May 2006 Creep and After-Creep Stress-Strain Behavior 1653 Table l. Physical Properties of the Composite Panels Tow ends per cm action CVI SiC fraction al-SiC(slurry) fraction Si fraction Panel density(g/cm) 0.2 0.14 2.83 7.1 0.21 0.21 0.13 2.74 B 0.33 0.14 2.79 sCatter in the thickness measurement was always less than 0.1 mm. CVI. chemical vapor infiltrated: SiC, silicon carbide: Si, silicon. Tensile test specimens were machined from the panels in the fore, the 0.002% offset strain stress was more useful for com- shape of a dog-bone (length= 152 mm; grip width= 12.6 mm; parison of all the loading curves. Selected specimens were cut gage width= 10 mm)with the 0 fibers aligned along the spec- and polished longitudinally and examined on an optical micro- imen length. Room temperature tensile tests were performe scope to determine the extent of matrix cracking Model 8562, Instron Inc, Canton, MA). Elevated temperature ensile tests were performed ing a screw-driven testing II. Results wrapped with wire( steel) screens in the grip region of the dog (1) Tensile Creep behavior bone test specimen (152 mm long tapered specimens: grip The tensile creep curves are plotted in Fig. I as total strain(creep width= 12.6 mm; gage width= 10 mm) in order to distribute strain plus loading strain) versus time for specimens from the grip pressure and alleviate stress concentrations from the hy three panels tested. In some cases, two specimens from the same draulic wedge grips(MTS 647). Room temperature strain meas- panel were tested per condition and some variation was urement was performed using a 25 mm gage length clip-on served for the two specimens. Table II lists the elastic moduli performed using a 25 mm gage length low-contact capacitance creep behavior of these composites for the most part exhibited a extensometer. For the elevated temperature tests, an MoSiz el ement. resistance-heated furnace was inserted in between the ed, 172 MPa, specimen ruptured and exhibited a rate of defor- two water-cooled grips. The furnace hot zone length was ap- mation increasing with time before failure. It is debatable proximately 15 mm. Room temperature tests were performed in whether"steady-state"creep is achieved. Figure 2 shows the load control, 4 kN/min. Elevated temperature tests(fast fracture strain rate versus strain for the creep experiments as determined nd creep) were loaded with a crosshead-displacement rate of by taking the instantaneous slope over 200 data points( 3.3 h) of the creep data. The composites do not appear to reach a 3.23 mm/min. The time to failure for the elevated temperature steady-state rate for 103 MPa stresses; however, for the 138 MPa tests was usually a few minutes under fast-fracture conditions similar to tests performed at room temperature. applied stresses a steady-state rate may have been achieved or at least nearly approached The 172 MPa specimen achieved a sec- ag, and reloading at predetermined loads to obtain hysteresis ondary-state strain rate; which perhaps was an inevitable arti- ops and determine residual stress in the specimen. Crack in- fact of a specimen experiencing an increasing creep rate before itation and propagation in the tensile specimens tested at room failure. The minimum strain rates and creep strains are listed in temperature was also monitored with wide band acoustic sen Table Il for each experiment. sors(B1025, Digital Wave Corp, Englewood, CO)attached di The stress-strain curves for several creep experiments as well rectly to the specimens. Two sensors were placed outside of the as the stress-strain curves for room temperature and 1315 fast fracture are plotted in Fig 3 for the composite panel from which gage section, and a third sensor was attached within the gage the most tests were performed. Several interesting observations section on the opposite side of the knife edges of the extenso- meter. This enabled accurate location of acoustic events. Only arise from the data in Fig. 3. First there is considerable non- the events that occurred in the gage section were used in the linearity in the 1315C fast-fracture curve at relatively low stress analysis (<100 MPa). For tests at room temperature, such non-linearity For elevated temperature tensile the furnace was first is almost always because of matrix cracking. At 1315C, this aised to 1315C and held at temperature for 10 min to insur could be because of either matrix cracking or composite creep temperature stability. The specimens were then loaded mon during loading or both. Note that the loading rate was relatively tonically to failure for fast-fracture tests or to the predetermined slow and as shown below much of this non-linear behavior is creep stress for the creep experiments. If specimens in the creep deed attributed to matrix creep although some minor micro- ived 100 h, they were rapidly unloaded to zero ress and immediately reloaded to failure at the cree ature under fast-fracture conditions. In some cases the mens were rapidly unloaded and immediately cooled to room temperature where a room temperature hysteresis to failure test A2-172 MPa was performed. In all cases, elastic moduli at room or elevat tures were determined using linear ssion of the 5- 50 MPa part of the loading curve. To determine the onset for B-138 MPa non-linear stress-strain (proportional limit), the 0.002% and 0.4 0.005% offset strain stresses were measured (ASTM C1275)and A2-138 MPa ompared for most of the fast-fracture tests on as-produced specimens, the initial loading curves of the creep tests, and the A1-103 MPa fast-fracture tests for the after-creep specimens. The 0.005% A1-103 MPa offset strain approach has been used at NASA Glenn Research Center for the Sylramic-iBN fiber reinforced MI composite sys- 0.1 B-103 MPa A2-103 MPa tem. 1 However as evident below the 0.005% offset strain stress was in some cases greater than the applied creep stress; there- 100 Time. hr Sylramic fibers were produced by Dow Coming(Midland, MD) foll a propn- etary in situ BN NASA-developed heat treatment. Fig 1. Tensile creep curves.Tensile test specimens were machined from the panels in the shape of a dog-bone (length 5 152 mm; grip width 5 12.6 mm; gage width 5 10 mm) with the 01 fibers aligned along the spec￾imen length. Room temperature tensile tests were performed using an electromechanical actuator universal testing machine (Model 8562, Instron Inc., Canton, MA). Elevated temperature tensile tests were performed using a screw-driven testing ma￾chine (Model 5569, Instron Inc.). The specimens ends were wrapped with wire (steel) screens in the grip region of the dog bone test specimen (152 mm long tapered specimens; grip width 5 12.6 mm; gage width 5 10 mm) in order to distribute grip pressure and alleviate stress concentrations from the hy￾draulic wedge grips (MTS 647). Room temperature strain meas￾urement was performed using a 25 mm gage length clip-on extensometer. Elevated temperature strain measurement was performed using a 25 mm gage length low-contact capacitance extensometer. For the elevated temperature tests, an MoSi2 el￾ement, resistance-heated furnace was inserted in between the two water-cooled grips. The furnace hot zone length was ap￾proximately 15 mm. Room temperature tests were performed in load control, 4 kN/min. Elevated temperature tests (fast fracture and creep) were loaded with a crosshead-displacement rate of 3.23 mm/min. The time to failure for the elevated temperature tests was usually a few minutes under fast-fracture conditions, similar to tests performed at room temperature. Room temperature tensile tests consisted of loading, unload￾ing, and reloading at predetermined loads to obtain hysteresis loops and determine residual stress in the specimen.14 Crack in￾itiation and propagation in the tensile specimens tested at room temperature was also monitored with wide band acoustic sen￾sors (B1025, Digital Wave Corp., Englewood, CO) attached di￾rectly to the specimens. Two sensors were placed outside of the gage section, and a third sensor was attached within the gage section on the opposite side of the knife edges of the extenso￾meter. This enabled accurate location of acoustic events. Only the events that occurred in the gage section were used in the analysis. For elevated temperature tensile tests, the furnace was first raised to 13151C and held at temperature for 10 min to insure temperature stability. The specimens were then loaded mono￾tonically to failure for fast-fracture tests or to the predetermined creep stress for the creep experiments. If specimens in the creep experiments survived 100 h, they were rapidly unloaded to zero stress and immediately reloaded to failure at the creep temper￾ature under fast-fracture conditions. In some cases the speci￾mens were rapidly unloaded and immediately cooled to room temperature where a room temperature hysteresis to failure test was performed. In all cases, elastic moduli at room or elevated temperatures were determined using linear regression of the 5– 50 MPa part of the loading curve. To determine the onset for non-linear stress–strain (proportional limit), the 0.002% and 0.005% offset strain stresses were measured (ASTM C1275) and compared for most of the fast-fracture tests on as-produced specimens, the initial loading curves of the creep tests, and the fast-fracture tests for the after-creep specimens. The 0.005% offset strain approach has been used at NASA Glenn Research Center for the Sylramic-iBNy fiber reinforced MI composite sys￾tem.15 However, as evident below, the 0.005% offset strain stress was in some cases greater than the applied creep stress; there￾fore, the 0.002% offset strain stress was more useful for com￾parison of all the loading curves. Selected specimens were cut and polished longitudinally and examined on an optical micro￾scope to determine the extent of matrix cracking. III. Results (1) Tensile Creep Behavior The tensile creep curves are plotted in Fig. 1 as total strain (creep strain plus loading strain) versus time for specimens from the three panels tested. In some cases, two specimens from the same panel were tested per condition and some variation was ob￾served for the two specimens. Table II lists the elastic moduli, creep, and retained strength properties for each experiment. The creep behavior of these composites for the most part exhibited a decreasing rate of deformation with time. Only the highly load￾ed, 172 MPa, specimen ruptured and exhibited a rate of defor￾mation increasing with time before failure. It is debatable whether ‘‘steady-state’’ creep is achieved. Figure 2 shows the strain rate versus strain for the creep experiments as determined by taking the instantaneous slope over 200 data points (B3.3 h) of the creep data. The composites do not appear to reach a steady-state rate for 103 MPa stresses; however, for the 138 MPa applied stresses a steady-state rate may have been achieved or at least nearly approached. The 172 MPa specimen achieved a sec￾ondary-state strain rate; which perhaps was an inevitable arti￾fact of a specimen experiencing an increasing creep rate before failure. The minimum strain rates and creep strains are listed in Table II for each experiment. The stress–strain curves for several creep experiments as well as the stress–strain curves for room temperature and 13151C fast fracture are plotted in Fig. 3 for the composite panel from which the most tests were performed. Several interesting observations arise from the data in Fig. 3. First, there is considerable non￾linearity in the 13151C fast-fracture curve at relatively low stress (o100 MPa). For tests at room temperature, such non-linearity is almost always because of matrix cracking. At 13151C, this could be because of either matrix cracking or composite creep during loading or both. Note that the loading rate was relatively slow and as shown below, much of this non-linear behavior is indeed attributed to matrix creep, although some minor micro- 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 20 40 60 80 100 120 Time, hr Total Strain, % A1-103 MPa A2-138 MPa A2-138 MPa B-138 MPa A2-172 MPa A2-103 MPa A1-103 MPa B-103 MPa Fig. 1. Tensile creep curves. Table I. Physical Properties of the Composite Panels Panel Tow ends per cm Average f BN fraction CVI SiC fraction a-SiC (slurry) fraction Si fraction Panel density (g/cm3 ) A1 7.1 0.30 0.04 0.25 0.23 0.14 2.83 A2 7.1 0.35 0.04 0.21 0.21 0.13 2.74 B 7.1 0.33 0.04 0.14 0.25 0.20 2.79 wScatter in the thickness measurement was always less than 0.1 mm. CVI, chemical vapor infiltrated; SiC, silicon carbide; Si, silicon. y Sylramic fibers were produced by Dow Corning (Midland, MI) followed by a propri￾etary in situ BN NASA-developed heat treatment. May 2006 Creep and After-Creep Stress–Strain Behavior 1653