Have heat, need electricity?


The thermoelectric effect has been used since the 1930’s to generate electric power, but it really hasn’t been used widely in consumer products to generate electricity.  While inefficient, this are suitable for some applications with relatively low electrical power requirements, combined with extreme environmental conditions.

Historical Background

In the 1940’s, the Soviet government produced kerosene-thermocouple generators to power radio amplifiers.  Fuel up, Plug in, Tune in!


The Soviets also have used radioisotope thermoelectric generators as power supplies in remote locations, such as lighthouses and navigation beacons in far Siberia.  The US used them for remotely located radar stations in Alaska and probably northern Canada.  If you see one of these, don’t bust it open.  Inside there’s a slug of strontium-90 or plutonium, and you’ll likely die of radiation poisoning from exposure to the source inside.


Thermoelectric generators have also been a mainstay of the space program, powering deep space probes where the solar flux won’t yield much energy.


The black tubes are the RTG’s

Principle of Operation



the principle of operation is incredibly simple.  With power supplied as shown above, one side gets cold and the other gets hot.  Run in reverse, heating one side and cooling the other side, a voltage is generated.  N and P typed doped bismuth telluride (Bi2Te3) is the most efficient semiconductor for this application, but other materials have been used in the past.





From the first version, I added a 3-Farad ultracapacitor and Pololu boost regulator, which acts as a DC-DC converter, stepping up voltage from 1.5V to 2.8V, the minimum forward voltage for the LEDs.  The capacitor smoothes out the voltage supplied to the boost regulator, and provides about a minute of light with no input from the thermoelectric elements.





As assembled, the stack is as thermally efficient as possible, with a copper base plate, and thermally conductive matting between the copper substrate and the thermoelectric element.  Unfortunately, a double-sided adhesive film was only available in a product with a relatively low thermal conductivity, but the impedance is not terrible considering its thickness.

The heat sinks used here are relatively dense for the application, being intended for use in forced-convection applications.  They do dissipate enough heat to provide significant energy flow under natural convection in this arrangement.


Most of the energy is dissipated as heat, but a small amount is converted to electricity.


Based on this chart from the manufacturer, CUI, the theoretical efficiency of the elements used is 1.25%, given the temperature difference.  That’s not very good compared with even low-efficiency PV panels, but PV’s don’t work at night, requiring battery or capacitor storage.  As the thermoelectric elements work directly from heat, this is a fine solution for the application.

Doing some calculus yields the theoretical voltage of the assembly, based on the electrochemical properties of doped bismuth telluride.


Experimental tests yielded a lower output than predicted, but that is expected due to system inefficiencies.


Cost Analysis




Gasoline Generator

PbA Battery

Marine Deep Cycle

Kerosene Lantern

Lighting system

LED lights

Fluorescent Lights

Fluorescent Lights, Inverter

Kerosene Lantern

Associated Hardware

LED lights




Recurring Needs


gasoline, oil

battery replacement

kerosene, mantles

Expected Life

20 years

10 years

3 years

(900 cycles)

20 years+

electrical Power Output (W)





Initial Cost





recurring costs


$2/day for fuel

$0.25/ day for recharging

$1/day for fuel

(US white gas price)

First year cost





Second+ year cost





5 Year Cost





toxic impact


gasoline, oil

Lead, Sulfuric Acid

Mantles, Hydrocarbon Fuel




 Sulfuric Acid



The total cost of the thermoelectric system is about $200 US, substantially less than the other options for portable electrical power/lighting used in food carts.  The cost breakdown above does not include lighting for the gasoline generator and lead-acid battery installations, which would total about $100-200 more.  The amount of electrical power generated is orders of magnitude greater for the gasoline generator and battery arrangements, making those suitable for more power-intensive situations like an ice cream truck.  But if all you have is a griddle, and all you need is some light, thermoelectric is the way to go.



This is a promising technology with many potential applications.  While not in wide use, it could be used to good effect to provide lighting, displacing other energy sources which are highly polluting and have relatively high costs.  The barrier to entry into the market is not insurmountable with a kit solution, and in volume the cost of devices would drop.

An issue is the fragility of the commercial elements I used – they can be overheated, which was a problem I ran in to, but there are higher temperature rated elements available.

Smartphone Ubicomp Device, Outlined

Environmental Impacts

  • Resource extraction

Exotic metals
Oil (plastics and energy)
Recycled materials, where possible
Virgin materials – especially in components are a real problem
  • Production

Subcomponent processing
Toxic materials use (BFRs, blowing agents, solvents)
Energy input
  • Transport

Diesel for the boats, trucks, and trains
Aviation fuel for air transport
  • Use

Energy use for charging
Toxic materials exposure
Replacement parts: battery, repair parts service
  • End Of Life

Developed nation reuse: ReCellular-type donation to needy individuals
Responsible recycling in developed nations for resource recovery
Transshipment to developing nations
Reuse as a mobile device
Resource recovery by backyard methods is highly hazardous, and often involves child labor 
  • Human Health impacts


Radiation has not been conclusively linked to carcinogenis, but studies are ongoing as to whether microwave radiation from mobile devices is a hazard. The Interphone study of 6000 individuals with brain cancer did not find any link, nor did a 420 000 person Danish study find any conclusive link.
As a marketing issue, radiation probably is not as large a concern as it was in the past, and as the radiation output is proportional to radio transmitter strength, the current emphasis on signal quality is likely to overrule concerns regarding radiation exposure.

Toxic Materials

Reducing the exposure to toxic materials by end users is an important design goal, and also a strong marketing position. Marker Ski Bindings slogan "where others use plastic, we use stainless steel" stands as a testament to the power of materiality. Making the shell out of metals and glass will reduce environmental impact at the end of life, extend the life of the product, and reduce toxic exposures. The additional embodied energy added in processing will be paid off in increased durability, and with the reduced need for internal components and fasteners, the burden will be similar or possibly reduced compared to conventional (2010) devices.

Reduction of toxic materials will improve worker health and safety as well, as their exposures to hazards will be reduced, and the risk of serious exposure will also decrease.

Reinvention – the Smart Ubicomp Device

The iPhone has shown the value of vertical integration of the device and the application services for it. Unfortunately, Apple’s MobileMe isn’t nearly as slick, and Microsoft’s ecosystem integration isn’t quite perfect, but synchronization with cloud data services and with your other computers really is the important part of the smartphone experience. I see the future being a hybrid-cloud model where there is local storage of data and cloud storage, and the smartphone is a bridge between the two. Sorry Google, the cloud isn’t perfect. Too many outages, too many places where there is no data connection (like on that transpacific flight …)


Videoconference (it’s coming, eventually)

Camera (still/video)

The book The Best Camera Is the One That’s With You makes the case for the iPhone as a primary camera. While the iPhone doesn’t have a flash, zoom lens, or particularly good optics, it is always there. Better cameras and noise suppression could make this even more viable substitute for a point-and-shoot.

Web access

It’s here. Mobile web will continue to improve, as will methods for navigating.

Augmented Reality

Augmented reality is an up-and-coming thing at ITP, which means it’s going to be a big thing soon. Context awareness is going to be pretty amazing. Crazy games, real time advice, and even better navigation platforms.

Application Support

Applications ranging from a virtual float level to Word Mobile make it possible to work and play anywhere, and in terms of convergence, the smartphone is where it’s going to be.  The App Store really revolutionized the concept of the smartphone as part of a vertically-integrated ecosystem, with an easy mode of getting quality software.  While the high-handedness of Apple is a subject for endless debate, the concept of simple access really made the smartphone as application platform a reality for most people


Smartphones already do audio and video, better outputs can replace heavier equipment like DVD players. I’d be pretty stoked to dock my phone to a pico projector and project some HD video anywhere, or connect digitally to a home receiver. DLNA is almost there, but people haven’t really been using it yet. Unfortunately for me, my late-70’s receiver doesn’t receive 2.4Ghz signals…

Data collection

Data collection from mobile devices with sensors and wireless network access has the potential to make many systems and services better. The possibilities are really incredible in terms of benign surveillance (one would hope benign)—tracking the speed of traffic on the highway, to detect traffic jams before they really cause a blockage, or just tracking my cycling would be good applications, as would applications like Project NOAH – Networked Organisms and Habitats (an ITP project) which uses mobiles for citizen science and education.

Persuasive Technology Platform

Smartphones are an ideal persuasive technology platform, as they are always with us, and offer an enormous amount of sensing and communication ability. Mobile platforms for health promotion and sustainability will become more prevalent and powerful in the years to come, pushing the

The PEIR project ( uses mobile sensing to analyze travel patterns, and encourage healthy eating and lower impact travel options (e.g. take the train rather than driving).

Replacement of other devices

The power of smartphones and their omnipresence makes it possible to replace a group of other devices with one ‘convergence device’. It’s a camera, it’s a reader, it’s all in your pocket! (hence the former name PocketPC)

The kindle, iPad, and similar devices bridge the gap between a smartphone and notebook PC, but are still a bit to carry around all the time. Improved screens and applications could make it more pleasant to read on a smartphone screen. I for one use my smartphone to read the news, which eliminates a rather large amount of paper and shipping for newspapers and periodicals. The networked nature of electronic devices opens a new frontier of interaction, which can be exploited in ways I cannot conceive.

Political dimensions

Mandate for return of used devices to the manufacturer

As mobile devices are almost completely composed of ‘technical nutrients,’ a closed loop cycle should be the paradigm of the future. Currently, the system promotes replacement of devices when the contract rolls over (generally every two years), but there is no promotion of recycling. Policy should be changed to require the return of the current phone before receiving a new one, and a penalty should be assessed if a new device is purchased without trade-in, to cover the additional cost to the system of not having a device to recycle. An exception for first time purchasers could be instituted, as smartphones are a positive force in modern life for sustainability, by convergence and as a persuasion platform.

Separate manufacturers from service providers

In most countries, wireless service is independent of device purchase. This arrangement avoids subsidies for devices, and therefore expensive devices should turn over less frequently. If you have to buy a device for $600, you should want to keep it longer than if the purchase price were $200 + 2 year contract.

Networks which profit heavily from restrictive contracts could be reined in by policy changes: making it illegal to subsidize devices by entering into a contract would change the wireless provision structure, and carriers would then be forced to compete more aggressively on service and price. The telecom lobby is very strong, it should be more heavily regulated by the government in order to promote better service and a more sustainable product-service system

The first mover under the new regulatory regime, instituting a closed materials loop and reducing taxed bads will enjoy a great market advantage. Where Apple achieved enormous market penetration with the iPod by being the ‘first best product’, a device manufacturer (Apple or someone else) could do the same in the mobile space if there were a regulatory shift. A company like HTC, RIM, or Nokia whose largest segment is mobiles would have a serious incentive to revolutionize their practices (even beyond WEEE compliance), Apple, Samsung, LG, and Motorola would be under less pressure, as they have other business divisions.

To this end, the first mover can jump the pack by becoming a force for change—If, say, Nokia has the most sustainable products and a highly flexible OS platform, they can lobby to raise the minimum standards for mobiles in terms of sustainability, and pressure competitors to match in terms of aesthetics, features, price, and compliance, in a race to the top paradigm. While a six year life as opposed to a two year life would reduce revenue from sales, sale of applications, media, and OS upgrades (if the market would support pay-for-major upgrades), could sustain a business model.

Smartphone Ubicomp Device

The reinvented Smartphone Ubicomp Device is intended to last six years, rather than the current design standard of two, and avoids the use of plastics and toxic materials where possible.  As an element of a technical nutrients cycle, as much as possible is recyclable material—currently it is not possible to achieve 100% reclamation of materials, especially of plastics and small amounts of exotic materials.


Impacts quantified by sustainable minds for the i910 and the concept device:



    Individual Impact (%) Category Impact (%)
Ecological Damage Acidification 0.01  
  Ecotoxicity 49.89  
  Global Warming 0.21 50.15
  Ozone Depletion 0  
  Water Eutrophication 0.04  
Resource Depletion Fossil Fuel 0.01 0.01
Human Respiratory 0.02  
Human Health Damage Human Carcinogens 30.11 49.84
  Human Toxicity 19.7  
  Smog 0.01