Absolutely. It is entirely feasible that near its ductile-brittle transition temperatiure, a steel, for example, can fail by transgranular cleavage in a test geometry with highly constrained stress-state, such as ahead of a notch or fatigue pre-crack in bending, whereas if the same material is tested in a uniaxial tensile test at the same temperature it would fail by ductile microvoid coalescence. This is because the test geometry can induce a much higher degree of triaxiality of stresses and strains in the vicinity of the notch/crack tip, as compared to that in an unnotched tesnile test, which can promote cleavage fracture, which is a local stress-controlled fracture mode, at the expense of the locally strain-controlled microvoid coalescence fracture mode.

Why triaxiality promote cleavage fracture?

A triaxial stress-state, such as exists ahead of a crack tip in, say in a thick bend sample in plane strain, maintains a high local tensile stress normal to the crack surface as the through-thickness strain is zero such that the stresses are unable to relax through the thickness. This is in contrast to a thin sample (i.e., thin in terms of being comparable to the plastic-zone size), where the stress-state is biaxial (as the through-thickness stress tends to zero) such that local tensile stresses normal to the crack now are lower due to stress relatation through the thickness. Since cleavage fracture is motivated by high tensile stresses, a triaxial stress field is much more potent in inducing such fracture.

In plain strain state , which properties of ductile(in non plain strain state)  steel alloy may change its fracture mechanism to brittle fracture?and why?

The mode of fracture depends on the damage which can be accommodated by the material. For BCC alloys the plastic deformation occurs by slipping mostly. The same can be imagined to be flow of material in favorable direction on favorable planes of crystals. Energetically, BCC crystal has varying number of favorable planes on which slipping becomes possible as the temperature increases (favorable/slip planes increases). This appears in form of a transition from a state where very small plastic deformation is permitted by the material to a state where good amount of energy is welcomed in form of plastic deformation. The mobility of the dislocations which results in plastic deformation can be enhanced/restricted by various means, such as; strain rate, temperature, triaxiality). Any of these three major condition change can result in change in fracture mode. For e.g. increase in temperature, decreasing strain rate and loss of constraint all three give same result i.e. more ductile deformation is possible and vice-versa.     A single specimen tested in the upper region of ductile to brittle transition can start with stable ductile tearing and as crack grows due to increment in triaxiality and increasing numbers of potential cleavage triggers, can end up in cleavage. There is no property of material which dictates the fracture mode; its the state of stress at the crack tip and the material response under the applied loading conditions. The in-homogeneity in any matrix suffers most from the strain incompatibility . This leads to void nucleation when material matrix is ductile and micro-cracking when its difficult. As the material when ductile i.e. it can slip; can accommodate more strain, voids grow and coalesce to form bigger void while in cleavage the micro-crack stays put till the tensile stress is not high enough and then dynamically propagates to end up in cleavage fracture.