memory Related Articles

Ⅰ IntroductionWith the improvement of computer technology, the demand for non-volatile memory is increasing, their read and write speed requirements are getting faster and faster, and the power consumption  are becoming smaller and smaller as required by users. But the traditional non-volatile memory such as EEPROM , FLASH, etc. have been difficult to meet these needs.Traditional mainstream semiconductor memories can be divided into two categories: volatile and nonvolatile. Volatile memory includes static random access memory (SRAM) and dynamic random access memory (DRAM). Both SRAM and DRAM lose their saved data when power off. Although RAM is easy to use and performs well, a big disadvantage of it is data loss.Non-volatile memory does not lose stored data in the case of a power failure, because all mainstream non-volatile memories are derived from read-only memory (ROM) technology. ROM, what is called a read-only memory is definitely not easy to write, in fact, it cannot be written at all. All memories developed by ROM technology are difficult to write data, including EPROM, EEPROM and Flash. And these memories not only have a slow writing speed, but also can only be erased and written in a limited number of times.Based on improving semiconductor technologies, ferroelectric memory, a new type of memories, has some unique characteristics. Ferroelectric memory is compatible with all the functions of RAM, and it is a non-volatile memory like a ROM. In other words, ferroelectric memory bridges the gap between these two types of storage, a type of non-volatile RAM. Compared with traditional non-volatile memory, it has attracted much attention due to its advantages such as low power consumption, fast read and write speed, and strong anti-irradiation capability. CatalogⅠ IntroductionⅡ TerminologyⅢ Working PrincipleⅣ FRAM Material FeaturesⅤ Circuit StructureⅥ Reading and Writing ProcessⅦ FRAM StructureⅧ Comparison of FRAM with Other Storage TechnologiesⅨ FRAM UsageⅩ SummaryⅪ One Question Related to FRAM and Going Further11.1 Question11.2 AnswerⅡ TerminologyFerroelectric Memory (FeRAM)Ferroelectric memory (FRAM), also known as F-RAM or FeRAM, is a type of random access memory with fast read and write speed, and the ability to retain data after power is turned off (such as read-only memory and flash memory) is combined, which is the most commonly used type of personal computer memory. Since it is not as dense as dynamic random access memory (DRAM) and static random access memory (SRAM), that is, it cannot store as much data as they do in the same space. In other words, it cannot replace DRAM and SRAM technologies. However, because it can store data quickly with very low power conditions, it is widely used in consumer’s small devices, such as personal digital assistants (PDA), mobile phones, power meters, smart cards, and security systems. FRAM’s read and write speed is faster than flash memory. In some applications, it may also replace electrically erasable read-only memory (EEPROM) and static random access memory (SRAM), and will become a key component of future wireless products. Ⅲ Working PrincipleFeRAM or ferroelectric RAM seems to indicate that an iron element exists within the memory this is not actually the case. A ferroelectric is a material containing a crystal that can spontaneously polarize. It has two states that can be reversed by an external electric field. When an electric field is applied to the ferroelectric crystal, the central atom moves in the crystal following the electric field direction. When an atom moving, it passes through an energy barrier, causing charge breakdown. Internal circuits react to the charge breakdown and set the memory. After the electric field is removed, the central atom remains polarization state, which makes the materials non-volatile, so the state of the memory is preserved. Because there is no atomic collision in the entire physical process, the ferroelectric memory has the characteristics of high read and write speed, ultra-low power consumption, and unlimited writes, making it very suitable to act as temporary storage memory in important systems to transfer various data between subsystems, for each subsystem to read and write frequently.Therefore, with an external electric field, the polarization characteristics of ferroelectric materials will change. When this electric field is removed, the data can still be saved. Without an external electric field, there are two stable states of polarization characteristics. Figure 1 is a hysteresis loop of a ferroelectric material capacitor, showing the different polarities of the ferroelectric capacitor under different applied electric fields. Among them, the two most important parameters are the degree of residual polarization Pr, and the coercive field Ec. In the absence of electric field effect, +/- Pr represents two states of “0” and “1”. To obtain these two states, the applied electric field must be greater than +/- Ec, at this time, the required threshold voltage is also determined.Figure 1. Ferroelectric Hysteresis LoopThe industry explores the use of ferroelectric materials for DRAM: using them as dielectric materials in DRAM capacitors. That is, ferroelectrics are used to replace high-K dielectric materials in standard logic devices, and finally non-volatile transistors are formed, which are FeFETs. The two stable polarization states of the ferroelectric gate oxide change the threshold voltage of the transistor, even when the supply voltage is removed. Therefore, the binary state is encoded in the threshold voltage of the transistor. The writing operation of the memory cell can be completed by applying a pulse on the gate of the transistor, which will change the polarization state of the ferroelectric material and affect the threshold voltage. For example, applying a positive pulse will reduce the threshold voltage, making the transistor in the “on” state. Reading is done by measuring the drain current. This memory mode is similar to the operating mode of a NAND flash: electrons are injected and drawn out of the floating gate, which adjusting the threshold voltage of the transistor.In contrast, the leakage current factor of ferroelectric capacitors is not as important as traditional non-volatile memories such as EEPROM and FLASH, because the information storage of FeRAM is realized by polarization, not free electrons. Ⅳ FRAM Material FeaturesIdeal ferroelectric materials need to meet the following characteristics:Small dielectric constantReasonable self-polarization degree (~ 5μC/ cm2)High Curie temperature (outside the storage and operating temperature range of the device)The thickness of ferroelectric materials should be thin (submicron) to make the coercive field EC smaller. Ferroelectric materials should stand a certain breakdown filed strength.Internal switching speed should be fast (nanosecond level)The ability to keep the data and the long-lasting ability will be good.If used by the military, it is also required to be able to resist radiation exposure. Good chemical stabilityGood processing uniformityEasy to integrate into CMOS processNo bad effect on the surrounding circuitsSmall pollution After years of research and development, there are currently two main types of mainstream ferroelectric materials: PZT and SBT.PZT is lead zirconate titanate PbZrxTil-xO3; SBT is strontium bismuth tantalate Sr1-yBi2 + xTa2O9. The structure of these two materials is shown in Figure 2. Figure 2. Schematic Diagram of PZT and SBT Material StructurePZT is the most studied and widely used. Its advantage is that it can be made at lower temperatures by sputtering and MOCVD. It has the advantages of large residual polarization, cheap raw materials, and low crystallization temperature.; its disadvantages are fatigue degradation problems, and lead pollution to the environment. Moreover, the film deposition process of these materials has proved to be very challenging. At the same time, the extremely high dielectric constant (about 300) of these materials is a big obstacle to their integration into transistors.In addition, scientists have discovered the presence of a ferroelectric phase in a less complex material, hafnium oxide (HfO2), which raise a new concept of storage concept. The researchers found that the ferroelectric phase) can be stabilized by doping silicon (Si) into HfO2. Compared with PZT, HfO2 has a lower dielectric constant and can deposit thin films in a conformal manner (ie, the atomic layer deposition (ALD) process). Most importantly, scientists are familiar with HfO2, because it is the HK gate oxide material in the logic device HKMG. By modifying this CMOS-compatible material, logic transistors can become non-volatile FeFET memory transistors.Functional verification of FeFETs has been implemented in a two-dimensional planar architecture. At the same time, the HfO2 conformal deposition process makes 3D stacking possible, for example, depositing ferroelectric materials on vertical “walls’ to stack transistors in a vertical direction.In terms of materials, 3D FeFETs can solve some of the challenges brought by 2D FeFET structures. One challenge is related to the polycrystalline nature of the HfO2. Scaling the thickness of the HfO2 film will significantly reduce the number of grains in this layer. Because not all the crystal grains have the same polarization direction, the reduction of crystal grains will affect the consistency of the transistor’s response to the external electric field, and eventually lead to large differences between the tubes. By 3D stacking, this drawback is overcome in physical filed. That is, HfO2 does not need to be compressed too thinly, thereby reducing tube-to-tube variation.These vertical FeFETs are expected to have more advantages than complex 3D NAND flash memory, including simple process, lower power consumption and faster speed. Compared to 3D NAND flash memory, vertical FeFET can be programmed at a lower voltage, which improves memory reliability and scalability.The biggest advantage of SBT is that it does not have the problem of fatigue degradation, and it does not contain lead, which meets EU environmental standards; however, its disadvantages are that the process temperature is higher, which makes the process integration difficult, and the degree of residual polarization is small. The comparison of the two materials is shown in Table 1.Table 1. Comparison between PZT and SBT  PZTSBTStructureABO3Layered structureDeposition technologySol-gel,MOCVDSol-gel,MOCVDProcess temperature450℃~700℃750℃~850℃Residual polarity3012Fatigue10101010Data hold85℃@10a- At present, from the perspective of environmental protection, PZT has been banned, but from the perspective of performance and process integration of ferroelectric memory and cost, SBT has no advantages compared to PZT. Therefore, the selection of ferroelectric materials is worth discussing. Ⅴ Circuit StructureThe circuit structure of the ferroelectric memory is mainly divided into the following three types: 2 transistors-2 capacitors (2T2C), 1 transistor-2 capacitors (1T2C), 1 transistor-1 capacitor (1T1C), as shown in Figure 3. The 2T2C structure has two opposite capacitors for each bit as a reference to each other, so the reliability is better, but occupies too much space, which is not suitable for high-density applications. The transistor / single capacitor structure can be used like a DRAM to provide a reference for each column of the memory array, compared with the existing 2T2C structure, they effectively reduce the required space of the memory cell by half. This design greatly improves the efficiency of ferroelectric memory and reduces the production cost of ferroelectric memory products. The 1T1C structure has a higher integration density (8F2), but its reliability is poor. And the 1T2C structure is a compromise between these two structures. Figure 3. Three FRAM StructuresAt present, in order to obtain a high-density memory, 1T1C structure is mostly used (as shown in Figure 4). In addition, a chain structure is also adopted, thus Chain FeRAM is made. This structure is similar to the NAND structure. Through this method, a higher storage density than 1T1C can be obtained, but this method will also greatly increase the access time. Chain FeRAM (CFeRAM) structure is shown in Figure 5. Figure 4. 1T1C Layout Figure 5. Chain FeRAM (CFeRAM) Circuit StructureⅥ Reading and Writing ProcessAccording to the polarity of the electronic memory cell, a small charge amount is “0” and a large charge amount is “1”. This charge is converted into a reading voltage, which is “0” when it is less than the reference voltage and when it is greater than the reference voltage represents “1”. The stored information is read out as shown in Figure 6. Figure 6. Reading and Writing Process of FRAMDuring the reading process, the word line voltage is increased to turn on the MOS transistor, and then the drive line voltage is increased as VCC, so that different charges of the storage capacitor are distributed to the bit line parasitic capacitance, so different voltages appear on the BL to identify the data. During a writing process, the word line is raised to turn on the MOS transistor, and a pulse is applied to the drive line, so that different data on the bit line are stored in two different steady states of the ferroelectric capacitor.By adding a positive voltage or a negative voltage, these two voltages can make the capacitor into two different polarities. In this way, the information is written into the memory. Ⅶ FRAM StructureAt present, the most common device structures of ferroelectric memories are planar and stack structures. The difference between the two is the location of the dry ferroelectric capacitor and the way in which the capacitor is connected to the MOS tube. In the planar structure, the capacitor is placed above the field oxide, and the electrode of the capacitor is connected to the active area of the MOS tube through metal aluminum. The process is relatively simple, but the unit spacing is large. In the stack structure, the capacitor is placed in the source region, the lower electrode of the capacitor is connected to the source terminal of the MOS tube through a plug based on CMP process, which has a high integration density. In addition, the stack structure can adopt the method of making ferroelectric capacitors on metal wires, thereby reducing the mutual influence during the formation process. The following schematic diagrams of the two structures are shown in Figure 7 and Figure 8. Figure 7. Planar Structure Figure 8. Stack StructureThe process of the planar structure is relatively simple. The isolation uses the LOCOS structure, and the planarization does not require the CMP. The stacked structure has a high degree of integration based on advanced technique, and STI is used for isolation, in addition, CMP is required for planarization, and copper wires can be used.In addition, there is a structure that uses a ferroelectric material as the gate. Such a device can eliminate the destructive problem of data readout, and theoretically it is more space-saving and can make more greater integration. However, there are still serious problems with this structure, that is, the data storage capacity is very poor, only one month or less, so it is far from practical. Figure 9 is a schematic diagram of such a structure. Figure 9. FeFET Structure DiagramAt present, the ferroelectric memory generally adopts a planar structure with the line width more than 0.5 μm, and generally uses a stack structure when the line width is less than 0.5 μm. Ⅷ Comparison of FRAM with Other Storage TechnologiesAt present, Ramtron’s FRAM mainly includes two categories: serial FRAM and parallel FRAM. Among them, serial FRAM is divided into I2C two-line FM24×× series and SPI three-line FM25xx series. Serial FRAM is compatible with the traditional 24xx and 25xx E2PROM pins and timing, which can be directly replaced.FRAM products have the advantages of RAM and ROM, and fast read and write speed, in addition, they can be used as non-volatile memory. Due to the shortcoming of ferroelectric crystals, the number of accesses is limited, beyond which FRAM is no longer non-volatile. The maximum access times given is 10 billion, but it not means FRAM will be scrapped when over this upper limit. In the terms of it, FRAM is not non-volatile, but it can still be used as an ordinary RAM.FRAM vs E2PROMFRAM can be used as a second option for E2PROM. Except the performance of E2PROM, the FRAM access speed is much faster. When using FRAM, it must be determined that once there are 10 billion accesses is down to FRAM in the system, there is no damage.FRAM vs SRAMIn terms of speed, price, and convenience, SRAM is better than FRAM; but from the perspective of the entire design, FRAM has certain advantages. Non-volatile FRAM can hold startup programs and configuration information. If the maximum access speed of all the memories in the application is 70ns, one piece of FRAM can be used to complete the system, making the system structure more simpler.FRAM vs DRAMDRAM is suitable for applications where density and price are more important than access speed. For example, DRAM is the best choice for graphics display memory. There are a large number of pixels to be stored, and the recovery time is not very important. If you don’t need to save the last content at the next boot, use volatile DRAM memory. The role and cost of DRAM are reasonable compared with FRAM. In short, it turns out that DRAM cannot be replaced by FRAM totally.FRAM vs FlashAt present, the most commonly used program memory is Flash, which is more convenient and cheaper to use. The program memory must be non-volatile, and easier to rewrite, but the use of FRAM is limited by access times.Ⅸ FRAM UsageData collection and recordingFeRAM allows designers to write data faster and more frequently, and at a lower price than EEPROM.Typical applications: meters (electric meters, gas meters, water meters, flow meters), RF/ID instruments, car black boxes, air bags, GPS, power grid monitoring systems, and so on. Parameter setting and storageFeRAM helps designers solve the problem of data loss due to sudden power failure by storing data in real time. Parameter storage in the FeRAM is used to track the changes of the system in the past time. Its purpose includes restoring the system state or confirming a system error when the power is on.Typical applications: photocopiers, printers, industrial controls, set-top boxes, network equipment  and large household appliances. Non-volatile bufferFeRAM can quickly store data before it is stored in other memory, so that the data in the buffer will not be lost when having power failure.Typical applications: industrial systems, ATM teller machines, tax control machines, commercial settlement systems (POS), fax machines, non-volatile cache memory in hard disk, etc. Ⅹ SummaryFerroelectric memory is an emerging non-volatile memory. It started early and realized industrialization. Because of its advantages such as low power consumption, fast read and write speed, and strong anti-irradiation capabilities, there is a market for small-scale storage areas with low power consumption and radiation resistance. Having the characteristic of anti-radiation, in the case of electromagnetic waves or radiation, the data is still safe, so it has important applications in space science, medicine and other specific fields. However, the ferroelectric memory also has the disadvantages that it is difficult to improve the integration, the process is more contaminated, and it is difficult to be compatible with the CMOS technique. So that it needs further research and solution. Ⅺ One Question Related to FRAM and Going Further11.1 QuestionWhat is FRAM used for?11.2 AnswerFerroelectric RAM is a random-access memory similar in construction to DRAM but using a ferroelectric layer instead of a dielectric layer to achieve non-volatility. It is one of a growing number of alternative non-volatile random-access memory technologies that offer the same functionality as flash memory. FRAM can be used in many fields, for example, with ultra-low power consumption, it is very suitable for intelligent water meters, gas meters and so on. Frequently Asked Questions about Ferroelectric RAM1. What is FRAM memory?Ferroelectric RAM (FeRAM, F-RAM or FRAM) is a random-access memory similar in construction to DRAM but using a ferroelectric layer instead of a dielectric layer to achieve non-volatility. 2. What is ferroelectric effect?Ferroelectricity is a characteristic of certain materials that have a spontaneous electric polarization that can be reversed by the application of an external electric field. ... Thus, the prefix ferro, meaning iron, was used to describe the property despite the fact that most ferroelectric materials do not contain iron. 3. How does FRAM work?FRAM is a nonvolatile storage memory that retains its data even after the power is turned off. However, similar to commonly used DRAM (Dynamic Random Access Memory) found in personal computers, workstations, and non-handheld game-consoles, FRAM requires a memory restore after each read. 4. What are the unique characteristics of FRAM?FRAM has the characteristics of both ROM (Read Only Memory) and RAM (Random Access Memory), and features faster write, great read/write cycle endurance, and low power consumption. 5. Which enables the read and write operation in Feram?Write Operation in Ferroelectric Random Access Memory (FRAM)Similar to read operation, a pre-charge operation follows a write access. The circuit applies 'write' data to the Ferroelectric capacitors. If necessary, the new data simply switches the state of the ferroelectric crystals.

ROM and RAM belong to the semiconductor memory. ROM is the abbreviation of read only memory, and RAM is the abbreviation of random access memory. ROM can keep data when the system powers off, and RAM is in the opposite, a typical RAM is the computer memory. But sometimes we are not clear: what is the difference between Flash Memory and ROM and RAM? How much do you really know about memory, or RAM, ROM, and flash memory specifically? Here we will tell them in different aspects, this article will act as a guide to basic memories.   Catalog     I What are types of RAM?  1.1 Static RAM (SRAM)  1.2 Dynamic RAM (DRAM) II What are types of ROM? III What is Flash Memory?  3.1 Performance Comparison  3.2 Interface Difference  3.3 Capacity and Cost  3.4 Reliability and Durability  3.5 Simple Operation  3.6 Software Support IV Conclusion FAQ     I What are types of RAM? RAM Upgrade Guide - What You Need to Know 1.1 Static RAM (SRAM) SRAM is very fast and the fastest storage device to read and write currently, but it is also very expensive, so it is only used in demanding places, such as the first and second level buffer of CPU. 1.2 Dynamic RAM (DRAM) DRAM keeps data for a short time and its operation is slower than the SRAM, but it's still faster than the ROM. In addition, it's much cheaper than SRAM. The computer memory is DRAM. DRAM is divided into many kinds, the most common ones are FPRAM / FastPage, EDORAM, SDRAM, DDR, RAM, RDRAM, SGRAM, WRAM and so on. DDR RAM, one of which is introduced here. DDR RAM (Double-Date-Rate RAM), also known as DDR SDRAM, the improved RAM and SDRAM, are essentially the same, except that they can read and write data twice in one clock, doubling the speed of data transmission. This is the most used memory in a computer today, and it has a cost advantage, which in fact defeats another Intel memory~ Rambus DRAM. Many high-end graphics cards are also equipped with high-speed DDR RAM to increase bandwidth, which can greatly improve the pixel rendering capability of 3D accelerator cards.     II What are types of ROM? PROM (programmable ROM), is programmed once that can not be modified, this is an early product, now it is gone. EPROM (erasable and programmable ROM), is a kind of universal memory which erases the original program by ultraviolet radiation. EEPROM, is electronically erased, very expensive, long writing time. For example, mobile phone software is generally placed in the EEPROM, when we call, some of the last numbers that are dialed are temporarily in the SRAM, but not immediately written in the call log (the call record is stored in the EEPROM).   Memory is used to storage data and programs that are currently in use (that is in execution), and the memory of the computer we normally used refers to dynamic memory (that is DRAM). The so-called "dynamic" in dynamic memory refers to when we write data to the DRAM, after a period of time, the data will be lost, so we need an extra circuit to refresh the memory. It works like this: whether a DRAM's memory cell stores 0 or 1 depends on whether the capacitor has a charge. 1 represents a charge, 0 represents no charge. But over time, the capacitance of 1 will discharge, and the capacitance of 0 will absorb the charge, which is why the data is lost. Taking refresh operation regularly checks the capacitor, if the charge is greater than 1/2 of the full charge, it is considered to represent 1, charging the capacitor fully; If the quantity of electricity is less than 1/2, it is considered to represent 0, and the capacitor is discharged to maintain the continuity of the data. ROM III What is Flash Memory? Flash memory combines the advantages of ROM and RAM. It not only has the electrically erasable and programmable capability (EEPROM), but also can read data quickly without power loss (the advantage of NVRAM). This memory is used in USB  and MP3. For the past 20 years, embedded systems have been using ROM (EPROM) as their storage device. However, in recent years, FLASH has completely replaced the position of ROM (EPROM) in embedded systems to store Bootloader and operating systems, or Program code or directly as the hard disk (U disk). At present, there are two kinds of NOR Flash and NAND Flash. NOR Flash reading is the same as the common SDRAM reading. Users can run the code loaded in NOR FLASH directly, which reduces the capacity of the SRAM and saves cost. NAND Flash doesn’t have memory random read technique, which reads in the form of a block at a time, usually 512 bytes at a time, thus it is cheaper. Users cannot run program code on NAND directly, so many developers using NAND have to use a small NOR Flash to startup and run code. NOR Flash, is generally used in small capacity because of its fast reading speed, and it is used to store important information such as operating system. The most common NAND FLASH used for large capacity, combines with embedded system DOC (Disk On Chip) and common "flash disk", which can be erased online. At present, the FLASH on the market mainly comes from Intel, AMD, Fujitsu and Toshiba, and the main manufacturers of NAND Flash are Samsung and Toshiba. NOR and NAND are the two main non-volatile flash memory technologies in the market. Memory IC Intel first developed NOR flash technology in 1988, which completely changed the situation in which EPROM and EEPROM dominated the world. Then, in 1989, Toshiba published its NAND flash architecture, emphasizing lower cost per bit, higher performance, and easy upgrades via interfaces like disks. But after more than a decade, a considerable number of hardware engineers still can not distinguish between NOR and NAND flash memory. Flash memory is often used interchangeably with phase NOR memory. Many in the industry are also confused about the advantages of NAND flash over NOR, because in most cases flash memory is only used to store a small amount of code, and NOR flash memory may be more appropriate, while NAND is an ideal solution for data needed high storage density. NOR is the main non-volatile flash memory technology in the market. NOR is generally used to store a small amount of code, especially in code storage media. NOR is characterized by simple application, no special interface circuit, high transmission efficiency, it belongs to the in-chip execution(XIP, eXecute In Place), so that applications can run directly in (NOR) flash memory. No longer need to read the code into the system RAM. Low write and erase speed can be very cost-effective but affect its performance for small capacity when it is 1~4MB. NOR flash comes with an SRAM interface and has enough address pins to access, it is convenient to storage and use each byte. NOR flash accounts for the majority of the 1~16MB flash market. NAND structure can provide extremely high cell density and high storage density, and fast write and erase speed. The difficulty of applying NAND lies in the management of flash and the need of special system interface. 3.1 Performance Comparison Flash memory is non-volatile memory and can be erased and reprogrammed on memory cell block. The write operation of any flash device can only be done in an empty or erased cell, so in most cases, the erasure must be performed before the writing. In general, the NAND devices perform the erasure operation easily. But NOR requires that all bits in the target block be written as 1 before erasing. Since the NOR device is erased as a block of 64~128KB, the time to perform a write / erase operation is 5s. In contrast, the erasure NAND device is performed as a block of 8~32KB, and the same operation only needs 4 ms at most. The performance gap between NOR and NADN is further widened by the difference of block size in the execution of erasure. It is shown that for a given set of write operations (especially when updating small files), more erasure operations must be performed in NOR based cells. Therefore, when choosing a storage solution, the designer must weigh the following factors: The read speed of NOR is a little faster than NAND, but the writing speed is on the opposite. The 4ms erasure speed of NAND is much faster than that of NOR. Most writing operation need running erasure. The NAND erasure unit is smaller and the corresponding erasure circuit is less. Flash memory 3.2 Interface Difference NOR flash has a SRAM interface, and enough address pins to address, and easy storage and program to every byte inside. NAND devices use complex I/O ports to serially access data, which may vary from product to product or from manufacturer to manufacturer. Eight pins are used to transmit and control address and data information. NAND read and write operations use 512-byte blocks, which is similar to the hard disk management. Naturally, NAND based memory can replace hard drives or other block devices. 3.3 Capacity and Cost The cell size of NAND flash is almost half that of NOR devices. Because of the simpler manufacturing process, the NAND structure can provide higher capacity within a given die size, thus lowering the price accordingly. NOR flash accounts for most of the 1~16MB flash market, while NAND flash is only used in 8~128MB products, which means that NOR is mainly used in code storage, and NAND is suitable for data storage, mainly used in CompactFlash, Secure Digital, PC Cards and MMC memory cards. 3.4 Reliability and Durability One of the key issues needed to consider is reliability when using flahs media. For systems that need to extend MTBF, Flash is a very suitable storage solution. So the reliability of NOR and NAND can be compared between bit switching, lifetime (durability) and bad block processing. - Service Life The maximum number of erasures per block in NAND flash memory is one million times, while that of NOR is 100,000 times. In addition, NAND memory has the advantage of 10: 1 erasure speed, 8 times smaller than that of the NOR device, and each NAND memory block has fewer deletions at a given time. - Bit Switching All flash devices are plagued by bit switching. In some cases (rarely, NAND occurs more often than NOR), a bit reverses or its reversion is reported. One changing bit may not be obvious, but if it happens on a critical file, this minor failure could cause the system to shut down. If existing wrong report, it may be solved by reading it a few more times. Of course, if the bit really changes, the error detection / error correction (EDC/ECC) algorithm must be used. The problem of bit inversion is more common in NAND flash memory, therefore, it is  recommended that using NAND flash memory and EDC/ECC algorithm at the same time. But this problem is not fatal when storing multimedia information in NAND. Of course, if you use a local storage device to store an operating system, configuration file, or other sensitive information, you must use the EDC/ECC system to ensure reliability. - Bad Block Processing The bad blocks in NAND devices are randomly distributed. There have been previous efforts to eliminate bad blocks, but they found that the yield is too low and the cost is too high. The NAND device needs to initialize the medium to detect the bad block and mark it. In the fabricated devices, the failure rate will be high if this treatment cannot be carried out by a reliable method. 3.5 Simple Operation NOR-based flash memory can be used very directly, can be connected as other memory, and can run code directly on it. NAND is much more complex because of the need for an I / O interface. Access methods for various NAND devices vary from manufacturer to manufacturer. When using NAND devices, you must write to the driver before doing anything else. Writing information to NAND devices requires that designers don’t write to bad blocks, which means virtual mapping is necessary all the time on NAND devices. 3.6 Software Support When discussing software support, a distinction should be made between basic read / write / erasing operations and higher-level software for disk emulation and algorithms of flash management, also including the performance optimization. Running code on a NOR device doesn't require any software support. But using a NAND device, you usually need a driver, that is MTD(Memory Technology Driver). Both NAND and NOR devices need MTD when writing and erasing operations. NOR devices requires less MTD, because many vendors offer more advanced software for NOR devices, including M-System 's TrueFFS driver, which is used by manufacturers such as Wind River System, Microsoft, QNX Software System, Symbian and Intel. It also used to simulate DiskOnChip products and manage the NAND flash memory, including error correction, bad block handling, and loss balancing. The main manufacturers of NOR FLASH is INTEL and MICRO, used to be the mainstream of FLASH products, but its market share now squeezed by NAND FLASH. Its advantage is that it can run programs directly from FLASH, but the process is complex, thus it is expensive. NAND FLASH's main suppliers are SAMSUNG and Toshiba, and it widely used in USB drives, memory cards, and MP3 players. With different processes, it has larger storage capacity and is cheaper than the NOR FLASH. But also has the drawback, is unable to address runs the program directly, only used to store the data. In addition, NAND FLASH is very prone to bad blocks, so it is necessary to detect it by the related algorithm.  NAND FLASH is used to store data and programs in laptops, but it must be started with NOR FLASH. In addition to SAMSUNG processors, other mainstream processors for laptops do not support NAND FLASH startup directly. Therefore, you must start the machine with a small piece of NOR FLASH, and run OS and other software through the NAND FLASH load into SDRAM. IV Conclusion The DRAM uses the charge on the gate capacitor of the MOS transistor to store the information. Once the power is down, all the information will be lost. Because the grid will leak, it needs refresh operation to replenish the charge on these gate capacitors regularly. And every time the data is read out, it requires running the same work. This is called dynamic refresh, so it is called dynamic RAM. Because it uses only one MOS to store information, it can be highly integrated and can do a lot of work. SRAM uses registers to store information, so once the power is down, the data will be lost. But when the power supply, its data will always exist without the need for dynamic refresh, thus it is called static RAM. The above is mainly used in the system with large capacity, do not need to recover data through refresh operation. Flash ROM uses the capacitor on the floating gate to store the charge to store the information, because the floating gate does not leak, so the information can still be saved after the power is off. Also because of its simple mechanism, it can be integrated highly, and the capacity can be very large. Flash rom needs to be erased by electricity before writing, but in different approach of erasure, EEPROM can be erased in bytes, and flash rom can only be done in sector. However, it can be written in byte units, mainly used for bios, U disk and Mp3 devices that require large capacity and data storage when power is down.   FAQ   1.What is ROM and its function? Read-only memory (ROM) is a type of storage medium that permanently stores data on personal computers (PCs) and other electronic devices. It contains the programming needed to start a PC, which is essential for boot-up; it performs major input/output tasks and holds programs or software instructions.   2. What are the four types of ROM? MROM (Masked ROM) ... PROM (Programmable Read Only Memory) ... EPROM (Erasable and Programmable Read Only Memory) ... EEPROM (Electrically Erasable and Programmable Read Only Memory) ...   3. What is a RAM in a computer? RAM stands for random-access memory, but what does that mean? Your computer RAM is essentially short term memory where data is stored as the processor needs it. This isn't to be confused with long-term data that's stored on your hard drive, which stays there even when your computer is turned off.   4. What is flash memory used for? Flash memory is a long-life and non-volatile storage chip that is widely used in embedded systems. It can keep stored data and information even when the power is off. It can be electrically erased and reprogrammed. Flash memory was developed from EEPROM (electronically erasable programmable read-only memory).   5. Is Flash memory a RAM or ROM? RAM is Read Only Memory. Unlike RAM, ROM is the persistent storage. ... Flash Memory is one category of ROM i.e Electrically Erasable Read Only Memory (EEPROM).   6. What is flash memory and its types? Flash memory is a type of erasable read-only memory (EEPROM) that clears and rewrites data in chunks for fast, energy-efficient access and rewriting. Flash memory, or flash storage, is non-volatile, which means it remains viable even without an active power source.   7. What is RAM and ROM? RAM, which stands for random access memory, and ROM, which stands for read-only memory, are both present in your computer. RAM is volatile memory that temporarily stores the files you are working on. ROM is non-volatile memory that permanently stores instructions for your computer.   8. Which is better RAM or ROM? RAM is random access memory and cannot hold the data without the power, whereas ROM is a read-only memory and can hold the data even without the power. ... With RAM, writing data is a much faster and lightening process, whereas ROM, writing data speed is much slower as compared to RAM.   9. What are the disadvantages of ROM? The disadvantages of Erasable Programmable ROM (EPROM) are: The static power consumption is high as the transistors used have higher resistance. It is not possible for a particular byte to be erased, instead the entire content is erased. UV based EPROM takes time to erase the content.   10.What are the disadvantages of RAM? If CPU wants to read data only from the RAM, then the data access from the cache and the registers is slow in comparison to ROM. RAM is volatile, which means it is difficult to store data for a lengthy period of time. Unplanned circumstances like a power outage can result in data loss.   11. What is the work of ROM? Read only memory (ROM) provides permanent storage for instructions needed during bootstrapping, or the process of turning on the computer. It does so by storing the BIOS and other firmware for the computer hardware. This firmware is very hardware-specific and rarely needs updating.   12. Why is more RAM better? Generally, the faster the RAM, the faster the processing speed. With faster RAM, you increase the speed at which memory transfers information to other components. Meaning, your fast processor now has an equally fast way of talking to the other components, making your computer much more efficient.   13. Is RAM is a volatile memory? Volatile memory is computer memory that requires power to maintain the stored information. Most modern semiconductor volatile memory is either Static RAM (see SRAM) or dynamic RAM (see DRAM).   14. Why is ROM so important? ROM provides the necessary instructions for communication between various hardware components. As mentioned before, it is essential for the storage and operation of the BIOS, but it can also be used for basic data management, to hold software for basic processes of utilities and to read and write to peripheral devices.   15. What is stored in ROM? ROM is memory that cannot be changed by a program or user. ROM retains its memory even after the computer is turned off. For example, ROM stores the instructions for the computer to start up when it is turned on again.  

Summary As the emergence of a range of electronic technologies appear,major changes in the design of real-time embedded systems like the internet of things,augmented reality,or artificial intelligence occurred. The unifying thread between all of them is a greater focus on the use of distributed systems coupled with a need for high performance to deal with the data they generate and consume. Different design direction There are tensions that pull the engineering of real-time devices employing such technologies in different directions. Edge devices such as IoT sensor nodes and gateways call for the lowest-power operation.However,It's not the only area that needs energy efficiency. Despite their reliance on high-performance graphics and responsiveness to movement, AR-enabled systems (such as head-up displays for machine operators) also have to preserve as much energy as possible, protecting battery life and preventing head-mounted displays from becoming uncomfortably warm. Similarly, versatile robots enabled by AI need to be able to operate away from mains power. Distributed processing allows intensive computational work to be moved to the cloud and so offload the embedded systems. However, the real-time nature of these applications calls for low latency. Applications such as motion control and AR suffer if the delay from input to response is too long. This issue is leading to the deployment of edge computing server or ‘cloudlets’ - efficient server blades located relatively close to the edge devices themselves. To support real-time applications such cloudlets are in a position to take advantage of changes in memory technology to better fit the real-time nature of the clients they serve than traditional server designs. Historically, engineers have been forced to choose between performance and persistence when designing bulk memories into real-time computer systems. DRAM is cost-effective for storing large amounts of data close to the processor but is volatile. To ensure data is not lost through power issues - which are more likely to occur in edge nodes - data often has to be copied to persistent storage, which have often much slower access times.The move from rotating disk drives to flash memory for larger applications has already helped significantly when it comes to read access times. But flash still has its drawbacks when it comes to write performance. The erasing and rewriting of data from/to flash memory takes multiple cycles during which high-voltage pulses are delivered to the target memory cells. That takes both time and energy that system designers do not want to waste.   Next generation memory technologies Next generation memory technologies are now appearing that overcome the write delays and power demands of flash. These technologies include ferroelectric memory, phase-change memory (PCM), magnetic random-access memory (MRAM) and resistive random-access memory (ReRAM). As devices based on these concepts become available, engineers can consider using them in novel memory hierarchies that optimise cost, increase resilience and improve real-time responsiveness. Here we may mention that PCM,which was first put forward as a possible memory material as long ago as the 1970s,It is based on the same group of chalcogenide materials as those used in rewritable optical disks. A useful feature of the chalcogenides is the way they react to heat. High-current pulses will melt the material. If left to cool quickly it turns to a resistive amorphous state. But the amorphous state can be converted to a crystalline form with a much higher conductivity by applying a small amount of heat. Thanks to this change in properties, readout circuitry can interpret the difference in resistivity between cells as representing ones and zeros. Though similar in behaviour to PCM, with the same core approach of switching between high-resistance and low-resistance states, ReRAM uses different materials to chalcogenide. Typically, the movement of ions within the cell under the influence of pulses of current forms conductive filaments. Reset pulses disrupt these filaments, greatly increasing resistance. One potential advantage of ReRAM is that a large number of candidate materials could be chosen to implement them. This provides the scope for manufacturers to introduce memories with different levels of resilience and storage time. Although these memories use current pulses, the total charge required to program a cell is much lower than that required for flash. In the memories being developed today, ReRAM requires less write energy than PCM but the write times are similar. However, endurance is better in PCM than ReRAM and PCM currently lies further ahead on the development path. Experts believe both PCM and ReRAM will scale better than flash in the long term and so could ultimately supplant flash entirely.   about Ferroelectric Memory Ferroelectric memory and MRAM use the spin properties of electrons for storage. The spin can be controlled with very little energy through a spin-valve structure similar to that used in high-density read heads for magnetic disks. In an MRAM, this spin valve is made from a sandwich of materials formed in a via that lies between two metal interconnect lines on the surface of an integrated circuit (IC). The valve alters the resistance of the via based on the spin states of different materials in the sandwich.Ferroelectric memory has been available for several decades but in comparatively low densities to those envisaged for the resistance-based memories. Ferroelectric memory requires both a capacitor and transistor to be formed on the base layer of the wafer. The other memories are all formed in the metal interconnect layers and, potentially, can be stacked for higher integration.What's more,a key advantage  for ferroelectric memory is its use of materials that polarise in two different directions based on an applied electric field. This polarisation requires even less power than is needed for MRAM, which makes it suitable for systems that need to be highly energy efficient.   The potential problem A potential problem for all the novel memories today is that they lack the cost-effectiveness and density of flash, which is now beginning to take advantage of 3D manufacturing techniques. In reality, for cloudlets and also edge devices themselves, the density is not a major issue as these memories can serve as the underpinning for persistent caches. The low-power and relatively fast write times of the novel memories provides applications with the ability to copy important data to the persistent cache. Data objects that need to be stored permanently can, from there, be copied to flash or disk storage. However, there is no longer any need to transfer data to flash or disk storage continually just to ensure that important but transient data is not lost. When the system restarts, it can recover its state from combining data in both the permanent and persistent arrays.As costs come down and performance improves, there is the potential for MRAM, PCM or ReRAM to begin to displace DRAM and so move the architecture to one in which only the caches on the processors themselves employ a volatile memory architecture (such as SRAM).Persistent memory technologies need not be isolated to cloudlets and high-performance systems. The use of ferroelectric memory by Texas Instruments in its MSP430 line of microcontrollers provides an example of the impact it can have in IoT edge nodes such as sensors. Many IoT applications will rely on energy harvesting to at least supplement a built-in battery. Some may dispense with the battery altogether. The problem with energy harvesting is one of reliability. There are situations, such as vibrational energy capture on heavily used industrial machinery, where the power source is predictable. But in many cases, even with the use of a supercapacitor for an energy reservoir, the system may run temporarily short of power and need to shut down. When enough external energy is supplied, it can resume normal duties.The use of ferroelectric technology provides the microcontroller with the ability to ensure data persists through unexpected power outages without incurring an energy penalty even when data is written to it frequently.  

SummaryRecently engineers discover the two-dimensional atomic sheets for memory storage when they were developing alternative ways to provide greater memory storage capacity on thiner computer chips. Most of us are curious about how engineers discover them?DiscoverA team of electrical engineers at The University of Texas at Austin, in collaboration with Peking University scientists, has developed the thinnest memory storage device with dense memory capacity, paving the way for faster, smaller and smarter computer chips for everything from consumer electronics to big data to brain-inspired computing. Discussion"For a long time, the consensus was that it wasn't possible to make memory devices from materials that were only one atomic layer thick," said Deji Akinwande, associate professor in the Cockrell School of Engineering's Department of Electrical and Computer Engineering. "With our new 'atomristors,' we have shown it is indeed possible." Made from 2-D nanomaterials, the "atomristors"—a term Akinwande coined—improve upon memristors, an emerging memory storage technology with lower memory scalability. He and his team published their findings in the January issue of Nano Letters. "Atomristors will allow for the advancement of Moore's Law at the system level by enabling the 3-D integration of nanoscale memory with nanoscale transistors on the same chip for advanced computing systems," Akinwande said.Memory storage and transistors have, to date, always been separate components on a microchip, but atomristors combine both functions on a single, more efficient computer system. By using metallic atomic sheets (graphene) as electrodes and semiconducting atomic sheets (molybdenum sulfide) as the active layer, the entire memory cell is a sandwich about 1.5 nanometers thick, which makes it possible to densely pack atomristors layer by layer in a plane. This is a substantial advantage over conventional flash memory, which occupies far larger space. In addition, the thinness allows for faster and more efficient electric current flow.Given their size, capacity and integration flexibility, atomristors can be packed together to make advanced 3-D chips that are crucial to the successful development of brain-inspired computing. One of the greatest challenges in this burgeoning field of engineering is how to make a memory architecture with 3-D connections akin to those found in the human brain. "The sheer density of memory storage that can be made possible by layering these synthetic atomic sheets onto each other, coupled with integrated transistor design, means we can potentially make computers that learn and remember the same way our brains do," Akinwande said. The research team also discovered another unique application for the technology. In existing ubiquitous devices such as smartphones and tablets, radio frequency switches are used to connect incoming signals from the antenna to one of the many wireless communication bands in order for different parts of a device to communicate and cooperate with one another. This activity can significantly affect a smartphone's battery life. The atomristors are the smallest radio frequency memory switches to be demonstrated with no DC battery consumption, which can ultimately lead to longer battery life. All in all,this discovery has real commercialization value as it won't disrupt existing technologies. Rather, it has been designed to complement and integrate with the silicon chips already in use in modern tech devices.  

SummaryMemory is one of the most important part for electronics. Computers and Smartphones woludn't be nearly as useful without room for lots of apps,music and videos. Devices tend to store that  information in two ways: through electric fields (think of a flash drive) or through magnetic fields (like a computer’s spinning hard disk). Each method has advantages and disadvantages. However, in the future, our electronics could benefit from the best of each. There are some questions put by Chang-Beom Eom, the Theodore H. Geballe Professor and Harvey D. Spangler Distinguished Professor of Materials Science and Engineering at the University of Wisconsin-Madison. “Can you cross-couple these two different ways to store information? Could we use an electric field to change the magnetic properties? Then you can have a low-power, multifunctional device. We call this a ‘magnetoelectric’ device.” In research published recently in the journal Nature Communications, Eom and his collaborators describe not only their unique process for making a high-quality magnetoelectric material, but exactly how and why it works. Physics graduate student Julian Irwin checks equipment in the lab of materials science and engineering Professor Chang-Beom Eom, where researchers have produced a material that could exhibit the best qualities of both solid-state and spinning disk digital storage. Magnetoeletric materialsMagnetoelectric materials,which have both magnetic and electrical functionalities,or "orders" already exist. Switching one functionality induces a change in the other.“It’s called cross-coupling,” says Eom. “Yet, how they cross-couple is not clearly understood.” Gaining that understanding, he says, requires studying how the magnetic properties change when an electric field is applied. Up to now, this has been difficult due to the complicated structure of most magnetoelectric materials. In the past,people studied magnetoelectric properties using very "complex" materials,or those that lack uniformity.In his approach,Eom simplified not only the research but the material itself. Drawing on Eom's expertise in material growth,he developed a unique process,using atomic "steps" to guide the growth of a homogenous,single-crystal thin film of bismuth ferrite. Atop that, he added cobalt, which is magnetic; on the bottom, he placed an electrode made of strontium ruthenate.  Bismuth Ferrite MaterialThe bismuth ferrite material was important because it made it much easier for Eom to study the fundamental magnetoelectric cross-coupling. Eom found that in their work,because of their single domain,they could actually see what was going on using multiple probing, or imaging, techniques.The mechanism is intrinsic. It’s reproducible — and that means you can make a device without any degradation, in a predictable way. To image the changing  electric and magnetic properties switching in real time, Eom and his colleagues used the powerful synchrotron light sources at Argonne National Laboratory outside Chicago, and in Switzerland and the United Kingdom. “When you switch it, the electrical field switches the electric polarization. If it’s ‘downward,’ it switches ‘upward,'” he says. “The coupling to the magnetic layer then changes its properties: a magnetoelectric storage device.” That change in direction enables researchers to take the next steps needed to add programmable integrated circuits — the building blocks that are the foundation of our electronics — to the material. While the homogenous material enabled Eom to answer important scientific questions about how magnetoelectric cross-coupling happens, it also could enable manufacturers to improve their electronics.Eom saied they can design a much more effective,efficient and low-power device now. 

A device that turns light into sound has allowed researchers to capture lightning in a bottle, in a sense, slowing down the light beams enough so that they can be easily stored and manipulated. Researchers at the University of Sydney in Australia, have figured out how to turn a light wave into a sound wave, creating an acoustic memory that they say will help data centers save energy by eliminating some electrical connections between processors. They reported their work in a recent issue of Nature Communications. “Our vision is to replace the electronic interconnects between different processors and computing machines with photonic ‘wires,’’’ said Birgit Stiller, a postdoctoral researcher who led the project. “So light transmission will be used instead of electronic connections.” The team built a chip that consists of a spiral-shaped waveguide made from a soft glass called chalcogenide, sandwiched between two stiffer pieces of silica glass. As a light beam travels through the chip, it is met by another pulse of light that has a slightly different frequency. The difference between the frequencies of the two light beams is a “beat,” a wave with a frequency 100,000 times lower, thus turning the light wave into a sound wave.  The sound wave lives for a brief time—several nanoseconds—in the spiral chalcogenide waveguide. To read it out, the device reverses the process, adding the beat frequency to a light pulse to recreate the original light wave. In standard optical fibers, light waves are prevented from leaking out of the fiber by a difference in refractive index between the core of the fiber and the cladding wrapped around it. In a similar way, the two types of glass keep the sound wave in place; the speed of sound is much slower in the chalcogenide than in the silica. Slowing down the waves provides time to synchronize different signals coming from different processors. That eliminates the need to convert the optical signal to an electronic signal. Electronics can produce excess heat and require more energy, which are important issues in the big data centers owned by Google, Amazon, or Microsoft, Stiller says. Further work with the design and materials might allow the sound waves to be stored longer, although the memory already lasts long enough for the use they envision. She and her team hope to refine the work further, with an eye to building a prototype of a manufacturable chip within the next few years.