Jan Linnros and Miao Zhang describe how they fabricate high-specification porous silicon membranes and use them for parallel optical DNA sequencing.

“The idea is that if you can make the DNA go through a pore in a membrane you can detect the base pair sequences, and by having an array of pores you can do that in a parallel fashion,” explains Jan Linnros, a researcher at KTH – Royal Institute of Technology in Sweden. “If you make 100 × 100 pores the capacity would be 10,000 times faster, so in the best case scenario you could sequence the full genome in a few hours.” While Linnros is quick to emphasise that such an estimate is optimistic, it still significantly outpaces the fastest equipment for full-genome sequencing at present, which takes a few days.

Typically electrical measurements are used to monitor the translocation of base pair sequences in DNA as they move through a pore, but for this kind of parallel sequencing optical techniques have the advantage. The DNA molecules are labelled with a fluorophore and then a wide-field optical microscope and camera tracks their translocation through the array of nanopores.

However, fabricating the nanopore arrays has proved challenging. The low-variation nanoscale pore dimensions required challenge even state-of-the-art lithography techniques. The work by Linnros , his PhD student Miao Zhang and the team at KTH – Royal Institute of Technology suggests that electrochemical etching after initial lithographic patterning may provide a solution. “The advantage of this kind of process is we can use regular top-down lithography to place pores, but we are not limited by regular lithography when it comes to the dimensions of the nanopores because they are set by the parameters of the electrochemical etching,” Linnros explains.

The team showed they could use the technique to fabricate nanopore arrays in silicon membranes. Unlike membranes made from silicon nitride, silicon emits little photoluminescence so there is less background noise obscuring the optical signals. This allowed the first optical detection of parallel translocations of fluorophore-labelled DNA on a conventional wide-field microscope, as opposed to more specialised equipment such as a total-internal-reflection fluorescence or confocal microscope.

Fabrication

The etching approach was inspired by work from another group in around 2005. “They showed that they could make very tiny pores in a membrane made of silicon,” says Linnros. “The pores were 2-3 nm wide but they were randomly placed. So we thought maybe we could do the same thing but incorporate e-beam [electron beam] lithography.”

The team start with a silicon device topped with first silicon dioxide and then a photoresist. They use e-beam lithography to pattern holes in the silicon dioxide, and then chemical etching to create pits shaped as inverse pyramids in the silicon. A final electrochemical etching step hollows out nanopores at the tip of each inverse pyramid.

“This is not perfect yet,” says Linnros. “We can get down to pore sizes of around 10 nm, and in the best case scenario a few pores go down to 8 nm, but there is a lot of room for improvement in the future.”

A tight squeeze

As a result of the high level of control achieved in the nanopore fabrication the researchers were able to investigate the effect of the nanopore width on optically detecting the DNA molecules with a fluorophore. They were surprised to find that the rate of detection of DNA translocating through a pore was greater when the pores were smaller.

The researchers explain these counterintuitive observations in terms of the different factors - the electro-kinetic forces and the molecule-pore interaction – affecting how quickly the molecule moves through the pore. For the larger pores the translocation was faster, so the optical signal was weaker and did not register above the background noise.

Full details of the fabrication and DNA translocation are reported in Nanotechnology. The articles are part of the Nanotechnology focus collection on DNA sequencing.