ 00 — o rv CN »O CN CO CO CO CO CN CO CM

»o »o CN CO

Os o • •

to o CN CO

rv. oo • •

K o

NO ON X

I I

to to < 5

112

0) 3 C

i o

U

> X X X xxt -j eo <

>

o u

"^ ^ 2^-

£ a

2 Ä p c —

"x O in U 8 f^-

— u w>

c o

ß -2 i S

•- i. -5 a.

^ S o o ^ Ä $ 3

■o

o>_, S -

o E

g-o 00

O «Q

K5

• • CM CM

S

o =

ON "t * • CO I^N CO CM

I I r^ CM • • ■*• 00

rs co

&? 0 S 10 o o o o

o p 00 fs

10 s^ o co tv s

• • CO CO

o o

"* CO lO CM • • • •

CO CM CO CO

o O s

CM O

CO 0

o — I I

CM —

X X I I

?>! W CM CM

CM ?^ CM CM

IV 00 X X

I I

to 10

8 R

R§ 85? O CN 5? 00 88 ^5^ CM CM

O* 00 CM CO >o ^r 0 ^ 3^

00 CO rv 00

8 0^ CM — ON ON §§i ^8 00 p

00 o> CM — CM CS CO CM CM CO CO —

00 o • * CO CM

CO lO ON O • • • •

^00 o> •— N* TT CM co II II

ON »— IV. 00 • • • t

00 o —

O CM • • IV ON CM CM

I I 00 o

ON ON —

CM CM CM CM to to

113

21?1- lie CO o- O 00

CN CO t^. o*

^- u-> rs CM K co

00 p- «o >o oo io tx 5s

CO

Cv,^ N 00 "* o^ >o 00 <N O o ^ tv -O "^ CO tx

. . • • • • • • • o • a a a

^ w o — (N CM — o '""

cs o IT) IV — — O

rP i «o — 00 00 o o> o CM CM

CM O • 1 • • • • • • a

«^ w o o O <— o o o o o o p- ■— o — o

1 0) 3 c

CO rs oo • • o o ■'t 00 • • o o

• 0 o -+ • •

IX IT) CO • • • uo o» • • +- ^ ~ o — CO ^ CN O <N — CM O >0 O^ CM CN4 O

c 1 o

U

> x x

1 ^C s s m — : = « _ § _

X UJ

i _l CO

<

ross

-Hea

d se

ed

n./m

'tn)

If) • a

U ^C o s *" — — " — ~ o =

« J 18 c co ^o O IV o o o o

Ave

rog

Bon

d-L

Thic

kn

(IQ

"3 i

^

• • o — CO CO

• • CO CM $4

t

O« CM • • • t 4o!, • • CM CM • • m %

r> 'O oo oo O CM CM O^ o >o CM hs

c •o O —• O 0) u o^ X 00 — E X 1 1 1 1 S M J J

Spec

! N

uml

M CO ^^

^- r— 66 -<■ CM CM CM CM •" ^ >> in ^—

■ ■ i J. ?J M <M ÖJ ^ J- ?4 ?>< pM ?«■ CM CM P^ PV CM CM F*- F*" CM CM wo ^0 CO IO U0 CO CO CO CO <s> U0 CO

i t 1 II

114

■a 0) D C *■ c o U

> X X X

<

£ « K to 4)

CO I

CO O

i CM O

CO « r- Ö

,4J « o

T3 O 0) I i, 8 w ft C

g)_j c —

c 0)

o »o

•o 00

.—' CO

00

o o

■* oo

CN CO*

o

o

o o

rv o

-o o* X

oo to < < $5

CO o

d —*

ir> oo • •

o o

o -«t

— CN

0- ^

m o CN CO

co rv CO CM

r^ oo IS. CM

"«t CO

K 00

CM CM to to

< 5

CO — IO hs

O CM O O in rs

CM O

CM* CO

&■ •— co m

d —* —' d

m CO

r-N CM

CO IT)

CO CO

o o

CM ^f

■* 00 • e

o o

N CO

o

CM O • o

00 vO

•o — I I

CM —

m to < <

co m

H • •

o o

X i

X I

CM CM CM CM

CM CM CM CM OO to

< 5

Qv o • •

o> — CM CO

I I N 00 • • o —

rs oo X X

i i

to oo <

IT) CO

00 00

o d

oo -«t

■—' CM'

O CM • «

rs. o CM CM

I i 00 o

o —

CM N CM CM OO 00 < < 55

115

O «O CO Q oo r- K US

CO CO CO — ^ rv ~o «o

CM — O S? •O >0 <3 ^

UO _l S^ 'I >l H » II " 7

m o

>l n II v

^^ n ■* ■* • m ■«»■ es — CM CO CO

co ■>* ■*

H • • ■ IO -H -H 4( ,.* ST H -H'

o o« rv •«»■ rs t CO vO CO

^C d d d o d d d d —' d d d

> o z ^ CN < ^^ • CO lO 00 . « K IN 00 — CN CO

CO -H +/ -M' -t1 CN • • •

M '|' -H -H

> O» >0 "0 (N U~ fN .- |N i.o co uo —

—' —' --' r-' ■*' . -' d --' V —' — CN UJ to < m Q "Q

o o in ■«<■ Q lO Q "^ iO CN Q UO

in Q uo Q IN ^ 00 üo

oo i- >o — CN m 35 CN ■* ^ IN CO CL

^«3 E £ u.

+< -*! ft H ■H -H -H -H O "> O "rt in IN S <N

^) ?> uo S

-H -H -H -H

2" o uo O "1 00 «0 tx o S tv oo —

in Q <o Q

oo •n «o m 3 u. l^ »- fv K O lO

z 5 3 —i ^ o < V

-H -H ^ V $s^^ CN cr>

■* r- — O. CM CO

Ö Q ^ 00 r5 * oü —

H H -H n ■^- CO CO CO o in ui — m ■* in •«

m 00 CN CN

Q z 0

US Jr ■t •* o- o <N — 00 —

CN CM 00 CD •o >o ^ ■Z

eo C o ^ _ 1- •S.ü K •- ^t ^ Ä^ftÄ _9S§: fcU

2 j> o 6 o -H -H -H ■« o >— rv >o oo >o o N •O "O •* CN

+1 -H -W -H 0~ ■O — CN ■* 00 O CN

+1 4( -H « CN Q »0 O O- * O- 00

>- Of

CO CN CN CN CM CO CN CN

.ä .|t ^ "* CO . . o • — CN m — CM <0 CM O

2 5 Ill's,

LU in 2 C

-H -H -H -U « -H -H -H ■H -H -H K CO 0> O K

CO CO 00 CO

— IN -- <0

CO CO CO CN

CO O» -^ f^-

CO CM CO CN Oi < UJ r «1 C Li,

CN s Q fx r r : r*~ E S ü ^— ESS

> S c

X X

in UJ 0 4> ' lO m o _i ."-. O- .c

< U oo w O S 5 S d : : r d : : :

c _0 *.

11 ^1 ?J z~ CM ~ CM —

Ot — O CN >- O CM — O

o»> o»> \CM — CS \CN — CM O CM ^ CN O CN — CM O CN — CN

116

»^ ^^mmm

7000

6000

5000

4000

3000

2000

1000

/ 720r

s

r

^

/ /

160oF

/

//

/

/

/

f

/

/

10 20 30 40 50 60

Strain (10"3)

Figure 17. Shear Tests, Metlbond 329, Aluminum, Cross-Head Speed - 0.5 in./min, 0700.

117

As wifh the tensile results, raising the test temperature lowers the modulus, the precision elastic limit and the microyield stress; increases the viscoelastic flow; and does not significantly affect the fracture behavior. The magnitude of these effects was smaller in the case of tensile tests than it was for the shear tests. Introducing a tapered bond line appears to raise the modulus, lower the strength levels, and diminish the viscoelastic flow. Based on average values, lowering the strain rate, in addition to raising the test temperature, has a negligible effect on the proper!les. The average shear modulus for the Phase VII results was 3.6 x 1(P psi. Raising the lest temperature to I600F decreased this value only 13 percent to 3.12 x 1CP psi. Lowering the strain rate by a factor of ten did not introduce any further changes in the shear modulus. Values for the P.E.L. and T decreojed In absolute value and percentage of the fracture stress when the test temperatuio was raised. At 72Pf, the P.E.L. was 7 percent and the T was 22 percent of T At 160 F, these percentages were 5 and 10, respectively, for the P. E.L. and the T At the higher temperature, the viscoelastic strain at unloading, y-,, was nearly doubled that for tests at 720F. The recovered strain, y«, as a percentage of y, increased at 160 F and was raised even more with the lower strain rate.

Lap Shear Tests

The base-line tests at 72 F with a cross-head speed of 0.5 in./min had an average fracture stress of 2375 psi, with 100 percent cohesive failure. When the test temperature was raised to 160 F, an average value of 2395psi was observed. Lowering the cross-head speed to 0.05 in./min at 160oF resulted in an average fracture stress of 2435 psi, again with 100 percent cohesive failure. These test results are presented in Table XXXVIII.

Within the ranges of temperature and strain rate used In these two phases, it was shown that test temperature was more effective in modifying the properties than was changing the strain rate. The property changes were relatively small for the increase of 88 F In test temperature.

PHASE IX, AGEING/ENVIRONMENT

For this phase of the investigation, two batches of FM1000 were stored In different en"ironments. One batch was kept in a controlled environment maintained at 680F with a relative humidity of 45 percent. Tensile specimens were prepared and tested after 4 months' exposure. After 8 and 12 months' exposure, additional tensile specimens and some lap shear specimens were prepared and tested. The second batch of FM1000 was stored in an uncontrolled environment in which both the temperature and humidity fluctuated from day to day over ranges of 72° to 750F and 20 to 60 percent relative humidity. After 4 months in this uncontrolled

118

Q Z <

> 1X1 1/1 < X

D Z

o 00 Q z o

< X

QL

<

> X X X

CO

<

0) u

-^- U

01 >

OJ _c o u

u p

CL

O 0) X

o a, U Q.,

00 .

0 ,—. t « _

c (1> 1_

u E 0)

oo Z

o o O O

o o 8 o o o o

8 S CN

O O 00 CN

00 O

CN CM

CN o §

IT)

o o =

IT) o

o o d d

o o CX3 d

o o

>- (N —I _J < < 2 5

OO

< < 2 ^

-I _J

11

119

environment, specimens were prepared for tensile and lap shear tests. The m< nufacturer's warranty for FM1000 is 6 months from dote of shipment when

red below 85°F. A l ! tests were made using the standard experimental conditions of Phase II wi th only uniform uorid-i ine thicknesses (no tapers).

Tensile Tests

The test results for the tensile specimens ere shown in Table XXXIX and summarized in Table X X X X . Wi th in the experimental scatter, ageing in the control led environment did not appreciably ai tsr the mechanical properties for up to 8 months' exposure. However, the material aged 12 months in the control led environment underwent marked decreases in properties. The tensile modulus dropped from 6 .3 fo?..5 x 10^ psi and the fracture stress was lowered to 7100 psi from 10,650 psi. It appears that the 4 months' exposure to the uncontrol led environment lowered the elastic tensile modulus but did not affect the other properties to any great extent.

Lap Shear Tests

The test results for lap shear specimens aged 8 months in a controlled environment and 4 months in an uncontrolled environment are shown in Table XXXXI . In contrast to the tensile results, there was a marked deterioration in the properties as a result of ageing. The fracture stress dropped 55 percent after 4 months' exposure to the uncontrolled enviionment. For the control led environment, after 8 months the fracture stress dropped 68 percent; after 12 months there was a 42 percen* decrease. The fracture mode was somewhat less cohesive in nature after the ageing. An addit ional set of experiments was performed (see Table XXXXI ) . Lap shear specimens were cured and then aged 4 months in the uncontrolled environment wi th no deterioration in properties. In fact , the fracture stress was higher than for freshly bonded specimens (8125 psi as compared to 7075 psi).

120

1/1 < X o.

z> z <

o o o

i

Z

X X X

CO

<

0) > V) 4)

JC

if 5 ^

o 8 Q. p u --

u_ oo b

2p a;

'>. p «i

•-um

c o

0) o .i '5 Q. UJ Li ^9?

•r ~ 3 "• « «a tyo o S o o

IXJ (- < ^^

Q.

p

O) i

«I c

c 0) k.

.§JJ u E

iÄ-z

18 c «.£

U I

^2

uo o m r^ o^ r^

S S S 00 o o

o o g ■>o o ^ o ro o CM o4 -^

o o c 00 Op oo CM ^ r^

o io in

<> K u")

CM N.

CM O

d = -

in — o • • • rx oo rv

CO — CM I

I I CM o o —

o

C 0) E c

> c

-o a)

c o

o o 00 K

CM IT) O —

o ^ o CM

E _c o ■^■

*- JD 0)

Ov —

.- Tt LO

o 00 -O CM —' tv

CM o * o :

■«t CM

b9

o X

E

u o

00 o

c w E c o

> c

a>

c o U

c c o o 5 5 ^ CM CO oo -o 1 1 "D <u t— (— 0) U) f- I— O) < LL. U. <

K s

8S 00 CN o r^-

^ ^S o

00 CO

o o o o 00 00

o ■o oo 00

$ 8 --,< K o Ch —

o o o 00

o o 00 CM

LO K t CM • • » I

-o »o IT) -tf

CM CM K = C

(1) E c

> c

N. :

CM LU CM O -D O

O = 0)

"2 •*- c o u c 3

• O r

(s. CM c ■* — • • •■— e »

m -^■ vo N K

c o

CM 5 — X i >- ^ X i

CM 1 1 CM — CN "D CM — 1— 1— 0) t- h- 1- H- O) t- 1- LL. LL. < LL. U.

121

cture

de

co

he

sive

)

2 2- V tt 5 ^ IT) O i

hv CO 1

0) i_ ^„

i l\ Q S5 1 E - ^

O Q K CO

u_ on b K -O il

-D -JJ

.9i 0' m >» t O </> u-

O o o \ CO CO

5 ^ Ä io m 1

C o 'E

■^ <-' D "O ^

r- <u o E « l — Q.

0) > 2 5

• — a. tu _i c? 4M ir> -<t | •*- _o o w> 7""* 0) U .ü^-f a. 0

t; ": -Dm J^ I X o £ o 0 00 CO

— <0 <■ ^. LU h- i ^^

• ^ • * ) 1 x ^ CN CM i X U- 1 X Q. 0«, ! LU *? r C

00 o j —i 4) (1)0 CM i i ^ t— H-u^_- ^ N =

< c i— 0)

E -o c

8 -c ^2

f-oi *> c CN

S (U \ LU O o «> • ^ Q. C

U CO ^ d : i

ID « ^

Ave

rag

e

Bond-L

in

Thic

knes

(10-3

in

c a U c

•4- C o

as in j

c 5 CO \ 4) L. CN co x i

" E ■a x 2 j •o ■— i

< H- 1—

UL LI-

122

2^ t c I .P Ä 5 J •«* ^ -o

00 C» 1^ O "1 00 00 00 00

o o 00 00

■D <J D

c o u

X X X

CO

<

CO I

en i

CN2

i o

a f-iT <rt

V * o ►— »- x^

■D O ^ ^ <U c X v/l

4» ^

u O-.E 00 ^

o o

o 00

IT) o

O I o

s o in

— — o

CN

(U c • c

o _J C ""• ■o c U 1

2°

c V L.

E J5 o F 0) -> # Z

CN O

d = =

IT) — O* • • • tv 00 K

— CN I I I CN

o o» —

o- o O CN

0

E

>

o "3 u

o

O^

d o

^ c^ LI. K =

0 00 O

c 0) E c 0 CN .t o > • c o =

LU

-a 0)

c o U C CO •"*

m •^ CN

c o

•< CN O

< U. U-

o CN o CN

o O

o' L) ^ ^- 0

X o CO

u o O

E D

-C

V > o CN

■t- <N CN

o ^_ o 0) u

o .c 4-

? CN u_ r^ -

o 00 >o *- c 0) t c o CN E o > c

• o =

LU

■u 0)

2 4- c 0

(J

c o

00 "D

CU O) <

tv. CN

c 0) E c

.2 *> r.

c o u c D

c o

o 00 00

d d

o o '-'

o 00 CO

d —*

CM

CN O

O :

CN 5 CO >- e5 ^ X ■ > ■t X i CN 1 1 CN — CN "D CN —

U- u_

a> <

123

>l ■ I IW^ ^■p

C

C O

X X X

CD

<

2 «< to

i

n i

CN2

i O

4) W

O 0) X

ö <u >

U i>i w

0) wi

CD OO

öü

in

o

E

5 c u U

C 0) u

" E

o 00 o

c 0) E c o

> c

0)

c o u

c o

CM

0)

<

o o

>

o —

CN

CM o

c — 0- IT) in

s: io NO

CO X

i -o

CO X

i

124

L

mam

IT) < X o.

Z <

o

<

D

z

X X X X

<

a! <u b ^ > o

uo

n i

CM O

i — o

C «- M «A

0)

5Ä s c o

g _9 #E *«, o. UJ 13 -S'

O (U 3 S»

— £ o o

■o

I I ^ S c

D E t) at

oo

oo

o

o

o rv oo

8 n s? CM -H O

oo O

CN

IN o

a>

in _E a»

li: <

o

8 1?

S

in CN

m CN m

o a -H

00

-H O

n

tz-o ^ C 0) c

E 2 I -** § T o .t: tt) u >

< .- 1>

n o in CO 00 oo

;.■

H

m H"

— m d d o

W) CN

JO

f( m" H M

N O r< d —' 00

8 CD

in

8 H in f

5? 8 8 o

8 r.

o o M H h

g g O TO in

m cö o

o Ul m H -H ro

8 O m m

03 CN hv ^r

"t >o CO

-H -H -H — OO IT)

•O V CN

4" c "5 a

" " " E tn <" "S 3

■£-o C 0) c c o a» c -8 o = (Ü 0 t (U E = 0) 0) E 2 E e c E CN 2 E c

0 00 *■ c "* o c r— •*- c w c o u o C 0 X •o 0

> <U 3 > -a o 0) u *> c

0

* ^.E c 0) Zs c

0) ^.E c 0)

125

mm^

c ■♦- 0) ^m ^ c 1) E

E 8

c w t

c

E

X

c .5 8

c c >

— c

c .E 2

c c

UJ

<

> 1 "i 2 "0

^ -o C 0)

3 t> I I a. =6

o

c

E fi

5^ c

11 I 8

o u ^ 1

O) o < o D

z <

0) ^8 ^i ^8 1

H;

»—

d" 8 F—

5 4) u. >

> £ | a: 3 -C < 4- a> O

^ 2 8 s? S o D i/»

0£ < o UJ g o o m I K CM o uo 0) n •— — — lO

a. 2 «> +i ■H ■H ■ti

-H

1 t3 S ^ g § 8 o (N 00 P^

• t^ (N n 00 -t

X X X X UJ _j CO -o < o ^ i— «> c

^ OO C •

o •

O d • o

IT)

d

126

SUMMARY OFRESUITS

The results of this research program will be summarized by phases, including major

property values (average) and the influence of the principal experimental variables. The effects of introducing nonuniform bond-line configurations will be listed

separately, again by phases.

Phase I, Base-Line Data

Properties of FM1000 on aluminum adherends (standard tests):

Tensile Shear Lap Shear

E* = 5.0xl0psi G = 1.43xl0psi fracture = 6750 psi P.E.L. =870psi (0.08a ) P.E.L. - 320psi (0.u3 •_) a = 1580psi (0.15a)'' r = 620psi (P.06T_) " mys r mys r

oj: = 10,800psi T = 9500psi

^ = 1.48x 10'3(@0.8a) y=5.2x 10-3(@0.5TF)

Deviations from elastic flow occurred at very low stress levels (for shear, about half that for tensile). Viicoelastic flow was considerably greater for shear than for tension.

Phase II, Adherend Material

Properties of FM1000 on titanium adherends (standard tests):

Tensile Shear Lap Shear

E* = 5.9x 105psi G = 1.48x 105psi fracture = 7075 psi

P.E.L. =810 psi P.E.L. =246psi

a =1785 psi mys

a =10,160 psi

€1 = 1.23x 10"3

T = 530 psi mys

T = 7900 psi

y =4.4x 10"3

Titanium adherends raised the tensile modulus 18 percent but did not appreciably

change the other properties.

127

Phase I, cryogenic

Properties of FM1000 on titanium adherends (-65 F):

Tensile

E* = 7,3x 105psi

P.E.L. =2500 psi

a = 5460 psi m/s

a = 15,000 psi

e. ■= nil

,5 Shear

G = 1.98x 10" psi

P.E.L. = 1360 psi

T =3120 psi mys

T = 13,400 psi

yi = nil

Lap Shear

fracture - 8435 psi

Properties of FM1000 aluminum adherends (-65 F):

Tensile Shear Lap Shear

E* = 7.15 x 10" psi G = 1.9 x 10" psi fracture = 8000 psi

P.E.L. = 3750 psi

a = 8030 psi mys r

a = 17,120 psi

(. = nil

At -65 F,all the strength properties were increased in value and there was no viscoelastic flow.

Shear

G = 1. ?x I05psi

P.E.L. = 525 psi

a mys

aF= 13

1470 psi

,260 psi

y] - nil

Phase IV, Strain Rate

Properties of FM1000 on aluminum adherends (low strain rate);

Tensile

E* = 5.0 x 105psi

P.E.L. -810 psi

a = 1570 psi mys r

a = 9100 psi

(] = 1.38x 10"3

Shear

G = 1.55x 105psi

P.E.L. -290 psi

r =510 psi mys

7"F--8100 psi

yf 10.7x 10"3

Lap Shear

fracture = 5600 psi

128

Umm^t^*^^

Properties of FM1000 on aluminum adherends (high strain rate):

Shear 5

G = 1.8 x 10 psi

P.E.L. =340 psi

Tensile

E* =6.1 x I05psi

P.E.L. = 1510 psi

Lap Shear

fracture = 7300

a = 3250 psi mys

a = 10,700 psi

c1 =0.52x lO-3

T = 770 psi mys

T = 8720 psi

y, = 8.1 x 10"

The lower strain rate did not appreciably change the Phase I properties; however, the higher strain rate raised E and G and the strength levels and diminished the viscoelastic flow.

Phase V, Cryogenic/Strain Rate

Properties of FM1000 aluminum adherends (low temperature, high strain rate):

Tensile Shear

E* =8.6x 105psi G = 2.15 x 105 psi

P.E.L. =3420 psi P.E.L. =700 psi

Lap Shear

fracture = 8200 psi

a =10,100 psi mys r

0^ = 17,900 psi

e. =nil

T = 2370 psi mys

r = 13,200 psi

y] = nil

This combination of experimental variables produced the highest strength levels for any of the FM1000 tests; the effects were the greatest for the tensile tests.

Phase VI, Bond-Line Thickness

Properties of EC2214 on aluminum adherends (thin bond-line);

Shear Tensile

E* =0.86x 106psi

P.E.L. =2350 psi

a = 4820 psi mys

a= 10,350 psi

^ =0.65x 10"3

G=3.4x 10 psi

P.E.L. =385 psi

T =1375 psi mys

TF = 9330 psi

^ =3.1 x 10"3

129

Lap Shear

fracture = 5950 psi

Properties of EC2214 on aluminum adherends (fhick bond-line):

Shear Tensile

E* -0.96x 106psi

P.E.L. =2230psi

a = 4290 psi mys r

a = 10,480 psi

f, = 0.32 x 10 -3

0=3,2x10 psi

P.E.L. =277 psi

r =940 psi mys r

r =8560 psi

y. =3.82x 10 -3

Lap Shear

fracture = 5550 psi

The effects due to bond-line thickness were indeterminate for this system.

Phases VII and VIII, High Temperature/Strain Rate

Properties of Metlbond 329 on aluminum adherends (standard strain rate, 72 F):

Tensile

E* = 8.3x 105psi

P.E.L. = 1840psi

cr = 2700 psi mys r

a = 4680 psi

f. =0.45x 10 -3

Shear 5

G = 3.6 x 10 psi

P.E.L. =486psi

r =1450 psi mys r

r = 6730 psi

y] = 1.6x 10"3

Lap Shear

fracture =2375 psi

Properties of Metlbond 329 on aluminum adherends (standard strain rate, 160OF):

Tensile

E*=6.5x 105psi

P.E.L. = 1270psi

a =2000 psi mys '

o = 5070 psi

e] =0.74x 10'3

Shear

G =3.12x 105psi

P.E.L. =287 psi

r = 538 psi mys

T = 5830 psi

y] =2.75x 10 -3

Lap Shear

fracture =2395 psi

130

■1 ■"" . ■- -. , m , lmt9

Properties of Metlbond 329 on aluminum adherends (low strain rate, 160 F):

Tensile Shear Lap Shear

E* - 5.95 x 105 psi G = 3.12 x 105 psi fracture = 2435 psi

P.E.L. =750 psi P.E.L. =304 psi

a = 1450 psi r = 630 psi mys r mys ^

a = 4450 psi T = 6400 psi

f^O.ZSxlO-3 y1=2.35xl0"3

The higher test temperature lowered the modulus values and decreased the strength levels. Reducing the strain rate accentuated that trend.

Phase IX, Ageing/Environment

Tensile properties of FMlOOOon titanium adherends:

Fresh Aged Aged Aged Aged Adhesive 4 Months 8 Months 4 Months 12 Months

Controlled Controlled Uncontrolled Controlled

E* (psi) 6.3 xlO5 5.0xl05 6.1 x 105 4.8 x 105 2.5 x 105

P.E.L. 1510 psi 890 psi 1800 psi 1250 psi 475 psi

a mys

CTF

3100 psi

10,030 psi

1525 psi

10,850 psi

3560 psi

9050 psi

3190 psi

10,075 psi

1530 psi

7100 psi

C1 1.26x10" 3 1.95xl0"3 0.71 xlO"3 l.OxlO"3 8.2x 10'3

lap shear fracture 7075 psi 2290 psi 3880 psi 4100 psi

The effects due to ageing were more pronounced for the lap shear specimens than tor the tensile tests.

Nonuniform Bond-Line Effects

The effects due to introducing nonuniform bond-line configurations are summarized in Table XXXXII. The most striking influence of nonuniformity was the large reduction in viscoelastic strain (both C, and y,). Generally the modulus values were raised, with only small changes in the strength levels. The fracture stress values were not significantly changed.

131

• — 0 1 -a Si

^- ■)■ * * * ^ —1 _l > " _i 1 -■ -' 1 _J

U. z —I zz z _l Z Z _J

% ^ KE z —J -2 z _l -J _i —i

u* X (/i

_J

Ui Qd —1 a:Z z /L _i z —i

a.' ■a

o a: Q: ^z a: Oi z z Qi. 8 to s z I! _l a:

Q z 0 to

2 -o 01

Of o> 0 u.

c 1

z 1 ,2> u

D — IE 0 z ^_ @ ® c

0 V j z i I e*. _J , * * _l _J _l

11

z Z u. 0 u. t/1 1- t> T z Za z z z z _l u TJ LU £ u. v Li. LU Ui SL 5

bE _l —1 zz z z z _. _I _o

—' z ;i

X LU —i

X X X

(— LU

_J —1 Zo. z —1 or _J _l

U4 O- _i CD -D < ft> t K- «

LU z z Zct a; a£ 1 1 « 1 t>

__^ ' ,r~" C w

LU to < X

<

ID C

— 1J 1

8!

c

=6 <

"I

(

Ti)

Cry

ogen

ic

(A

l)

<

>l c

E *- to

V

,

(Al)

C

ryo

ge

nic

/Str

ain

R

ate

VI

(Al)

T

hin

Thi

ck

Bon

d-Li

ne T

hick

nes

VII

(Al)

H

igh

Tem

p.

VII!

(Al)

H

igh

Tem

p/S

trai

0) c e II

«

132

CONCLUSIONS

Based on the test results and their interpretation, the following conclusions have been drawn:

1. Tapered bond lines do not degrade the basic properties of the adhesive, the fracture stress is not reduced, the mode of failure is not changed, and the values for the precision elastic limit and microyield stress remain largely unchanged. Some property values are Improved, such as the modulus, which is usually slightly increased. The largest improvement is a marked reduction of viscoelastic strain following a load-unload cycle.

2. Poor bonding and curing practices and out-of-date adhesive have the most damaging effects on bonded joint properties.

3. Deviations from perfectly elastic behavior occur at surprisingly low stress values. However, much of that strain is recovered within two minutes after unloading.

4. The elastic limits are much lower for shear loading than for tensile loading, and the viscoelastic strain is greater for shear (y.) than for tension (f,).

5. V/ithin the range of variables used here, temperature had a larger influence on the properties than did the change in strain rate.

6. Raising the test temperature has the same general effects as decreasing the strain rate. Conversely, a lower test temperature causes the same kind of changes as raising the strain rate. Both sets of effects are due to modifications in the viscoelastic strain mechanisms.

7. Temperature excursions below the curing temperature do nor change the room-temperature property values.

8. Ageing of FM1000 for 12 months at 68 F with a relative humidity of 45 percent produced a marked decrease in the tensile properties.

133

^

LITERATURE CITED

1. Hughes, EJ., Rutherford, J.L., Bossier, F.C., CAPACITANCE METHODS FOR MEASURING PROPERTIES OF ADHESIVES IN BONDED JOINTS, Rev. Sei. Insf., Vol. 39, No. 5, May 1968.

2. Bossier, F.C., Franzblau, M.C., Rutherford, J.L., TORSION APPARATUS FOR MEASURING SHEAR PROPERTIES OF ADHESIVE BONDED JOINTS, Brit. Jour. Sei. Inst. (J. of Phys. E), Vol. 1, Series 2, 1968.

3. Rutherford, J.L., Bossier, F.C., Hughes, E.J., ON MEASURING THE PROPERTIES OF ADHESIVES IN BONDED JOINTS, Society Aerospace and Material Process Engineers, 14th Nat. Symp., Nov. 1968.

134

Unclass i f ied Secu r i t y C l a s s i f i c a t i o n

DOCUMENT CONTROL DATA - R& D (Security classification ol title, body of abstract and indexing annotation must be entered when the overall report Is clasi

1 O R I G I N A T I N G A C T I V I T Y (Corporate author)

Kear fo t t D iv is ion

S inger -Genera l Precis ion, Inc . 1 i t t l e Fal ls . New Jersev

2a. R E P O R T S E C U R I T Y C L A S S I F I C A T I O N

Unclass i f ied 1 O R I G I N A T I N G A C T I V I T Y (Corporate author)

Kear fo t t D iv is ion

S inger -Genera l Precis ion, Inc . 1 i t t l e Fal ls . New Jersev

2b. G R O U P

3. R E P O R T T I T L E

Study of the Performance o f Adhesive Bonded Joints

4 . D E S C R I P T I V E N O T E S (Type of report and Inclusive dates)

Final Report; 12 June 1969 through 12 June 1970 5- A U T H O R I S ) (First name, middle initial, last name)

Haro ld K . Shen and John L . Rutherford

« R E P O R T D A T E

November 1970 7 a . T O T A L N O . O P P A G E S 7b. N O . O F R E P S

1 4 7 3 I 8a . C O N T R A C T O R G R A N T N O .

DAA J 0 2 - 6 9 - C - 0 0 9 4 6 . P R O J E C T N O

Task 1F162204A17001 c.

d.

9a. O R I G I N A T O R ' S R E P O R T N U M B E R ( S )

USAAVLABS T e c h n i c a l Repo r t 70 -65 [

8a . C O N T R A C T O R G R A N T N O .

DAA J 0 2 - 6 9 - C - 0 0 9 4 6 . P R O J E C T N O

Task 1F162204A17001 c.

d.

9b. O T H E R R E P O R T N O ( S ) (Any other numbers that amy be assigned \ this report)

F i na1 Repo r t RC 70-3 10. D I S T R I B U T I O N S T A T E M E N T

T h i s document i s s u b j e c t t o s p e c i a l e x p o r t c o n t r o l s , and each t r a n s m i t t a l t o f o r e i g n governments o r f o r e i g n n a t i o n a l s may be made o n l y w i t h p r i o r app rova l o f U .S . Army A v i a t i o n M a t e r i e l L a b o r a t o r i e s , F o r t E u s t l s , V i r g i n i a 23604.

11. S U P P L E M E N T A R Y N O T E S 12. S P O N S O R I N G M I L I T A R Y A C T I V I T Y

U . S . Army A v i a t i o n Ma te r i e l Laboratories Fort Eustis, V a .

13 . A B S T R A C T

The purpose of this work was to obcerve the ef fects of se lected exper imenta l and mater ia l var iables on the performance o f me ta l /me ta l adhesive bo-.ded jo in ts . Three adhesives were used: FM1000, EC2214 and M e t l b o n d 329 w i t h a luminum and t i t an ium adherends. The major exper imenta l var iables were test temperature and strain ra te . Ma te r i a l var iables inc luded composi t ion of adhes ive, adherend mate r ia l , bond - l i ne thickness and tapered bond l ines. The f o l l o w i n g propert ies were determined for tensi le and shear loading: modulus, m ic roy ie ld stress, prec is ion e last ic l i m i t , v iscoelast ic f l o w , and f racture behav io r . Lap shear propert ies were also measured. The strains were determined using a high sens i t iv i ty capac i t ance - t ype extensometer.

It was found that the tapered bond lines d id not degrade the propert ies; in f a c t , the v iscoelast ic f l ow was considerably reduced. The adhesives at 72 F, and h igher , were e last ic on ly up to about 5 to 15 percent o f the f racture stress. The e last ic l imi ts were lower and the v iscoelast ic f low was h igher for shear load ing than for tens ion. Raising the test temperature and lower ing the strain rate produce s imi lar e f fec ts . Temperature cycles below the cur ing temperature d id not in f luence the room-temperature proper t ies . Poor bonding procedures and o u t - o f - d a t e adhesives p rov ided the most damaging ef fects on the proper t ies .

P O f t M 4 J M P L A C I I O O r M M U T I . I « « . W H I C H I I t • r • i D D . « o 7 . . l 4 7 3 n H L r i m u w u n . U n c l a s s i f i e d

Security Classification

Unclassified Security C lass i f i ca t ion

t 4 . K E Y W O R D S

L I N K A L I N K B L I N K C t 4 . K E Y W O R D S

R O L E W T R O L E W T R O L E W T

Adhesives Adhes ive-bonded jo in ts M ic romechan ica l propert ies Epoxy adhesives Extensometer Aerospace techno logy Mic ros t ra in techniques Thin f i l m jo ints Tensi le modulus Sheer modulus Precision e last ic l im i t Fracture stress Mode of f racture Elastic deformat ion Tapered joints V iscoe las t i c i t y Lap shear Strain rate Bond- l ine thickness Cross-head speed

Unclassified Security Classi f icat ion