Strange phenomenon in Tension of Branched Nanotube

Dear all,

   I run a simulation for tensiling a branched silicon nanotube
(Bsint) based on lmp_windows.exe.
I found a strange phenomenon. When we tensile a single silicon
nanotube (SWsint), we can get a perfect result, but for Bsint, we
can't tensile it, as shown in the attached.
   Could you give us some advices for this. Thanks a lot in advance.

Best Regards,
Mike

bnt-lmp.rar (1.73 MB)