{"id":1799,"date":"2022-01-16T21:13:53","date_gmt":"2022-01-16T20:13:53","guid":{"rendered":"https:\/\/lms.nanoproject.eu\/lms\/?post_type=unit&p=1799"},"modified":"2022-01-16T21:13:53","modified_gmt":"2022-01-16T20:13:53","slug":"size-matters-especially-in-electronics","status":"publish","type":"unit","link":"https:\/\/lms.nanoproject.eu\/lms\/unit\/size-matters-especially-in-electronics\/","title":{"rendered":"Size matters especially in electronics"},"content":{"rendered":"

Last but not least we want to get an impression of how small those ICs and transistors actually are and what challenges come along with it. After the schematic observation of the MOSFET the desire to have a look at one on a chip alleviated. For better understanding: In the very first text we learned how small a human hair (50 \u2013 80 \u00b5m) is, to understand the nanoscale. When we lay two hairs on top of each other it gets very difficult to even see the cross section. In fact, behind that cross section fit thousands of MOSFETs.<\/p>\n

Let us examine other possibilities for making very small structures visible e.g. by a light microscope. This discussion will also lead us to the problems we encounter by the manufacturing of MOSFETs. Fundamentally, in light microscopy a specific organized ensemble of lenses is creating the magnified picture of a smaller object. In this context a light beam shines through the object and the lenses into the eye. The question you may ask now is: How small could an object possibly be in this regard? For the answer to that question, you have to know that light microscopes are operating in the visible wavelength-area. This means that there is only a defined range and the half of the shortest wavelength is the smallest object you can see with such a microscope. Or expressed in other words: Two lines have to be half of the smallest wavelength distanced to each other, so they can be distinguished. This law is known as the Abbe limit and is described by the following equation:<\/p>\n

\"\"<\/p>\n

In this case is the wavelength, NA the numeric aperture (containing the refraction index and the angle of the light). Thus, you can come to the resolution limit , which tells you how far the lines have to be apart, to be distinguishable. The equation also shows how the performance of a microscope can be improved: Either by increasing the numerical aperture or by using shorter wavelengths.<\/p>\n

The wavelength of visible light ranges between 380 and 780 nm, therefore, by making use of the Abbe equation, we can estimate the smallest possible resolution of a light microscope to 200 nm. For this estimation we assume that the smallest wavelength is about 400 nm and the numerical aperture (NA) of air is about 1.<\/p>\n\n\n\n\n
\n

Remember<\/strong><\/p>\n<\/td>\n<\/tr>\n

Limitation of resolution<\/strong><\/p>\n

The Abbe Limit is the formula for the resolution in dependence of the wavelength: In this case smaller wavelength enable higher resolution.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

This is an appropriate time for an examination of the manufacturing of chips. How can we develop something that is too small to be seen? The modern era is using a procedure called \u201cphotolithography\u201d. In this procedure a pattern of a mask is transferred onto the surface of a silicon wafer. Latter is the substrate we already saw in the scheme earlier. The concept is very similar to the operating principle of your printer at home. The difference is that the inkjet is replaced with a light beam. This turns out to be very helpful when we create our extremely small structures, because the dye molecules exceed the wavelength of the light multiple times by their size.<\/p>\n\n\n\n\n
\n

Definition<\/strong><\/p>\n<\/td>\n<\/tr>\n

Photolithography<\/strong><\/p>\n

The process of printing very small structures on a silicon surface. The wavelength of has to be short to obtain higher resolution.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

\"\"<\/p>\n

The steps of photolithography are shown in the scheme above. In the beginning the silicon substrate is covered with a silicon dioxide layer. A Photoresist is applied on top of this layer afterwards by a process called \u201cspin-coating\u201d. Now the actual printing-process can take place: A mask with a certain pattern is placed above the system and is exposed to light beams. Hence, only some spots of the photoresist are meeting light. At this point there two different types of photoresist: The first option (left side of the scheme) is \u201cnegative resist\u201d. In this case the area, which is exposed to light, changes its structure in a way that it becomes stable towards a solvent. So, the masked areas can be washed away. Now the unprotected areas are etched by chemicals, and we already transferred our pattern onto the wafer. The final step is a washing-procedure, where the remaining photoresist is removed. The \u201cpositive resist\u201d can be seen on the right side of the scheme. Here, the areas, which are exposed to light, become soluble and are washed away, which tells both procedures apart. With this example we can understand how we bring the insulating SiO2<\/sub> in the MOSFET structure, which we saw earlier.<\/p>\n

\"\"<\/p>\n

Now we can put the puzzle pieces together, so a bigger picture emerges. We have seen that the length of the insulating layer is approximately the gate length. We also have seen the manufacturing of this component by photolithography. This makes us understand that the performance of the photo-processing is essential for building very small surface structures, respectively our tiny MOSFETs. In conclusion: We need to improve photolithography to still fulfil Moore\u00b4s Law with more transistors on less space. The limitation of photolithography in general is also given by the Abbe limit.<\/p>\n

The most important factor is therefore the wavelength. Repeatedly we notice that a smaller wavelength is a main factor in the optimization of building nanostructures. Thus, the state of the art is the use of \u201cdeep ultraviolet\u201d light, which is about 193 nm. So, in this case we notice that it is possible to leave the visible range in comparison to the light microscope. The aim for the future is at 13.5 nm and is called \u201cextreme ultraviolet\u201d light. Nevertheless, Moore\u00b4s Law is present up to this day. Basically, in all applications of integrated circuits the number of transistors on a chip has increased rapidly.<\/p>\n

Last but not least we can immerse ourselves into the subject of the \u201ctechnology nodes\u201d. This term defines the state-of-the-art in the manufacturing processes. For example, the \u201c45 nm technology node\u201d describes a certain procedure. In this context it includes a specific way the photolithograph is built with all its components, the wavelength used in the system, etc. The number 45 nm describes the \u201chalf pitch\u201d. Latter is half of the distance between two single structures like contacts or conducting lines. So, it is not the gatelength (the gatelength can be even smaller). When we explore the history of this process development briefly, we get an idea of how important the optimized manufacturing is up to this day: 1971 to 1998 was the time of 10 \u2013 0,25 \u00b5m nodes. 1999 was the year, where we came from micro to nanometres with the 180 nm node. The following decades showed huge improvements up to the 5 nm node in 2020, which lead to the Apple A14 Bionic chip (shown below). In this context only TSMC was able to place 11,8 billion transistors on 88,45 mm2<\/sup>. By the way: If you are reading this text on an iPhone 12 or iPad of the 4th<\/sup> generation, one of those chips is operating in your hand right now.<\/p>\n

And to keep pace with the prediction of Moore, TSMC is willing to come up with the 4 nm node in 2022. IBM on the other hand already introduced a prototype of 2 nm node chip in May of 2021, which is the first of its kind.<\/p>\n\n\n\n\n
\n

Definition<\/strong><\/p>\n<\/td>\n<\/tr>\n

Technology nodes<\/strong><\/p>\n

A description of the manufacturing process of modern chips. New generations of technology nodes are essential for keeping up with the pace of Moore\u00b4s Law.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

But can this miniaturization be infinite? It is interesting that the answer to this extremely modern topic can be found in an era, which was a very long time ago. The Greek philosopher Democritus already claimed that nothing can be downsized without a limit. At some point you will end up at an undividable unit \u2013 the atom. If we think of the 2 nm node chip, there aren\u00b4t more than 20 silicon atoms across a single transistor. So, one could think, that the end of Moore\u00b4s Law cannot be avoided.<\/p>\n\n\n\n\n
\n

Important<\/strong><\/p>\n<\/td>\n<\/tr>\n

Where are the Limits of Moore\u00b4s Law?<\/strong><\/p>\n

A gatelength, which comes close to the size of a single silicon atom cannot be downsized any further.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"comment_status":"open","ping_status":"closed","template":"","format":"standard","meta":{"_vibebp_attr":"","_vibebp_dimensions":"","_vibebp_responsive_height":"","_vibebp_accordion_ie_support":"","footnotes":""},"module-tag":[],"class_list":["post-1799","unit","type-unit","status-publish","format-standard","hentry"],"_links":{"self":[{"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/unit\/1799","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/unit"}],"about":[{"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/types\/unit"}],"author":[{"embeddable":true,"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/comments?post=1799"}],"version-history":[{"count":1,"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/unit\/1799\/revisions"}],"predecessor-version":[{"id":1840,"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/unit\/1799\/revisions\/1840"}],"wp:attachment":[{"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/media?parent=1799"}],"wp:term":[{"taxonomy":"module-tag","embeddable":true,"href":"https:\/\/lms.nanoproject.eu\/lms\/wp-json\/wp\/v2\/module-tag?post=1799"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}