Rate of Dissolution Experiment

Discussion: The rate of dissolution experiment is an excellent hands-on exercise for simple factorial experimentation. Taking less than one hour (for conducting and analyzing the data), participants will learn about factors that impact the mean of the response as well as those which influence variation. The concept of non-linearity and interaction are also well demonstrated in this exercise.

Raw Materials: 50 glasses (plastic, clear), antacid pills (at least 50), one pint of cola, measuring spoon (tablespoon), and thermometer (range from approx 40 deg F to 200 deg F).

Directions: Split group into four teams. Declare one person on each team the captain. Ask the captains to gather together and dissolve one table in warm water. Ask them to reach agreement on what is defined as a “fully dissolved” pill (it is very important that all teams use the same criteria). Once the captains have reached an understanding, they are to use it as basis for all measurements within their groups. The factors for the experiment are:

Factor / Low / High
Water temperature / Cold / Hot
Cola / None / 1 tablespoon

The four runs for the experiment are:

RUN / TEMP / COLA / Time to Dissolve
1 / Cold / 0
2 / Cold / 1
3 / Hot / 0
4 / Hot / 1

Assign one run to each team. Have each team dissolve 7 pills. Run cold water from the nearest available tap until the temperature seems to stabilize. Record the temperature with the thermometer. Do the same with the hot water tab. Record the actual as the water temperature levels.

Typical data: Typical data is as follows (your data will be different):


Typical results:

Exercises: Review the main effects plot for the mean and standard deviation. Review the contour plot. Notice for our contour, .5 tablespoons of coke in conjunction with 50% cold/ 50% hot water are predicted to have a dissolve time of approximately 45 seconds. Ask each team to dissolve a pill and record the times at this point. All times will typically be much lower than the prediction at the center. This suggests our linear model is not adequate.

Some attendees in the class may recall the following rule of thumb from H.S. chemistry: For every 10 degree C increase in temperature, the rate of reaction is doubled (non-linear). In order obtain better predictions from our experiment; three (or more) levels would have been more effective.

Option Exercises: 1. Time permitting; a central composite design (CCD) could be run so as to evaluate its predictive ability. We found the CCD provided good predictive ability over the experimental range. 2. Heat water to near boiling, record the temperature, and dissolve 4 different pill in separate container. See if the contour plot is able to provide a good prediction (extrapolation).

Analysis software:

Analysis was completed using DOE Wisdom software

References on DOE and Variance reduction factors:

“Straight Talk on Designed Experiments”,

“Experimental Design for Injection Molding”,

“Understanding Industrial Designed Experiments”

all the above are available from Launsby Consulting.