back to S1S home

Sunday, October 25, 2020

#450 Thirty years ago: October 1990

 The table of contents of the October ’90 issue of IEEE Electron Device Letters indicated strong interest in Heterojunction Bipolar Transistor (HBT) technology using complex ternary III-V compounds, AlGaAs in particular. It looks like the need to move HBT-based analog circuitry solidly into a GHz territory was well recognized at that time.


In terms of process technology, feature article published in October ’90 issue of the monthly trade journal Solid-State Technology demonstrated potential of MERIE (Magnetically Enhanced Reactive Ion Etching) in the definition of patterns smaller than 0.5 µm. Yes, 0.5 µm (or 500 nm, except that “nm” unit was at that time not used yet) pattern definition was an ultimate goal then.


Another feature article in Oct.’90 issue of SST emphasized integrated processing and cluster tools. The cluster tools market was predicted to hit $2.2 billion by 1994. I don’t know whether this prediction was correct in terms of timing and $$, but, as we all know, with time cluster tools  took over advanced semiconductor manufacturing for good.


Posted by Jerzy Ruzyllo at 10:41 AM | Semiconductors | Link

Sunday, October 11, 2020

#449 Quantum has been around for a long time

Quantum computing, quantum bits, or qubits, quantum information, these are all terms commonly used these days to reflect on the things to come. But quantum computing is just one aspect of the “quantum story”. Quantum phenomena, or in other words phenomena described by the quantum physics, rather ten classical physics, were not only known and understood, but also routinely exploited in commercial and exploratory semiconductor devices for decades.


For instance, consider the effect known as quantum tunneling allowing electrons to cross potential barrier without changing energy upon which tunnel diodes and Zener diodes, available for some 50+ years, are based. The dark side of tunneling is its parasitic effect on the leakage current of the MOS gates featuring ultra-thin (less than 3 nm) gate oxides. The need to control this effect prompted high-k gate dielectric materials introduction into the MOSFET technology some fifteen years ago.  


On top of it come numerous papers presenting results of the research on all possible aspects of quantum tunneling in semiconductor devices published in the last three decades of the 20th century. My own small contribution in this area came exactly 40 years ago! (see J. Ruzyllo, “Lateral MIS Tunnel Transistor”, IEEE Electron Device Letters, vol. 1(10), 1980).


Quantum tunneling is just one example. How about quantum wells, 2D electron gas, ballistic transport, and other quantum physics-based phenomena ubiquitous in state-of-the-art electronic and photonic semiconductor devices? Some more on these issues next time. 


Posted by Jerzy Ruzyllo at 08:26 PM | Semiconductors | Link

Sunday, October 4, 2020

#448 More on semiconductor material choices

Few more comments to conclude a brief overview of criteria guiding selection of semiconductor materials suitable for specific device applications (see entries #445 and #446). In addition to those to which a number can be assigned (e.g. width of the energy gap in eV, electron mobility in cm2/Vs, etc.) there are others, often equally important, which escape numerical classification.


Radiation hardness is one among them. We don’t want threshold voltage of the millions of transistors comprising an integrated circuit to be altered as a result of circuit’s exposure to high energy radiation (electronics in the proximity of nuclear reactor, exposure to nuclear explosion, cosmic rays, gamma rays, X-rays…) 


In addition, manufacturability related characteristics of any given semiconductor material are coming to play. Is the material thermally stable enough to allow elevated temperature processes during device manufacturing? Is it mechanically sturdy enough to allow robotic handling during device manufacturing? On the other end of the spectrum, is semiconductor material flexible enough to be compatible with flexible electronics and photonics?


Last, but certainly not least is the cost of semiconductor material and device fabrication processes which need to be weighed against the range of applications of commercial devices manufactured and expected profits.  


As you see the list of criteria is long, and expectations high. The good news is that if you will ever have a problem finding ideal semiconductor for your application, you can always count on silicon as there is a very good chance it will do a trick, may be not perfectly, but it will (ok, granted, as long as temperature at which device is expected to operate will not exceed some 150oC...)


Posted by Jerzy Ruzyllo at 08:50 PM | Semiconductors | Link

Sunday, September 27, 2020

#447 Thirty years ago: September 1990

 September 1990 was a month when my colleagues and collaborators did demonstrate Penn State’s leadership in exploration of the wide bandgap, carbon-based semiconductors. Papers on “High-Temperature Schottky Diodes with Thin-Film Diamond Base” by G. Gildenblat et al., published in September 1990 issue of the IEEE  Electron Device Letters,  and “Oxidation of Single-Crystal Silicon Carbide” by Z. Zheng, R.E. Tressler, and K.E. Spear published in the same month issue of  the Journal of The Electrochemical Society testify to this effect.


Other than that, out of eleven papers published in “Silicon Devices” section of the September 1990 issue of the IEEE Transactions on Electron Devices, six were devoted squarely to SOI MOSFET technology. Silicon-on-Insulator technology was getting significant attention thirty years ago. And rightly so…


Posted by Jerzy Ruzyllo at 08:47 PM | Semiconductors | Link

Sunday, September 20, 2020

#446 Semiconductor material choices: what else matters?

 In addition to those most obvious listed in the previous blog, there are other considerations influencing selection of semiconductor materials for various device commercial applications.


An important one is susceptibility to doping as both n-type and p-type doping should be possible with all semiconductor materials. Diamond for instance has an “n-type problem”. Being naturally p-type, diamond cannot be easily converted into n-type semiconductor which makes fabrication of p-n junction based devices in diamond a challenge. Too bad, as other than that, diamond is an excellent semiconductor. 


Thermal conductivity is another important parameter as it determines ability of semiconductor material to dissipate heat generated during device operation. By the way, diamond is an excellent conductor of heat.     


Oxidation characteristics of semiconductor material also come to play. Ability to form its own, device-grade native oxide is a major advantage of semiconductor material. Silicon excels in this regard and its unmatched by other semiconductors oxidation characteristics (growth of high-quality native oxide SiO2 by means of thermal oxidation) effectively opened the door to MOS (Metal Oxide Semiconductor) technology without which, as we all know, we wouldn’t be even close to where we are in terms of the impact of electronics on our lives.   

Posted by Jerzy Ruzyllo at 11:12 AM | Semiconductors | Link

Sunday, September 6, 2020

#445 Semiconductor materials: properties and applications

Whenever it comes to the selection of semiconductor material for any given semiconductor device application, width of the energy gap (or bandgap) is the first consideration. The reason is rather straightforward - if the material features no naturally occurring energy gap (for instance graphene), then it cannot be used to fabricate devices efficiently performing some mainstream electronic functions such as for instance switching.


Another important characteristic is the type of the bandgap, direct, or indirect with the former allowing superior in terms of efficiency emission of light as compared to the latter. This criterion, though, is less clear-cut now as it used to be because of the success in enforcing light emission from the indirect bandgap silicon. Not very efficient, but still, commercially viable.


Then comes electron mobility which decides on how good of the conductor of electricity given semiconductor is, and also determines speed at which device fabricated using given semiconductor can operate. It is convenient from some device applications point of view to have a material featuring at the same time high mobility of electrons and (relatively) holes. But here again quantum phenomena are blurring the picture because when the ballistic transport takes over, electron mobility is no longer a factor determining device’s speed of operation.


Look for more details regarding device implications of physical characteristics of semiconductor materials in my book “Guide to Semiconductor Engineering”.

Posted by Jerzy Ruzyllo at 10:00 AM | Semiconductors | Link

Sunday, August 30, 2020

#444 One more time on questionable predictions

 This is going back to the post of July 7, 2014 entitled “Predictions, predictions” in which I expressed my disdain with regard to misleading, misinforming, technically frivolous predictions regarding semiconductor engineering often meant to attract attention to a specific self-serving cause.


Not much has changed since then. The problem is of course not with technically sound, well founded in industrial/business reality, and overall, very needed forecasting. The problem is with predictions which can be found in some broadly disseminated semi-technical publications which are meant to grab our attention.  For instance, statement about “chemical computers replacing in about ten years traditional silicon microprocessors-based computers”, or about graphene replacing in the near future silicon in all electronic device applications. And how many times over the years we read about “insurmountable physical barriers”, “red-brick walls”, “limits of technology”, and “ends of the road” among others?


So, what’s the conclusion? Just don’t take for granted everything that you read about the future of semiconductor electronics and photonics and use common sense to see through the often grossly misleading proclamations.

Posted by Jerzy Ruzyllo at 09:26 PM | Semiconductors | Link

Sunday, August 16, 2020

#443 Thirty years ago: August 1990

Not surprisingly, six out of eight papers published in August 1990 issue of IEEE Electron Device Letters were devoted to research on innovative transistors solutions. Whether SOI MESFET, heterostructure FET, high performance BJT, or GaAS FETs, the urge to develop innovative transistor technologies able to meet anticipated future challenges was evident.


At the process side, as some papers published in the August ’90 issue of the Journal of the Electrochemical Society indicate, low-temperature, or rather low-thermal budget processing continued to be of interest. Also, papers devoted to plasma etching, including remote µwave plasma which was a technique attracting significant attention at that time, and RIE were noticed.

Posted by Jerzy Ruzyllo at 11:26 AM | Semiconductors | Link

Sunday, August 9, 2020

#422 Equivalent Gate Length (EGL)

Scaling of the gate length of the MOSFETs comprising advanced logic ICs was for the last fifty years a main tool used to improve transistor’s performance in terms of speed of operation, on/off ratio, circuit density and others. Obviously, shortening of the gate length cannot continue indefinitely and at certain point needed improvements in transistors’ performance started to accomplished employing other “tools”.


In the blog #379 posted in December 2017, I proposed a concept of the Equivalent Gate Length (EGL) meant to establish connection between gate scaling as a way to improve transistor performance, and alternative solutions involving modifications of transistor architecture and/or material choices potentially resulting in the comparable to gate scaling performance improvements, but implemented without reducing gate length.


Recently announced by Intel expected 20% improvement in transistor performance accomplished by re-engineering of the FinFET while maintaining the gate length unchanged (whatever it is in terms of actual distance between S and D) is an example of the latter.


I wonder how much the gate length would have to be reduced to accomplish the same 20% improvement in transistor performance. In other words, what would be EGL in the case of the proposed innovative 10 nm (this number refers to the technology node and not to the actual gate length) “SuperFin” transistor architecture? 7 nm, which would mean full-node improvement, or may be 8 nm? It would be interesting to know this number to better understand the extent of this improvement and how does it translate into the improvement scenario old timers among us are used to.

Posted by Jerzy Ruzyllo at 10:40 AM | Semiconductors | Link

Sunday, July 26, 2020

#441 Thirty years ago: July 1990

As indicated in my previous “Thirty years ago” pieces, papers published in 1990 demonstrate “awakening” of semiconductor research community to the fact that the continuation of transistor scaling, basically undisturbed until then, is bound to face challenges moving well below 0.25 µm technology which was more or less a standard at that time.


Various papers in IEEE’s Transactions on Electron Devices and Electron Device Letters point to the anticipated problems with transistor’s drive current in the scaled down devices. The terms “hot-carrier degradation” or “DIBL” (drain-induced barrier lowering) were ubiquitous in the device-related research papers thirty years ago.


In terms of the mainstream silicon manufacturing technology lowering of the temperature of some notoriously high temperature processes was of interest. As an example, paper in the July ’90 issue of the Journal of the Electrochemical Society reported on the successfully implemented epitaxial growth of high-quality silicon at 850oC. Also, wet cleaning of deep trenches was getting attention because of growing at that time need to improve trench isolation technology.

Posted by Jerzy Ruzyllo at 11:20 AM | Semiconductors | Link

‹‹ ›› is a personal blog of Jerzy Ruzyllo. He is Distinguished Professor Emeritus in the Department of Electrical Engineering at Penn State University. With over forty years' experience in academic research and teaching in semiconductor engineering he has a unique perspective on the developments in this technical domain and enjoys blogging about it.

This book gives a complete account of semiconductor engineering covering semiconductor properties, semiconductor materials, semiconductor devices and their uses, process technology, fabrication processes, and semiconductor materials and process characterization.

With over 2000 terms defined and explained, Semiconductor Glossary is the most complete reference in the field of semiconductors on the market today.


  GEKA Associates

Metrology Experts

Copyright © 2021 J. Ruzyllo. All rights reserved.