A nanopore is an extremely tiny hole (a few nanometres in diameter) in an insulating membrane separating two ionic reservoirs. When a voltage is applied across the membrane, ions and charged biomolecules, such as DNA, traverse or translocate through the pore. The translocation of DNA blocks the ionic current, thus transforming the translocation to an electronic signal. If the nanopore encodes different bases with different current amplitudes, the DNA sequence could be readily identified.

However, current experimental platforms separate the nanopore from the electronics required for sensing the signal, which results in lower bandwidth, increased noise and prevents massive parallelization. Using state of the art fabrication techniques, a new platform that co-integrates the nanopore within a CMOS chip has been developed. This allows the realization of highly compact and massively parallel biosenors on-a-chip.

Suspended oxide layer

The major challenge to the development of a CMOS integrated nanopore has been the tight thermal budget (>350 °C) of post-CMOS processing. Conventional approaches require high-temperature deposition of thin silicon dioxide or silicon nitride films, which are used as the insulating membrane for the nanopore. The team overcame this issue by utilizing the existing pristine oxide layer normally used as the insulator for on-chip capacitors. By suspending the oxide layer, using microfabrication techniques, freestanding membranes were formed.

To their surprise the group discovered that oxides available in CMOS on-chip capacitors were very leaky, preventing their direct use. The researchers overcame this by using atomic layer deposition (ALD) of aluminum oxide as a barrier layer. The aluminum oxide, due to its low surface charge, additionally lowers the flicker noise of the pore.

The nanopore was drilled using a tightly focused electron beam in a TEM and then subsequently shrunk using ALD. Quantitative DNA translocation experiments performed with 48 kbp λ-DNA proved the functionality of the on-chip pore.

Path to massively parallel biosensors

The approach holds the promise of realizing massively parallel nanopore biosensors. Since each nanopore can be intimately connected to its own sensor electronics, thousands of translocation experiments can be performed in parallel. Furthermore, the work also demonstrates the fabrication of a nanopores using e-beam lithography enabling wafer-scale processing of nanopore sensors.

The team is currently conducting translocation experiments using custom designed on-chip electronics and hopes to apply the results to DNA and protein sensing.

The researchers presented their results in the journal Nanotechnology.