Pure tilt boundaries around c-axis

 

 

One of the characteristic patterns of etch pit distribution close to the center of good quality 2” wafers is shown in Fig. 1.  It consists of lines extending along <1-100> directions forming hexagonal or triangular shapes.  At higher magnifications, it is easy to see that the lines are made of hexagonal and/or circular etch pits characteristic of dislocations in SiC.  Thus, the lines are traces of dislocation arrays.  The linear etch pit densities along the lines were estimated to be in the order of 103 cm-1, while the etch pit densities in areas without these arrays were on the order of 105 cm-2.  The corresponding average spacings between neighboring dislocations are 1 mm and 30 mm, respectively.

 

 

Fig. 1  Optical micrograph of the etch pit arrays on KOH etched Si(0001) face of a 4H SiC wafer.

In the figure, most small etch pits are in regular circular shape, which implies that these are due to threading dislocations extending approximately normal to the basal plane and wafer surface. In particular, they are different from the shell etch pits usually associated with basal plane dislocations. These arrays were analyzed by transmission electron microscopy and by high resolution x-ray diffraction.

Fig. 2(a) is a plan view conventional TEM image showing a part of an array shown in Fig. 1.  Four dislocations are visible and are marked with arrows in the figure.  They form a straight array along a direction of the <-1100> type.  The wide white contrast features are thickness fringes due to changing thickness across the sample foil.  This image was taken in a two beam diffraction condition slightly tilted from the c-axis in an angle smaller than 5°.   The fact that the dislocation images are nearly point shaped in this projection implies that the dislocation lines are almost parallel to the c-axis.  High resolution TEM allowed to determine the line directions and Burgers vectors of the individual dislocations more precisely. Fig. 2(b) is a plan view high resolution lattice image around a dislocation in the array shown in Fig. 2(a). The image was taken by choosing the c-axis as the zone axis. The dislocation is of pure edge type and threading along the c-axis without tilt. Its core is at the intersection of the two extra half planes marked with two rows of dots. The corresponding Burgers vector, determined by drawing a Burgers circuit around the core, is of the a/3<11-20> type with a direction marked with an arrow.

 

        

 

Fig. 2  (a) Plan view bright field conventional TEM micrograph showing a part of a [-1100] dislocation array shown in Fig. 1. (b) C-axis plan view lattice image around a dislocation in the array shown in (a).  The two extra half planes are marked with two arrays of dots and the corresponding Burgers vector direction with an arrow.

 

The Burgers vectors of the individual dislocations in the array were determined to be same and perpendicular to the array.  In general, edge dislocations with same Burgers vectors are known to align themselves along the direction perpendicular to the Burgers vectors to minimize their overall strain energy in a process called polygonization (R. E. Reed-Hill and R. Abbaschian, Physical Metallurgy Principles, 3rd ed., PWS Publishing Company, Boston, USA (1994) 233-239).  The average spacings between neighboring dislocations in the arrays were estimated to be in between 0.3 and 1 mm.  These correspond to the misorientations in between 60 and 200 arc seconds across the arrays, assuming the dislocations are pure edge ones with Burgers vectors of the a/3<11-20> type.  The type of misorientation would be a tilt of which the rotation axis is parallel to the c-axis, that is, the crystals on both sides of a boundary are rotated around an axis in the boundary, normal to the basal plane. Such misorientation should not affect the basal plane reflection of x-rays.

 

Electron microscopy allows analysis of only a very small fragment of the dislocation array.  The average large scale misorientations have been measured by high resolution x-ray diffraction (HRXRD).  Fig. 3 shows the morphology of one of the samples.  This sample was selected in such a way as to isolate a single well-defined array in 4 mm ´ 4 mm area.  One side of the sample was cleaved to expose the prism plane of the {-1100} type, perpendicular to the array direction.  The horizontal boundary in the top portion of the figure between the gray background and the white sample surface is the cleaved edge.  The well-defined etch pit array is marked by an arrow and is clearly seen to be the only array that intersects the prism plane edge. 

 

 

Fig. 4  Optical micrograph of a high resolution XRD sample having a well-defined <-1100> array and a cleaved edge perpendicular to it.

Fig. 5(a), (b) and (c) are the three w scan curves collected in three different configurations of the diffraction plane with respect to the array. Each configuration gives one of the three possible misorientation components: two tilts and one twist. The measured component is defined by the diffraction plane normal (w axis).  The diffraction plane normal was in the basal plane and parallel to the etch pit array for the case (a), while perpendicular to the array for the case (b).  These two components represent the misorientation of the basal plane.  The former is one of the two tilt components of the low angle boundary and the latter is the twist component.  The second tilt component was measured in the case (c) where the diffraction plane normal was parallel to the c-axis.  In each diffraction experiment the beam footprint straddled the dislocation array.  In configurations (a) and (b) one can see only one peak in x-ray reflection with shoulder separated from the main peak by approximately 10 arc seconds.  In scan (c), the results show two peaks of comparable intensity separated by 140 arc seconds. From this, it is clear that the dominant misorientation is the rotation of basal plane around c-axis.   This is in agreement with TEM results and attests that the domain boundary is made almost exclusively of pure edge dislocations as shown in Fig. 2.

 

        

(a-c)                                                                 (d)

Fig 5  Three w scan rocking curves corresponding to the three components of the misorientation associated with the array in Fig. 4 (a) tilt with the axis in the basal plane, parallel to the array; (b) twist with the axis in the basal plane, perpendicular to the array; (c) tilt with the axis parallel to the c-axis (d) geometry of the sample, diffraction experiment, and misorientation.

 

 

The origin of the threading edge dislocations is discussed in Threading edge dislocations page.  One of the mechanisms responsible for their appearance is prism plane slip which results is bands of etch pits extending along the <11-20> directions on Si(0001) faces of SiC wafers.  An example of two such bands about 50 and 30 mm wide (marked with arrows) are shown in Fig. 6.   It is apparent that in addition to the general orientation of the slip band there is a fine structure consisting of dislocations forming short arrays along [11-20] direction and perpendicular to the slip band direction.  It is likely that the fine structure represents initial stages of the polygonization process.

 

 

Fig. 6  Optical micrograph of the detail of two prismatic slip bands (marked with arrows).

 

The schematic diagram of dislocation distribution and Burgers vector orientations are shown in Fig. 7.  With prolonged annealing of boules during the growth process, the polygonization will progress further and well defined <-1100> low angle tilt boundaries will form.

 

 

Fig. 7 Distribution of dislocations in a prismatic slip band (projection along [0001]) (a) as created and (b) at the initial stages of polygonization.

 

It should be stressed here again, that the domain structure shown in Fig. 1 and discussed above does not affect the basal plane x-ray diffraction and cannot be responsible for commonly observed multiple peaks in basal plane reflections.  Other types of low angle grain boundaries with different morphologies and origin must be present in SiC boules. 

 

Summary

Threading dislocations forming etch pit arrays along <-1100> directions were revealed by KOH etching on Si(0001) faces of [0001] grown crystals.  Conventional and high resolution TEM have identified the dislocations as pure edge ones with dislocation line direction along the c-axis and Burgers vector of the a/3<11-20> type. The Burgers vectors of the individual dislocations in an array were same and perpendicular to the array. The misorientations due to the arrays were estimated in the range of 60-200 arc seconds.  The type of misorientation was determined to be pure tilt around the c-axis.  From these observations, the arrays were interpreted as formed by polygonization of the threading edge dislocations.

 

Return to:

 

[ CMU SiC Page ] [ MSkowronski Home Page ]