Sei sulla pagina 1di 59

Thermoelectric Effects 1

Running head: THERMOELECTRIC EFFECTS

A Study of the Seebeck and Peltier Thermoelectric Effects

Alan Pnakovich

American Heritage School, Plantation, Florida

In partial fulfillment of the requirements for Third Period Honors Science

Mrs. Page

December 1, 2009
Thermoelectric Effects 2
55th State Science & Engineering Fair of Florida
OFFICIAL ABSTRACT and CERTIFICATION
A Study of the Seebeck and Peltier Thermoelectric Effects
Alan Pnakovich Category
American Heritage School, Plantation, Florida, Broward County, USA Pick one only- -
Mark an “X” in
This experiment explored two main principles of thermoelectricity, the
Box at right
Seebeck and Peltier effects. The Seebeck Effect generates electricity from
temperature differences between two junctions of different metals or
Behavioral and
semiconductors. The reverse, Peltier Effect, creates temperature differences by Social Science
applying electricity to junctions of different metals or semiconductors. Eight Biochemistry
hypotheses were tested varying the types and thicknesses of metal wire, amount of Botany
electricity and temperature difference applied, and exploring how modern
Chemistry
semiconductor devices could amplify Seebeck and Peltier effects.
Computers
The Seebeck Effect was investigated by applying temperature differences to Earth and
both metal thermocouple wire and semiconductor-based “Peltier modules.” As Planetary
Science
hypothesized, these semiconductor devices generated several orders of magnitude
Engineering
more voltage than metal wire. Tests with thermocouple wires proved hypotheses
regarding proportionality of voltage generated to temperature difference and Environmental
Sciences
Seebeck coefficients of the metals, and the lack of effect of wire thickness.
Mathematics
The Peltier Effect was investigated by subjecting Peltier modules to different Medicine and
current levels, which proved hypotheses regarding their cooling capability using Health
measurements from infrared and thermocouple thermometers. A DC power supply Microbiology
and multimeter with data logging capabilities were used in most tests. Physics and
Astronomy XX

Tests with the Peltier modules confirmed their effectiveness in capturing


Zoology
waste heat and converting it into electricity, an exciting innovation that auto
companies are exploring to convert exhaust heat into electricity to increase gas Team

mileage. Thermoelectricity is now becoming a direct competitor to solar as an


alternate energy source. Peltier modules are also an effective means of cooling
computer CPUs, not possible with compressive cooling in such limited space.

1. As a part of this research project, the student directly handled, manipulated, or interacted with (check ALL
that apply): human subjects Potentially hazardous biological agents
vertebrate animals microorganisms rDNA tissue

2. Student independently performed all procedures as outline in this abstract. X Yes X No


3. Student worked or used equipment in a site other than school, field, or home. Yes X No
X

4. This project is a continuation of previous research? Yes X No X

5. My display board includes non-published photographs/visual depictions of


humans (other than myself ): Yes X No X

I/We hereby certify that the above statements are correct and the FOR INTEL ISEF
information provided in the Abstract is the result of one year’s research. OFFICIAL USE
I/We also attest that the above properly reflects my/our own work. ONLY
___________________________________ __________________
Finalist or Team Leader Signature Date
This embossed seal attests that this project is in compliance with all federal and state laws and regulations and that all appropriate reviews and
approvals have been obtained including the final clearance by the SSEF/FFFS Scientific Review Committee.
Thermoelectric Effects 3

INTRODUCTION

Statement of Purpose

The purpose of this project is to investigate “thermoelectricity” using the Seebeck and

Peltier Effects to observe the impact of heat and electricity on metals and semiconductors, where

temperature differences can be converted directly into electricity, or electricity into a temperature

difference, with no moving parts or use of other machines such as compressors or generators.

Depending on which effect is being tested, temperature difference and amount of electricity (i.e.

voltage and current) could be either a control (independent variable) or a variable (dependent

variable), in addition to other independent variables such as type of metal wire used, the

thickness of the wire, the type of Peltier device being tested, and how long the test runs.

Background Research

In everyday life, devices like toasters, ovens, hairdryers, and home heaters are commonly

used to convert electricity into heat. Although most people probably do not know this, you can

convert heat directly into electricity with no moving parts or use of other machines such as

generators. Another little known fact is that electricity can also produce cooling without any

moving parts or machinery required. There are many metals and other materials that can do this,

among them semiconductors have the most dramatic effects.

The two facts above are referred to as thermoelectric effects. If a closed circuit is made

of two different metals and one junction is held at a higher temperature than the other, a current

will flow as long as the difference of temperature is kept. This is known as the “Seebeck effect.”

The Peltier effect is the inverse of the Seebeck effect. If a current is sent through the junction of

two different metals, heat is absorbed by the junction when the current flows in one direction,

and is emitted by the junction when the current is reversed (Eshbach, 1952).
Thermoelectric Effects 4

Tomas Seebeck discovered thermomagnetism in 1821. While studying the effects of heat

on certain metals, he took two different metals and connected them in a loop and applied heat to

one end, which he found could repel a compass needle. Seebeck himself did not fully understand

the cause of this event. He assumed that a magnetic field was created and called the effect

thermomagnetism. The electric current generated by the Seebeck effect is actually caused by

heat flowing from the hot metal to the colder, and also by the thermoelectric properties specific

to each metal. The thermoelectric potential of a metal is measured by what is now called the

“Seebeck Coefficient.” Though it made relatively little impact upon the scientific world for

nearly a century, the Seebeck Effect eventually became the basis upon which all future work in

thermoelectricity was built (McGrath & Travers, 2007). Figure 1 below (Salman, 2009)

illustrates how a simple Seebeck Effect circuit could be set up to measure voltage generated by

applying heat from a candle (at the left side) and cooling from ice (near the center) to the

junctions of two different metal wires (Materials 1 and 2), and connecting these to a multimeter

(at the right side). The larger the temperature difference, and the greater the difference in

Seebeck coefficients between the two metals, the larger the voltage generated.

Figure 1: Simple Circuit to Measure Seebeck Effect Voltage Generated

Jean Peltier discovered the thermoelectric effect bearing his name in 1834. Peltier found

that the junctions of different metals were heated or cooled depending on the direction an electric
Thermoelectric Effects 5

current passed through them. The “Peltier effect” is found to be proportional to the first power

of the current, not to its square, as is the permanent generation of heat caused by resistance

throughout the circuit (Duckworth, 1960). The Peltier effect was an accidental discovery made

during an experiment in which Peltier joined copper wire and bismuth wire together and then to a

battery. When he switched the battery on, one of the junctions of the two wires got hot, while

the other junction got cold. If the cold junction was put inside an insulated box, it became a low-

efficiency refrigerator. Modern students are often introduced to Peltier as a physicist, but might

be surprised to find that he didn't study physics until his retirement from the clock-making

business at age thirty. Figure 2 below (Salman, 2009) illustrates how a simple circuit could be

set up to create a temperature difference (Tc and Th) at the junctions (A and B) between two

different metals (Materials X and Y) by connecting this to a battery or DC power supply

(designated by Vin at the bottom). The larger the current supplied to the circuit, and the greater

the difference in Seebeck coefficients between the metals, the larger will be the temperature

difference created.

Figure 2: Simple Circuit to Create a Temperature Difference from the Peltier Effect

For many years after Seebeck's discovery of the thermocouple circuit, it was used as a

sensitive thermometer of incomparable accuracy and range, from a few degrees above absolute

zero to several thousand degrees Fahrenheit. Beginning in the early 20th century, scientists
Thermoelectric Effects 6

experimented to determine if more heat could generate a more powerful current, possibly

powerful enough to run machinery. For years, scientists worked to find the best combination of

alloys to maximize the output of thermocouples. After World War II, physicists experimented

with thermocouples that ran on the heat from decaying radioactive isotopes. These nuclear

thermocouples are used to power deep-space probes, devices that must run unattended for many

years and are too far from the sun to use solar panels (McGrath & Travers, 2006). With modern

techniques, thermoelectric “modules” can now be produced using semiconductors that deliver

efficient solid state heat-pumping for both cooling and heating. A practical thermoelectric

Peltier module generally consists of two or more elements of n- and p-type doped semiconductor

material that are connected electrically in series and thermally in parallel (Ferrotec, 2009). Many

of these units can be used to generate DC power in special circumstances, such as conversion of

waste heat into useful electricity. New and often elegant uses for thermoelectrics continue to be

developed each day (Tellurex, 2006).

Figure 3: Seebeck Effect with Semiconductors Figure 4: Peltier Effect with Semiconductors

Figure 3 above (Whyte, 2009) illustrates how n- and p-type semiconductors, sandwiched

between heat and cooling sources, are used to create the Seebeck effect. The heat would push
Thermoelectric Effects 7

both the negative and positive chare carriers downward, and create an electric current moving in

a clockwise direction through the semiconductors around the circuit. Figure 4 (Whyte, 2009)

shows how a DC current moving clockwise in series through the semiconductors around the

circuit would create the Peltier effect. The charge carriers in the n-type semiconductor would be

drawn downward (along with their heat) by attraction to the positive side of the DC power

supply, while the charge carriers in the p-type semiconductor would be repelled and pushed

downward (along with their heat).

Figure 5: Peltier Effect Cooling with Two Semiconductor Pairs

Figure 5 (Ferrotec, 2009) shows a more realistic view of how a Peltier cooling module is

constructed, with multiple pairs of n- and p-type semiconductors, a “heat sink” near the middle,

and the object being cooled (e.g. a CPU chip) at the top of the diagram. As in Figure 4, electric

current moves in series counterclockwise around the circuit, and all heat is “pumped” downward.

Figure 6 below (Tellurex, 2006) shows a more complete version of a Peltier module with

many pairs of n- and p-type semiconductors, sandwiched between ceramic substrates with the

positive and negative connecting wires shown at the right side of the diagram.
Thermoelectric Effects 8

Figure 6: Realistic Diagram of Peltier Cooling Module with Many Semiconductor Pairs

Figure 7: Actual Peltier Module with 127 Semiconductor Pairs Opened Up

Figure 7 above (Noll, 2008) shows the insides of an actual Peltier Module with 127 n-

and p-type semiconductor pairs, which is a typical number used in industry. The module is

opened up revealing the semiconductor pellets welded to their connecting plates with their

positive and negative connecting wires on the right side, and the top connecting plates on the left

side. Both top and bottom plates are glued to the white colored ceramic substrate that holds

everything together.
Thermoelectric Effects 9

The Peltier effect has also found significant value in recent years. The main use of the

Peltier effect is as a refrigerator; thermocouples can be cooled to a temperature low enough to

liquefy nitrogen and helium (McGrath & Travers, 2006). It was only after mid-20th Century

advancements in semiconductor technology, however, that practical applications for

thermoelectric devices became realistic. New and often elegant uses for thermoelectrics

continue to be developed each day (Tellurex, 2006). A common application of Peltier modules

is in the cooling of computer CPU’s. Others include low-cost, moderate- and high-capacity,

general-purpose modules for cooling sensitive equipment such as instrumentation, laboratory

equipment, consumer appliances, and for commercial and military applications (Melcor, 2009).

Travel coolers/warmers operating Peltier modules are now produced under a variety of brand

names (e.g. Igloo, Black & Decker, and Koolatron) and marketed through well-known sources

such as Amazon and Sharper Image (Warner, 2004).

Hypotheses

1. If two junctions are made between two different conductive metals and are held at different

temperatures, then a voltage will be created in proportion to the temperature difference and

the Seebeck coefficients for the metals.

2. The amount of voltage created in the above, should be independent of the thickness of the

metal/wire used, similar to connecting batteries in parallel instead of series, and should not

decline over time after reaching a “steady state.”

3. The amount of time required to reach steady state voltage should be directly proportional to

the thickness/surface area of the metal wire.

4. If commercially available Peltier modules made from ‘p’ and ‘n’ semiconductors are used to

generate electricity through the Seebeck Effect, and to create heating and cooling from the
Thermoelectric Effects 10

Peltier Effect, then the results, in terms of voltage generated or temperature difference

created, should be many times greater than that for metal wire under the same conditions.

5. The voltage generated from #4 above should be proportional to the number of pairs of

semiconductor pellets in the Peltier module.

6. The temperature difference created from a Peltier cooling module should be proportional to

the current flowing through the circuit, if the voltage is held constant.

7. A Peltier cooling device should maintain a voltage after disconnecting the power source, and

should decline in a measurable, nonlinear way as the device approaches room temperature.

8. If two Peltier units are “stacked,” then the resulting temperature difference should be close to

the total of each unit alone.


Thermoelectric Effects 11

METHOD
Materials
Table 1: Materials Used in This Project

Item Quantity Size


Mastech Variable DC Power Supply, Model HY3010E 1 30v/10a
Tekpower Digital Multimeter with Data Logger, Model DT9602R 1 na
Mastercool High Temp Infrared Thermometer Model 52225A-SP 1 na
Digital Thermometer, Extech Model 39240, -40 °C – 200 °C 1 na
CT* Thermoelectric Peltier Module, 15.4v/9a, 12711-5L31-09CQ 1 40x40cm
CT Thermoelectric Peltier Module, 15.2v/5a, 12711-5L31-05CL 1 30x30cm
CT Thermoelectric Peltier Module, 8.6v/9a, 07111-5L31-09CL 1 30x30cm
CT Thermoelectric Peltier Module, 0.8v/6a, 00711-5L31-06CB 1 10x10cm
CT Aluminum Heat Sink, Model HS-5.4-5.4-1.4-01 1 13.7 cm
CT Fan 134 CFM, DC 12 volts, 120mm sq. x38mm, FN-120-134 1 120 mm
CT Thermal Paste Compound, Ceramique Model TG-CMQ-22G 1 22 gm
Laptop Computer (to log temp. vs. time with Peltier modules) 1 na
DC 12 volt Power Supply for Fan, Spectrum Model 44-6681 1 12 volt
Type J Thermocouple: Constantan-Iron, Nanmac # A13-5-12 1 0.51 mm.
Type T Thermocouple: Constantan-Copper, Nanmac A13-14-12 1 0.51 mm
Type K Thermocouple: Chromel-Alumel, Nanmac A13-20-12 1 0.51 mm
Type E Thermocouple: Constantan-Chromel, Nanmac A13-27-12 1 0.25 mm.
Type E Thermocouple: Constantan-Chromel, Nanmac A13-28-12 1 0.38 mm.
Type E Thermocouple: Constantan-Chromel, Nanmac A13-29-12 2 0.51 mm
Type E Thermocouple: Constantan-Chromel, Nanmac A13-30-12 2 0.81 mm.
Soldering Gun, Weller Model 8200 N 1 140 watts
Solder, Electrical Rosin Core, 60% Tin/40% Lead, 1.27 mm. Dia. 1 21 gms.
Soldering Paste Flux, Alpha Fry No. 51012 1 56 gms.
Dewalt Electric Drill, Model DW104 1 na
Alligator Clip-End Connecting Wires 10 24 gauge
Alligator Clip Manipulator Arms (for holding parts for soldering) 1 na
Propane Torch, Bernzomatic, Model TX9 1 0.5 kg
Dry Ice Block 1 4.5 kg.
Isopropyl Alcohol (99+% pure), Iso-Heat (from auto supply) 1 375 ml.
Antifreeze (Ethylene Glycol), Mobil Permazone 1 500 ml.
Hot Plate or Gas Grill w/Side Burner (to heat antifreeze outside) 1 na
Wire Strippers 1 na
Safety Equipment (goggles, protective mask, leather gloves) 2 each na
* CT above stands for the company “Custom Thermoelectric”
Thermoelectric Effects 12

Procedure

First, the Seebeck Effect will be used to study the electricity generating potential of two

different base metals connected at two junctions with each exposed to different temperatures.

This will test hypotheses 1-3 above. The independent variables will be four temperature

differences created by hot and cold liquids, the types of metal that the thermocouple wires are

made from, and the thickness of the wires. The dependent variable is the voltage generated.

1) Prepare the seven different thermocouples by cutting one of the wires in the middle and

soldering the loose ends to form a second junction between the two metals. Thermocouples

come welded only at one end of the two 12-inch long wires.

2) Connect each of the remaining loose ends of the cut wire to an alligator clip connector and

then connect these to the positive and negative sockets of the multimeter. The thermocouples

have one of the ends colored red, which should be connected to positive.

3) Bend the uncut 12-inch long wire into a “U” shape, so that one end can be dipped into the hot

liquid and the other into the cold liquid.

4) Prepare a pot of boiling water and an insulated container of ice cubes in water. Have these

sitting side by side.

5) Perform two tests with each of the seven metal wire combinations. The first will put one

junction in the boiling water and the other junction in ice water, for close to a 100 °C

difference. The second will put one junction in ice water and leave the other junction

exposed to room temperature, for a temperature difference of around 23 °C.

6) Set the multimeter to measure in millivolts, and for each test, record the voltages reached

once the reading on the multimeter has stabilized. For each, check and record the

temperature of the ice water and the boiling water as well.


Thermoelectric Effects 13

7) For the type ‘E’ thermocouples, of which there are four different thicknesses, connect the

multimeter to a computer using a data logging device, measure the times taken to reach the

final “steady-state” voltage with one junction in the ice bath and the other in the boiling

water, which will test hypothesis #3 above.

8) Select one of the seven samples and hold it in the ice bath and boiling water for an extended

period, recording voltage using a data logger to measure if there is any decline in voltage

over time, which tests part of hypothesis #2.

9) Move outside and prepare an extremely cold ice bath using chunks of dry ice mixed with Iso-

Heat (near 100% isopropyl alcohol) added to almost fill an insulated container. This mixture

will reach about -64 °C after it has stopped “boiling” from the sublimation of the dry ice.

10) Prepare an extremely hot liquid bath by heating a pot of antifreeze on a hot plate or BBQ grill

side burner. Stay on the “up-wind” side of the burner and wear safety masks (to prevent

inhaling any vapor) and safety glasses and gloves, since this liquid will reach about 190 °C.

The adult sponsor should be the only one handling this pot of hot antifreeze.

11) Perform two tests with each of the seven metal wire combinations. The first will put one

junction in the hot antifreeze water and the other junction in dry ice bath, for about a 250 °C

difference. The second will put one junction in the antifreeze and leave the other junction

exposed to outdoor temperature, for a temperature difference of around 160 °C.

12) For each test, record the voltages reached once the reading on the multimeter has stabilized.

Also for each, check and record the temperatures of the hot and cold liquid baths using the

immersible thermocouple temperature wire.

In the next part of the experiment, four different commercial Peltier modules will be used

to test their maximum potential to generate electricity when one side of each device is exposed to
Thermoelectric Effects 14

very hot temperature and the other side to a very cold temperature. A propane torch to heat the

heat sink on one side, and dry ice to chill the other side, will be used to create the maximum

temperature differences. These will test hypotheses four and five above.

13) Use metal polish on the surface of the heat sink to make it as smooth as possible to increase

thermal conductivity. When complete, it should look close to a mirror-like surface.

14) Prepare the heat sink to measure its temperature just below the surface where the Peltier

modules will be placed, by drilling a 5/32 inch hole parallel to the surface into which a digital

thermometer will be inserted.

15) Suspend the heat sink over the edge of a countertop by attaching the heat sink to the side of a

large, heavy, metal, rectangular cooking pan. This can be done easily if the gap between the

heat sink “fins” is a little larger than the thickness of the cooking pan. The last gap in the

heat sink fins can be slipped over the edge of the pan, and hang out like a cantilever beam.

16) With some “thermal paste” covering the tip of the digital thermometer, insert it into the hole

in the heat sink. The tip should be located around the middle of the heat sink surface.

17) Cover the bottom (hot side) of a Peltier module with a very thin layer of thermal paste.

18) Secure the Peltier module to the center top of the reheat sink (directly over the end of the

digital thermometer) by pressing and rotating it on the heat sink until there is a complete

layer of thermal paste between the Peltier module and the heat sink, and the module feels as

if it is almost glued to the heat sink.

19) If available, attach a very thin thermocouple wire to the top of the Peltier module, to attempt

to measure the temperature difference between the two sides of the Peltiers. This may not be

possible since the surface contact should be as smooth and complete as possible.
Thermoelectric Effects 15

20) Have a piece of dry ice nearby between several towels for easy access. The dry ice should

have sides that are as flat as possible, as it will be pressed to the flat surface of the Peltiers.

21) Using the propane torch, heat the heat sink from underneath until it reaches a temperature of

110-115 °C. Do not go above this temperature, as the Peltier devices have a maximum

operating temperature of 125 °C, after which the solder connecting and holding the pellets

will melt and the device will be completely destroyed.

22) Wearing leather gloves, pick up the piece of dry ice as quickly as possible and press it firmly

to the top side of the Peltier device, while an assistant is carefully watching the voltage meter

(or it is connected to a data logger).

23) As soon as the dry ice is applied, the Peltier module will vibrate vigorously, making an

almost squeaking noise as the ice vaporizes, going straight from a solid state to a gas through

sublimation. It is the formation of the gas between the ice and the flat Peltier surface that

causes the noise and vibration. This should be done carefully and only once if possible, as

the harsh vibration will eventually damage the Peltier.

24) Connect the data logging device from the multimeter to the RS232 port of the computer to

record voltage readings. Remove the dry ice once the researcher has noted on the multimeter

that the maximum voltage has been reached. Make sure that the scale of the multimeter is set

to measure volts, since the larger Peltier modules (with 127 “pellet pairs”) should produce as

much as five or more volts. A millivolt setting is required for the thermocouples, since they

will only produce a few millivolts, even with a much larger temperature difference.

25) Repeat the procedure for each of the four Peltier modules.

In the final part of the experiment, the Peltier Effect is used to study the reverse of the

above, with electricity applied to Peltier modules to generate a temperature difference between
Thermoelectric Effects 16

one side of a module and the other. This will test hypotheses 6-8. The independent variables

will be the four different types of Peltier modules, and the amounts of voltage and current

applied to them. The dependent variable will be the temperature difference created.

26) Connect the fan to the bottom of the heat sink, so that the fan blows air up through the vanes

of the heat sink. The connection can be made simply using duct tape around the outside

edges near the lower ends of the heat sink fins, since this will not become hot like it might at

the surface on which the Peltier modules will be attached.

27) The fan will then need to be mounted to some structure that keeps it elevated several inches

above countertop level, with its intake side having access to free air flow (otherwise it would

not be able to suck in fresh cool air to blow on the heat sink). A simple way to do this is by

taking four extra thick books, and stacking two on each side of the bottom of the fan intake.

Only about a quarter inch of the fan base would need to overlap the edge of the books, and

this could be taped to the books using duct tape.

28) As previously done, the 5/32 inch diameter digital thermometer should be inserted, with

some thermal paste smeared on its tip, into the hole drilled near the surface of the heat sink.

29) Each Peltier module will then be tested in the same manner as follows.

30) First, a thin layer of thermal paste will be spread on the hot side (side to which the leads are

connected) of the Peltier, and it should then be pressed firmly onto the center of the heat sink

(directly above the digital thermometer) while rotating back and forth, until the film of paste

has been evenly distributed around all edges, and it almost feels like it has been glued down.

31) One test method will be to use a flat piece of Styrofoam, about one inch thick, to place on top

of the cold side of the Peltier module, on which about ten pounds in weights (e.g. books)

should be placed to create a firm contact between the Peltier and the heat sink.
Thermoelectric Effects 17

32) A thin wire thermocouple connected from the side of the infrared thermometer should be

inserted between the Styrofoam and the top of the Peltier module.

33) The second method will be to allow the top of the Peltier module to remain open to room

temperature air, to test if more cooling is achieved. Using this method, the infrared

thermometer would be used to measure the temperature on the cold side of the Peltier.

34) Under both methods, the hot side temperature of the Peltier would be measured from the

digital thermometer inserted in the top of the heat sink.

35) For each test with the four Peltier modules, the maximum voltage would be fixed on the

variable power supply according to the specifications for each Peltier.

36) Each test with the different Peltiers would be conducted at several different current levels,

during which the hot and cold side temperatures would be recorded every ten seconds.

37) To test hypothesis seven above, voltage would be recorded with a data logger after the power

supply has been cut off from one of the Peltier modules, and this would be allowed to decline

until near zero to test graphically if the decline is nonlinear, as hypothesized.

38) To test hypothesis eight, the smallest Peltier module would be attached using thermal paste to

either the largest (40 mm sq.) or medium (30 mm sq.) size Peltier module having 127 Pairs of

semiconductor pellets. Calculations will need to be performed to determine whether to use

the large or medium Peltier in the stack.

39) The larger Peltier module would be thermal pasted to the heat sink, and the two Peltiers

would need to be connected in electrical series.

40) A voltage and current would be applied and the overall temperature difference created with

the stacked approach measure and recorded.


Thermoelectric Effects 18

RESULTS
Tables of Data, Charts, and Graphs

Table 2: Seebeck Effect: Thermoelectric Generating Tests Using Metal Thermocouple Wires

Dependent
Control Variables Variable Comparison to Theoretical
Wire Temp Difference Differ-
Diam- Hot Cold Differ- Measured Seebeck Theoretical (measured - ence
*T/C eter Bath Bath ence Voltage Coeffi- Voltage** theoretical (Percent
Type (mm) °C °C °C (millivolts) cient (millivolts) millivolts) of theor.)
Type E 0.25 24.2 1.4 22.8 1.40 68 1.55 -0.15 -9.7%
Type E 0.38 24.2 1.4 22.8 1.40 68 1.55 -0.15 -9.7%
Type E 0.51 24.2 1.4 22.8 1.40 68 1.55 -0.15 -9.7%
Type E 0.81 24.6 1.5 23.1 1.50 68 1.57 -0.07 -4.5%
Type E 0.25 98.5 1.4 97.1 6.20 68 6.60 -0.40 -6.1%
Type E 0.81 99.0 1.5 97.5 6.25 68 6.63 -0.38 -5.7%
Type E 0.38 99.0 1.4 97.6 6.20 68 6.64 -0.44 -6.6%
Type E 0.51 99.0 1.4 97.6 6.50 68 6.64 -0.14 -2.1%
Type E 0.38 181.0 30.6 150.4 10.00 68 10.23 -0.23 -2.2%
Type E 0.51 184.0 30.6 153.4 10.30 68 10.43 -0.13 -1.3%
Type E 0.25 192.0 30.6 161.4 11.10 68 10.98 0.12 1.1%
Type E 0.81 192.0 30.6 161.4 11.10 68 10.98 0.12 1.1%
Type E 0.38 181.0 -64.0 245.0 16.00 68 16.66 -0.66 -4.0%
Type E 0.51 184.0 -64.0 248.0 16.20 68 16.86 -0.66 -3.9%
Type E 0.25 192.0 -64.0 256.0 16.70 68 17.41 -0.71 -4.1%
Type E 0.81 192.0 -64.0 256.0 16.70 68 17.41 -0.71 -4.1%
Type J 0.51 24.2 1.1 23.1 1.25 55 1.27 -0.02 -1.6%
Type J 0.51 98.8 1.0 97.8 5.20 55 5.38 -0.18 -3.3%
Type J 0.51 192.0 30.6 161.4 8.70 55 8.88 -0.18 -2.0%
Type J 0.51 192.0 -64.0 256.0 13.90 55 14.08 -0.18 -1.3%
Type T 0.51 25.1 1.0 24.1 1.05 43 1.04 0.01 1.3%
Type T 0.51 98.7 1.0 97.7 4.15 43 4.20 -0.05 -1.2%
Type T 0.51 192.0 30.6 161.4 7.10 43 6.94 0.16 2.3%
Type T 0.51 192.0 -64.0 256.0 11.00 43 11.01 -0.01 -0.1%
Type K 0.51 25.2 1.4 23.8 1.00 41 0.98 0.02 2.5%
Type K 0.51 98.8 1.4 97.4 4.00 41 3.99 0.01 0.2%
Type K 0.51 188.0 30.6 157.4 6.30 41 6.45 -0.15 -2.4%
Type K 0.51 188.0 -64.0 252.0 10.30 41 10.33 -0.03 -0.3%
* T/C Denotes Thermocouple Type & Metal Combinations
Type E - Constantan & Chromel Type T - Constantan & Copper
Type J - Constantan & Iron Type K - Alumel & Chromel
** Calculation of Theoretical Voltage is explained in the Calculations subsection of the Results section.
Thermoelectric Effects 19

Figure 8: Seebeck Effect (Supports Hypothesis 1)


Thermoelectric Generating Results Using 0.51 mm Thermocouple Wire

16
Type E: Seebeck Coeff. = 68
14 Type J: Seebeck Coeff. = 55
Type T: Seebeck Coeff. = 43
12 Type K: Seebeck Coeff. = 41
Voltage (millivolts)

10

0
0 50 100 150 200 250
Temperature Difference C

The data in Table 2 and the graph in Figure 8 fully support the statements in Hypothesis

1, that the voltage created from a temperature difference between the junctions of two different

metals will be in proportion to the temperature difference and the Seebeck coefficients for the

metals. All measured voltages differ less than 10% from the theoretical, with the average

difference being only 2.8%. The graph in Figure 8, in which all data is based on 0.51 mm wire,

shows that the voltage increases in a linear manner proportional to the temperature difference

between the two junctions. The graph also shows clearly that the larger the difference in

Seebeck coefficients between the metal wire combinations, the larger the voltage produced.
Thermoelectric Effects 20

Figure 9: Test of Voltage Generated for Wires of Different Thicknesses


at Various Temperature Differences Between Junctions (Hypothesis 2)

16
Type E - 0.25 mm
Type E - 0.38 mm
14
Type E - 0.51 mm
Type E - 0.81 mm
12
Voltage Generated (millivolts)

10

0
0 50 100 150 200 250
Temperature Difference ( C)

The graph above in Figure 9 illustrates the first part of Hypothesis 2, that the Seebeck

voltage generated should be independent of wire thickness, assuming that the metal wire

combinations remain the same. For this test, the wires used are from Type E thermocouples,

which are made of constantan and Chromel. The graph visually confirms the hypothesis, since

the voltage generated is almost the same for each of the wire thicknesses viewed at the same

temperature difference, which makes the lines appear to overlay each other. This proves that

wire thickness is irrelevant to voltage generated from the Seebeck effect.


Thermoelectric Effects 21

Table 3: *Sample Data - Test of Seebeck Voltage Stability over Extended Period (Hypothesis 2)

Voltage Voltage
Seconds Minutes (millivolts) Seconds Minutes (millivolts)
0 - - 515 8.58 6.1
1 0.02 1.6 516 8.60 6.0
2 0.03 5.2 517 8.62 6.1
3 0.05 6.0 518 8.63 6.0
4 0.07 6.2 519 8.65 6.0
5 0.08 6.3 520 8.67 6.0
6 0.10 6.2 521 8.68 6.1
7 0.12 6.3 522 8.70 6.0
8 0.13 6.2 523 8.72 6.0
9 0.15 6.2 524 8.73 6.0
10 0.17 6.3 525 8.75 6.0
*Only the first and last 10 rows of data are shown above for brevity.

Figure 10: Test for Seebeck Voltage Decline over Time (Hypothesis 2)
7

6
Voltage (millivolts)

5
0.51 mm diameter thermocouple
4 wire was used for the test

0
- 1 2 3 4 5 6 7 8 9
Elapsed Time in Minutes

The graph in Figure 10 above illustrates the second part of Hypothesis 2, that the amount

of Seebeck voltage created not decline over time after reaching a “steady state.” This supports

the hypothesis since, after quickly rising to steady state, the graph gives visual evidence that the

voltage did not go up nor down except for minor fluctuations.


Thermoelectric Effects 22

Table 4: Raw Data from Tests of Response Time versus Wire Diameter (Hypothesis 3)

Time in Voltage Time in Voltage Time in Voltage Time in Voltage


Trial Sec. (milli V) Trial Sec. (milli V) Trial Sec. (milli V) Trial Sec. (milli V)
Trials for 0.25 mm Wire Trials for 0.38 mm Wire Trials for 0.51 mm Wire Trials for 0.81 mm Wire
Trial 1 0 - Trial 1 0 - Trial 1 0 - Trial 1 0 -
1 2.7 1 2.6 1 4.9 1 3.6
2 4.1 2 3.7 2 5.5 2 4.9
3 5.5 3 5.1 3 6.1 3 5.8
4 5.8 4 5.8 4 6.2 4 6.2
5 5.9 5 5.9 5 6.3 5 6.2
6 6.0 6 6.0 6 6.3 6 6.2
7 6.0 7 6.0 7 6.3 7 6.2
Trial 2 0 - 8 6.0 Trial 2 0 - Trial 2 0 -
1 3.1 Trial 2 0 - 1 2.6 1 4.9
2 5.2 1 2.5 2 5.5 2 5.2
3 5.8 2 5.2 3 6.1 3 6.0
4 6.0 3 5.6 4 6.2 4 6.1
5 6.1 4 5.9 5 6.2 5 6.2
6 6.1 5 5.9 6 6.2 6 6.2
Trial 3 0 - 6 5.9 7 6.2 7 6.2
1 1.3 7 5.9 Trial 3 0 - Trial 3 0 -
2 4.9 Trial 3 0 - 1 3.3 1 4.6
3 5.8 1 2.7 2 5.2 2 5.5
4 6.1 2 4.2 3 5.6 3 5.7
5 6.2 3 5.5 4 6.0 4 5.8
6 6.2 4 5.9 5 6.0 5 5.8
Trial 4 0 - 5 5.9 6 6.0 6 6.0
1 4.6 6 5.9 Average 4.7 7 6.1
2 5.4 7 5.9 8 6.1
3 5.9 Average 4.7 Trial 4 0 -
4 6.1 1 2.9
5 6.1 2 5.2
6 6.1 3 5.9
Average 5.0 4 6.1
5 6.2
6 6.2
7 6.2
Average 5.3
Thermoelectric Effects 23

Table 5: Summary for Tests of Response Time versus Wire Diameter (Hypothesis 3)

Voltage Seconds Voltage Seconds


Trial (millivolts) to Peak Trial (millivolts) to Peak

Trials for 0.25 mm Wire Trials for 0.51 mm Wire


Trial 1 6.0 6.0 Trial 1 6.3 5.0
Trial 2 6.1 5.0 Trial 2 6.2 5.0
Trial 3 6.2 5.0 Trial 3 6.0 4.0
Trial 4 6.1 4.0 Average: 0.51 mm wire 4.7
Average: 0.25 mm wire 5.0
Trials for 0.81 mm Wire
Trials for 0.38 mm Wire Trial 1 6.2 4.0
Trial 1 6.0 6.0 Trial 2 6.2 5.0
Trial 2 5.9 4.0 Trial 3 6.1 7.0
Trial 3 5.9 4.0 Trial 4 6.2 5.0
Average: 0.38 mm wire 4.7 Average: 0.81 mm wire 5.3

Ovarall Average: 4.9 seconds

The data in Table 4 shows the results of the multiple trials testing response time for four

wire thicknesses made from the Type E thermocouple. The results collected were the times

required to reach steady state voltage when one wire junctions was held in boiling water and the

other was in an ice bath. Table 5 is a summary of all of the final results as well as the averages.

The data collected did not support Hypothesis #3, since the hypothesis stated, “The amount of

time required to reach steady state voltage should be directly proportional to the

thickness/surface area of the metal wire.” The reason why this was disproven is that it took

about the same time, between 4.7 and 5.3 seconds averaging 4.9 seconds, to reach steady state no

matter what the thickness was. Possible reasons are discussed in the Conclusions section below.
Thermoelectric Effects 24

Figure 11: Graphs for Tests of Response Time versus Wire Diameter (Hypothesis 3)

Reaction Time for 0.25 mm Wire Reaction Time for 0.38 mm Wire
7 7
6 6

Voltage (millivolts)
Voltage (millivolts)

5 5
4 4
Trial 1
3 3 Trial 1
Trial 2
2 2 Trial 2
Trial 3
1 1 Trial 3
Trial 4
0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Seconds Seconds

Reaction Time for 0.51 mm Wire Reaction Time for 0.81 mm Wire
7 7
6 6
Voltage (millivolts)

Voltage (millivolts)

5 5
4 4
Trial 1
3 Trial 1 3
Trial 2
2 Trial 2 2 Trial 3
1 Trial 3 1 Trial 4
0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Seconds Seconds

Figure 11 is a graph of table 4. This figure is another way of showing that Hypothesis 3

is not supported. Hypothesis 3 is not supported since all of the wires, no matter how thick or

thin, reached a peak steady state in about the same amount of time. Every wire thickness

reached a maximum voltage of around six millivolts in about five seconds. Possible reasons are

discussed in the Conclusions section below.


Thermoelectric Effects 25

Table 6: Maximum Seebeck Voltage Generated by Temperature Difference (Hypothesis 4-5)

Type of Peltier Module: Size and Number of Semiconductor Pairs


(units in all four columns below are volts)
Time 40 mm sq. 30 mm sq. 30 mm sq. 10 mm sq.
(sec) 127 pellet pairs 127 pellet pairs 71 pellet pairs 7 pellet pairs
0 0.00 0.00 0.00 0.00
1 2.99 1.95 0.25 0.16
2 3.63 4.44 0.91 0.22
3 5.01 4.87 1.31 0.38
4 5.31 5.70 1.44 0.42
5 5.46 5.82 1.98 0.42
6 5.64 6.10 2.34 0.42
7 5.67 6.15 2.58 0.41
8 5.73 6.10 2.94 0.40
9 5.74 6.09 3.24 0.40
10 5.72 6.14 3.30 0.40
11 5.77 6.05 3.49 0.40
12 5.78 6.03 3.59 0.40
13 5.78 5.75 3.65 0.41
14 5.80 5.67 3.66 0.41
15 5.56 5.67 3.69 0.41
16 5.40 5.68 3.71 0.41
17 5.38 5.70 3.72 0.41
18 5.37 5.89 3.73 0.41
19 5.32 5.71 3.72 0.42
20 5.27 5.17 3.71 0.41
21 5.25 3.95 3.70 0.41
22 5.45 3.67 3.71 0.41
23 5.42 2.89 3.72 0.41
24 5.25 2.42 3.72 0.41
25 5.22 2.16 3.72 0.40
26 5.14 1.78 3.73 0.40
27 5.13 1.53 3.73 0.40
28 5.15 1.33 3.73 0.39
29 5.28 1.16 3.75 0.39
30 5.31 0.98 3.76 0.38
31 5.25 0.88 3.77 0.38

Maximum voltages were generated by exposing one side of each Peltier module to

110°C, and the other to dry ice at -78°C. For brevity, data above is only through peak voltages

reached, highlighted in yellow. The full set of data was graphed in Figure 12, which compares

results from the four Peltier modules. The results confirm Hypothesis 4, since the Peltier

modules generated several hundred times more voltage than the metal thermocouple wires.
Thermoelectric Effects 26

Figure 12: Peak Voltage Generated by Peltier Modules (Hypotheses 4-5)

To Use the Seebeck Effect to Generate Electricity (voltage)


6 - One side of Paltier Module subjected to dry ice at -78 C
- Other side thermal pasted to heat sink heated to 110 C
40 mm/127 pellet pairs: 5.80 volts max.
30 mm/127 pellet pairs: 6.15 volts max.
5 30 mm/71 pellet pairs: 3.77 volts max.
10 mm/7 pellet pairs: 0.42 volts max.

4
Voltage (volts)

Voltage declines quickly


3
once dry ice is removed

0
0 10 20 30 40 50 60 70 80 90 100
Time in Seconds

Two tests were conducted for each Peltier module. The test producing the maximum

voltage was graphed in Figure 12 above, comparing results for all four modules. This graph

supports Hypothesis 5, that the voltage generated from a temperature difference should be

proportional to the number of pairs of semiconductor pellets in the Peltier module.” For

example, seven pairs generated 0.42 volts, while seventy-one pairs generated 3.77 volts, almost

ten times greater. The results were not an exact linear proportional ratio, but fairly close to one.
Thermoelectric Effects 27

Table 7: Test Data for 40mm/127 Pair Peltier Modules Subjected to Various Current Levels

40 mm / 127 @ 3 amps 40 mm / 127 @ 6 amps 40 mm / 127 @ 9 amps


Time Temp Temp Temp Temp Temp Temp Temp Temp Temp
(sec.) Low °C Hot °C Delta Low °C Hot °C Delta Low °C Hot °C Delta
0 24.8 24.8 0.0 24.6 24.6 0.0 25.0 25.0 0.0
5 22.5 26.4 3.9 -18.0 28.0 46.0 -5.9 35.0 40.9
10 20.1 27.9 7.8 -19.0 29.0 48.0 -20.0 37.0 57.0
15 18.6 28.0 9.4 -23.0 30.0 53.0 -21.0 38.0 59.0
20 17.1 28.0 10.9 -24.0 31.0 55.0 -22.0 39.3 61.3
25 15.4 28.1 12.7 -24.0 31.5 55.5 -22.0 40.5 62.5
30 13.7 28.2 14.5 -24.0 32.0 56.0 -22.0 41.6 63.6
35 10.1 28.3 18.3 -24.0 32.0 56.0 -22.0 42.7 64.7
40 6.4 28.4 22.0 -24.0 32.0 56.0 -22.0 43.8 65.8
45 -2.3 28.5 30.8 -24.5 32.5 57.0 -22.0 45.1 67.1
50 -11.0 28.6 39.6 -25.0 33.0 58.0 -22.0 44.4 66.4
55 -11.5 28.7 40.2 -25.5 33.0 58.5 -22.0 47.4 69.4
60 -12.0 28.7 40.7 -26.0 33.0 59.0 -22.0 49.3 71.3
65 -12.5 28.8 41.3 -25.5 33.5 59.0 -22.0 50.3 72.3
70 -13.0 28.8 41.8 -25.0 34.0 59.0 -22.0 51.8 73.8
75 -13.0 28.9 41.9 -25.0 34.5 59.5 -22.0 52.7 74.7
80 -13.0 28.9 41.9 -25.0 35.0 60.0 -22.0 52.7 74.7
85 -13.0 29.0 42.0 -11.0 34.8 45.7 6.2 46.0 39.8
90 -13.0 29.1 42.1 3.1 34.5 31.4 20.2 42.0 21.8
95 -13.0 29.1 42.1 15.6 34.1 18.5 34.3 41.0 6.7
100 -13.0 29.1 42.1 28.0 33.6 5.6 35.3 39.0 3.7
105 -13.0 29.2 42.2 29.7 33.1 3.5 35.6 38.0 2.4
110 -13.0 29.2 42.2 31.3 32.6 1.3 34.7 37.0 2.3
115 -13.0 29.7 42.7 31.3 32.2 0.8 33.7 36.3 2.6
120 -13.0 30.1 43.1 31.3 31.7 0.4
125 2.5 29.0 26.6 30.7 31.3 0.5
130 17.9 27.9 10.0 30.1 30.8 0.7
135 21.3 27.9 6.7 29.9 30.6 0.7
140 24.6 27.9 3.3 29.7 30.3 0.6
145 25.4 27.9 2.5 29.5 30.1 0.6
150 26.1 27.8 1.7 29.2 29.8 0.6
155 26.2 27.8 1.6 29.0 29.7 0.7
160 26.2 27.7 1.5 28.7 29.5 0.8
165 26.2 27.7 1.5 28.7 29.4 0.7
170 26.2 27.6 1.4 28.6 29.2 0.6
Thermoelectric Effects 28

Figure 13: Temperature Difference Created in 40 mm/127 Pellet Pair


Peltier Modules Subjected to Various Current Levels
80

70 Delta @ 9 amps
Delta @ 6 amps

60 Delta @ 3 amps
Temperature Difference ( C)

50

40

30

20

10

0
0 20 40 60 80 100 120 140 160
Time (seconds)

This graph in Figure 13 above supports Hypothesis 6 “the temperature difference created

from a Peltier cooling module should be proportional to the current flowing through the circuit, if

the voltage is held constant.” The graph shows that the larger the current (in amperes), the larger

the temperature difference becomes. However, the change is not in a linear proportion. For

example, three amps creates about a forty-three degree temperature difference, while six amps

only creates a sixty degree difference, not an eighty-six degree linear proportional difference.

This would mean the change is in a non-linear proportion.


Thermoelectric Effects 29

Table 8: Test Data for 30 mm/127 Pair Peltier Modules Subjected to Various Current Levels

30 mm / 127 @ 2.5 amps 30 mm / 127 @ 5 amps


Time Temp Temp Temp Temp Temp Temp
(sec.) Low °C Hot °C Delta Low °C Hot °C Delta
0 26.3 26.3 0.0 27.0 27.0 0.0
5 26.7 26.4 -0.3 -7.2 34.0 41.2
10 12.9 26.9 14.0 -16.0 34.5 50.5
15 0.0 27.4 27.4 -21.0 35.0 56.0
20 -5.9 28.0 33.9 -22.0 36.0 58.0
25 -14.0 28.5 42.5 -22.0 36.8 58.8
30 -17.0 29.0 46.0 -22.0 37.8 59.8
35 -17.0 29.2 46.2 -22.0 38.4 60.4
40 -18.0 29.3 47.3 -22.0 38.8 60.8
45 -18.0 29.5 47.5 -22.0 39.2 61.2
50 -18.0 29.6 47.6 -23.0 39.6 62.6
55 -19.0 29.8 48.8 -23.0 40.5 63.5
60 -18.0 30.0 48.0 -23.0 42.3 65.3
65 -17.0 30.2 47.2 -23.0 42.6 65.6
70 -17.0 30.9 47.9 11.8 37.0 25.2
75 4.4 30.9 26.5 24.3 35.0 10.7
80 12.9 30.9 18.0 29.3 34.0 4.7
85 21.6 30.9 9.3 29.3 33.4 4.1
90 23.4 28.6 5.2 30.6 33.0 2.4
95 25.8 28.5 2.7 30.3 32.0 1.7
100 26.3 28.4 2.1 30.2 31.8 1.6
105 26.4 28.3 1.9 29.4 31.6 2.2
110 26.6 28.2 1.6 29.4 31.0 1.6
115 26.9 28.1 1.2 29.3 30.6 1.3
120 27.0 28.0 1.0 29.3 30.3 1.0
125 27.0 27.1 0.1

Above was only tested to 5-amps since this was the maximum operating limit for the

device.
Thermoelectric Effects 30

Figure 14: Temperature Difference Created in 30 mm/127 Pellet Pair


Peltier Modules Subjected to Various Current Levels
70

60 Delta @ 5 amps
Delta @ 2.5 amps

50
Temperature Difference ( C)

40

30

20

10

0
0 20 40 60 80 100 120
Time (seconds)

The temperature difference shown in Figure 14 above is not linear, since increasing the

voltage from 2.5 to 5 (doubling it) did not double the temperature difference and only increased

it by about 15°C. Although the temperature to voltage change was not linear, it still supported

Hypothesis 6, as in the previous test, and is just a non-linear relation.


Thermoelectric Effects 31

Table 9: Test Data for 30 mm/ 71 Pair Peltier Modules Subjected to Various Current Levels

30 mm / 71 @ 3 amps 30 mm / 71 @ 6 amps 30 mm / 71 @ 9 amps


Time Temp Temp Temp Temp Temp Temp Temp Temp Temp
(sec.) Low °C Hot °C Delta Low °C Hot °C Delta Low °C Hot °C Delta
0 24.9 24.9 0.0 24.0 24.0 0.0 24.7 24.7 0.0
5 5.1 24.7 19.6 -17.0 27.6 44.6 -5.6 30.8 36.4
10 -4.9 25.1 30.0 -22.0 27.8 49.8 -19.0 32.0 51.0
15 -10.0 25.1 35.1 -28.0 28.1 56.1 -26.0 33.0 59.0
20 -12.0 25.1 37.1 -28.0 28.3 56.3 -27.0 33.5 60.5
25 -13.0 25.2 38.2 -29.0 29.0 58.0 -27.0 34.0 61.0
30 -14.0 25.3 39.3 -29.0 29.2 58.2 -27.0 34.5 61.5
35 -14.0 25.4 39.4 -30.0 29.4 59.4 -28.0 35.5 63.5
40 -14.0 25.5 39.5 -30.0 29.6 59.6 -28.0 36.0 64.0
45 -14.0 25.6 39.6 -30.0 30.0 60.0 -28.0 37.2 65.2
50 -15.0 25.8 40.8 -21.0 31.0 52.0 -28.0 38.0 66.0
55 1.6 25.4 23.8 3.4 27.0 23.6 -10.0 35.0 45.0
60 12.1 25.0 12.9 9.9 26.6 16.7 17.0 31.6 14.6
65 18.7 25.6 6.9 19.3 26.4 7.1 24.6 30.0 5.4
70 20.6 25.3 4.7 22.2 26.3 4.1 26.6 29.6 3.0
75 23.1 24.0 0.9 24.3 26.1 1.8 27.1 29.1 2.0
80 23.4 24.3 0.9 26.0 27.1 1.1
85 23.7 24.6 0.9 26.0 26.4 0.4
90 23.9 24.6 0.7
Thermoelectric Effects 32

Figure 15: Temperature Difference Created in 30 mm/ 71 Pellet Pair


Peltier Modules Subjected to Various Current Levels
70

60 Delta @ 9 amps
Delta @ 6 amps
Delta @ 3 amps
50
Temperature Difference ( C)

40

30

20

10

0
0 10 20 30 40 50 60 70 80 90
Time (seconds)

As with the previous two tests for different Peltier modules, the results shown in Figure

15 are the same in that they indicate a proportional relation between current and temperature

difference, just not a linear one.


Thermoelectric Effects 33

Table 10: Voltage Decline Data after Power Shut-Off to Peltier Modules (Hypothesis 7)

Time 40mm/127 30mm/127 40mm/127 30mm/71 30mm/127 30mm/71 40mm/127 30mm/71


(sec) at 9 amps at 5 amps at 6 amps at 9 amps at 2.5 amps at 6 amps at 3 amps at 3 amps
0 20.05 16.63 12.39 10.54 8.32 6.87 6.63 3.71
1 6.16 2.24 2.74 2.40 1.65 1.30 1.75 0.95
2 2.68 1.71 2.17 1.37 1.42 1.07 1.49 0.79
3 2.04 1.38 1.74 1.07 1.29 1.00 1.30 0.74
4 1.55 1.17 1.52 0.90 1.08 0.77 1.11 0.61
5 1.52 0.98 1.31 0.81 0.95 0.67 0.94 0.52
6 1.25 0.79 1.24 0.68 0.78 0.57 0.84 0.46
7 1.04 0.69 0.98 0.60 0.60 0.51 0.72 0.39
8 0.93 0.56 0.84 0.50 0.51 0.43 0.59 0.33
9 0.79 0.45 0.71 0.40 0.46 0.34 0.56 0.30
10 0.63 0.37 0.64 0.36 0.39 0.32 0.46 0.24
11 0.54 0.33 0.54 0.30 0.32 0.26 0.40 0.21
12 0.51 0.28 0.44 0.26 0.30 0.22 0.32 0.19
13 0.44 0.22 0.40 0.21 0.26 0.20 0.31 0.15
14 0.36 0.20 0.35 0.20 0.24 0.17 0.27 0.13
15 0.32 0.17 0.29 0.16 0.20 0.14 0.22 0.12
16 0.29 0.15 0.27 0.14 0.20 0.12 0.20 0.10
17 0.26 0.13 0.23 0.12 0.18 0.11 0.18 0.09
18 0.22 0.12 0.20 0.11 0.16 0.09 0.15 0.08
19 0.21 0.11 0.19 0.09 0.15 0.08 0.13 0.07
20 0.19 0.10 0.16 0.08 0.13 0.08 0.13 0.06
21 0.17 0.09 0.15 0.08 0.12 0.07 0.11 0.06
22 0.16 0.08 0.14 0.06 0.11 0.06 0.10 0.05
23 0.15 0.07 0.12 0.06 0.10 0.06 0.10 0.05
24 0.14 0.07 0.10 0.06 0.09 0.05 0.09 0.04
25 0.12 0.07 0.10 0.05 0.09 0.05 0.08 0.04
26 0.12 0.07 0.09 0.05 0.08 0.04 0.07 0.04
27 0.11 0.08 0.09 0.04 0.08 0.04 0.07 0.04
28 0.10 0.09 0.08 0.04 0.07 0.04 0.06 0.03
29 0.10 0.09 0.08 0.04 0.06 0.04 0.06 0.03
30 0.09 0.09 0.07 0.04 0.06 0.03 0.06 0.03
Thermoelectric Effects 34

Figure 16: Voltage Decline Profile after Power Shut-Off to Peltier Modules (re: Hypothesis 7)
20
18 40mm/127 at 9 amps
16 30mm/127 at 5 amps
40mm/127 at 6 amps
14 30mm/71 at 9 amps
Voltage (volts)

12 30mm/127 at 2.5 amps


30mm/71 at 6 amps
10
40mm/127 at 3 amps
8 30mm/71 at 3 amps
6
4
2
0
0 1 2 3 4 5 6 7 8 9 10
Time from Power Shut-Off (seconds)

Figure 17: Voltage Decline Profile from One Second after Power Shut-Off to Peltier Modules
6
40mm/127 at 9 amps
5 30mm/127 at 5 amps
40mm/127 at 6 amps
4 30mm/71 at 9 amps
Voltage (volts)

30mm/127 at 2.5 amps


30mm/71 at 6 amps
3 40mm/127 at 3 amps
30mm/71 at 3 amps
2

0
1 2 3 4 5 6 7 8 9 10 11
Time from Power Shut-Off (seconds)

Figures 16-17 are based on the same data. Figure 16 begins at full voltage drawn by the

Peltier modules at peak temperature differences. Figure 17 begins one second after power is shut
Thermoelectric Effects 35

off, and this is the graph that would be used to judge the correctness of Hypothesis 7. The graph

shows that voltage is retained for a while after power shut-off, as stated in the hypothesis. Then

the graph shows a curved line of declining voltage that becomes closer to zero over time. These

two observations confirm that Hypothesis 7 is correct.

Table 11: Data from Stacking Two Peltier Modules Together at Various Currents (Hypothesis 8)

Data for 3 amps Data for 6 amps Data for 9 amps


Time Cold Hot Temp Cold Hot Temp Cold Hot Temp
(sec) Temp Temp Diff °C Temp Temp Diff °C Temp Temp Diff °C
0 26.0 26.0 0.0 25.0 25.0 0.0 25.0 25.0 0.0
1 24.6 26.3 1.7 24.6 25.4 0.9 24.6 26.6 2.0
2 23.6 26.6 3.0 22.5 26.3 3.8 21.5 28.5 7.0
3 21.5 27.0 5.5 21.5 27.3 5.7 18.5 30.4 11.9
4 19.5 27.3 7.8 18.5 28.2 9.7 16.4 32.3 15.9
5 16.4 27.6 11.2 14.4 29.1 14.7 13.4 34.2 20.8
6 14.4 27.7 13.3 11.3 29.5 18.2 9.3 35.1 25.8
7 12.4 27.9 15.5 6.2 30.0 23.8 8.3 36.0 27.8
8 9.3 28.0 18.7 5.2 30.4 25.3 6.2 37.0 30.8
9 7.2 28.2 20.9 1.0 30.9 29.8 3.1 37.9 34.8
10 5.2 28.3 23.1 0.0 31.3 31.3 2.1 38.8 36.7
11 3.1 28.4 25.3 -1.0 31.6 32.6 1.0 39.4 38.4
12 1.0 28.5 27.5 -3.1 31.9 35.0 1.0 40.0 39.0
13 0.0 28.6 28.6 -5.2 32.1 37.4 1.0 40.7 39.6
14 -1.0 28.7 29.7 -7.3 32.4 39.7 1.0 41.3 40.2
15 -3.1 28.8 31.9 -9.4 32.7 42.1 1.0 41.9 40.9
16 -3.1 28.8 32.0 -10.5 32.9 43.4 1.0 42.5 41.5
17 -5.2 28.9 34.1 -11.5 33.1 44.7 1.0 43.1 42.1

137 -24.4 42.1 66.4


138 -24.4 42.1 66.5
139 -24.4 42.2 66.5
140 -25.4 42.2 67.6
141 -25.4 42.2 67.7

For brevity, only the first seventeen data rows, up to the peak for the 9-amp test, and the

last five rows, up to the peak for the 6-amp test were shown in Table 11 above. All data is

graphed in Figure 18 below.


Thermoelectric Effects 36

Figure 18: Temperature Differences from Stacking Two Peltier Modules (Hypothesis 8)

70

60

50
Temperature Difference C

Temp Diff C @ 6 amps


40 Temp Diff C @ 3 amps
Temp Diff C @ 9 amps

30

20

10

0
0 20 40 60 80 100 120 140
Time (seconds)

The Peltier modules used for the stacking test were the 40 mm sq. 127 pellet pair (for

which results were graphed earlier in Figure 13) and the 30 mm sq. 71 pellet pair modules

(graphed earlier in Figure 15), because both were capable of handling nine amps of current. The

resulting temperature differences were not as expected in Hypothesis 8, that they would be close

to the total of each module alone. For example, the expected results for three amps would have

been would have been 83.9 °C (total of 43.1° + 40.8°), instead of 56.6 °C as in Figure 18 above.
Thermoelectric Effects 37

Also, the total for six amps should have been about 120 °C (60° for each module), instead of

67.7 °C for the stacked modules. Without further research, it is difficult to say exactly why

stacking modules does not give simple additive results. It is suspected that the modules were too

close in size, and that the cooling effect from the larger one was not sufficient to act as an

adequate heat sink for the hot side of the other module. The smallest 10 mm module could not

be used, as it was broken in earlier tests. The most unexpected result was for the nine amp test,

where the temperature difference only reached 41.5 °C, and then declined over time. This result

indicates that there are a peak number of amps that gives the greatest temperature difference.

Exceeding or being lower than this amount causes the temperature difference to decline.

Pictures

Photo 1: Top wire is original thermocouple, connected only at one end. Bottom wire had

one of the two different metal wires cut and soldered to the other to form two junctions.
Thermoelectric Effects 38

Photo 2: Soldering two ends of the different metal wires to form a second junction.

Photo3: Immersing the wire junctions into boiling and ice water and to produce 6.2 millivolts.
Thermoelectric Effects 39

Photo 4: Creating an extreme temperature difference. Dry ice in pure alcohol at -63°C on left,

Photo 5: And the hot bath of near boiling antifreeze at 195°C, for 258°C difference.
Thermoelectric Effects 40

Photo 6: Peltier module on “heat sink” being preheated to 110°C for Seebeck test.

Photo7: Applying dry ice to other side of Peltier module to generate Seebeck voltage.
Thermoelectric Effects 41

Photo 8: Preparing a Peltier module to test Peltier Effect - creating a temperature difference.

Photo 9: Ready to apply current and record temperature using infrared thermometer.
Thermoelectric Effects 42

Photo 10: Peltier test just after recording -14°C temp on the cold side and 25.2°C on hot side.

Photo 11: Preparing to perform the Peltier Effect test with “stacked” Peltier modules.
Thermoelectric Effects 43

Photo 12: Stacked test with multimeter temperature probe under Styrofoam at left.

Photo 13: Data logging from multimeter to computer to record voltage and temperature.
Thermoelectric Effects 44

Calculations

The only calculations that needed to be done for this experiment were to calculate the

theoretical voltages in Table 2 for metal wire combinations with two junctions exposed to

temperature differences, which were compared to the actual measured voltages. For the

calculation, eFunda, an online reference for Engineers, states that “If the Seebeck coefficients are

nearly constant across the targeted temperature range, VOUT = (SA – SB) × (TTip – TRef)” (eFunda,

2009). SA and SB are the Seebeck coefficients for the metals being tested, and TTip – TRef for this

experiment are the temperatures at the two junctions of the metal wires. The metal wires used in

this experiment were from four thermocouple types, for which the Seebeck coefficients were

provided by the manufacturer, Nanmac Corporation. The actual voltages measured in the tests

came out very close to the theoretical, with the maximum difference being 9.7% for three tests at

the minimum temperature difference of 22.8 °C. The average difference was only 2.8%.

Error Analysis

Errors could have entered into this experiment due to inaccuracy in instruments and due

to human error. However, the researcher does not believe that the combined effects of both had

any impact on proving or disproving the hypotheses or on conclusions drawn. There were three

electronic devices used in this experiment, a DC power supply, a data-logging multimeter

(including a temperature probe), and an infrared thermometer including a thin wire thermocouple

thermometer. For the DC power supply, the accuracy of the current setting used in the Peltier

tests for Figures 13-15 is +/- 2%. For the current range used of 3-9 amps, this would only have

produced an error of 0.06-0.18 amps, which would have an insignificant effect on the

temperature differences produced, especially since a relative comparison was being made

between results at different current levels. Both the infrared thermometer and its thermocouple
Thermoelectric Effects 45

have an accuracy of +/- 2%, and this amount of error would not have a meaningful effect on the

results, especially in terms of proving or disproving a hypothesis. The multimeter has a voltage

reading accuracy of 0.5% and a temperature probe accuracy of 0.75%, and neither is considered

even close to being able to produce an error that would affect the interpretation of results.

There were several ways that human error could have occurred, such as placing the dry

ice in a level even position on the Peltier module, reading temperatures and voltages when these

had to be done manually, placing the thermocouple probes in the proper spot in the ice and hot

baths, and holding the torch under the center of the heat sink. Great care was taken to avoid any

inaccuracies, and only one is suspected to have occurred. In figure 12, the 40 mm 127 pair

Peltier module only reached 5.8 volts, while the 30 mm 127 pair module reached 6.15 volts. The

40 mm module may not have reached the higher voltage of the 30 mm because of its larger size,

since the dry ice may not have been covering the entire module as effectively as the smaller one.

DISCUSSION

Conclusions

This project explored two main aspects of thermoelectricity, the Seebeck Effect and the

Peltier Effect, and eight hypotheses were created to investigate some of the interesting aspects of

these effects. Hypotheses related to the Seebeck Effect were tested using both base metals from

thermocouple wires and semiconductor-based Peltier modules. The Peltier Effect was tested

using only Peltier modules, as difficulty was encountered trying to impose a voltage and current

on a circuit made up basically of two types of wire, without creating a short circuit tripping

circuit breakers. Almost all of the hypotheses were proven correct, as discussed below.

Hypothesis 1 relates to the Seebeck effect, and the data in Table 2 and graph in Figure 8

confirm that the hypothesis is correct. The graph in Figure 8 shows that voltage increases in
Thermoelectric Effects 46

almost a perfect linear proportion to the temperature difference, and that the larger the difference

in Seebeck coefficients between the metal wire combinations, the larger the voltage produced.

All measured voltages were very close to calculated theoretical voltages.

Hypothesis 2 has two parts, the Seebeck voltage generated should be independent of wire

thickness, and Seebeck voltage should not decline over time after reaching a “steady state.” The

graph in Figure 9 confirms the Seebeck voltage to be independent of wire thickness, as the lines

for each wire thickness are almost identical across the temperature range. Table 3 and Figure 10

show that after the voltage bolts up to its peak, it then declines slightly until stabilizing at a

steady state around six millivolts. Therefore, both parts of Hypothesis 2 were proven correct.

Tables 4-5 and the graph in Figure 11 did not support Hypothesis 3, that thinner wire

should have a quicker response time, since all four wire thicknesses reached a peak voltage

around six millivolts in about five seconds. Although the results did not support the hypothesis,

the researcher suspects that the wires used were all too thin to make any observable difference in

the temperature range tested. The wires tested were all less than one millimeter thick. To see a

significant difference, the range of thicknesses would probably have to be much wider, likely in

the 1-6 millimeter range. The wires tested are all considered to be “fine gauge,” and there was

no real measurable difference in response time. In industry, thermocouples may need to stand

the intense heat from a furnace for example, often requiring much thicker wire than used here.

The data in Table 6 and graph in Figure 12 proved Hypothesis 4, that the Seebeck effect

voltage produced from semiconductor-based Peltier modules would produce many times more

electricity than base metal wire, even made from thermocouples. This was proven by imposing

several hundreds of degrees temperature difference to both, and the metal wire produced a
Thermoelectric Effects 47

maximum of only 16.7 millivolts (Table 2), while the Peltier modules produced a maximum

voltage of 6.15 volts, 368 times more.

Figure 12 proves Hypothesis 5 to be correct, that Peltier modules with more pairs of

semiconductor pellets should produce proportionally more voltage than modules with less,

although it was not a perfect linear proportion. For example, seven pairs produced 0.42 volts

while seventy-one pairs produced 3.77 volts, less than the 10:1 ratio of a linear proportion. One

unexpected result was that the 40 mm Peltier device produced less voltage than the 30 mm one,

even though both have 127 pellet pairs. This was most likely caused by the difficulty of

perfectly covering the cold side of the Peltier devices with dry ice. The whole surface of the

large Peltier device probably was not fully covered, thus lowering the temperature difference and

causing the largest Peltier device to produce less voltage than it could have.

Tables 7-9 and Figures 13-15 prove Hypothesis 6, that the temperature difference created

from a Peltier device is proportional to the current flowing through it. The data confirmed

Hypothesis 6 since, whenever the current increased, the temperature difference increased along

with it. Although the relation was not a direct linear proportion, Hypothesis 6 only stated that

the temperature difference created was proportional to current, so the hypothesis was confirmed.

Table 10 and Figures 16-17 support Hypothesis 7, that Peltier devices should maintain a

voltage after the power supply is turned off, and this should decline in a non-linear way over

time. This conclusion is best viewed in Figure 17, which begins one second after power off,

since voltage is held for a while because of the temperature difference created. It then drops off

in a non-linear curve as the temperatures approach room temperature.

Hypothesis 8, if the Peltier devices are stacked, then the temperature difference will be

the sum of both devices, was proven largely incorrect. Except for the nine amp test, Figure 18
Thermoelectric Effects 48

does show that stacking two modules produces an increased temperature difference compared to

what each one can generate individually (see Figures 13 and 15), but the results were not even

close to the sum of both of the temperature differences. The detailed discussion under Figure 18

provides a possible explanation, which concerns the modules being too close in size for the cool

side of the larger module to sufficiently cool the hot side of the smaller module. Unfortunately

the smallest module had already broken down in an earlier test for it to be used. The test at 9

amps was a real failure, since the temperature difference barely reached 42 °C, while the devices

individually reached differences of between 66 °C and 75 °C. It is thought that the modules used

were only effective in a stacked manner at low current levels, since even the six amp test reached

a difference of only 67 °C, compared to about 60 °C for each individually.

Applications

Thermoelectric (TE) technology has evolved to the point where lab-based systems are

reaching efficiencies of more than 6%, with the possibility of attaining 20% conversion

efficiency in the future. That's still too inefficient to be used for applications like auxiliary

power generation or general cooling and heating, but it's ideal for some smaller niche

applications (Ouellette, 2007). Peltier modules are already used in applications that are too small

to use compressor cooling, such as cooling CPU’s in high speed computers and for beverage

coolers that plug in to cigarette lighters in automobiles. Their ability to switch instantly from

heating to cooling by reversing the flow of electricity is finding use in a desktop heater/cooler,

now being developed by a company called Herman Miller; infrared sensor coolers; and

scavenging waste heat to both heat and cool car seats, as needed.

Ford, Nissan, and Hyundai are among the automotive companies using TE devices to

capture and recycle heat in some car models. NREL has a $16.2 million program to further
Thermoelectric Effects 49

develop thermoelectric waste heat recovery schemes for passenger cars to improve fuel

efficiency, thereby reducing the use of fossil fuels and associated CO2 emissions (Ouellette,

2007). BMW and Ford are said to be exploring similar waste exhaust heat recovery and

conversions into electricity to increase gas mileage. Thermoelectric technology is also becoming

a direct competitor to solar power as an alternative for residential, commercial and industrial

fuel-fired heating systems, not to mention diesel-powered electric generators

Recommendations

This experiment can be modified to include two related thermoelectric concepts, the

“Spin Seebeck Effect” and the “Thompson Effect.” The Spin Seebeck effect was discovered by

Japanese scientist Eiji Saitoh of Keio University in October 2008 (Spintronics-Info.com, 2008).

This effect can be thought of as another form of the Seebeck effect. Instead of producing

electricity, this effect generates a magnetic field (ironically what Thomas Seebeck though he had

done almost 200 years ago). The reason why it is called the Spin Seebeck effect is because to

generate a magnetic field the electrons rearrange themselves according to their spins. Possible

applications for the Spin Seebeck effect are on the cutting edge such as power spintronic devices

(an emerging field of science). These spintronic devises could be used in a new computer chip

design that could reduce waste heat generated or even eliminate this. Japanese researchers

found that the new Spin Seebeck Effect can be used to make a “magnetic battery.” The

Thomson effect can be thought of as a combination of the Seebeck and the Peltier effect. The

Thomson effect is the emitting or absorption of heat when electric current passes through a

circuit composed of a single material that has a temperature (Thomson Effect, 2009).

One suggestion for conducting this project would be to test Hypothesis 3, regarding

response time of the wire in producing voltage from a temperature difference, using a range of
Thermoelectric Effects 50

wire thicknesses reaching up to several millimeters. In addition, subjecting the wire junctions to

a higher temperature (e.g. using a torch) would be possible with the thicker wire, and a

difference in response time would likely then be measurable. Another second recommendation

would be to find a way to test the Peltier effect with base metal wire, rather than with just the

Peltier modules. It might be possible to avoid tripping circuit breakers by adding a load, such as

a properly sized resistor, to the circuit. This might allow electricity to be used to create a

temperature difference at the junctions of the metal wires, which would probably have to be

measured with a thin thermocouple wire probe held against the junctions using Styrofoam. A

third recommendation would be to use a thermocouple probe held against the cold side of the

Peltier modules using thermal paste with a slightly grooved piece of Styrofoam to allow data

logging all of the cold side temperatures to a computer, rather than reading it manually with an

infrared thermometer. This was only done in the final stacked Peltier test, and worked very well.

Some advice to anyone trying the experiment named “Cool Junctions” from the Science

Buddies website is that a more sensitive multimeter is needed than the one recommended with a

200 millivolt scale. Using combinations of copper, iron or aluminum wire and a typical 100 °C

difference (ice and boiling water), without resorting to dry ice and boiling antifreeze, voltage

generated would only be about 1.5 millivolts, barely perceivable on a 200 millivolt scale. Using

constantan will help to reach about 6 millivolts, but this is not available at local hardware stores,

only from thermocouple suppliers. This experiment used a multimeter that measures to one-

tenth of a millivolt.

Some other advice: Before handling dry ice, be sure to use gloves, since dry ice is about

74 °C below zero (-109 F). Do not put the Peltier module over its design temperature limit, as

this melt the solder holding it together, which occurred in testing the 10 mm module. Before
Thermoelectric Effects 51

soldering, be sure to put flux on the wire, because without it the solder would not stick properly.

Always use a heat sink with a fan below it to prevent overheating the device. Watch the Anti-

freeze carefully, since it is flammable. Before using thermal paste, make sure it is placed in hot

water as it is very think.

BIBLIOGRAPHY

Eshbach, O. (1952). Handbook of Engineering Fundamentals (2nd Ed.). New York, NY: John

Wiley & Sons, Inc.

Ed. McGrath, K., & Travers, B. (2007). "Seebeck Effect." World of Scientific Discovery. At

galenet.galegroup.com/servlet/SciRC?ste=1&docNum=CV1648500546. Accessed

9/15/2009.

Salman, A. (2009). Cool Junctions. At http://www.sciencebuddies.com/science-fair-

projects/project_ideas/Elec_p031.shtml. Accessed 11/29/2009

Duckworth, H. (1960). Electricity and Magnetism. New York, NY: Holt, Rinehart & Winston.

Ed. McGrath, K., & Travers, B. (2006). "Thermocouple." World of Invention. At

galenet.galegroup.com/servlet/SciRC?ste=1&docNum=CV1647500806. Accessed

9/15/2009.

Society for Science and the Public. (2009). International Rules for Precollege Science Research:

Guidelines for Science and Engineering Fairs 2009-2010. Washington, DC: SSP.

Ferrotec Corporation. (2009). Thermoelectric Modules. At

http://www.ferrotec.com/technology/thermoelectric. Accessed 10/06/2009.

Tellurex Corporation. (2006). Introduction to Thermoelectrics. At

www.tellurex.com/cthermo.html. Accessed 9/7/2009.


Thermoelectric Effects 52

Whyte, D. (2009). Icy Hot Electricity: The Thermoelectric Effect. At

http://www.sciencebuddies.org/science-fair-projects/project_ideas/Elec_p055.shtm.

Accessed 11/29/2009

Noll, S (2008). Physics 51b Lab 12: Peltier Coolers. At

www.physics.pomona.edu/courses/Phys51b/Lab12_PeltierCooler.pdf. Accessed

11/29/2009

Melcor Corporation. (2009). Thermoelectric Handbook. At

http://www.melcor.com/handbook.html. Accessed 9/20/2009.

Warner, F. (2004). "Hot Snacks, Cold Snacks." (Cars)(GADGETS: Keeping You Warm, Well

Fed and Safe). The New York Times, Oct 27, 2004 pG2.

eFunda. (2009). Thermocouple Theory. At

http://www.efunda.com/designstandards/sensors/thermocouples/thmcple_theory.cfm.

Accessed 9/23/2009.

Ouellette, J. (2007). Waste Not, Want Not. At

http://blogs.physicstoday.org/industry07/2007/10/waste_not_want_not.html. Accessed

11/28/2009.

Spintronics-Info.com (2008). Japanese researchers find new "spin Seebeck effect", can be used

to make a 'magnetic battery.' At http://www.spintronics-info.com/japanese-researchers-

find-new-spin-seebeck-effect. Accessed 11/29/2009.

Thomson Effect (2009). Thomson Effect Encyclopædia Britannica Online. At

http://www.britannica.com/EBchecked/topic/593136/Thomson-effec Accessed

11/30/2009
Thermoelectric Effects 53

APPENDIX – PROJECT LOG

Week beginning August 30, 2009

Today the researcher searched the Internet for a science fair project. The researcher found

many projects and began listing possible materials and devices that would be needed for the

projects that had potential. After this the researcher decided to base his project on a project

found on Science buddies called “Cool Junctions”. Even though his project was based on cool

junctions the researcher decided to add multiple different hypotheses to make the project his

own. After finding this project the researcher began collecting information on the two

thermoelectric effects he would be testing, the Peltier and Seebeck effect.

Week beginning September 6, 2009

The researcher and adult sponsor began looking up websites that sold the supplies needed

to do the experiment chosen and made a list with the devices. The researcher and adult sponsor

read the information collected to understand the effects that were going to be tested better. The

adult sponsor contacted several suppliers to ask questions about the products sold. The adult

sponsor and the researcher went to several electronic stores (e.g. Radio Shack, Warrick’s hobby

shop, Lafayette Electronics, and Alpha electronics) to see what they had. The researcher and

adult sponsor met with a retired master electrician who worked in his career at his store to

discuss the project.

Week beginning September 13, 2009

The researcher tested several times to see if vital parts of the experiment such as the

Seebeck effect (which did not come close to producing the amount of electricity hoped for)

worked as planned. In an attempt to get measurable results the researcher twisted three metal

wires together (hoping this would triple the voltage) and tested the Seebeck effect again using
Thermoelectric Effects 54

the lighter as a heat source and an ice bath to cool another junction to produce maximum voltage.

The results were too small to be read by the multimeter so the researcher took the data logger

from his previous experiment last year and found the Seebeck effect only was producing less

than 2 millivolts. The researcher decided to get bigger results he needed to find an alloy called

constantan (about 55% copper and 45% nickel) since it had a large negative Seebeck coefficient

while most other metals had a positive Seebeck coefficient. This is helpful because the amount

of voltage produce is determined by the temperature difference between the two junctions and

the difference in Seebeck coefficients between the two wires. The greater both of those numbers

are the more voltage produced. The researcher also decided he needed to find semiconductors

since they have a high Seebeck coefficient. The researcher, finding himself at a dead end,

decided to base his project on ideas from both the Cool Junctions project and another project

called Icy Hot Electricity, which uses a commercial Peltier module from an Igloo car cooler.

The researcher found many articles and printed them and organized them into a larger binder

with tabs. The researcher worked on a write up for doctor Golden, worked on I.S.E.F. forms,

and the researcher’s adult sponsor went to Dr. Golden’s class too.

Week beginning September 20, 2009

The researcher began to come up with ideas on how to expand his project and made four

aspects of his experiment. The researcher and his adult sponsor began researching the materials

and devices they might need for the project. The researcher began to discuss attending Dr.

Golden’s class with his parents.

Week beginning September 27, 2009

The researcher continued to search for suppliers for the materials and devices needed.

The researcher and Adult sponsor decided to get their supplies from amazon.com. These devices
Thermoelectric Effects 55

included a power supply, thermometer, multimeter, and semiconductors. The researcher and his

parents decided to leave Dr. Golden’s research class.

Week beginning October 4, 2009

The researcher decided that he did not want to follow the advice and buy a whole Peltier

cooler and take it apart just to get the small Peltier devices. Instead he decided to order Peltier

devices without the cooler to save money and time. The researcher and adult sponsor found a

supplier that had constantan in the form of thermocouples, which contain two metals with a high

Seebeck coefficient difference. The multimeter arrived on Friday 10/10 and the infrared

thermometer arrived Saturday 10/11.

Week beginning October 11, 2009

The researcher had no time to read the instructions on Sunday because his adult sponsor

wanted him to revise his right up so it would fit all of Mrs. Page’s requirements. In the end the

researcher had eight hypotheses to cover everything he wanted to test.

Week beginning October 18, 2009

This week the researcher decided to test the power supply and Peltier devices received to

make sure they worked. The adult sponsor polished the heat sink until one could see his or her

own reflection. The researcher also received thermal paste to increase electrical conductance so

the Peltier devices wouldn’t over heat. The thermocouples arrived from U.P.S. The researcher

and the adult sponsor labeled each of the thermocouples by the type of metal it was, the Seebeck

coefficient, wire thickness, and the maximum temperature it can handle. Each wire was labeled

with a letter. The information on each letter wire was put into a spreadsheet.

Week beginning October 25, 2009


Thermoelectric Effects 56

The researcher began testing. The wires are 12 inches long. The researcher and adult

sponsor began soldering, the metal wires together.

When soldering, first one must hold the wire by using alligator clips. After this is done

one must add flux to make sure the solder sticks on. Then one must use the soldering gun to heat

the wire enough so the solder could melt on it. After applying the solder one must simply let the

solder cool, which happens very quickly. After the wires were soldered the researcher tested the

Seebeck effect’s data found on the Internet compared to the results one could get from testing it

him or herself. The researcher and adult sponsor did two tests per wire type. One was boiling

water and ice bath; and the other was the ice bath and open air. After testing the researcher and

adult sponsor put all of the data into an excel spreadsheet. In the spreadsheet the researcher

multiplied the Seebeck coefficient difference and the temperature difference to get the theoretical

voltage then found the results to be remarkably accurate. After testing the boiling water and ice

bath the researcher increased the temperature difference by increasing the temperature difference

by boiling anti freeze 194 degrees centigrade and creating a dry ice bath with dry ice and pure

alcohol -64 degrees centigrade. The researcher and adult sponsor wore glasses and gas masks at

all times. To boil antifreeze, the burner on a barbecue was used so there are no poisonous vapors

in the house. Then the researcher finished his research plan, which took two days.

Week beginning November 1, 2009


Thermoelectric Effects 57

The researcher conducted tests with type E wire at the gages: 0.010, 0.015, 0.020, 0.032

inches. The test was done to see if wire thickness would change how quickly the wire reached

steady state. The expectations were the thicker wire would take longer to reach steady state.

The results may have been partially affected by the burner turning on and off to keep the water at

the right temperature. What was found was the wire thickness did not seem to affect the

electricity generated.

Week beginning November 8, 2009

Today the researcher tested the Peltier devices’ ability to produce electricity via the

Seebeck effect. To do this, the heat sink was suspended over the counter top by attaching the

heat sink to the side of a regular, heavy, cooking pan. When the heat sink was set up thermal

paste was applied to the spot the Peltier device was going to be placed. A digital thermometer

was inserted into the hole drilled into the heat sink and the tip of the thermocouple was right

underneath the Peltier devices. Since handling a blowtorch can be dangerous the adult sponsor

held the flame underneath the heat sink right at the spot the Peltier device was to heat the other

side of the Peltier device. While the adult sponsor heated the heat sink the researcher monitored

the temperature of the heat sink to make sure it reached 110 degrees Celsius so the maximum

amount of voltage could be created without melting the Peltier device. With one side of the

Peltier device hot the adult sponsor placed the dry ice to make one side cold while the researcher

multimeter to find and tell the adult sponsor when the peak of the voltage was reached since,

after the peak was reached, all the Peltier device would do was slowly lose voltage unit it ceased.

Another reason why the voltage was only recorded until it reached its peak was the fact that

eventually all of the dry ice would go through sublimation (the process matter goes through

when it skips the liquid state and directly goes from a solid to a gas.
Thermoelectric Effects 58

Each device was tested three times and there were four devices. The two devices with

127 pairs reached a maximum voltage of about six volts, while the device with 71 pellets reached

3.8 volts, and the device with 7 pairs peaked at 0.4 volts. The six-volt maximum of this test was

several hundred (368) times higher than the highest voltage (16.7 millivolts) for the wire tests.

After this the data was made into charts and the data was also put into tables.

Week beginning November 15, 2009

With the remaining 3 Peltier devices (one of the device’s solder melted causing the whole

thing to fall apart) the Peltier effect was tested. For safety precautions a fan was taped to the

bottom of the heat sink to make sure it did not overheat. The fan with the heat sink on top of it

was taped between two stacks of books so the fan could draw air. When testing each device the

Peltier modules were again pasted onto the heat sink with thermal paste so the heat sink could

absorb the heat from the hot side of the Peltier. The temperatures on the cold side of the Peltier

were measured using the infrared thermometer. After each test the Peltier device had to be

wiped off since ice was building up on the device. This was probably caused by condensation

freezing on the devices. Every five seconds the researcher read off the temperature of the cold

side of the Peltier device, and the adult sponsor read the temperature of the hot side, and these

were immediately typed into an excel spreadsheet. Each device was tested at different amounts

of amps and each combination of amps and Peltier module was tested multiple times.

Week beginning November 22, 2009

To test stacking Peltier devices the second largest Peltier device was placed on top of the

largest Peltier device with thermal paste in between each device and on top of the heat sink to

absorb heat. The fan was on at all times to keep the devices from heating up. The temperature

of the Peltier device was planned to be monitored with the infrared thermometer and the data was
Thermoelectric Effects 59

recorded by hand. Even with the fan on the bottom so much heat was generated from the Peltier

devices that the thermal paste melted causing the devices to move. To counter act this, the

researcher and adult sponsor bought some Styrofoam and taped to the top of a Peltier device to

keep it from moving. Since there was some Styrofoam on top of the Peltier device it was

impossible to monitor the device with the infrared thermometer. Instead a probe was placed

underneath the Styrofoam and was attached to the Peltier device with thermal paste. The thermal

paste also gave the probe a more accurate measurement since thermal paste conducted heat

better. Data was pulled from the digital thermometer into the computer directly for the cold side

and the how side was measured manually every five seconds. After this the data was pulled into

an excel file were graphs and data tables were made. The files were saved and with every part of

the science fair project tested the devices were put away.

Potrebbero piacerti anche