A Test Vehicle for RF/​DC Evaluation and Destructive Testing Of Vertically Grown Nanostructures (VGCNS)

We have devel­oped an RF test vehi­cle suit­able for mea­sur­ing DC and microwave per­for­mance of ver­ti­cal­ly grown car­bon nanos­truc­tures (VGCNS) as via-inter­con­nects. A first ver­sion of the inter­con­nect test vehi­cles devices were designed, fab­ri­cat­ed and char­ac­ter­ized. The RF small sig­nal (S‑parameter) and large sig­nal mea­sure­ments show that car­bon nanofi­bres can be used as inter­con­nects in microwave cir­cuit, even for pow­er devices. The design of test vehi­cle employs a three met­al lay­er con­cept, form­ing sequen­tial­ly the ground, sig­nal and device under test struc­tures for char­ac­ter­i­za­tion in a microstrip con­fig­u­ra­tion. The struc­tures as such con­sist­ed of inter­con­nects of dimen­sions rang­ing from 50 nm to 100 µm diam­e­ter made of VGCNS. In the first ver­sion of the inter­con­nect test vehi­cles, the inter­con­nects were made of car­bon nanofibers grown at 450 C. From SEM mea­sure­ment we found that the result­ing height was around 1.5–2 µm. Epoxy poly­mer SU‑8 was employed by spin­ning on the device and a sub­se­quent etch back process was car­ried out to open up the tip of the fibres to con­nect to con­sec­u­tive inter­con­nects with the third met­alli­sa­tion lay­er. After grow­ing the nanofi­bres, it was observed, using SEM, that inter­con­nect sizes small­er than 10 µm diam­e­ter suf­fered from par­a­sitic growth and there­fore the effec­tive device dimen­sion devi­at­ed from the ini­tial design. We car­ried out small sig­nal mea­sure­ments using a vec­tor net­work analyser for fre­quen­cy rang­ing from 1 to 25 GHz, in order to char­ac­terise the trans­mis­sion and reflection/​absorption of the devices as func­tion of their diam­e­ter size. The large sig­nal eval­u­a­tion was per­formed by mea­sur­ing the gain com­pres­sion of the devices. In addi­tion destruc­tive tests, aim­ing at test­ing the cur­rent car­ry­ing capa­bil­i­ty of the inter­con­nect, have also been per­formed. The resis­tiv­i­ty of inter­con­nects was mea­sured to vary varies from 0.2–1.3 mΩ·mm. Appar­ent­ly, the device per­for­mance is con­sid­er­ably influ­enced by the fill fac­tor of the inter­con­nect with VGCNS. Small vari­a­tions in fill fac­tor (in %) pro­vid­ed a large vari­a­tions in device resis­tiv­i­ty. Fur­ther­more, it was also observed that the resis­tance drops at high­er pow­er lev­els. RF con­duc­tiv­i­ty of inter­con­nects ranges from 5×10³ S/​m to 7×10⁵ S/​m. The aver­age input pow­er before inter­con­nect destruc­tion is larg­er than 25W with effec­tive device diam­e­ter rang­ing from 3 µm to 100 µm inter­con­nects. In addi­tion, the aver­age gain com­pres­sion before inter­con­nect destruc­tion was found to be 0.6 dB. It was not pos­si­ble to extract the con­duc­tiv­i­ty val­ue of an indi­vid­ual nanofiber using com­par­i­son to sim­u­la­tion data, since the devices might have suf­fered from par­a­sitic growth as well as pin­hole met­al dif­fu­sion dur­ing top met­al con­tact for­ma­tion. This cer­tain­ly affects the actu­al device dimen­sion and prop­er­ties. Nev­er­the­less, the proof of con­cept of design and man­u­fac­tur­ing a test vehi­cle for RF mea­sure­ments of ver­ti­cal­ly grown nanos­truc­tures was achieved. We will report the find­ings and anom­alies in the mea­sured devices. Fur­ther improve­ment is expect­ed in the com­ing test vehi­cle version.