« September 2006 | Main | November 2006 »

October 31, 2006

Essential Chemistry

When I was a child, science was synonymous with chemistry. Scientists in movies were always peering into test tubes, and they were surrounded by bubbling solutions. I had several chemistry sets as a child, but later in life I discovered I was more interested in electronics and physics. Chemistry is still a large part of what Honeywell does, and Honeywell is a member of the American Chemistry Council, the trade association representing US chemical companies. There are 125 member companies in the American Chemistry Council.

The American Chemistry Council is attempting to put a more positive spin on chemistry through a "rebranding" campaign called "essential2", variously reported as costing $20 - $35 million. I'm reminded of the Dupont slogan, "Better Living... Through Chemistry". Part of the essential2 campaign is a web site that lists quite a few examples of Chemistry's impact on life and industry.

• Safety
..... Child Safety
..... Food Safety
..... Security
..... Automotive Safety
..... Workplace Safety
• Health
..... Children's Health
..... Disease Prevention
..... Medical Equipment
..... Pharmaceuticals
• Environment
..... Air
..... Water
..... Agriculture
..... Renewable Resources
• Innovation
..... Knowledge
..... Discovery
..... Materials
• Economy
..... Jobs
..... Business
..... Trade
..... Energy
• Everyday
..... Families
..... Fun
..... Celebrations

An interview of some chemical company executives revealed that the purpose of the campaign is two-fold; namely, to increase investor confidence through a better industry image, and to ameliorate an increasingly hostile regulatory environment.

On the topic of rebranding, have you ever eaten a Chinese gooseberry? You may have eaten a few with the newer name, kiwifruit. Kiwifruit (Actinidia deliciosa), native to China and originally called the Chinese gooseberry, was introduced into New Zealand in 1906. The fruit was rebranded kiwifruit in 1974 to dissociate it from China during the Cold War.

1. Joyce Gannon, "A Roundtable Discussion: 'Quiet' chemical giants move to raise profile -- and image" (Pittsburgh Post-Gazette, October 29, 2006).
2. "US trade group rebrands chemistry," Nature, vol. 443 (21 Sept. 2006), p. 257.

October 30, 2006

Exponential Vampires

As a consequence of commercial exploitation, Halloween has evolved into a major US holiday. One popular Halloween character is the vampire, a humanoid creature that feasts off human blood, and its bite can change a human into a vampire. The vampire concept has led to numerous films and television shows.

Costas Efthimiou, a Professor of Physics at the University of Central Florida, has developed a theory as to why vampires do not exist. His theory is purely mathematical, of the type called reductio ad absurdum. From literary records, he assumes that the first vampire came into existence in the year 1600, when the population of the earth was about 500 million. He finds that if vampires feed on a human just once each month, changing that human into a vampire, the world's entire population would be vampires in just thirty months, or just two-and-a-half years. The reasoning is simple - The growth is exponential, and 230 is slightly more than a billion.

Nearly ten years ago, I contributed an exponential growth problem to Math Counts, an organization established to encourage mathematics in middle school education. Its current sponsors are Lockheed Martin, Raytheon, Texas Instruments, General Motors, 3M, Northrop Grumman, and NASA. My problem is simply stated. I have two parents, who also had two parents (my four grandparents), who also had two parents (my eight great-grandparents), etc. Civilization existed at least as far back as the Egyptian pyramids (4500 years ago), and a human generation is about 20 years. How many ancestors (parents, grandparents, and greatn grandparents) did I have at that time? The answer, 2225 = 5 x 1067, indicates that there were many more people alive in the past than today.

Of course, there must be a flaw in my ancestor model, but I can also point out a flaw in Efthimiou's vampire-non-existence proof. As every fan of "Buffy the Vampire Slayer" knows, vampires have the option of killing their prey before feeding on their blood, so new vampires are not necessarily produced.

1. Sara Goudarzi, "Vampires a Mathematical Impossibility, Scientist Says (LiveScience)".

October 27, 2006

The Day 1000 Problem

Many of you will remember the Year 2000 Problem, often called Y2K. An entire bureaucracy was built around the idea that mission-critical programs would fail at the new millenium because programmers took the shortcut of only using the last two digits of the year in databases and calculations. It has been estimated that the Y2K problem cost $US 300 billion. One reason the problem existed is that early mainframe computer programs worked so well that they were regularly modified instead of being replaced, and they were used over a period of time unimagined by their original programmers.

On New Year's Eve, December 31, 1999, I fully expected a power failure at my house, but it never happened. Everyone did their homework. Some computers waited until February 29, 2000, to fail. They did not follow the leap year rule that adds a 29th day to February in years divisible by 4, but only for centennial years if they are divisible by 400. Thus 1900 and 2100 are not leap years, but 2000 was. The Y2K problem was featured in an episode of the Simpsons. Homer had neglected to update the computer at his nuclear power plant, causing a bug to propogate across the internet, resulting in the end of civilization.

The Mars Exploration Rovers, Spirit and Opportunity, were only expected to have a lifetime of 90 martian days (called sols). Instead, they've gone far beyond wildest lifetime expectations. Spirit reached 1000 sols on Wednesday (October 25, 2006). Earlier in the year, the rover's computer code was updated to avoid a "Day 1000" problem. Although programmers were fairly certain that their new code would work, there was always some apprehension. Opportunity will reach 1000 sols in late November. Because of their extended life, the cost of the Mars Exploration Rover Mission is now less than $350,000 a day.

Have you heard about the Year 2038 Problem? A typical Unix timestamp stores a date and time as a 32-bit signed integer representing the number of seconds since January 1, 1970. This timestamp will roll over in 2038.

1. Marooned Mars rover returns stunning panorama (New Scientist, 25 October 2006).
2. Victoria Crater Marks Giant Cake For 1000 Sols On Mars (marsdaily.com).
3. Spirit survives sol 1000 (Nature).

October 26, 2006

Mechanical Math

My generation was the last to use a slide rule for calculation. I used a slide rule from my high school days in the early 1960s, and I used it throughout college. The slide rule had been in use for more than 300 years at the time. It was invented by William Oughtred at Cambridge in 1630, and at that time it was a pair of loose rods you would slide against each other. The whole process was made possible by the discovery of logarithms by John Napier, a Scottish physicist and mathematician. As we all know, the addition and subtraction of logarithms is a proxy for multiplication and division.

The slide rule met its demise shortly after 1972, when Hewlett-Packard introduced its HP-35 handheld scientific calculator as a way to increase sales of its seven segment LED numerical displays. The HP-35 was no bargain, since it sold for $395 (about $1,800 in today's money). Also, the operator needed to input data in a strange sequence called Reverse Polish Notation (RPN). It may have been strange to the operators, but RPN is an obvious and efficient way to program a stack calculator.

Hewlett-Packard's hegemony in the calculator arena ended in January, 1974, with the introduction of the Texas Instruments SR-50 scientific calculator at $169.95, less than half the price of the HP-35. It used ordinary "infix" notation, rather than RPN, and this was a big selling point for some customers. It is conjectured that the "SR" in SR-50 was an homage to the slide rule. What followed was the precipitous decline in calculator prices. A typical "four-banger" is now just a couple of dollars, and an exquisite scientific calculator can be purchased for under ten dollars.

1. The Museum of HP Calculators.
2. The Calculator Museum Web Page.
3. Forth, one of my favorite programming languages, is stack-oriented.

October 25, 2006

Materials in China

In a previous post, I discussed the decline in US competitiveness in science and engineering. A recent editorial [1] by Cordelia Sealy, the Editor of Materials Today, addresses the decline of US research in materials science and engineering, and a corresponding increase in materials research in China. Here are some of her findings.

• Although the total number of authors of materials-related papers is increasing, materials-related publications by US authors have declined since the year 2000.

• Chinese authors of materials publications are increasing.

• China has overtaken the US, Japan, and the UK in materials publications in ceramics, composites, and metals

• Although the US and Japan still maintain a significant lead in publications in polymer science, electronics, and electrical engineering, China is catching up.

Should this be that surprising? China has vast mineral deposits, especially of the rare earth metals, and a long history of materials technology. Porcelain originated in China at about 25-220 AD. In a previous post, I mentioned the pigment, Han Purple (BaCuSi2O6), which was developed in China around that same time.

1. Cordelia Sealy, "Editorial: As one of the oldest applied sciences, what does the future hold for materials science?" Materials Today vol. 11, no. 9 (Elsevier Ltd., November, 2006).
2. CIA Factbook - China

October 24, 2006

Supply and Demand

In yesterday's post, I commented on the high price of copper, a price so high that a penny, which is only 2.4% copper, is now worth more than a penny. In this case, there are economic forces at work in both the supply side (decreased production) and the demand side (more copper needed in China). Sometimes, an increase in a commodity price arises from manipulation of the market.

Titanium is a useful aerospace material. It is strong, ductile, oxidation resistant, and refractory. The ultimate tensile stress of titanium is the same as that of high strength, low alloy steels, but it is significantly lighter. It is so important to aerospace that the Soviet Union tried to corner the world market for titanium during the Cold War to prevent the production of US military aircraft. The CIA established a European front company to obtain titanium, and the SR-71 reconnaissance aircraft was produced from Soviet titanium.

Tantalum is another commodity prone to wild price fluctuations. Its oxide, Ta2O5, is extremely useful for fabrication of capacitors, since it has a high dielectric constant (about 25) and low electrical loss. Tantalum has had two price bubbles, one in 1979-1980, and the other in 2000, when its price soared to five time the usual value. Rhodium price fluctuations have been more severe. In one period, 1990-1991, the price jumped from $2k to $5k per Troy Ounce. Rhodium was as low as as $225 in 1996, and is $5k once again. The high price of rhodium in the late 1980s caused increased production in South Africa, and this led to the price decline to the 1996 level. Russia is the major rhodium producer, and interruption in the Russian supply has caused the sharp price increase to date.

Here are some current metal prices.

• Aluminum - $1.14/lb.
• Copper $3.61/lb.
• Tin - $4.52/lb.
• Lead - $0.50/lb.
• Zinc - $1.61/lb.
• Nickel - $15.15/lb.
• Gold - $634.20/troy oz.
• Silver - $11.325/troy oz.
• Platinum - $1241.60/troy oz.
• Mercury - $7.68/lb.

What's a Troy Ounce? It's 480 grains, or 28.349523125 g, which is 10% more than an avoirdupois ounce. The name derives from Troyes, France, an important trading city in the Middle Ages. There are twelve ounces in a troy pound. The likely reason - twelve is divisible by 2, 3, 4, 6, and 12.

1. Historical Statistics for Mineral and Material Commodities in the United States (USGS).
2. Historical Metal Prices (kitco.com).

October 23, 2006

In for a Dime, in for a Dollar

How much is a penny worth? That sounds a lot like the question, "Who's buried in Grant's Tomb?" With the current surge in copper prices, a penny is worth more than a penny. Last week, the metals market value of a penny was 1.00983 cents. A penny has been more than just copper for some time. A penny was 95% copper from 1944 - 1982, but pennies have been a 97.6% zinc core with 2.4% copper plating since 1982.

The ravages of inflation have forced changes in the composition of other coinage over the course of time. Here are the present metals market values of some historic coinage:

• 1909-1982 Cent, $0.0232822 (232.8200%)
• 1946-2006 Nickel, $0.0711509 (142.3000%)
• 1942-1945 Nickel (35% silver), $0.6598 (1319.6000%)
• 1946-1964 Dime, $0.8485 (848.5000%)
• 1932-1964 Quarter, $2.1201, (848.0400%)
• 1948-1963 Half Dollar, $4.2402( 848.0400%)
• 1964 Half Dollar, $4.2402 (848.0400%)
• 1965-1970 Half Dollar (40% silver), $1.7345 (346.9000%)
• 1921-1935 Peace Dollar, $9.0665 (906.6500%)
• 1971-1976 Eisenhower Dollar (40% silver), $3.7046 (370.4600%)

Copper items are priced at all time highs. Copper wire prices doubled in the first quarter of this year, and copper pipe price has risen almost 75 percent in the last six months. The cause of these high prices is attributed to strikes at african copper mines, and the expansion of markets in China. Copper is not the only metal demanding high prices this year. Nickel is at its highest price in nineteen years, and titanium is in short supply.

Who's buried in Grant's Tomb? Technically, no one, since Grant's body is entombed above ground and is not really buried. This riddle was made popular by Groucho Marx in an early television game show, "You Bet Your Life". He would accept any answer, so each contestant would walk away with some prize.

1. Metals values of coins (coinflation.com).
2. Jennifer Marrs, "Penny for your thoughts: Cost of copper way up."
3. Chanyaporn Chanjaroen, "Nickel Climbs to 19-Year High as Mine Blockade Curbs Supply."

The Pac Man Defense

William Bendix (1906-1964) was an actor who played Chester Riley, an aircraft assemblyman and father of a typical 1950s american middle-class family in the television series, "The Life of Riley". I remember him also in Don Juan Quilligan (1945). In that film, he played a barge captain who had a wife at each end of his weekly run between Brooklyn and my hometown, Utica, New York.

Yesterday, our Honeywell Aerospace compatriots from South Bend, Indiana presented an overview of their history in a teleconference for our Advanced Technology organization. Their presentation featured another Bendix, Vincent Hugo Bendix (1882-1945), who is an important part of Honeywell's history. Vincent Bendix was an automobile manufacturer who invented the Bendix drive, an automatic clutch that made electric starters practical for automobiles and aircraft. Bendix founded the Bendix Brake Company in 1923, and the Bendix Aviation Corporation in 1929. These were merged into Bendix Corporation.

Bendix specialized in hydraulic systems for braking and flap activation, developed a pressure carburetor used in most aircraft prior to World War II, and manufactured many electrical and electronic aircraft instruments, including radar systems. Bendix manufactured almost all avionics for US planes in World War II. After the war, Bendix launched several consumer products, including clothes washers, radios, phonographs and televisions. It aslo produced car radios for Ford and other automobile manufacturers. Bendix even made a minicomputer, the Bendix G-15. The Bendix computer division was sold to Control Data Corporation in 1963.

In 1982, Bendix Corporation was acquired by Allied Corporation during a heated and protracted attempt by Bendix CEO William Agee to acquire Martin Marietta. Bendix had acquired a majority stake in Martin Marietta, but Martin Marietta's management used the short time before change in control to sell some of its businesses and launch a hostile takeover of Bendix. This strategy has become known as the Pac-Man Defense; namely, eat your enemies before they eat you. All this leads us back to "The Life of Riley." At the end of each show, Chester would look into the camera and say, "What a revolting development this is."

Martin Marietta eventually merged with Lockheed Corporation to form Lockheed Martin in 1995. AlliedSignal acquired Honeywell in 1999 and took the Honeywell name. The Bendix name lives on in Honeywell's Bendix/King brand of avionics.

By the way, my maternal grandfather worked as a machinist for Bendix.

1. William Bendix on the Internet Movie Database.
2. Bendix/King Web Page
3. A somewhat biased perspective on the Bendix-Martin Marietta affair.

October 19, 2006

Mpemba Effect

Many years ago, my sister-in-law asked me why ice cubes froze faster when starting with hot water, rather than cold water. I told her I didn't think that was true, but if it were true, it would likely be caused by the one or more of the following:

• The hot water could trigger the thermostat in her freezer to supply excessive cooling.

• Evaporative cooling of the water.

• Since hot water evaporates, there is less water to cool.

• Convection caused by the temperature gradient between the hot ice cube tray and the cool surface of the water would make heat transfer more efficient.

This was done in true scientific fashion - I gave her a lot of technical information, but didn't answer her question. What I didn't know at the time was that hot water does freeze faster than cold water, and this had been observed thousands of years ago. Aristotle, in his Meteorology (Book I, Chapter 12), writes

"The fact that the water has previously been warmed contributes to its freezing quickly: for so it cools sooner. Hence many people, when they want to cool hot water quickly, begin by putting it in the sun. So the inhabitants of Pontus when they encamp on the ice to fish (they cut a hole in the ice and then fish) pour warm water round their reeds that it may freeze the quicker, for they use the ice like lead to fix the reeds."

Except for some observations by Francis Bacon and René Descartes, the existence of this effect languished for all those centuries, until Erasto B. Mpemba, a Tanzanian high school student, noticed the effect while making ice cream from hot mixes. Mpemba subsequently published his experiments with Denis G. Osborne in 1969, and since that time it has been known as the Mpemba Effect. An extremely detailed review of the Mpemba Effect has been published by Monwhea Jeng of the Department of Physics, University of California.

Scientists have proposed some explanations for the Mpemba Effect beyond the few I detailed for my sister-in-law:

• A different definition of freezing, perhaps looking just for the point when a surface layer of ice forms, rather than waiting for the volume to become a solid block of ice.

• The insulating effects of frost.

• The boiling process releasing dissolved gases.

• Cold water supercools, and hot water does not.

• The effect of impurities, such as calcium carbonate and magnesium carbonate.

Jonathan Katz of the University of Washington, St Louis, says the Mpemba Effect is all a matter of impurities - heating precipitates out solutes. In an explanation all Materials Scientists can relate to, Katz reasons that cold water, as it freezes, concentrates solutes in the remaining liquid fraction. Because of the well known principle of freezing-point depression, the remaining liquid water becomes harder to solidify.

1. Marcus Chown, "Why water freezes faster after heating," New Scientist (3 June 2006).
2. Eric Berger, "Does hot water really freeze faster than cold?"
3. Physics Web, "Does hot water freeze first?"
4. Monwhea Jeng, "Can hot water freeze faster than cold water?"

October 18, 2006

Ice-Nine and Metallic Water

In Kurt Vonnegut's novel, Cat's Cradle, a phase of water ice is discovered to have unusual properties. A crystal of this novel phase, ice-nine, acts as a seed that spontaneously crystallizes water under normal temperatures and pressures. The problem is that ice-nine will only melt at high temperature (45.8 degrees Celsius), slightly above Earth ambient. Ice-nine eventually leads to the solidification of all the world's water, and destruction of life on earth.

There actually is an "ice nine," but it's called ice IX (Roman numeral nine). Unlike Vonnegut's ice-nine, ice IX only exists below 140K at pressures from 200 - 400 MPa. It has a density of 1.16 g/cc. There are twelve crystallographically distinct phases of ice.

Recently, Thomas R. Mattsson and Michael P. Desjarlais from Sandia National Laboratory (Albuquerque, New Mexico) have used supercomputer simulations to refine the phase diagram of water. One thing they found is that the phase boundary for metallic water occurs at lower temperature (4,000 instead of 7,000 Kelvin) and lower pressure (100 GPa, down from 250 GPa). As its name implies, metallic water behaves as a metal, and the electrical conductivity of metallic water may contribute to the magnetic field of the planet Neptune. Their work has been reported in Physical Review Letters [2].

1. Phase diagram of water revised by Sandia researchers
2. Thomas R. Mattsson and Michael P. Desjarlais, "Phase Diagram and Electrical Conductivity of High Energy-Density Water from Density Functional Theory," Phys. Rev. Lett. 97, 017801 (2006)

October 17, 2006

Ununoctium Discovered

A paper last week in Physical Review [1] by Russian physicists at the Joint Institute for Nuclear Research (Dubna) and their american collaborators at the Lawrence Livermore National Laboratory announced the discovery of Element 118, Ununoctium. This name, the IUPAC systematic name, is only a placeholder, built from the combination un-un-oct, or 118. There is a tradition (not without controversy) that the discoverers are allowed to name their element, so the element will be given another name within a few years. Ununoctium sits below Radon in the Periodic Table.

The Dubna team bombarded californium (element 98) with a high energy beam of calcium (element 20) to produce Ununoctium (118 = 98 + 20). Sightings of heavy elements are fleeting, and the evidence is sparse. The only evidence of the element in this case was its decay products.

The history of Ununoctium is as interesting as its discovery. Physicists at the Lawrence Berkeley National Laboratory (LBNL) reported the synthesis of three nuclei of Ununoctium in 1999 through the impact of krypton (36) on a lead (82) target, but the discovery was retracted in July, 2001, when the discovery could not be reproduced [2-4]. Worse still, one member of the research team had apparently fabricated the data, and when a formal inquiry was held, no record of the discovery data could be found. Berkeley Lab Director Charles Shank had this observation.

"Science is self-correcting. If you get the facts wrong, your experiment is not reproducible. In this case, not only did subsequent experiments fail to reproduce the data, but also a much more thorough analysis of the 1999 data failed to confirm the events. There are many lessons here, and the lab will extract all the value it can from this event." [5]

1. Yu. Ts. Oganessian, et al., "Synthesis of the isotopes of elements 118 and 116 in the 249Cf and 245Cm+48Ca fusion reactions," Phys. Rev. C 74 (October 9, 2006).
2. Controversy-Plagued Element 118, the Heaviest Atom Yet, Finally Discovered
3. "Element 118 Bows Out," Physics Today (September 2001), page 20
4. Lawrence Berkeley Lab Concludes that Evidence of Element 118 Was a Fabrication, Physics Today (September 2002), p.15
5. Results of Element 118 Experiment Retracted

October 16, 2006

Down the Tubes

When I first moved to New Jersey, I enjoyed eating at Victoria Station, a local restaurant named after the eponymous London railway station. The restaurant closed shortly thereafter, but it introduced me to British Railway memorabilia, which served as the restaurant's decor. One of the most interesting pieces of such memorabilia is the London Underground Tube map created by an engineering draftsman (or, should we say, draughtsman) named Harry Beck.

What's so special about Beck's map? Beck realized that traditional maps, laid out geographically and superimposed on a street map, were not the best representation of the London Underground. Stations in the heavily populated central regions were too close together to read, and the peripheral stations were isolated at far corners of the map. Beck decided to draw a map in the style of an electrical schematic, in which the connections are important, and not the actual positions of the components. It was a logical engineering decision that transcended conventional wisdom, and it is now copied by many rail lines, such as SEPTA in Philadelphia. It is not copied by the New York Subway System, presumeably because New York City is uniformly dense.

Beck's map eventually fell out of his control and, for a time, strayed from the original design concept. However, the logic of his original concept persevered and it is maintained in the present map, which still credits Beck. Beck, who died in 1974, is remembered in an exhibit at the London Transport Museum, and by a commemorative plaque at his home station, Finchley Central. Beck's map came in second only to the Concorde as the best twentieth century design in the Great British Design Quest of 2006.

1. Ken Garland, "Mr Beck's Underground Map," (Harrow Weald, Middx: Capital Transport, 1994), ISBN 1-85414-168-6.

October 13, 2006

Superconductors Super Again

In a previous post, I reported the prediction by two German scientists, Werner Marx and Andreas Barth, that research in high-temperature superconductors (HTSC) is in decline, and the field will be dead sometime between 2010-2015. The prediction arose from an extrapolation of the citation history of the field.

As could be expected, researchers in HTSC had serious complaints about the paper, which existed only in preprint form [1] on the arXiv web site. As a consequence, the authors have revised their paper to remove this prediction. Paul Grant, a noted HTSC scientist, says that there was some "heavy leaning on them" by the HTSC community to remove the prediction [3]. Marx, however, says that there is no argument about the data. "The data and the conclusions concerning the significant decrease of basic-research publication output (activity) around cuprate high-temperature superconductors between 1985 and 2006 remain the same." Everyone agrees that since the study was based on citations in physics and chemisty, applied research areas such as engineering were missing from the analysis.

1. Andreas Barth, Werner Marx, Mapping High-Temperature Superconductors - A Scientometric Approach (23 page PDF file)
2. Slow death for a hot topic (Matin Durrani)
3.Superconductivity fights back (Nature, October 10, 2006).

October 12, 2006

Genetic Engineering of a Different Kind

Jeroen Domburg is a student at the University Twente (De Universiteit Twente) in Enschede, Netherlands. The University is described as an "entrepreneurial research university," that focuses on both technology and its sociocultural aspects (after all, this is The Netherlands). It is the only campus university in the Netherlands.

Domburg has published details of how he made an inexpensive, hand-held ink jet printer. This printer is made from an individual printer cartrdge, and it is capable of writing on irregular surfaces. The HP 51640M cartridge he used is not just an ink supply. It contains a MEMS print head with more than a hundred nozzles. Although many of you may be interested in the printer project per se, what interested me was his approach to determining the unpublished connection diagram of the printer cartridge.

Although the cartridge has more than a hundred nozzles, there are only 32 connecting pads. Since there are many circuit options, from shift registers to a simple matrix, that could be used by the integrated circuit inside the cartidge to drive the MEMS printer element, finding the proper connections seems like an impossible task. Domburg used an approach he borrowed from genetic engineering. The idea is to make a "gene" defective, and observe what changes occur in the "organism." He covered connector pads with tape and observed the affect on a printer. With only three pins exposed, the printer drew a single, thin line. Through more experiments, he found that the cartridge is organized as eight sectors with seventeen nozzles each. Using this non-biological version of genetic engineering, he discovered the best connection of nozzles for his printer.

By the way, the main page for Jeroen Domburg's web site has an example of a captcha, as described in a previous post.

October 11, 2006

Carbon Neutral Computers

In a previous post, I commented on the 130 watt peak power requirement of the Pentium D 840 computer chip from Intel. Now, VIA Technologies has introduced what they call the World's First Carbon Free Processor. VIA has calculated the carbon equivalent of the energy used by their VIA C7-D processor over a presumed three year lifetime under usual operating conditions. They then fund reforestation, alternative Energy, and energy conservation projects that offset the amount of carbon generated by the processor. VIA has also introducing a carbon benchmark, TreeMark, which is essentially a graphic image of the number of trees needed to remove the carbon generated by a product over its lifetime.

The VIA energy calculation, which can be applied after some modification to nearly every electric system or component, is as follows:

• The peak wattage is multiplied by a 36.8% usage factor (an average of 27% standby mode, 3% sleep mode, 67% idle mode, and 3% peak performance mode)

• This is then multiplied by the number of hours in its lifetime (three years = 26,280 hours) to get watt-hours; and divided by 1000 to get kilowatt-hours (kWh).

• This is then multiplied by a global average of carbon dioxide produced per kWh (0.501 kg/kWh)

• This is then divided by an estimated carbon dioxide sequestration rate for the average broad leaf tree (7.3 kg/year, or 21.9 kg over the three year lifetime).

VIA needs to "plant" only four trees for its VIA C7-D processor. The Pentium D 840, at 130 watts, needs 28 trees (VIA seems to give Intel a small break, listing the processor at only 26 trees). Another popular chip, the AMD Athlon 64, needs twenty one. Of course, the VIA C7-D processor is not as fast as either the Intel or AMD chips, but it's the only zero carbon chip as far as the end user is concerned. VIA also lists other energy uses equivalent to the lifetime carbon dioxide emissions from the VIA C7-D processor (97 kg):

• Driving 520 kilometres (323 miles) in a typical gasoline fueled car.
• Driving 930 kilometres (578 miles) in a Toyota Prius hybrid car.
• Flying 320 kilometres (199 miles).

Morristown, NJ, to Richmond, Virginia is a "one VIA" trip by automobile (350 miles on Google Maps), or four trees.

1. VIA energy calculation white paper (four page pdf file).
2. "CPU pins growth on green projects," EE Times (September 18, 2006), p. 22.

October 10, 2006

100,000 Digits of Pi

I've wanted to write an article on pi, and now I have an excuse. Last week, Akira Haraguchi, a sixty year old mental health counselor in Japan, recited 100,000 decimal places of pi from memory. This feat took sixteen hours, and it was videotaped so that it could be authenticated by the Guinness Book of Records. Haraguchi breaks a previous record of Hiroyuki Goto, also from Japan, who recited 42,195 decimal digits in 1995.

How difficult is this? Homer's Iliad has 15,693 lines of dactylic hexameter verse of about six words on each line, for an appoximate total of 94,000 words. The Iliad was recited from memory by Greek bards for centuries before it was written down, so I would argue that the tasks are roughly equivalent. This may be the limit for memorization. Nevertheless, this task would have been impossible half a century ago, before the advent of the digital computer, since pi was known to only a few hundred decimal places. For example, D. F. Ferguson was able to calculate pi with a mechanical desk calculator to just 808 decimal places in 1947. The present record for computer calculation is more than a trillion digits.

The decimal expansion of pi never ends. Pi is a transcendental, irrational number with a number sequence that does not repeat. How many decimal places of pi are really needed? With only fifty digits you can calculate the circumference of the universe to within a proton's width. I've memorized eleven places, which is definitely overkill for all practical calculations. Indeed, a float in the C programming language has just seven digit accuracy. The IEEE 754 floating point standard for 64 bit arithmetic has only 16 digit accuracy.

1. Man recites pi to 100,000 places (Yahoo! News)
2. Rice balls power pi record
3. Man Recites Pi to 100,000 Places (happynews.com)

October 09, 2006

Quality Functional Deployment

On Friday, I attended a performance by Philip Glass in Allentown, PA. Philip Glass has performed often in New York City (NYC), and since Morristown NJ, is often thought to be just a suburb of NYC, most would think that my travelling an hour and a half by car to hear Philip Glass is unusual. My decision was based on several personal preferences.

• Travel - Highways around NYC are terribly congested. The bridges, or the one tunnel into the city from New Jersey cause considerable delay, and travelling inside the city is difficult. The alternative, a commuter train, still involves driving to, and parking at, a train station, and the train ride takes an hour. Allentown is a direct route via interstate highways. Yes, there is one toll, but it's $0.75, quite a bit less than the $6.00 entrance fee into NYC.

• Venue - New York City is horribly expensive. A concert ticket there would be at least twice that of Allentown. Symphony parking in Allentown is free. There are the tolls just to enter NYC from New Jersey, and once you arrive, parking would cost about $50.00. Dinner for two, if you can get reservations, would be at least double what it would cost in Allentown.

• Other Factors - Our daughter lives within an hour's drive of Allentown, so she and her boy friend joined us for dinner.

This decision-making process of weighing factors and deciding an outcome is done by people every day. Some decisions, such as which car to buy, or deciding whether to retire, involve many factors and take a considerable time. Some business and technology decisions are even more complicated. Is there a good process for making decisions, whether small or large?

In my mind, one of the most powerful Six Sigma tools is Quality Functional Deployment (QFD). QFD is essentially this:

• You decide what factors are important.

• You decide how to score these factors, usually on a 0-5 or 0-10 scale; for example, when ranking some thing as qualitative as the beauty, 0 = Downright Ugly, 5 = Agreeable, and 10 = Beautiful.

• You assign relative weights to the importance of these factors. Sometimes cost is very important, and that may have triple the weight of the next most imporant factor.

• Then you just rank each factor for each alternative, multiply by the weight, and sum the scores for each alternative. Often there is such a wide gap of scores, the best alternative is fairly obvious

• You might also normalize the scores to give 100% for the best alternative. This makes comparisons easy to see.

Several years ago, I was a member of a Morristown team that decided which of two alternative locations would be best for the relocation of our technical library. I advocated using QFD to make the decision, and it became an easy decision to make [1]. You can download the QFD chart we used, here, and perhaps use it as a template for your own decisions.

The next time you need to decide which mobile telephone service to use, try QFD. By the way, Theodore Roosevelt and Woodrow Wilson spoke in the Allentown Symphony Hall during the campaign for the 1912 Presidential Election.

1. QFD Chart Example - Morristown Technical Library Relocation (Excel Spreadsheet).

October 06, 2006

American Wins Chemistry Nobel Prize

Roger D. Kornberg, a Professor of Structural Biology at Stanford University, has been awarded the 2006 Nobel Prize in Chemistry for his research on transcription, how information stored in DNA leads to protein synthesis via messenger RNA. The 2006 Nobel Prize in Physics was awarded earlier in the week to two American physicists.

In a previous post, I listed two Nobel Prize winning physicists whose fathers had also won the Nobel Physics Prize. This phenomenon is not just limited to Physics. Roger Kornberg's father, Arthur Kornberg was awarded the Nobel Prize in Medicine in 1959. Roger and Arthur Kornberg are the sixth father and son winners of a Nobel Prize. The elder Kornberg, like his son, was a professor at Stanford. There was also one father and daughter combination - Pierre Curie and his daughter Irène Joliot-Curie. Pierre's wife and Irene's mother, Marie Curie, quite significantly won two Nobel prizes, one in physics, the other in chemistry.

Kornberg, as the sole winner of this year's award, gets a cool $1.4 million. Kornberg was born in 1947, the same year I was born. My Nobel must be a little late.

October 05, 2006

Safety Day

Today is Safety Day at the Morristown laboratories - a time to hear inspirational safety talks; and also a time to clean your office and lab.

There is a story about Albert Einstein's tour of a university department. The department chairman gave orders for a very thorough laboratory clean-up in preparation for the visit of this famous scientist. One vacant laboratory was designated as the dumping point for old equipment and other debris.

Einstein arrived, and the department chairman showed him the various labs. Einstein seemed unimpressed. Towards the end of the tour they were in the corridor outside the designated dumping lab, when a student opened the door to exit the lab. Einstein looked inside and exclaimed, "So, this is where all the work is done!"

1. Occupational Safety and Health Administration.
2. Morristown HS&E Web Site.

October 04, 2006

The Woodstock of Physics

The March, 1987, meeting of the American Physical Society in New York City is often called the "Woodstock of Physics." That meeting launched the field of High Temperature Superconductivity (HTSC) to standing-room-only crowds. It was reported that so many copies were made of the preprints of the most important manuscripts that they were bleached nearly white by the high intensity lights of the copy machines. Now it appears that the field is in decline.

A preprint [1] posted on the arXiv web site with the innocuous title, "Mapping High-Temperature Superconductors - A Scientometric Approach," is being interpreted as a prophecy of the death of HTSC by 2015. The authors, Andreas Barth of the Fachinformationszentrum - Karlsruhe, and Werner Marx of the Max Planck Institute for Solid-State Research - Stuttgart, analyzed citation numbers for the field, which peaked around 1990. Citations for HTSC papers were about 8500 in 1990, but they have eroded to 4400 in 2005. An extrapolation of the data gives zero sometime between 2010-2015. One possible reason for this decline is that researchers are being drawn to more promising fields, such as nanotechnology. Also, there is a complaint that the many theoretical papers in HTSC have given no useful guidance to physicists and materials scientists hoping to synthesize new materials.

1. Andreas Barth, Werner Marx, "Mapping High-Temperature Superconductors - A Scientometric Approach" (23 page PDF file)
2. Slow death for a hot topic (Matin Durrani)

October 03, 2006

Nobel Physics Prize Awarded to Two Americans

John C. Mather of the NASA Goddard Space Flight Center, and George F. Smoot of Lawrence Berkeley National Laboratory, were awarded the 2006 Nobel Prize in Physics for investigations into the origin of the universe. Using the NASA Cosmic Background Explorer (COBE) satellite, they observed fluctuations in the cosmic microwave background radiation. They will share $1.4 million.

A Starbucks on Every Corner

The late 1960s was the start of a downwards spiral in physics employment. Many physicists endeavored to apply their skills to other areas, and some became quants. A quant, short for "quantitative analyst," is Ph.D. physicist or mathematician who applies his modeling skills to the financial markets. Who better to apply things like stochastic calculus to market models?

Pablo Jensen, a computational physicist who has a joint appointment with the Laboratoire de Physique and the Laboratoire d'Economie des Transports, Lyon, France, has published a paper in the current issue of Physical Review [2] that expands this tradition. He has applied models used for atomic interaction to the problem of finding the optimum location for a business. Jensen reasoned that, just as atoms of one species attract or repel atoms of another, he could model how stores of different types seem to cluster. Of course, all of us have noticed that fast food stores tend to cluster, but a Starbucks drives away all other coffee shops.

Jensen examined the types of stores within 300 feet of each other, and he found, for example, that jewelry stores prospered near other clothing accessory stores, but not around grocery or hardware stores. Confirmation of his model came from a comparison of bakery locations in 2005 and 2003. Bakeries that existed in 2003, but were found by his model to have a poor location, tended to disappear by 2005. Jensen has published, on his personal web site, a map showing the best locations to open a bakery in Lyon.

While on the topic of French bakeries, a "boule," from the french word "ball," is a round loaf of white bread. It is also the word crystal growers use to describe large, synthetically grown, single crystals.

1. Charles Q. Choi, Atomic Physics Predicts Successful Store Location (Live Science).
2. Pablo Jensen, Network-based predictions of retail store commercial categories and optimal locations, Phys. Rev. E 74, 035101(R) (2006).

October 02, 2006

Little Things Getting Bigger

The September, 2006, issue of R&D Magazine contains a mini-market study on nanotechnology from the Helmut Kaiser Consultancy. As could be expected, the nanomaterials market is strong, and it's getting bigger. Here are some highlights.

• Presently, the US nanotechnology R&D effort is $26 billion.

• The National Science Foundation projects a worldwide nanotechnology market of more than $1 trillion by 2015.

• The present nanomaterials market is $108 billion. It is expected to become more than $275 billion by 2015.

• The present Aerospace nanotechnology market is $12 billion, and will reach $58 billion in 2015.

• The biggest market segment is Electronics, with $129 billion presently, surging to $246 billion in 2015.

• There were nearly 90,000 patents with nanotechnology content in the period 1976-2002.

• Nanotechnology research does not generate only de novo materials. It is an enabling technology that allows the application of existing materials in novel ways (e.g., new drug delivery systems).

There are many applications of nanotechnology in my own field of Aerospace materials. One principal application is coating of components for oxidation protection, corrosion protection, and as a thermal barrier. One technique, Atomic Layer Deposition, allows coating of components with deep, internal passages - something that is impossible by vapor phase techniques because of shadowing, or by liquid coating because of surface tension.

Nanotechnology has had its critics. David M. Berube, who is not a scientist, but a professor of English at the University of South Carolina, published a book, "Nano-hype: The Truth Behind the Nanotechnology Buzz," Prometheus Books, 2006 (ISBN 1-59102-351-3). His opinion is that much of what is called nanotechnology is just standard materials science in new clothing. Berube thinks that the nanomaterials business will evolve into a handful of suppliers selling low margin products in huge volumes, much like the chemical industry of today. There is also the potential problem of nanomaterial toxicity. Studies of airborne particles have shown that particles become more toxic as they get smaller. This may be more of a safety problem in manufacturing than in deployment of nanotechnology products.

1. Wikipedia article on nanotechnology.
2. The National Nanotechnology Initiative.
3. Chemical & Engineering News Special Report on Nanotechnology.