editor's blog
Subscribe Now

A Different TSV

Through-silicon vias (TSVs) are a key technology for stacking chips in so-called 3D (or 2.5D) configurations. And there’s lots of talk about how to use deep reactive ion etching (DRIE) methods to get a nice deep high-aspect-ratio hole and then fill it with metal to make this connection.

But in a conversation with Silex at the MEMS Executive Congress, they pointed out a different approach they take to TSVs. And, it turns out, this approach has been in production for five years, so it’s not exactly new – yet I haven’t seen it discussed in the mainstream (and maybe that’s just me).

What’s different here is that it’s not a metal via: it’s a silicon via. Which gives rise to the clever name SIL-VIA.

It’s predicated on highly-doped bulk silicon. As a pure-play MEMS company, wafers may have characteristics different from that of CMOS wafers. Instead of etching out the entire hole for a pillar of metal, they etch a border around the hole (typically cylindrical, but it can actually be any shape). They then grow oxide in that gap, which now isolates the silicon remaining inside from that outside.

The inside and outside silicon are still connected, since the etch doesn’t go all the way through the wafer. The central core is “released” by grinding away the back. Now the inside portion is isolated, and can form a ½-1-Ω via. This contrasts with the 10-20 mΩ that are more typical of metal vias.

The primary benefit they tout is that the via is thermally matched to the surrounding bulk material, which they say is a problem with metal vias. They also note, of course, that this has been in production for several years; they claim experience on over 50,000 wafers implementing over 100 different products.

There’s a bit more information mid-page on their technology page

Leave a Reply

featured blogs
Oct 14, 2019
Simon Segars opened Arm TechCon with a new look, having discovered that real men have beards. This is the 15th Arm TechCon. In this post I'm going to focus on the new things that Arm announced... [[ Click on the title to access the full blog on the Cadence Community sit...
Oct 13, 2019
In part 3 of this blog series we looked at what typically is the longest stage in designing a PCB Routing and net tuning.  In part 4 we will finish the design process by looking at planes, and some miscellaneous items that may be required in some designs. Planes Figure 8...
Oct 11, 2019
The FPGA (or ACAP) universe gathered at the San Jose Fairmount last week during the Xilinx Developer Forum. Engineers, data scientists, analysts, distributors, alliance partners and more came to learn about the latest hardware, software and system level solutions from Xilinx....
Oct 11, 2019
Have you ever stayed awake at night pondering palindromic digital clock posers?...
Oct 11, 2019
[From the last episode: We looked at subroutines in computer programs.] We saw a couple weeks ago that some memories are big, but slow (flash memory). Others are fast, but not so big '€“ and they'€™re power-hungry to boot (SRAM). This sets up an interesting problem. When ...