Kress 1

Experimental Determination of Absolute Zero (Celsius) Using the Ideal Gas Law

O. Kress and S. Bhatta

Abstract:

Our goal was to find absolute zero on Celsius scale and test if the gas law applies for reasonable

gas temperatures. We recorded temperature and pressure data from a thermocouple and pressure

sensor of a fixed volume of gas and used the ideal gas law to extrapolate the expected

temperature when the pressure equaled zero. This temperature was out experimental value for

absolute zero on the Celsius scale. Our value of -269.639 differed from true (-275.15) by only

3.511 degrees, or 1.285%. Our linear regression results for Table 1 indicate a very uniform,

linear group of data points. The correlation coefficient (R), the coefficient of determination (R2),

and the adjusted R2 all show values of 0.999 where 0 indicates an unpredictable set of data points

for which a linear regression does not represent or predict the data, while 1 shows a perfect

representation and prediction. Although we obtained 90 data points, we only calculated a single

experimental value of the absolute temperature, so a standard error of multiple values of this was

not possible.

Introduction:

This experiment tested the theory behind the ideal gas law, first stated by Clapeyron in 1834. The

ideal gas law is:

pV =nRT
 Kress 2

p Pressure
V Volume
n Number of moles
R Universal Gas Constant
T Absolute Temperature

This equation is familiar to anyone who has taken a course in science during their pre-collegiate

academic career. One notices the simplicity and linearity of the formula. For this experiment we

take advantage of the linearity between the fixed volume of ideal gas, the pressure of the gas, and

its absolute temperature. This relation is commonly used in chemistry to solve stoichiometric

problems involving gases and conversions to and from moles of gas to the volume of a gas.

Technique:

We realized that our experiment would have to correlate the temperature of a fixed volume of

gas with its pressure. This was accomplished by placing a sealed thermocouple midway down

into a glass bulb filled with air. The bulb was also attached to a set of valves. One of these

allowed a pressure reading from a transducer in terms of voltage while the other valve was

closed just before readings began in order to assure the air temperatures and pressures within and

outside the bulb were equalized.

Then, the bulb was lowered into a large glass beaker filled with ice water. Important details

arose regarding this method. One of these included making sure the entire bulb was covered in

isothermal water. Allowing part of the bulb to be exposed to air would mean the air mixture

within would be warmer near the top and cooler below. Also, the temperature sensor would be

reading only the air temps of the gas in contact with the thermocouple. This would mean that the

pressure readings would be too high for the given temperature reading.
 Kress 3

After a few mistrials, we were able to correctly submerge the air bulb into the ice bath. The

beaker containing the ice bath was put onto a hot plate to allow for warming to near boiling.

Also, a magnetic stirrer was employed to insure consistent water temperatures throughout the

beaker. Temperature readings were taken in 1 degree Celsius intervals while the appropriate

voltages from the pressure sensor were recorded. Our temperatures ranged from 3 degrees to 90

Celsius, while our voltage readings ranged from 3.3750 to 4.141 volts.

During the experiment, the local barometric pressure was recorded from a mercury barometer

located within the lab. This pressure was used to determine the correct relative voltage reading

from the pressure sensor. The current air temperature was also recorded for reference. The air

pressure was recorded to be 14.194 psi and the corresponding voltage reading was 3.5630. The

pressure sensor was calibrated to range from 1 volt at vacuum to 6 volts at 30 psi with an input

voltage from a DC power source of 8 volts.

Data:

Table 1 shows directly the data obtained during the experiment in Celsius and Volts. Graph 1

plots Celsius vs. Volts. Solving the ideal gas law equation for p:

nRT
p=
V

p=0 when T =0. Our goal is to prove this result by extrapolation since we are restricted to

reasonable temperatures. To relate pressure and voltage, we used fact that the barometric

pressure of 14.194 equaled 3.5630 Volts on the voltmeter reading from the transducer. Using

this relation and the given calibration of 0 PSI =1Volt, linearity was exploited to form
 Kress 4

P=f ( V )=aV + b

Using 0 PSI =1Volt:

0=a+b

a=−b

P=aV −a

Using 14.194 PSI =3.5630 Volts:

14.194=a∗(3.5360)−a

a=5.538

So the convert Volts to PSI:

P= (5.538 ) V −5.538 (Eq. 1)

Using SigmaPlot 11.0, we applied (Eq.1) to the Volts column of Table 1 and obtained the

relative pressures (Table 2). We then graphed pressure (PSI) vs. Celsius (Graph 2). SigmaPlot’s

analysis showed when PSI = 0, Celsius = -269.639. This is our experimental value for absolute

zero on the Celsius scale. The accepted value for absolute zero is -273.15 Celsius.

Uncertainties:

The confidence interval of 99.999%, so the data interval is 99.999% sure to include the best-fit

regression line. The sum of the squares came out to be 0.00477 and the mean sum is 0.0000542.

There are unknown uncertainties in the temperature gauge, the voltmeter, the pressure sensor

transducer, and the mercury barometer’s accuracy and precision. Also, it is possible the valves

were not completely airtight.