The temperature dependence of critical strain energy release rate (G(c)') a
nd standardized Charpy notched impact strength (CNIS) were measured for a t
hermoplastic polyurethane (TPUR) reinforced with 30 wt% of short glass fibe
rs (SGF) over a temperature interval ranging from -150 degrees C to 23 degr
ees C (RT) at two strain rates, 70 and 150 s(-1), respectively. Fractograph
ic observation of fracture planes was used to qualitatively assess the frac
ture modes and mechanisms. Adhesion between the reinforcement and the matri
x was excellent and the integrity of the fiber-matrix interfacial contact w
as relatively insensitive to exposure to hydrolysis during the immersion in
boiling water for 100 hours. At temperatures above -30 degrees C, there wa
s a large extent of plastic deformation in the vicinity of crack planes whi
le at temperatures below -50 degrees C, the extent of plastic deformation w
as substantially reduced. This resulted in a change in the major energy dis
sipation mechanism and led to a decrease of both CNIS and G(c)' values for
SGF/TPUR composites. It was suggested that the plastic deformation of TPUR
matrix in the immediate vicinity of glass fibers was the primary source of
energy dissipation at temperatures above -30 degrees C, while the friction
and fiber pull-out was the main dissipative process below -50 degrees C. Ov
er the whole temperature interval investigated, greater G(c)' values were o
btained at higher strain rate of 150 s(-1), without any significant change
in the fractographic patterns observed on the fracture planes. The CNIS/G(c
)' ratio, used to assess suitability of CNIS for comparison of materials, c
hanged with temperature substantially suggesting that the functional depend
ences of CMS and G(c)' on temperature differ substantially. Hence, CNIS dat
a do not provide a reliable base for material selection and for design purp
oses in this case.