Lecture #5 - Overview
Analytical Response Statistics - Part 3
Calibration Curves
Statistical Tools in Quantitative Analysis The Method of Least Squares Calibration Curves Using a Spreadsheet for Least Squares
Measure of Unknown
Amount/Concentration of Unknown
= Known Amount/Concentration of Standard
Concentration of Standard
Construction of Calibration Curves
Standard Solutions = Solutions containing known concentrations of analyte(s) Blank Solutions = Solutions containing all the reagents and solvents used in the analysis, but no deliberately added analyte
Construction of Calibration Curves
Step 1: Prepare known samples of analyte covering a range of concentrations expected for unknowns. Measure the response of the analytical procedure for these standards.
Serial Dilution e.g.
1x
1/5x
1/25x
1/125x 1/625x
Blank
Measure response with analytical procedure
Construction of Calibration Curves
Step 1: Prepare known samples of analyte covering a range of concentrations expected for unknowns. Measure the response of the analytical procedure for these standards. Step 2: Subtract the (average) response of the blank samples from each measured standard to obtain the corrected value.
Construction of Calibration Curves
Step 1: Prepare known samples of analyte covering a range of concentrations expected for unknowns. Measure the response of the analytical procedure for these standards. Step 2: Subtract the average response of the blank samples from each measured standard to obtain the corrected value. Step 3: Make a graph of corrected versus concentration of standard, and use the method of least squares procedure to find the best straight line through the linear portion of the data. Step 4: To determine the concentration of an unknown, analyze the unknown sample along with a blank, subtract the blank to obtain the corrected value and use the corrected value to determine the concentration based on your calibration curve.
Corrected
Measured - Blank
Calibration Curves
Analytical Response
Measure of Unknown
Amount/Concentration of Unknown
= Known Amount/Concentration of Standard
Concentration of Standard
Method of Least Squares
to draw the best straight line through experimental data points that have some scatter and do not lie perfectly on a straight line
Method of Least Squares
Vertical Deviation = di = yi - y = yi - (mxi + b) di2 = (yi - y)2 = (yi - mxi - b)2
y = mx + b y !y !x y-intercept (b) x Slope (m) = !y !x
We wish to minimize to minimize the magnitude of the deviations (regardless of sign) so we square the terms. This is where Method of least Squares takes its name.
Method of Least Squares
"(xiyi) "xi Slope: m= "yi "(xi2) Intercept: b= "xi "xi n n "(xiyi) "yi D
Determinants
A B D AD - BC
"(xi2) D= "xi
Method of Least Squares
m = n"(xiyi) - "xi"yi n" (xi2) - ("xi)2
"(xi2)"yi - ("xiyi)"xi n" (xi2) - ("xi)2
Method of Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line. Amount Protein (mg) 0 5.0 10.0 15.0 20.0 25.0
Absorbance 0.099 0.185 0.282 0.345 0.425 0.483
Corrected* 0.000 0.086 0.183 0.246 0.326 0.384
* Absorbance - Average Blank (=0.0993)
Method of Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line.
Method of Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line.
m=
xi
" n=6
0 5.0 10.0 15.0 20.0 25.0 75
0 0.086 0.183 0.246 0.326 0.384 1.225
yi
0 0.43 1.83 3.69 6.52 9.60 22.07
x iy i
0 25 100 225 400 625 1375
x i2
n"(xiyi) - "xi"yi n" (xi2) ("xi)2
xi
= (6)(22.07) - (75)(1.225) (6)(1375) - (75)2 m = 0.015445714 " n=6
0 5.0 10.0 15.0 20.0 25.0 75
0 0.086 0.183 0.246 0.326 0.384 1.225
yi
0 0.43 1.83 3.69 6.52 9.60 22.07
x iy i
0 25 100 225 400 625 1375
x i2
b=
"(xi2)"yi - ("xiyi)"xi n" (xi2) - ("xi)2
= (1375)(1.225) - (22.07)(75) (6)(1375) - (75)2 b = 0.01109524
6 data points
Method of Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line.
m = 0.015445714 b = 0.01109524 y = (0.015445714)x + (0.01109524)
Method of Least Squares
to draw the best straight line through experimental data points that have some scatter and do not lie perfectly on a straight line
y = mx + b #y y
(xi,yi)
Vertical Deviation (di) = yi - y
di = yi - y = yi - (mxi + b) (di)2 = (yi - mxi - b)2 x
Uncertainty and Least Squares
#y ! sy = "(d1 - d)2 (degrees of freedom) "(d1)2 (degrees of freedom) "(d1)2 n-2
sy =
sy =
Uncertainty and Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line. Calculate the uncertainty associated with this line.
Uncertainty and Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line. Calculate the uncertainty associated with this line.
xi
" n=6
0 5.0 10.0 15.0 20.0 25.0 75.0
0 0.086 0.183 0.246 0.326 0.384 1.225
yi
0 0.43 1.83 3.69 6.52 9.60 22.07
x iy i
0 25 100 225 400 625 1375
x i2
di (=yi - mx - b) -0.0111 -0.0022 0.0174 0.0032 0.0060 -0.0132
0.00012321 0.00000540 0.00030442 0.00001036 0.00003589 0.00017525 0.00065442
d i2
sy =
"(d1)2 n-2 (0.00065442)/(6-2) 0.0001636
= =
= 0.012790808
Uncertainty and Least Squares
Uncertainty and Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line. Calculate the uncertainty associated with this line.
sm 2 =
sy2n D sy2"(xi2) D
sb2 =
0 5.0 10.0 15.0 20.0 25.0 " 75.0
xi
yi
0 0.086 0.183 0.246 0.326 0.384 1.225
0 0.43 1.83 3.69 6.52 9.60 22.07
xi yi
0 25 100 225 400 625 1375
xi 2
0.00012321 0.00000540 0.00030442 0.00001036 0.00003589 0.00017525 0.00065442
d i2
"(xi2) "xi
"xi n 75 6
D=
1375 D= 75
n=6 sy = 0.012790808
= (1375 x 6) - (75 x 75) = 2625
Uncertainty and Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line. Calculate the uncertainty associated with this line.
Uncertainty and Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line. Calculate the uncertainty associated with this line.
0 5.0 10.0 15.0 20.0 25.0 " 75.0
xi
yi
0 0.086 0.183 0.246 0.326 0.384 1.225
0 0.43 1.83 3.69 6.52 9.60 22.07
xi yi
0 25 100 225 400 625 1375
xi 2
0.00012321 0.00000540 0.00030442 0.00001036 0.00003589 0.00017525 0.00065442
d i2
sm2 = sy2n D = (0.012790808)2 (6) (2625) = 0.000000373954 sm = 0.000611518
0 5.0 10.0 15.0 20.0 25.0 " 75.0
xi
yi
0 0.086 0.183 0.246 0.326 0.384 1.225
0 0.43 1.83 3.69 6.52 9.60 22.07
xi yi
0 25 100 225 400 625 1375
xi 2
0.00012321 0.00000540 0.00030442 0.00001036 0.00003589 0.00017525 0.00065442
d i2
sb2 =
sy2 "(xi2) D
= (0.012790808)2 (1375) (2625) = 0.0000856977 sb = 0.009257307
n=6 sy = 0.012790808, D=2625
n=6 sy = 0.012790808, D=2625
Uncertainty and Least Squares
Example: To analyze protein levels, you use a spectrophotometer to measure a colored product which results from chemical reaction with protein. To construct a calibration curve, you make the following measurements of absorbance (of the colored product) for several known amounts of protein. Use the method of least squares to determine the best fit line. Calculate the uncertainty associated with this line.
Linearity
m = 0.015445714 0.000611518 = 0.0154 0.0006 b = 0.01109524 0.009257307 = 0.011 0.009
Linear Range vs. Dynamic Range
Dynamic Range
Linear Range
Determining Linearity
Square of Correlation Coefficient R2 = ["(xi - x)(yi - y)]2 "(xi - x)2 "(yi - y)2 R2 close to 1 (e.g. " 0.99, 0.98, 0.95) R2 High (>0.95) R2 Low (<<0.95)