Epinoya – Knowledge Applied is Power

The Magic of Moore’s Law

In 1965, Moore’s Law was an observation. By 1975 Moore’s Law had become a benchmark to which semiconductor companies strove. It provided a metric and a trend that created new industries, changed cultures, and enhanced our lives. Moore’s Law made information and communication more democratic in that the common person now has more power to influence markets, politics and culture.

Why has Moore’s Law brought such amazing technological and societal advances? I would submit that it is due to a shared vision. A global consortium of companies, universities, research institutions and governments is focused on meeting the expectations of Moore’s Law. This synergy has enabled advances that no individual or single institution could accomplish. This is, in a way, similar to President Kennedy’s vision in 1961, outlining the goal to putting a man on the moon. The institutions throughout the USA focused on the goal to put a man on the moon “before this decade is out”. The goal was accomplished in 1969. Since this was the only goal outlined, the country lost focus on space exploration for a while. Therein is the difference between Moore’s Law and the “moon shot”. The benchmark/goal was a rate of development and not an end point.

But, now, the trend is slowing and appears to be approaching the end. I had assumed that Moore’s Law would be supplanted by a new computing paradigm before it became too expensive to maintain the development rate. Instead, the industry is looking at ways to keep going. The new paradigm that the semiconductor industry is touting is termed “More than Moore”. This has a different approach than with Moore’s Law. “More-than-Moore” (MtM) is characterized by functional diversification of non-digital semiconductor-based devices. It is acknowledged that “they do not necessarily scale at the same rate” as that of digital functionality. The intent is to provide additional value by migrating non-digital functionality from system board-level to ultimately integrating with the digital functions on the chip (SoC).

The problem is that MtM, now, has no singular focus. There is a diversification of domains. Each domain intends to have its own “law of expected performance” (LEP). While Moore’s Law “digital trend” is not abandoned, it is not the primary focus. I see an industry struggling to maintain value growth in the face of mounting, unsustainable costs because of its paradigm-lock on 2D scaling. I have outlined the importance of Moore’s Law in previous articles here and here and proposed that the trend can be sustained by 3D integration. The added benefit is that MtM can be implemented more simply through heterogeneous 3D integration.

Does it Matter if Moore’s Law Dies of “Old Age”?

The cleanroom at GLOBALFOUNDRIES' Fab 1 in Dresden, Germany — The cleanroom at GLOBALFOUNDRIES’ Fab 1 in Dresden, Germany

Moore’s Law appears to be coming to the end. It is getting too expensive to double transistors, particularly at the historic rate of every two years. Does that matter? What happens if it does?

The semiconductor industry is one of the few industries that produce a product that enables the advancement of the industry. The chips that are produced are used to improve design and manufacturing for the next round of products. Under the Moore’s Law paradigm, the improved chips enabled increased performance while reducing cost. Consumers felt that they needed to upgrade their computers on a regular basis just to keep up. As a result, demand continued to climb. As Moore’s Law faltered, hardware improvements slowed. The consumer did not feel the need to upgrade as often and demand slowed. The mobile market provided a boost to chip demand, but trading computing power for mobility. That market, now, seems to have reached its apex. So, the next hope is to build more servers for Cloud applications, trying to share cost of more expensive chips and IoT to massively deploy low-end chips to leverage the economies of scale.

The problem is if the cost of the improved chips goes up, the next round cost will also go up. Instead of a downward cost spiral enabled by Moore’s Law, that spiral will turn upward. The ROI will continue to shrink for semiconductor companies and negatively impact the electronics market. This, in turn, will hurt global economies.

Whether you believe in Moore’s Law or not, its multi-decade trend has produced incredible capabilities for businesses and customers. It has enabled emerging markets and economies and has improved cultures. I contend that the cost/performance improvements in the semiconductor industry are a necessary driver for continued economic growth.

Now is the time to change semiconductor manufacturing to affect a return to the Moore’s Law trend. It is time to drive 3D integration.

The Death of Moore’s Law

For the last several years, people have predicted the end of Moore’s Law. The reasoning is that there is a limit at which one can’t shrink transistors any further. A reoccurring comment has been “You can’t divide an atom.” I had assumed that its demise would be at the hands of a new paradigm like quantum computing. Now, with Intel’s announcement that the next doubling of transistors will take 2 ½ years, it looks like it may die of old age.

I, personally, do not believe that Moore’s Law needs to die of old age. Having worked within and in support of the semiconductor industry, I believe that the scaling argument is based on a faulty assumption; that one must only use two dimensions. I also believe that the industry is finally waking up to this fact with the surge in interest in 3D integration. But it has come too late to keep the industry on the Moore’s Law curve.

I have watched as the increased cost of scaling has forced the formation of collaborative research organizations, e.g. Sematech. Chip companies have shifted market and business strategies like the fabless ecosystem. And continued M&A has resulted in massive organizations with deep pockets that make barriers to market entry by new players almost impossible. As a result, I believe that the Semiconductor industry is ripe for disruption.

When I worked at the Hughes Technology Center in the early 90’s, we were working on enabling technologies for 3D integrated circuits (3DIC). Our strategy was to freeze scaling at 0.25 micron (that’s 250 nm folks!) and build another active layer on top, doubling the circuit density. There were several technologies that we were developing to do this. For example, HRL had developed a TSV on which I was able to grow high quality silicon epitaxy. This was used to build a 3D version of a Pentium-based PC in a “cube” as demonstrator. We filed for a patent disclosure, but corporate declined to pursue. Another development was wafer bonding and thinning. We developed a scanning plasma process that flattened the device wafer while thinning it. We had a 200mm demonstrator wafer bonded to a handle wafer that was 10nm thick with +/- 1nm variation. Obviously, 10nm is not very useful, but it meant that FDSOI was comparatively easy. Our bonding technique allowed conductors and dielectrics to be bonded, simultaneously. Our university collaborator used this process to demonstrate the fabrication of a CMOS circuit by bonding NMOS and PMOS circuits. There other technologies developed that I won’t go into for lack of reader attention. But these were only steps toward the ultimate goal, which was monolithic 3D integration.

Monolithic 3D integration was not to be the stacking of processed layers, but depositing and processing layers on a continuous process. Think in terms of transistors along with other components embedded in a matrix of dielectric with interconnects routed for optimal distances. This would require different equipment and different chemistries. One enabler we were working on was atomic layer deposition (ALD). The sub-category, atomic layer epitaxy (ALE) was the process we believed would provide the embedded transistor structures. I submitted a proposal the develop ALE silicon, which was declined just prior to GM Hughes Electronics’ demise. With Hughes’ breakup, all of these technologies have fallen into disuse. I believe that it is time to resurrect some of these concepts and develop the necessary equipment and processes to revitalize Moore’s Law.

I have an initial product concept that I would like to develop that would be an enabler to control the new processes. I am interested in finding investors who would fund the startup. If you are one or know of one, please contact me.

Why is Fostering Innovation Hard?

There have been many articles and books addressing how to “do” innovation and enumerate the various barriers. Innovation has been identified with increased competitiveness and even survival in today’s economy. However, the lack of innovation today has been decried by pundits and even governments. Companies, heralded in their early days as innovators, seem to flounder after a while. Other companies, who see the danger of being made irrelevant and try to become innovative, struggle to even break out of their own box. What is wrong? Why is innovation so hard and why is it so hard to sustain?

The answer is that an innovative company is in a metastable state. It is occupying a higher energy state. The barrier that prevents the company from falling back into a lower state of status quo is fairly low; lower for some than others. The barriers to innovation are the steep wall of energy one must scale to reach the metastable state. All the ways to kill innovation are the shallow wall of energy that the company can fall over back to the ground state. Maintaining the innovative state requires vigilance and work. It requires building a culture from the BoD and CEO to the janitorial staff. It requires retraining thought processes and breaking long held paradigms. Innovation has to become part of a company’s DNA. When it does the barrier to status quo gets higher, but not insurmountable. What does an innovative company look like? The profile I have is a composite of the innovative companies I have come into contact with. Most have most of the qualities in common. Several no longer exist because innovation was killed and subsequently the company, as well.

So, what does an innovative company look like? Is it the no-dress-code, flex-time, free-food, perks-plus environment? These might help with the current generation employee, but there are more fundamental characteristics of an innovative organization.

Company leadership is committed to innovation. The leadership invests in the employees and trusts them. In turn, the employees develop a trust in the leadership.
There is a “no fear”, “can do” mind-set throughout all functions in the company. The answer to the question, “Is it possible?” is not “No”, but “Not yet”.
Many in the company are well educated, but all are constantly learning.
There is a pervasive curiosity; consistently asking “Why?” and “What if?”
There is also a “maker” mentality and the discipline to execute the processes to launch the new product or process.
An innovative company will listen to the voice of the customer, but can discern real versus perceived needs.
An innovative company is not efficient. Failure costs money; as does finding your way to discovery. That’s why you need high profit margins.

Typically, we think of Apple, Google, or Facebook as being exemplary. But I also think of the company I joined, fresh out of school, Hughes Aircraft. Hughes was a defense contractor with all of the restrictive regulations one associates with working with the DoD. Yet, they gave us practical laser technologies, wire bonding for semiconductor packaging, LCD technology employed in the first digital desktop projectors, 3D surround sound, back-of-the-seat in-flight entertainment, satellite TV, digital X-ray imaging, and ion propulsion engines for space travel. Well, the ion propulsion was a little early. It was repurposed into reactive ion etching used in semiconductor processing. One of the characteristics of Hughes leadership is that the CEO, up until the last one, was an engineer, often one who rose through the ranks. Hughes was a large company with about 90,000 employees at its peak. Not all the divisions fit the above profile. But two I worked at and others I worked with did. I remember a senior Air Force manager wishing aloud at a project review meeting for a particular microwave component with characteristics that were unprecedented. One of the project engineers slipped out during a break and returned with some samples and data. He presented the package to the slack-jawed manager with the question, “Like this?” We were provided means to explore and push boundaries, which enabled the engineer to anticipate needs before they were even voiced.

M&A has been identified as one of the ways to kill innovation. This was certainly a danger for Hughes when it was acquired by GM. There was some “brain-drain” when some of the more creative engineers left the company. The CEO was able to maintain the course and provided an additional direction toward automotive challenges. Some of the projects started were: heads up displays, digital instrument clusters, remote key locks, rear object detection, and collision avoidance.

However, it took a change in leadership to bring the company to its knees. The new CEO was not technically trained and was focused solely reducing costs. As cuts were made, more creative engineers left the company, fewer contracts were won, the bottom line suffered and therefore more cuts were made. The company was in a death spiral until it was sold off in parts to competitors. The loss of intellectual property was incalculable.

This is just one case I can cite from experience of an innovative company that was brought down because of a couple of the innovation killers that have been identified.

Sometimes You Need Transitional Products That Also Differentiate

DOA Launch
When I joined Brooks Instrument, they were just completing a product launch for their new digital mass flow controller (MFC). The product launch received no response. The customer interface was digital and the semiconductor customers had not yet transitioned from analog interfaces to digital. I then had the opportunity to restart the development cycle to correct the short-comings of the DOA product. Some of the things we addressed were the lack of modular design in both the hardware and software. We spec’d out a new, more powerful processor to handle the expected performance bandwidth we needed to handle the new product definition. The main key was a dual interface that could be analog or digital. This provided the customer performance and feature improvement over the traditional analog MFCs while maintaining the legacy interface. But we also had a new technology coming down the pipe that would leverage the full digital design.
Leverage New Technology for Customer Benefit
Brooks had engaged a technology partner to develop multi-calibration capability. At the time, all MFCs were calibrated to a single gas. If you needed a MFC for a different gas you had to buy another one. This created challenges trying to find the best mix of spares in inventory. If the customers could get multiple calibrations on a single MFC they could reduce the number of spares in stock. Prior to releasing the product I wrote an article exposing the lack of accuracy in the traditional calibration technique. We were going to calibrate using a much more accurate method and provide multiple calibration capability. The market received the product well and we were first with a market differentiator. The new product enabled us to penetrate companies that we were not able to penetrate before.
Differentiation That Makes a Difference
An example of this occurred when my lead engineer and I accompanied our sales rep to a semiconductor fab in Texas. Up until this point, we had zero sales with this customer. The reception was cold, if not almost hostile. I don’t know how our sales guy managed to get the meeting, but we were there. We filed into a conference room with a long table. The Brooks contingent sat at one end nearest the projector and screen. The fab manager and two of his engineering managers sat at the other end. It did not take an expert in body language to see that we had our work cut out. I led off the presentation introducing the product, outlining the benefits in general. Then I turned the presentation over to my engineer. As he explained the technology and how it would benefit the fab operations, I noticed that the engineers began leaning forward and then chair-hopping closer, all the while asking questions. By the time my engineer finished, the fab engineers were sitting next to us. I stood and asked if there were any more questions. The fab manager turned to our sales guy and asked “So, how much would it cost to convert the fab?”

Where Do You Get Your Ideas?

In my previous post I talked about motivations for innovation. The example I focused on was Brooks Instrument’s Coriolis MFC (mass flow controller). So how did we come up with this idea and why did we think it would work?

Where do you get ideas?

I am convinced that there is no magic formula that works in every case. There are a variety of good approaches for getting ideas that won’t work sometimes. The key is to stay flexible and if the results are unsatisfying try a different approach. In Brook’s case we used a “seeded” brainstorming method. I call it “seeded” in that we stablished a context and knowledge baseline before we launched into the brainstorming to expand and then filter ideas to the mostly likely to succeed.

I invited technologists from our sister divisions and subsidiaries to the seminar and included some consultants that had already been screened by Emerson corporate. Several presented technologies that were known and some potential from external ideas. I presented my idea that I had originally pitched to Unit Instruments. Obviously, most of us came with a predisposition toward a particular idea. To keep the effects of prejudice to a minimum, we exposed the prejudices and requested that the participants keep open minds. Here is where we conducted the brainstorming collecting as many ideas as we could. We then discussed product feasibility, time to market, expected development costs, risk factors, among other topics to drill down to a final candidate.

One of our sister subsidiaries was Micro Motion whose founder had invented the Coriolis flow sensor. Coriolis sensors respond directly to mass flow. There are no conversions or inferences required to obtain a measurement of mass flow as with other mass flow sensing technologies. There was considerable IP protection and the technology could obsolete the multi-calibration capability that we had just developed. This would not cannibalize our current product line initially since our initial application would be for liquid flow. Our main products were for gas and vapor flow. The liquid flow products throughout the industry were known to be weak in meeting customer needs. A Coriolis based MFC for liquid applications would expand our product offerings and allow for a reasonable return on investment from the Digital Thermal MFCs. Eventually, the Coriolis MFCs would expand to include gas and vapor applications displacing the digital thermal MFCs. Ok, so we had a technical direction to head in, but what was the product going to be like?

Is the New Product You Just Launched Obsolete?

New product development is about competition; trying to provide a superior product that will attract more sales and satisfy the customer’s needs. Too often, product development is reactive. A competitor comes out with something new that forces you to either meet or exceed their offering. We can see this in the smart phone and tablet markets recently. But this has been going on in the automotive industry for some time. It’s getting more and more difficult to distinguish between different car makes in the main stream markets.

Thermal Mass Flow Controller

The Mass Flow Controller (MFC) market has the same issues. Having worked with three different MFC companies, I have seen the intensity of competition first hand and the tendency to be reactive. Part of the reason for this is the narrowness of the field. When I was at Unit Instruments, the then CEO illustrated our situation using a picture of a foot race. There were six runners all bunched together at the front. He indicated that although we were the leader, we had three competitors breathing down our necks and two more not far behind. It’s not a comfortable situation. It would be far better if we could separate from the pack. It is far better to be proactive, leading the way to better products.

Such was the case at Brooks Instrument with their new Coriolis mass flow controller (CMFC) product line. At the time Brooks was owned by Emerson Electric. Brooks was part of the pack mentioned above, but toward the back so they were in a better position to take some risk. Emerson owned several other divisions and subsidiaries that were involved in process instrumentation. One of the major events every year was that each business unit would make the pilgrimage to St. Louis where Emerson’s headquarters and corporate retreat were located. The purpose was to review with the CEO and the senior staff the previous year’s results and plans forward. The event was held at the corporate retreat (which has since been divested). As a part of the day-long presentation and discussions, our VP of Engineering presented our new product activities. We had successfully launched our digital MFC with multi-calibration capability. We were disrupting the market. At the end of the presentation the CEO asked our VP of Engineering how old the technology was that our mass flow sensor was based on. Brooks and all their competitors used technology that was developed in the mid-1960s. It is based on the transfer of heat from a heater to a sensor by the flow stream. The mass flow rate is inferred from the heat capacity of the fluid. Our VP explained this to the CEO. The CEO asked, “Is there a chance that one of our competitors would develop a superior flow sensing technology?” Our VP acknowledged that it would be possible. The CEO then asked, “What are you going to do about it?”

Obviously, this would be a strong motivator to launch another NPD project. Our VP of Engineering asked me to put together an internal seminar on flow sensing technologies. I had been keeping abreast of potential technologies that could be incorporated into our products and had an idea for a potential flow sensing technology that I had originally proposed to Unit Instruments. They had initially explored it, but then dropped it when they were acquired. But this would be a focused search for technology that would displace the core MFC sensor.

In later posts I will explore our approach, methods, and challenges in our effort to launch a new product line.

Welcome to Epinoya

Welcome to the Epinoya site, dedicated to ideas, invention and innovation.

My name is Michael Barger. I have more than 30 years of experience in semiconductor, sensors and instrumentation development serving the defense, industrial, and consumer markets. Disruptive technologies my teams developed brought to market innovative products that combine high performance and high value for semiconductor, process instrumentation, imaging and business systems.

My services include product development, project and program management, technology scouting, product/technology road-mapping, benchmarking, process development/improvement, and ideation.

Contact me if you have needs in any of these areas.