Professional Documents
Culture Documents
Central Experimental: Composite Designs
Central Experimental: Composite Designs
Response Surfaces
Figures 1-3 show typical response surfaces that are fre-
quently encountered in chemistry. A response surface is a
plot of the system response (e.g., percent yield of a reac-
tion) versus each of the factors or variables that have an
influence on the response (e.g., temperature and pressure)
(2, 3).
A Maximum
For Figure 1, suppose that the following factors are rep-
resented by the following variables. Figure 2. Response surface for the full two-factor second-order poly-
nomial model.
The response y\ is the percent yield.
y2 = 96.65 1.683X, 2.183x2 + 0.01250(^)2
- -
A Minimum
1
Author to whom correspondence should be addressed. (713) 743-
-
0.01 750(x2)2 -
0.008660x1x2 2809.
25
(0, +1.41)
319
,00 20 -
6 0
*0
5 -
!°
0 -'-:-1-1-1-
0 5 10 15 20 25 30
Drops H202
Experimental
To demonstrate the use of a central composite design, the
absorbance response from a constant volume of a solution
of vanadyl sulfate (VOSO4) can be investigated as a func-
tion of the volume in drops of hydrogen peroxide (H202)
and sulfuric acid (H2S04) that are added (16, 17). Figure 5. Graph of the fitted full two-factor second-order polynomial
The first factor xj is the number of drops of 1% H202. The model, eq 4.
second factor x2 is the number of drops of 20% H2S04. The
measured responseyq is the absorbance. The experimental
factor space can be bounded between 0 drops and 30 drops yl = 334.75 -
31.84*1 -
11.46x1 + 13.13(x*i)2
of H202 and H2S04.
+ 3.63(x2)2 + 24.50x4x2 (3)
When using c4 = 15, di = 5, and eq 2, the following equa-
tions give the real (uncoded) values chosen for both factors. The following is the fitted model in uncoded factor space
lowest (-a) = 8 in real units.
36.82*! -
21.34x2 + 0.53(*i)2