Go to the CD home page
Newsroom
-News archive
-Press contacts
Newsletters & Articles
In focus
Subscribe to eNews
Shows & Events
Relevant Links
-Shows & Events
Relevant Links
-Microcontrollers
-Sales & distribution offices

1997-07-07 , E/IC-804/451

Philips creates ultra fast OTP memory process for microcontrollers


Philips Semiconductors has announced that it has developed an ultra fast One Time Programmable (OTP) memory process that is now available in its XA 16 bit and 80C51 8 bit microcontrollers. This process to create EPROM (Electrically Programmable Read Only Memory) results in access times of less than 60 nanoseconds. It was specially developed for Philips Semiconductors by the Philips Research Labs to be closely integrated with Philips' 0.5µ CMOS process, requiring only two more mask steps making it relatively inexpensive to implement. Each memory cell measures only seven square micrometers, giving a high packing density, and can operate between 1.8 and 3.3 V. Memory blocks of up to 4MB in size can be created at a time.

The key advantage to customers of OTP memory is that software can be easily loaded into ICs by the customers themselves. Thus code can be quickly tested and improved by simply loading new versions of the code into the ICs rather than having to get a mask change at the manufacturing stage of the ICs. This results in fast software development, fast time to market and the ability to upgrade software easily. In addition, rather than having to stock variants of an IC each with different ROM codes, customers just stock one master IC and then load the appropriate ROM software according to requirements. The initial uses are for consumer communications applications such as cordless telephones and pagers.

Capacitors can be easily added
Because the OTP memory process uses two layers of polycrystalline silicon (or poly for short) in its creation, it is easy to add capacitors to the ICs as they also use two layers of poly and just one more mask stage. These capacitors are required for EMC reduction and the creation of analog output devices on the IC. The latter are particularly important when the OTP memory is used on microcontrollers as mixed signal interfaces are required between the digital microprocessor and the analog world.

Patented way to avoid expensive high voltage transistors
Normally, OTP memories require expensive, high voltage transistors to be laid down on the periphery of the memory matrix to handle the 13 V used to program the memory cells. Philips realised that this expense could be avoided because normal transistors can operate at these high voltages long enough for programming to be done. In other words, at 13 V, normal transistors can survive for about 1000 seconds but the programming operation takes only a second after which the transistors are never used again.

Patented overcoat enables Salicide to be laid down for fast operation
The conventional design of an OTP memory cell (see diagram 1) places the control gate on top of the floating gate with an insulating layer of IPD (Inter Poly Dielectricum) in between to separate them and prevent charge leakage.


Any attempt to put a layer of Titanium Salicide (TiSi2), or Salicide (Self Aligned Silicide) for short, on the top of the control gate runs into problems with the Salicide going down the sides of the structure and causing shorting between the gates.

In the Philips' patented memory cell shown in diagram 2, the control gate overlaps the floating gate completely and so a Salicide layer can be laid down on top with no possibility of shorting occurring. This Salicide layer provides much improved conductivity to the gate, resulting in extremely fast memory access time of less than 60 nanoseconds that Philips believes to be the fastest currently available.


Programming an OTP memory cell
When made, all the memory cells in the matrix have zero charge in the floating gate. The programming process can change this by the application of 10 V on the control gate, 5 V on the drain and 0 V on the source. This excites hot electrons into the floating gate to create a charge that will remain in there for at least ten years - Philips calculate that it equates to an average charge loss of less then one electron per week.

The cells are read by applying a small voltage on the control gate to induce conduction in the channel between the source and the drain. But, if the floating gate has a charge, this charge prevents the control gate from inducting conduction in the channel between the source and the drain and so no current flows. By convention, an unaltered cell that allows current to pass equates to a "1" while a programmed non-conducting cell equates to "0".

Philips Semiconductors, a division of Philips Electronics NV, headquartered in Eindhoven, The Netherlands, is the ninth largest semiconductor supplier in the world. Philips Semiconductors' innovations in digital audio, video, and mobile technology position the company as a leader in the consumer, multimedia and wireless communications markets. Sales offices are located in all major markets around the world and are supported by regional customer application labs.

Copyright © 2000
Royal Philips Electronics
All rights reserved.
Terms and conditions.


Go to Philips Semiconductors' home page Philips home Feedback your comments and questions Go to semiconductors.com search / index