40TH ANNIVERSARY grain growth depended on the temperature but the time to stabilisation corresponded to the period of primary creep observed at higher loads. At temperatures above 1150oC, for the Nicalon 200 fibres and above 1050. for the served that it was possible to measure an initial shrinkag followed by positive creep showing that two mechanisms were in competition. The variation of steady state creep rate with temperature, T, and applied stress, o, can often be modelled by: 5um where A is a constant depending on the material. The stress exponent, n, and the activation energy for creep, 2, can be deduced from creep experiments and their val- Figure 5 Fracture morphology of a first generation Nicalon fibre [15]- ues can suggest the processes involved. Below 1200 C, the stress component is near to unity for Nicalon 200 fi interior of the fibre. Failure surfaces remain brittle at high bres revealing Newtonian creep, most probably caused by temperature but new types of defects are seen compared grain boundary sliding and controlled by the oxygen rich to those found at room temperature. Local chemical in- homogeneities at the fibres surfaces such as carbon rich increases to 2 as the intergranular phase decomposes and grain sliding becomes more difficult. zones are preferentially decomposed or oxidized giving The two mechanisms which were in competition were rise to porous weak regions The fibres were seen to creep from around 1000C[17] therefore grain growth and perfection and grain bound although, when first observed, this was considered a con- ary sliding. The results are that the creep curves of first troversial observation as most studies indicated that the generation fibres show primary creep which lasts several fibres shrank on heating above this temperature. These hours followed by secondary creep with no tertiary creep stage up to 1250C Shrinkage could be reduced in these latter conclusions were based on fibres heat treated un- fibres by heat treatment at 1200 C in Argon for 5 h which der no imposed load however the creep observations were induces densification and an increase in Youngs modulus made on fibres subjected to loads at temperature. This of around 15 GPa. The creep activation energy was found still leads to some confusion in the literature with some researchers referring to fibre stability as reflecting the to be around 250kJ/mol characteristics of the fibres, tested at room temperature As the applied stress was increased, at a given tem- after heat treatment. The present authors prefer to consider perature, the period during which shrinkage was ob- the characteristics under load at high temperature, which served decreased until a stress was reached at which only positive creep was measured although, as inti better reflects possible end use conditions. The strength mated above, the processes inducing shrinkage still and elastic moduli characteristics of a material are or- occurred and controlled the period of primary creep ten altered at high temperature with respect to those at Table v shows the stresses as a function of temperature room temperature, because of reversible mechanisms at below which shrinkage has been observed for the two the level of the finest structure of the material and so do not become apparent when the temperature is lowered. fibres Exposure to high temperatures can also, of course, induce Creep curves of the first generation fibres revealed the irreversible changes in the fibre structure which also need existence of stress level thresholds, defined as a creep rate to be understood. The testing conditions for defining sta bility at temperature will be explained as necessary in this TABLE V Maximum stress at which shrinkage has been detected in first ge eneration SiC based fibres Further studies on the first generation fibres revealed that. under low loads. the fibres did shrink but, under Temperature(C) Nicalon 200 Tyranno LOX-M higher loads they crept. This behaviour was observed from around 1000oc. studies on the nicalon 100 series fibres which first revealed this behaviour showed that shrinkage 150 0.34Gl could be attributed to B-Sic grain growth in the fibres 1350 0.18 0. 08 GPa hich stabilised with grain sizes around 3 nm. The rate of 1450 006Gl 0.19 82740TH ANNIVERSARY Figure 5 Fracture morphology of a first generation Nicalon fibre [15]. interior of the fibre. Failure surfaces remain brittle at high temperature but new types of defects are seen compared to those found at room temperature. Local chemical in￾homogeneities at the fibres’ surfaces such as carbon rich zones are preferentially decomposed or oxidized giving rise to porous weak regions. The fibres were seen to creep from around 1000◦C [17], although, when first observed, this was considered a con￾troversial observation as most studies indicated that the fibres shrank on heating above this temperature. These latter conclusions were based on fibres heat treated un￾der no imposed load however the creep observations were made on fibres subjected to loads at temperature. This still leads to some confusion in the literature with some researchers referring to fibre stability as reflecting the characteristics of the fibres, tested at room temperature, after heat treatment. The present authors prefer to consider the characteristics under load at high temperature, which better reflects possible end use conditions. The strength and elastic moduli characteristics of a material are of￾ten altered at high temperature with respect to those at room temperature, because of reversible mechanisms at the level of the finest structure of the material and so do not become apparent when the temperature is lowered. Exposure to high temperatures can also, of course, induce irreversible changes in the fibre structure which also need to be understood. The testing conditions for defining sta￾bility at temperature will be explained as necessary in this paper. Further studies on the first generation fibres revealed that, under low loads, the fibres did shrink but, under higher loads they crept. This behaviour was observed from around 1000◦C. Studies on the Nicalon 100 series fibres which first revealed this behaviour showed that shrinkage could be attributed to β-SiC grain growth in the fibres which stabilised with grain sizes around 3 nm. The rate of grain growth depended on the temperature but the time to stabilisation corresponded to the period of primary creep observed at higher loads. At temperatures above 1150◦C, for the Nicalon 200 fibres and above 1050◦C, for the Tyranno LOX-M fibres, and under low loads, it was ob￾served that it was possible to measure an initial shrinkage followed by positive creep showing that two mechanisms were in competition. The variation of steady state creep rate with temperature, T, and applied stress, σ, can often be modelled by: ε˙ = Aσn exp
−Q RT where A is a constant depending on the material. The stress exponent, n, and the activation energy for creep, Q, can be deduced from creep experiments and their val￾ues can suggest the processes involved. Below 1200◦C, the stress component is near to unity for Nicalon 200 fi- bres revealing Newtonian creep, most probably caused by grain boundary sliding and controlled by the oxygen rich intergranular phase. Above 1200◦C, the stress component increases to 2 as the intergranular phase decomposes and grain sliding becomes more difficult. The two mechanisms which were in competition were therefore grain growth and perfection and grain bound￾ary sliding. The results are that the creep curves of first generation fibres show primary creep which lasts several hours followed by secondary creep with no tertiary creep stage up to 1250◦C. Shrinkage could be reduced in these fibres by heat treatment at 1200◦C in Argon for 5 h which induces densification and an increase in Young’s modulus of around 15 GPa. The creep activation energy was found to be around 250 kJ/mol. As the applied stress was increased, at a given tem￾perature, the period during which shrinkage was ob￾served decreased until a stress was reached at which only positive creep was measured although, as inti￾mated above, the processes inducing shrinkage still occurred and controlled the period of primary creep. Table V shows the stresses as a function of temperature below which shrinkage has been observed for the two fibres. Creep curves of the first generation fibres revealed the existence of stress level thresholds, defined as a creep rate T A B L E V Maximum stress at which shrinkage has been detected in first generation SiC based fibres Temperature (◦C) Nicalon 200 Tyranno LOX-M 1050 - 1.00 GPa 1150 - 0.65 GPa 1250 0.34 GPa 0.40 Gpa 1350 0.18 GPa 0.08 GPa 1450 0.06 GPa 0.19 827