Millipede

from Wikipedia, the free encyclopedia

Millipede (from Latin thousand feet) is a storage technology for digital data developed by IBM in Zurich in 1999 to a functional model , in which individual craters are positioned in a thermoplastic with the help of 1024 pointed arms, read out and leveled again. It is based on atomic force microscope technology developed by Nobel Prize winner Gerd Binnig .

Basic principle

The basic principle is comparable to that of the cardboard punch card that still appeared in the 19th century , but is applied to structure sizes in the nanometer range . Another crucial difference is that the technology used can be used to delete and overwrite the bits . Tiny small lever (engl. Cantilever ) with a fine tip of the silicon melt as pinholes in a polymer - medium to write bits. The same tips are also used to detect these holes, i.e. to read out the bits again. This is done by bringing the tip close to the polymer film and heating it. If the tip dips into a bit crater, the heat exchange between it and the storage medium increases, as a result of which the electrical resistance of the lever decreases. In order to overwrite a bit, new recesses are created with the tip on the crater rim, the edges of which overlap the old recess and thus force the polymer material towards the crater.

Because the holes are very small, they can be placed very close to each other and thus achieve high data densities . With this technology, IBM scientists in the Rüschlikon laboratory have succeeded in penetrating the nanometer range. When storing data, a recording density of one terabit per square inch could be achieved, which corresponds roughly to the content of 25 DVDs on the area of ​​a postage stamp . This density was achieved with a single silicon tip that creates pits around ten nanometers in diameter. In order to increase the data rate , i.e. the writing and reading speed , not just a tip is used, but a whole matrix of little levers that work in parallel. A prototype has more than 4,000 such tips, which are arranged in a small square 6.4 mm on a side. These dimensions make it possible to pack a complete high capacity storage system into the smallest standardized format for flash memory.

A lever arm in the Millipede writes and reads from a 100  µm  × 100 µm cell assigned to it. While the read and write head and the storage medium move on hard disks , for example , only the medium is moved on Millipede memories. Two coils , which are placed between magnets , drive the movement of the platelet: the microscanner can be positioned with an accuracy of up to two nanometers. The position can be determined from the overlapping area of ​​strip-shaped sensors , but these sensors consume a comparatively high amount of energy .

Building Millipede

Lever arm field

The core of the Millipede technology is a two-dimensional arrangement of V-shaped silicon spring tongues (lever arms) that are 70  µm (thousandths of a millimeter) long. At the end of each lever arm there is a sensor for reading and a resistor above the tip for writing. The tip is just under a micrometer long and the radius is only a few nanometers. Lever arms are arranged in fields on an integrated circuit (IC, chip). The chip is 7 mm × 14 mm in size. In the center sits a field of, for example, 4096 (64 × 64) lever arms that are etched out of the silicon. The actual data carrier consists of a polymer film that is only a few nanometers thick on a silicon substrate . The heads read, write or delete the desired bit, individually controlled via multiplexers . Up to 100,000 write and overwrite cycles are said to have been successfully tested so far. And although mechanics are used in the construction , a transmission speed of up to 20 to 30 megabits per second can be achieved. An IBM publication from 2010 names 10,000 read-write cycles as the lifespan.

Microscanner

The movement of the storage medium relative to the cantilever array is realized with the aid of a silicon-based x / y microscanner. The scanner consists of an approx. 6.8 mm × 6.8 mm scan table that carries the polymer medium and two electromagnetic triggers. The scanner chip is mounted on the silicon plate, which serves as the mechanical base of the system. The distance between its surface and the surface of moving parts of the scanner is approx. 20 µm. The scanning table can be moved by 120 µm in x and y directions using a trigger. Each trigger consists of two permanent magnets built into the silicon plate and a small coil located between the magnets . In order to erase the vibrations from the outside, a so-called pivot is used, which is coupled with the triggers.

Positioning

The information about positioning is provided by four thermal sensors. These sensors are located directly above the scanning table, on the cantilever array. The sensors have thermally insulated heaters . Each sensor is positioned over one edge of the scanning table and is heated by electricity. Some of this heat is conducted through the air to the scanning table, which now serves as a cooler . Moving the scanning table causes a change in the efficiency of this cooling system , which leads to a change in the temperature of the electrical resistance of the heater.

A sophisticated design ensures the exact leveling of the tips above the storage medium and dampens external vibrations and shocks. Time division multiplexing electronics , as used in a similar way in memory chips ( DRAM ), enable each individual tip to be addressed in parallel. Electromagnetic actuation moves the substrate with the storage medium on its surface very precisely in the x and y directions, so that each tip can read and write in its storage area of ​​100 µm side length. The short distances contribute significantly to low energy consumption.

For the functions of the device, i.e. reading, writing, erasing and overwriting, the tips are brought into contact with the polymer film on the silicon substrate.

Writing technology

The writing of bits is carried out by heating the integrated in the cantilever resistor to typically 400  ° C . The tip, which is also heated as a result, softens the polymer , sinks in and leaves a depression of a few nanometers. For reading, the reading sensor of the cantilever is heated without softening the polymer film. If the tip “falls” into a recess, the reading sensor cools down slightly because of the larger contact area between substrate and tip and thus greater heat dissipation, but this leads to a measurable change in resistance. To overwrite data, the tip makes indentations in the surface. Their outer edges overlap the old wells and thus erase the old data. More than 100,000 write and overwrite cycles have proven that the concept is suitable for a rewritable memory type.

In order to get the data in and out of the memory faster, a complete matrix arrangement of levers processes the medium at the same time. However, it turned out that it is extremely difficult to manufacture mechanics and electronics in one piece on one chip. The researchers therefore decided to implement the construction in two pieces:

  1. The matrix of the microtips is provided with tiny contact pins that look like the knobs on Lego bricks under the electron microscope.
  2. These studs are then contacted with their counterparts on the electronic circuit board .

The prototype shown recently demonstrated the technical feasibility of a product in terms of storage density , performance and reliability . While the storage technologies used today are gradually reaching their fundamental limits, the nanomechanical approach has enormous development potential for a storage density that is a thousand times higher. This nanomechanical data carrier generates almost no heat , consumes little electricity and is shock-resistant.

At the moment, IBM is looking for SD card manufacturers who are interested in using the technology and in licensing it. IBM is already in talks, but has not yet wanted to name any potential partners. It should not take more than two to three years to reach market maturity - provided that partners can be found.

Web links

Individual evidence

  1. IBM's 'Millipede' Project Demonstrates Remarkable Trillion-Bit Data Storage Density ibm.com, June 11, 2002, accessed June 23, 2017.
  2. 1000 Tips for Ultrahigh-Density Data Storage ibm.com, October 11, 1999, accessed June 23, 2017.
  3. Pharma 2010: Silicon reality ibm.com, IBM Global Business Services, brochure / PDF, p. 10 side column. 2010, accessed June 23, 2017.