U. S. FOREST SERVICE RESEARCH PAPER FPL 75 OCTOBER 1967

COMPRESSIVE AND SHEAR PROPERTIES OF AND POLYIMIDE FILM HONEYCOMB CORE

FOREST PRODUCTS LABORATORY FOREST SERVICE

U. S DEPARTMENT OF AGRICULTURE MADISON, WIS.

This Report is One of a Series Issued in Cooperation with the MIL-HDBK -23 WORKING GROUP ON COMPOSITE CONSTRUCTION FOR AEROSPACE VEHICLES of the Departments of the AIR FORCE, NAVY, AND COMMERCE ABSTRACT

This report presents the results of compression and shear evaluations of two film honey­ comb cores. Properties determined will enable structural engineers to arrive at rational designs of sandwich construction using these cores. The cores were of polyester and polyimide plastic films and of 2-pound-per-cubic-foot density. Cores were evaluated in thicknesses from 0.4 to 1 inch, and it was determined that com­ pressive strengths for cores 1 inch thick were about 65 percent of those for cores 0.4 inch thick. Core shear strengths parallel to the core ribbon direction were twice those perpendicular to that direction. Core shear strengths for 1-inch-thick core were about 80 to 95 percent of those for 0.4-inch-thick core for polyester cores and 60 percent for polyimide cores. COMPRESSIVE AND SHEAR PROPERTIES OF POLYESTER AND POLYIMIDE FILM HONEYCOMB CORE 1

BY PAUL M. JENKINSON, Engineer

FOREST PRODUCTS LABORATORY 2 FOREST SERVICE U. S. DEPARTMENT OF AGRICULTURE

INTRODUCTION CORE MATERIALS USED

Sandwich constructions comprised of strong Cores were produced commercially, The poly­ thin facings bonded to a thick lightweight core ester core was received from the manufacturer can be used to produce stiff light weight structural in sheets of the required thicknesses, and the panels for use in aircraft and other flight vehicles. polyimide core received cut to specimen Sandwich of polyester and polyimide plastic films sizes. Both cures were of 0.003-inch plastic may be utilized for special applications where film formed to 3/8-inch hexagonal cells. Average the chemical and physical characteristics of density of the polyester core was 2.0 pounds per these are advantageous. As an exploratory cubic foot. Density of the polyimide shear speci­ study, the mechanical properties of honeycomb mens averaged 2.0 pounds per cubic foot. while cores of these plastic were evaluated at the compression specimens averaged 2.3 pounds 75° F. per cubic foot. Honeycomb core shear and compressive strengths tend to decrease as core thickness increases; to determine the magnitude of this PREPARATlON OF SPECIMENS effect, specimens were evaluated in thicknesses of 0.40, 0.62, and inch. Core received from the manufacturer was The Military Handbook 23 Working Group conditioned to constant weight in a room main­ this project and experimental and ana­ tained at 73° F. and 50 percent relative humidity. lytical work was conducted at the Forest Products Specimens were cut from the sheets of polyester Laboratory in and 1967. core using a bandsaw.

1This paper is another report in the series (MIL-HDBK-23)prepared and distributed by the Forest Products Laboratory under U. S. Naval Air Command Order IPR and U. S. Air Force Contract F 33 615-67M 5001. Results here are preliminary and may be revised as additional data become available. 2Maintained at Madison, Wis., in cooperation with the University of Wisconsin. the resin formed fillets, which bonded the cell walls to the plates. The loading plates were stiff enough to resist bending. A stiffness of at least 600,000 pounds per square inch for each inch of width and core thickness is recommended.3 On this basis, required loading plate thicknesses were calculated as follows:

After the had cured, specimens were removed from the conditioning room and tested immediately. Samples of each type of core were and the moisture content determined.

EVALUATION OF SPECIMENS

Compression Specimens

Compression specimens were mounted between Figure 1.--Core shear apparatus showing two loading plates, and this assembly was placed specimen with attached loading plates, between a spherical bearing. block and the upper spherical bearing block, and dial gage assembly used for measuring deformations. platen of a testing machine. A specimen and M 132 010 loading plates are shown in figure 2. The loading plates were of magnesium tooling plate 4-1/2 by Compression specimens were 4 inches square 6 inches by 1/2 inch thick. These plates were and 0.40, and 1.00 inch thick. Ends of chosen because their flatness and thickness are specimens were dipped in epoxy resin to form held to very close tolerances in manufacture. reinforcing fillets about 1/32 to 1/16 inch deep A load of about 50 to 100 pounds was applied at each end of the core cells. while the specimens were alined. Screw jacks Shear specimens were 4 inches wide with the were then placed under each corner of the bearing length 12 times the core thickness, except fox block to prevent further movement of the block. the 0.40-inch-thick specimens which were 6 inches Marten’s mirror compressometers were attached long (15 times their thickness). Steel loading to the loading plates to measure deformations, plates were bonded to the core with a room­ as shown in figure 2. Load was then applied at temperature-setting epoxy resin (fig. 1). The a rate such that failure occurred in 3 to 6 minutes. core cell ends contacted the loading plates and Specimens failed by buckling of the cell walls,

3 American Society for Testing and Materials. Shear test in flatwise plane of flat sandwich con­ structions or sandwich cores. ASTM Standard C 273-61.

FPL 75 2 Figure 2.--Core compression apparatus showing specimen, magnesium loading plates, and Marten's mirror compressometers. M 132 011 followed by crushing and crinkling of the cell Failure occurred by progressive buckling of corners. Most visible evidence of failure dis­ the core cell walls, followed by crinkling of the appeared after removal of load. cell corners. Most visible evidence of failure would disappear after load was removed. Shear Specimens

Shear specimens were evaluated using the PRESENTATlON OF DATA AND apparatus shown in figure 1. The specimens, DISCUSSION OF RESULTS with attached loading plates, were mounted between notched blocks between the platens of a Average values and standard deviations of shear testing machine, with the tower end supported on and compression properties of the a spherical bearing block initial load of 50 to honeycomb core are summarized in table 100 pounds was applied while the loading plates Properties of individual specimens as well as were firmly and evenly seated in the notched average values and standard deviations are pre­ blocks by tapping the spherical bearing block sented in the appendix in tables A1 and A2. with a hammer. Screw jacks were then placed Moisture contents of the cores at 73° F. and under the four corners of the spherical bearing 50 percent relative humidity were 0.4 percent block (fig. 1) to prevent further movement of the for the polyester core and 1.6 percent for the block during application of load. The movable polyimide core. platen of the testing machine was driven at a The effect of thickness on flatwise com­ constant speed such that failure of the specimens pressive strength of polyester and polyimide occurred in 3 to 6 minutes. honeycomb cores is shown in figure 3. The Movement of loading plate with respect to strength of a core 1.0 inch thick was about the other was measured to 0.0001 inch by a collar 65 percent of that of a 0.4-inch-thick core. and dial gage, as shown in figure 1. The steel Typical compressive stress-strain curves are collar was attached to one loading plate with set shown in figure 4. These are drawn for the screws, while the dial gage was similarly fastened 5/8-inch-thick cores. to the other plate. The spring-loaded dial stern The effect of core thickness OR flatwise shear maintained contact with the collar. The collar strength is shown in figure 5. The strength of and dial gage were mounted so that they moved polyester core loaded perpendicular to the core away from each other as the specimen deformed, ribbon direction was not markedly affected by preventing damage to the dial when failure core thickness. Strengths of 1-inch-thick poly­ occurred. ester core loaded parallel to the core ribbon

3 2 Table 1--Mechanical properties1 of polyester and polyimide fiIm honeycomb core

1Numbers in parentheses are standard deviations. 20.003-inch polyester or polyimide film formed to 3/8-inch hexagonal cells.

direction, and for 1-inch-thick polyimide core loaded parallel perpendicular, were about 80 and 60 percent of 0.4-inch-thick core strength, respectively. The shear strength for specimens loaded parallel to the core ribbon direction was about twice that perpendicular to the core ribbon direction. Typical shear stress-strain curves are shown in figure 6 for the polyester core and in figure 7 for the polyimide core.

Figure 4.--Typical compressive stress- strain curves for plastic film Figure 3.--Effect of thickness of flat- honeycomb cores (5/8-inch thick). wise compressive strength of pIastic fiIm honeycomb core. M 133 468 M 133 469

FPL 75 4 For the polyester core, the proportional limit was about half as high and the modulus of rigidity about 40 percent as much for core loaded per­ pendicular the core ribbon direction as for core loaded parallel to the core ribbon direction. Strain to maximum load twice as high for core loaded perpendicular to the core ribbon direction. The proportional limit stress was about half the maximum shear stress. For the polyimide core, proportional limit was about 40 percent as high and the modulus of rigidity about 20 percent as much for core loaded perpendicular to the core ribbon direction as for core loaded in the parallel direction. to maximum load was about twice as high for core loaded perpendicular to the core ribbon direction. The proportional stress was only about 25 percent of the maximum shear stress, Figure 7 shows the two-stage behavior of the polyimide core, After the initial straight portion of the stress-strain curve, buckling of the cell walls became pronounced shear stiffness dropped markedly. A tension field then developed in the cell walls and a second linear portion of the stress-strain curve was observed. Finally, Figure 5.--Effect of thickness on flat- crinkling of the cell wall corners occurred and wise shear strength of plastic fiIm shear stiffness dropped again until the maximum honeycomb core, for loads applied shear stress was reached. Figure 8 shows a failed parallel and perpendicular to core polyimide shear specimen, showing the pro- ribbon direction. M 133 473

Figure 6.--Typical shear stress-strain curves far polyester honeycomb core (5/8-inch thick) with loads applied parallel and perpendicular to core ribbon directions. M 133 472

5 Figure 7.--Typical shear stress-strain curves for polyimide honeycomb core (5/8-inch thick) with loads applied parallel and perpendicular to core ribbon direction. M 133 471 nounced buckles which cause the core to carry load a tension action and thus provide shear resistance after initial cell wall buckling. The second “proportional limit” occurred at about half the maximum shear stress, comparing closely with results for the proportional limit of the polyester core. The slope of the second linear of the polyimide stress-strain curve is about half that of the initial slope which was used to calculate the modulus of rigidity. Because no abrupt failures were observed for the polyimide cores tested in shear and little damage was visible after removal of load, a few specimens were subjected to multiple cycles of loading and unloading. The results are plotted in figure 9. For specimens loaded parallel to the core ribbon direction, the unloading curve for each cycle was roughly parallel to the loading curve. The maximum load for the second loading cycle was somewhat lower and the shear stiffness less than for first cycle, indicating some damage due to initial loading. Practically no residual shear strains remained after load was removed. For specimens loaded perpendicular to the core Figure 8.--Polyimide spear specimen after ribbon direction, results were similar except that failure, showing shear buckles and some plastic deformation occurred, resulting in wrinkles associated with tension field residual shear after load was removed. effects. M 132 670

FPL 75 6 SUMMARY OF RESULTS

Compressive strengths polyester and polyimide cores 1 inch thick averaged about 65 percent of those for cores 0.4 inch thick. Care shear strengths parallel to the core ribbon direction were twice those perpendicular to that direction. Shear strengths for 1-inch­ thick core were about 80 to 95 percent of those of 0.4-inch-thick core for the polyester core and 60 percent the polyimide core. The ratio of the modulus of rigidity for loading perpendicular to the core ribbon direction to that for parallel loading was about 0.4 for the poly­ ester core and 0.2 for the polyimide core. The proportional limit stress was about 50 percent of the maximum stress for the polyester core and 25 percent of maximum stress for the polyimide core.

Figure 9.--Typical shear stress-strain curves for polyimide honeycomb core (I-inch thick) showing cycles of loading and unloading. Material is loaded either paraIIeI or perpendicu­ lar to the core ribbon direction. M 133 470

7 APPENDIX

Table A1--Compressive properties of polyester and polyimide honeycomb core.

FPL 75 8 1.5-9 Table A2-- Shear properties of polyester and polyimide honeycomb core.

1 Denotes direction parallel or perpendicular to core ribbons.