Metrohm 757 VA Computrace - Manual (Page 133)

5.8 Data processing and evaluation
757 VA Computrace – Software
123
5. Calculation of result deviation MC.dev.
The total deviation of the calculated substance mass concen-
tration Mass.conc. is determined using a linear error calcu-
lation which takes into account both the error contribution
from the measurement and that from the calibration. Inde-
pendent of the number of measurements, the total deviation
MC.dev. is always calculated in a way that Mass.conc. ±
MC.dev. gives the range in which the mass concentration
may be expected with a probability of 68.3%.
Rules for calibration curves
The result determination with the aid of a calibration curve
saves time compared with standard additions, but is reliable only
• if the matrix of all samples and calibration solutions is identi-
cal or has no influence on the measurement
• if all measurement parameters (capillary, temperature, etc.)
remain unchanged during measurements
• if the accuracy of the results obtained is checked regularly
with the standard addition method.
With regard to optimum accuracy and scatter, a number of rules
must be observed with calibration curves:
• Check linearity range
In development of the method, the linearity range of the cali-
bration curve should be checked for each substance by re-
cording the curve over a wide concentration range. Using the
calibration curve shown in the
DETERMINATION CURVES win-
dow, you can then determine the region in which the curve is
linear and that in which it is nonlinear.
• Working in the linear range
Performing determinations in the linear range, to keep the
scatter as low as possible it is advisable to calibrate above all
in the lower and upper part of this range and select as many
replications as possible.
• Checking the offset
The size of the offset indicates a possible systematic error or
blank value. To convert this error into the effective mass con-
centration in g/L, Y.offset must be divided by Slope.