A microcontroller unit (MCU) is a small computer on a single integrated circuit that typically contains a central processing unit (CPU) core, static random access memory (SRAM) modules, embedded flash memory modules, a system integration module and peripheral modules including a timer, an analog-to-digital converter (ADC), serial communication and networking.

In contrast to microprocessors, which are used in personal computers and other high-performance applications, microcontrollers are designed for small applications. You can find them used in automotives, consumer electronics, computers and peripherals, wired and wireless communications, smart cards and other simple applications using repeatable programs. According to market research firm iSuppli, in 2008 the total revenue for MCUs was close to $16 billion (USD), with the top three manufacturers (Renesas Technology, Freescale Semiconductor and NEC Electronics) each having more than 10 percent of the MCU market share.

Microcontrollers with embedded flash memories (flash-MCUs) are widely used in real-time control application markets. The programmable code storage provided by on-chip flash memories contributes to the reduction of production costs and expansion of real-time adaptive control applications, realizing a value innovation with remarkable cost/value advantage. In 2007, flash-MCUs accounted for more than 30 percent of the MCU market and are predicted to make up more than 50 percent of the MCU market by 2010. Almost all the MCU market segments now use embedded flash solutions. Automotive is the single largest user of flash-MCU as it accounts for more than 50 percent of flash-MCU applications in 8-, 16- and 32-bit products (Kevin Zhang, Embedded Memories for Nano-Scale VLSIs, Springer, 2009).

Embedded flash memory technologies deviate significantly from dominant discrete flash memory technologies due to their specific requirements such as host-logic CMOS process compatibility, performance, cost and reliability. Minimizing costs is a major technology requirement for embedded flash memory. This is achieved by fewer process steps and voltage reduction for less area penalty to the base CMOS logic, rather than mere cell-size scalability. The extra masking steps should be as small as possible and the area penalty reduction in small capacity of memory frequently encountered in embedded applications is focused on smaller periphery circuitry. These requirements for embedded applications often do not allow the direct import of technologies from discrete high-density flash memories, like NOR and NAND. Embedded uses often require faster access time to match the on-chip processing speeds, which further deviates from the discrete flash memory in technology and design.