INTRODUCTION
Failure in steels can occur in a variety of mechanisms. Brittle fracture, particularly in cleavage mode is of a priority concerned, since such failure induces severe catastrophe with the least sustained load. Cleavage initiates from the stress raiser by opening mode loading tends to propagate a crack along a specific crystallographic plane that involves the breaking of the atomic bonds. Through the Linear Elastic Fracture Mechanics modeling (Inglis, 1913; Griffith, 1920; Irwin, 1957), a material property called the fracture toughness has been derived to measure the resistibility of material against such crackinduced fracture. For a known material category and fracture toughness value, the fracture by brittle or elastic deformation through cleavage mode can be subsequently defined (Wells, 1961). A standard approach has been well accepted in a testing methodology as in described in the ASTM E 39990 (1991).
In fracture processes, telltale marks appearing on the cracked surfaces can offer a wealth of information related to the fracture mechanism and the material characteristics. For polycrystalline solids like steels, the grains are randomly oriented with respect to each other. Under cleavage loading, a cracking path can be induced to change its direction when come across the grain boundaries, which results in the creation of distinct faceted cracked surfaces. At the onset of crack propagation in particular, cleavage failure leaves unique marks that appears like river and feather patterns (Anderson, 1995). The achievement to trace these trademarks will be the evidences to verify the cleavage mechanism on brittle material fracture.
FRACTURE TOUGHNESS MODEL
Irwin (1957) has derived the K factor to characterize the intensity of local stress such that the stress state of a material is expressed by
where, the K_{I} is the ModeI stress intensity factor, r is the distance of any entity originating from the crackfront, θ is an angle measured from the crack plane and f_{ij}(θ) is a known function of θ. From the dimensional analysis (Anderson, 1995), the only parameters defined in a fracture loaded body is the loading stress σ and the crack length α, which is related as
where:
γ 
= 
An undetermined dimensionless factor. 
Considering the relationship between K_{I} and the ratio of crack length
over the width of a body a/W, the dimensionless factor is thus related to the
geometric function f(a/W). Since the function consists of a factor
for any size of a/W, the equation is alternately written as
where:
for Compact Tension geometry (Sih, 1973; ASTM E 39990, 1991). These relations are defined for prediction of elastic fracture by cleavage mode, which utilizes a geometric body confined by the plainstrain deformation of Compact Tension specimen. Therefore, the K_{I} is defined as measurement of a fundamental material fracture toughness or resistance to crack propagation.
MATERIAL AND PROPERTIES
The elemental composition of the carbon steel was obtained by performing the
Glow Discharge analysis. The main elements concern was accurate up to a coefficient
of variation of no more than 5% and the average data composition is tabulated
in the Table 1. Microstructural details of the material was
also obtained by selecting, grinding, polishing and etching to review the pearlitic
and base iron grains as shown in the Fig. 1.
Table 1: 
Elemental
material composition obtained by glow discharge analysis 


Fig. 1: 
Pearlitic
(light) and baseiron (dark) microstructure 
Table 2: 
Mechanical
tensile properties derived from tests conducted using the Universal Tensile
test machine 

The pearlitic
structure constitutes up to about 36% of the material. Further mechanical tensile testing, determined by ASTM E801 (2002), of the
material was conducted using a servo hydraulic tensile testing machine. The
material deformation characteristic results were also obtained and analyzed
to yield the average tensile properties as tabulated in the Table
2 below; coefficient of variation ranges from 3.7% to 5.6%.
TESTING COMPACT TENSION GEOMETRY
The Compact Tension geometry for modeI plainstrain linear elastic fracture characterization was adopted as determined by the ASTM E39990 (1991). The geometrical dimensions of the compact tension specimen were as shown in the Fig. 2. The standard dimensional profiles had been derived based on the constraint of elastic material characteristic properties in order to achieve plainstrain fracture toughness criterion. In particular, to ensure validity of linear elastic fracture mechanics application, the critical sharp crack tip is necessary to constraint the initiation of crack by separation of elastic atomic bonds. At the macromechanical level, a macro scale Vshape profile of the crack tip was machined out using the wire cutting technique. In order that fatigue precracking could be conducted to create fine hairline crack ahead of the Vnotch tip.

Fig. 2: 
ASTM
E39990 compact tension specimen geometry 
A total of five specimens were prepared following the proportional dimensions
for one critical thickness of about 26 mm such that the effective elastic fracture
toughness could be achieved at plain strain constraint of B≥2.5 (K_{IC}/σ_{ys})^{2}.
Characteristic loaddisplacement at fracture initiation and reduction of data
were obtained via standard unidirectional tensile loading by cleavage to the
specimen using a servo hydraulic tensile machine. The applied loads and corresponding
Crack Opening Displacements (COD) were recorded by the machineintegrated load
cell and an externallymounted clip gage spanning at the starter notch mouth.
DETERMINATION OF K_{IC}
The load versus COD curve (PCOD curve) provides the characteristic signature
of the fracture resistibility to crack development under cleavage load. The
reduction of data based on ASTM standard derives a 5% deviated slopeline that
intersects the characteristic curve provides the fracture load point at initiation
is presented in Fig. 3. At the same time, initial linear slope
must be within the constraint of 0.7≤ liner slope ≥ 1.5 or P_{max}/P_{Q
}≤1.10, in order to satisfy plain strain condition. As a result of determining
the fracture load point at initiation, the stress intensity factor K_{IC}
or fracture toughness value can now be determined. The resultant experimental
K_{IC} for a range of thickness is shown in the Fig. 4.

Fig. 3: 
Characteristic
P versus COD plot 

Fig. 4: 
Fracture
toughness versus thickness characteristic plot 
Plain strain elastic fracture toughness free from thickness effect is experimentally
derived at the value of about 33 Mpavm with a coefficient of variation of 5.9%. While the limiting thickness for this K_{IC} value is estimated at
a minimum value of 23 mm for this particular carbon steel material.
FRACTOGRAPHIC ANALYSIS
The fractographic analysis of the fractured surface provides important insights as to the mode and mechanisms of failure. Three characteristic fractured zones, fatigue precrack, crack initiation and gross fracture, were prepared and subjected to Scanning Electron Microscopic (SEM) observation. Their characteristic microfailure behaviour and mechanism were obtained as shown in the Fig. 5. In the precrack
zone, wavelike striation patterns were observed had been a result of planar slippage deformation due to the fatiguedloading effect. At the onset or crack initiation zone, flat and faceted microstructure surfaces were distinctly observed. Following the onset, gross plastic failure by separation of specimen produces microstructures with characteristic dimpled together with coalescence of microvoids were wide spread. The interest of the SEM analysis was focused at the instant onset of crack propagation as this provides an indication of material characteristic and cleavage. Corresponding trademarks of cleavage; were identified by river and feather deformation patterns observed at the onset zone as shown in the Fig. 6 and 7, respectively.

Fig. 5: 
Fractographs
of distinctive zones in precrack, initiation and gross failure 

Fig. 6: 
River
patterns at crack onset 

Fig. 7: 
Feather
marks at crack onset 
CONCLUSIONS
Planestrain fracture toughness value of about 33 MPavm had been obtained experimentally at a limiting thickness of at least 23 mm for the carbon steel material utilized in the investigation. In addition, the fractography analysis had illustrated in the microstructural deformation patterns, distincted to river and feather marks at the crack onset, as a commonly trademark found on the cleavage and brittle failure surfaces. Hence the characterization of cleavage in carbon steel was successfully validated by the determination of planestrain fracture toughness characteristic and the fractography analysis.
ACKNOWLEDGMENT
The authors wish to express their sincere appreciation to the Center of Materials and Minerals (CMM), Universiti Malaysia Sabah, for their support in this project.