Page 384 - Fiber Fracture
P. 384

366                                                        J. Bemholc et al.

                                       ab initio vs. Classical MD
                               lo '1

















                                    - graphene, classical MD
                                    H (5,5)tube, ab initio
                               -5  '   - graphene, ab initio            J
                                                     I
                                  0                  5                 10
                                                 Strain (%)
               Fig.  8.  Formation energy of  the  off-axis (5-7-7-5)   defect in  (n,O) tubes of  various diameters. In  (n,O)
               tubes, D = 0.078~ nm. Inset: formation energy of the off-axis (5-7-7-5)  defect in (10,m)  tubes of different
               chiralities. Note that in  (n,m) tubes the chiral angle x = arctan[&/(2n   +m)]  is zero in zigzag tubes and
               30" in armchair tubes. All data refer to 10% strain.


               nanotube, the same C-C  bond will be parallel to the applied tension, which is already
               the minimum energy configuration for the strained bond. The formation of the Stone-
               Wales defect is then limited to rotation of the bonds oriented 120" with respect to the
               tube axis. Our analysis shows that  the  formation energy of  these defects  is  strongly
               dependent on curvature and thereby on the diameter of the tube. In fact, a number of
               different behaviors have been observed when constructing the brittle vs. ductile map of
               stress response of  carbon nanotubes. The results of  static energetics calculations and
               molecular dynamics simulations for (n,O) tubes of  various diameters D at  10% strain
               are summarized in Fig. 8. Remarkably, the formation energy of the off-axis (5-7-7-5)
               defect (obtained via the rotation of  the C-C  bond oriented 120" wrt. tube axis) shows
               a crossover with respect to the diameter, and it is negative for (n,O) tubes with n < 14
               (D < 1.1  nm).  Similarly, the  formation  energy of  this  defect  in  chiral  tubes  of  the
               (10,m) family (chosen as a particular example) is always negative, although it changes
               with the  chiral  angle  x. This  result  implies that  the  bond-rotation transformation is
               still efficient in releasing the strain energy of the tube. This effect is clearly due to the
               variation in curvature, which in the small-diameter tubes makes the process energetically
               advantageous. Therefore, above a critical value of  the curvature a plastic behavior is
               always possible and the tubes can be ductile.
                 From these calculations one can identify the full range of elastic responses in strained
               carbon nanotubes. In  particular, under high-strain and low-temperature conditions, all
   379   380   381   382   383   384   385   386   387   388   389