Commercial nanotechnology
Introduction
At this forum, there is a remarkable interest in modern semiconductor (p / p) technology and an attempt to understand what stage Zelenograd Micron, Russian silicon microelectronics, is at, whether it is possible to "catch up and overtake" the west in this area of ​​high tech (we liked the habrahabr article in many ways) . ru / post / 218171 , we continue the reasoning on this topic, give some answers and clarifications to the information contained in it). In this text, we present the opinion of a person who worked 20 years in the American p / p industry and several years ago tried to return to the Russian Federation and share his experience and knowledge with his Russian colleagues. The important thing is that he knows both sides well, both Russian and American.
Commercial nanotechnology
Let's divide the tasks right away. To begin with, we leave out the military theme, p / n electronics for military tasks, applications, and orders. It has its own features.
Start with the definition. What is commercial nanotechnology? This is an economically viable, large-scale (millions of pieces) production of products containing nanoscale elements with desired properties. For example, modern semiconductor chips, each of which contains billions of nanoelements with critical dimensions of several tens of nm, formed with a reproducibility of several nm and measured with sub-nanometer accuracy. Thus, technology is not only a way of production or a 'knowhow', but also the highest reproducibility, a high percentage of usable products (in the p / p industry in a competitive situation, this percentage is usually above 90%), the cost of products is competitively low.
The ability to make single samples is not yet a technology. Otherwise, we can say that many universities in Canada or America have 10 nm technology and less. How so? Yes, in their laboratories can make an integrated circuit of several dozen transistors with a critical size of several nanometers. There are universities in which transistors working on CNT (Carbon Nano Tubes) have been creating for many years, etc. But no one would even think of calling it 'technology'. So they will say “they made a working transistor on CNT” or “built a logical element & based on CNT transistors”. And here is the technology? This is only the first step towards technology. In world practice, this is called 'research'. R & D follows research. R & D is followed by 'productization'. 'Productization' is followed by 'fanout' (placing technology in several factories and increasing production). After the successful completion of all these difficult and time-consuming stages, we can talk about creating a 'technology'. Many experts believe, by the way, that 'research' is far from the most difficult (or expensive) element of this chain.
')
Let us now look at the basic elements of technology. After all, it is precisely the nanoscale elements created in the manufacturing process that ultimately must provide the required electrical properties of the product. In semiconductor chips, these elements are transistors, and the critical size of the technology is usually close to the size of the transistor gate half of the period of the first metal level (half pitch of MET1) or, in other words, half of the minimum period available for lithography. Thus, the minimum period of the MET1 integrated circuit technology 65 nm will be close to 130 nm. From the point of view of lithography (and, therefore, of technology in general), it is the minimum attainable period that is important, and not the size of the transistor gate (as some companies claim sometimes). Using various processing tricks (for example, variations in the dose of photons during lithography or additional etching of the shutter material), almost any size of shutter can be obtained with a fixed period of lithography. The authors know examples when the nominal gate size of the technology 65 nm was 40 nm, while the minimum achievable lithography period of that time was 130 nm. This approach does not allow to change the packing density of the transistors of the microcircuit, but it is useful for improving the electrical characteristics of the transistor and, therefore, the product.
As an example, we can use published micrographs of transistors manufactured at Micron (see Fig. 1 or habrahabr.ru/post/213373 ). Assuming that the size of the gate is 54 nm (as follows from the data of Micron himself), one can easily measure the period of the structure, which is equal to 190 nm. Thus, this structure is representative of the technology 95 nm, but not 65 nm. This is not surprising, since it is known that this structure was made on the lithographic equipment of the technology of 90 nm (the minimum attainable period of 180 nm). As stated earlier, Micron is not the only company that assesses technology by the size of the shutter. Many eminent companies do the same thing in pursuit of the fame of a technology leader.
<img src = "
"alt =" image "/>Figure 1. Micrograph of 65 nm transistors manufactured at Micron. Shutter size = 54 nm. Period = 190 nm. [Mikron: Period = 2.71 au * 54 nm / (2 * 0.38 au) = 190 nm, where 'au' (arbitrary units) is the size of the arrows in inches.
The above definition of nanotechnology should be taken seriously and very carefully, without losing a single detail. This is significant because such a definition or, more precisely, an understanding adopted by the international community is significantly at variance with that adopted and existing at the “intuitive level” in the Russian Federation, where technology means only a method of manufacturing a product (a sequence of certain operations for manufacturing a product). According to this still Soviet definition, technology was called (and is still called) the production of the world's largest large integrated circuits (there was such a joke in the 70s) with a yield of <0.1% (and this is not a joke). This understanding of technology reigns in the minds of high Russian leadership, which makes a decision on the development and financing of the industry, this is essential. Let's do this practice “we will not stop at the price” of the existence of some production processes with the yield of a few percent recognize possible in military production (but undesirable even there) and we will not call it technology. Just as we will not call the technology and the manufacture of single samples / prototypes. Only reproducible, mass, cost-effective, competitive serial production of products (see the definition given above) can be called technology.
The history of Russian science and technology is such that technocrats were never systematically included in the elites. The true reasons for this are most likely political economics. There simply was not a long period of capitalism, free enterprise in the history of Russia, and it is he who “is friends” with technologies, he needs and perceives them. Attempts to create technology from a stick can be considered successful only in the near-war industries (nuclear industry, space technology). Since with the development of these industries there is a certain understanding and experience, we allow ourselves some comparisons in terms of their organization. Moreover, the financing of the industry in our country is carried out in a Soviet-style centralized, whatever it may be said about the private investments of individual oligarchs. The reason for this is correctly indicated (“ habrahabr.ru/post/218171 ) in a comparative analysis of simple“ buy-sell ”and high-tech business. Therefore, in modern Russia, the oligarch, forced by the high authorities to “nanoelectronic duties”, investing its money in a private-state high-tech program, is jealous of the simpler and profitable business of its clan / class friends and equalizes profitability by means of a corruption abduction of funds from the program. This is a common practice, as we see it.
The production of modern microcircuits in an already-launched version is a multistage, strictly regulated, sensitive to deviations process, consisting of thousands of complex operations. For example, a deviation of more than 10% in the critical size of one of the billion transistors can lead to the failure of the entire electronic circuit. Marriage. An insignificant deviation of the physicochemical conditions of any of these thousands of operations can lead to such a failure. In this case, the entire batch of substrates that have passed the problematic process can go into marriage, and these are hundreds of thousands of scrap chips, these are discarded materials, electricity, working time, disruption of expected supplies, that is, huge losses. In this sense, continuous production is many times more complicated and more responsible than single-piece production.
It is possible to accelerate the creation of a high-tech product by borrowing, buying or extracting, by describing its production processes in developed countries. But in any case, you need to have sufficient in number and quality of your own qualified team (personnel), infrastructure, necessary equipment and materials (and, therefore, related industries). This is a hard, slow and very resource-intensive task. The launch of each of the technologies (180, 130, 90, 65, 45, 32 nm) in the world required the involvement of thousands of qualified specialists / experts in the components of technology, materials science, metrology, circuit design, and so on. Each time these were people who already had experience with such technology launches. These people had impeccable engineering support, the ability to instantly deliver any necessary equipment and materials. With this perfectly organized work, it needed billions of dollars and about 2-3 years of work of such a team on each of the technologies. And when we hear the
In the history of the development of breakthroughs in the atomic and space industries in the USSR, the personalities of managers, Kurchatov and Korolev, who directly interacted with the customer / owner, who made decisions on finance and other resources, played a particularly important role. They pressed the necessary decisions with their authority, argued, proved, convinced. And not like now, when between the technical management of the project and the main representative of the customer are the inhibiting layers of technically illiterate officials, economists and financiers. Of course, another important difference from the Stalin era is the presence of powerful corruption, this introduces resistance to movement towards the goal, and often this friction becomes an absolute hindrance, completely slowing down the process. What has been said in this paragraph is not an excuse for Stalin’s “effective management.” Simply, our system has been transformed externally, but in fact is controlled by the same methods. Therefore, either it should be demolished as a brake on a high-tech, or, for reasons of the case, at least use these methods from the past.
If you have a V-2 rocket, then you are still very far from a flight to Mars. If you have a technology of 250 nm, then you will not be able to jump all steps with one jerk (180 nm, 130 nm, 90 nm, 65, at, 45 nm, 32 nm ..) and immediately start, say, 22 nm. These are crazy fairy tales, to believe that this is possible. In general, in exact sciences and in engineering, it is necessary to do without lies and exaggeration of achievements. A characteristic lie is a report on the launch of technology when single samples are obtained. Or give out for their achievements products made in cooperation with China, in which the Chinese have made all the significant nanotechnological parts. Well, and much, much more ... We love and are able to trumpet about achievements. About the launch of the 65 nm technology on Micron, messages in the press appear quite regularly and every time as the primary message.
So let's discuss the real business. In the Russian Federation, the actual (launched or debugged) technology is 90 nm. Talking about 22 nm or 3D microchips is fantastic. They can be maintained only in a theoretical manner, in order to increase their technical outlook.
But do not fool yourself. Recognizing the status quo is an important part of the problem statement. A competent formulation of the problem is the way to solve it.
Features of the Russian nanotechnology market
Another important aspect. Above, we essentially defined nanotechnology through the market. In the Russian Federation, an ambiguous situation with this. Consumer market in the Russian Federation on computers + smartphones + ... exists, yes ... But something is not visible domestic product, fully produced by its nanoprom. What sector is Micron products aimed at? The head of Rosnano does not understand much about technology, physics, etc. But you cannot refuse him that he boldly and globally knows how to influence the Russian market. Just forms it with his own hands. Well, everyone remembers voucher privatization, which made (?) Everyone (?) Owners (?) At one beautiful moment. All three questions in brackets require reflection. So, approximately the same creates a market for the products of the Russian nanotech. As far as we understand, Mikron has a large renewable order from the Moscow metro: cards with RFID chips pasted in, replacing the previously successfully used tokens and patches. Despite the massive, large renewable capacity, this sector of the market is not open, competitive. Well, IBM or Intel cannot offer their chips to the Moscow metro, even if their products are better and cheaper. That is, this is a specific order with cartel signs. The following global idea of ​​Mr. Chubais was presented at the Rosnanoforum 2010 (or 2011 ??). Replace throughout the country bar-code labeling of electronic goods, using RFID markers, “sewn” in each package with the goods. Let us leave aside the joys of accounting control from such an innovation, they exist. But we will note the same "firm" approach of the forced formation of consumption. They built the population and imposed a high-tech product on it for its money, without asking. Attempts at such an approach were with tablets for schoolchildren, etc. Why is it important for us to mention? - because in fact such orders do not differ from the military, for example, navigators or walkie-talkies for the military. So, find the "honest" products of Micron, which competes with the world in the free market and compare consumer and economic characteristics. We cannot do it yet, we don’t find ...
Industrial metrology. Nanometrology in Russia
Next, we want to talk about industrial metrology in nanotechnology as applied to well-established 180-32 nm technologies. For smaller scales 28-11 nm, new tools and techniques are being developed and applied - we will not go there today. Industrial metrology provides measurements of critical parameters of manufactured structures and their elements, used both during production, in-line process control, and others, including electrical testing of output products (for example, wafer acceptance testing). Focus on in-line control. This is the key that allows you to raise the quantity and quality of suitable products. The control is practically carried out after each operation, which can lead to a change in the topological dimensions of the elements on the substrate. Some operations (for example, all operations with a photoresist) upon detection of a size mismatch or, for example, unsatisfactory adhesion, can be redone, repeated. The control of elements is carried out selectively under a reasonable assumption about the identity of elements created in identical conditions. A sample of test objects, their location on the substrate, the statistical reliability of monitoring, ... and much more are also the most complex technical and conceptual procedures. In this case, in-line control should not delay the flow of substrates along the production line (create a narrow throat) and should be simple enough for the operator (and ideally fully automated). Control of the critical dimensions of the elements of microcircuits is one of the most important metrological processes of modern nanoelectronics. The main tool for controlling critical sizes today is a scanning electron microscope, SEM. It allows you to make such measurements without contact with sufficient speed and accuracy (although for some applications the ultimate accuracy of the EMS at 3-5 nm may not be completely satisfactory). There are difficulties in interpreting the results of SEM measurements for critical object sizes <100 nm (the problem is related to determining the true position of the edge of the line or wall of the trench). To accurately measure the size of nanoscale elements, special methods for calibrating or correcting EMS should be applied (" www.gcrm.info/Home/publications-1 ). Neglect of these techniques can cause reproducible errors and serial defects. Inaccuracy (or bias) of measurement of critical dimensions with the help of SEM varies with the geometry of the object, the general topology of the structure and the physicochemical properties of the materials. The world n / a industry at the beginning of the century faced this problem and recovered from it. Serious attempts to realize 90 nm and especially 65 nm production in the Russian Federation should take this experience into account.
Here we see some obstacles, a local flavor, so to speak. Any measurements are reduced as a result to the procedure of comparison of the measured object and a certain standard, standard. So, the national Russian standard size sample for the scale of several dozens of nm in interest to us and the GOST instructions (technique) for its use for the EMS calibration have serious problems and are created without taking into account the above important points. Note that we can talk about errors in determining the critical dimensions of elements of integrated circuits of several nm, comparable to the tolerances of technology 90 and 65 nm (as a rule, the tolerances of critical dimensions are close to 10%). As is known, measurement uncertainty should not exceed 10–15% of deviations of dimensions (or tolerance) controlled by metrology. See, for example, 'Handbook of Silicon Semiconductor Metrology', ed. Alain Diebold, 2001. , 1-2% (0.9-1.8 0.7-1.3 , , 90 65 ). , (sample-to-sample bias variation ), 3-5 90 65 , , . , sample-to-sample bias variation () (measurement uncertainty) . ( APC) / . , '' '' '' , . .
, , . , , , ( ..), .
Conclusion
. , , , , / . , , / . , , , , , .
Source: https://habr.com/ru/post/220843/
All Articles