Philip Garrou

Semprius
January 16, 2008

Semprius ! – wasn’t that the chant the crowd gave the Russell Crowe character in Gladiator every time he took out the competition? Maybe so, but that’s not what we’re going to talk about today.

The Semprius I’m familiar with is a Research Triangle Park NC startup, commercializing technology invented by the John Rogers group at the University of Illinois. ( In my full disclosure mode let me indicate that I am involved with consulting for this company) It was launched in 2005 with Joe Carr, former CEO of OSRAM Opto Semiconductors, at the helm. In April of 2007 they secured  $4.1 MM  in Series A financing with Intersouth Partners and ARCH Venture Partners co-leading the round and Illinois Ventures also participating. Applied Ventures, the venture capital operation of Applied Materials, invested an additional $500K in Semprius a few months later.

In the fall of 2006 the company won top honors in The Wall Street Journal's Sixth Annual Technology Innovation Award in the semiconductor category and in 2007 Semprius received the 2007 Spin-out of the Year award given by the Council for Entrepreneurial Development, the largest entrepreneurial support organization in the nation.

This past summer Semprius received a SBIR Phase I project from NSF to demonstrate printing of thin-film transistors (TFT) on plastic substrates for use as high-performance backplanes in flexible displays and another SBIR, this time from DARPA, to develop light weight flexible photovoltaics.

So why is Semprius causing all this stir in the microelectronics community ? Semprius is developing unique patented technology called  micro Transfer Printing ( µTP), which allows them to remove semiconducting devices from a source wafer and transfer them onto virtually any surface, including glass, flexible or rigid plastic, metal or non silicon semiconducting materials.

Circuits or devices to be µTP are fabricated in normal IC foundries using standard  fab sequences with the exception that there must be an etchable “release layer” beneath the circuit. In the case of silicon they use an SOI wafer and use the buried oxide as the release layer. In the case of GaAs circuits this can be an AlAs layer. These wafers then go through a process to delineate and release the individual chiplets by a vertical etch around the chiplet and subsequent horizontal etch of the oxide release layer. Micro tethers (bridges), which are designed to break in a controlled manner during the pick-up process, are used to hold the released devices in place on the surface of the source wafer until they are ready to be removed.

The released devices are removed from the source wafer using an engineered elastomeric stamp which is the key to their technology. The elastomeric stamp populated with devices is then aligned to a target substrate, brought into contact with the substrate and controllably releases the chiplets onto the substrate. This process sequence is shown below.

The transfer printing step exhibits both massively parallel transfer and “geometric magnification.” Depending on the specific configuration of the circuit design, in each transfer step hundreds to thousands of chiplets can be “printed”. The inventors term “printed” refers to being “transferred” or in our packaging language “picked-and–placed”. By appropriately designing the stamp, it is possible to pick up only one of every “n” chiplets from the source wafer. As a result, when operated in a “step and repeat” fashion, a dense array of chiplets on the mother wafer can be transferred into sparse arrays on the device substrate, as shown below.

The key feature of this approach is that all of the demanding fabrication process steps necessary to fabricate high performance IC chips or other devices are performed on the “mother” substrate and not on the final device substrate. As a result, the inherent mechanical or chemical instabilities of the receiving substrate do not limit the choice of semiconductor manufacturing processes or materials for fabricating the devices. ( Flexible substrates such as flexible backplanes for OLED displays become a natural application space. )

Due to the elastomeric nature of the transfer stamp, transfer printing is ideally suited for handling ultrathin (less than 10 µm) and small (less than 250 µm,( although several mm are possible) ) devices. Indeed , the elastomeric transfer stamp can move thousands of devices in a single printing step.
It is this ability to handle tiny structures and to move thousands of them accurately, in a massively parallel fashion, that will enable new technological breakthroughs such as massive parallel pick and place of the driver circuits for display pixels. When it comes to photovoltaics, Semprius μ-transfer printing in combination with the right micro lens technology enables : (1) small cell size (0.01 mm2); (2) easier thermal management; (3) easier current collection; (4) uniformly illuminated cells and (5) high concentration ratios, all of which result in high performance concentrator photovoltaic modules.

So if you have an application that needs ICs on an unusual surface –or- an application that could profit from massively parallel pick-and-place of small ICs or passive elements –or- you’re designing a circuit in one semiconductor material that could benefit from a subcircuit made of another semiconducting material think Semprius. 

Next week an exclusive interview with EVG discussing their perspective on the evolviong 3D IC marketplace.....If you want to stay up on whats happening on the leading edge stay linked to "Perspectives From the Leading Edge"....................................

Posted by Philip Garrou on January 16, 2008 | Comments (0)