Evaluation of Concentration Uncertainty and Propagation Calculation

In titrimetric analysis, achieving high-precision concentration data relies not merely on stoichiometric accuracy but on a rigorous assessment of uncertainty throughout the measurement process. The evaluation and propagation of concentration uncertainty serve as the cornerstone for quantifying the reliability of analytical results. This systematic approach enables researchers to identify dominant error sources, optimize experimental protocols, and derive statistically valid conclusions. This guide outlines the fundamental principles of uncertainty evaluation and the mathematical framework for its propagation.

Fundamental Components and Sources of Uncertainty

In titrimetry, the uncertainty associated with the final concentration is rarely a single entity; rather, it is a composite result derived from multiple independent or correlated measurement components. Adhering to the principles outlined in the Guide to the Expression of Uncertainty in Measurement (GUM), these components typically include:

Uncertainty of Standard Solutions: This stems from the purity of primary standards, weighing errors, and volumetric deviations during solution preparation.
Burette Reading Errors: These encompass instrument calibration limits, meniscus reading deviations, and operator estimation errors.
Sample Weighing Errors: Influenced by the precision of the balance, hygroscopic nature of the sample, and potential mass gain or loss due to environmental exposure.
Endpoint Judgment Errors: Arising from the sharpness of indicator color changes and the subjective interpretation of these shifts by the operator.
Temperature Correction Errors: Caused by uncorrected thermal expansion effects altering solution volumes.

Each component is generally expressed as a standard uncertainty ($u$). These values are determined either through Type A evaluation (statistical analysis of repeated measurements) or Type B evaluation (using manufacturer specifications or literature data divided by an appropriate coverage factor).

Mathematical Model for Combined Uncertainty

When an output quantity $y$ (the concentration $C$) is a function of multiple input variables $x_1, x_2, ..., x_n$, the combined standard uncertainty $u_c(y)$ is calculated using the law of propagation of uncertainty. If the input variables are uncorrelated, the formula relies on the summation of variances:

$$u_c(y) = \sqrt{\sum_{i=1}^{n} \left( \frac{\partial y}{\partial x_i} \cdot u(x_i) \right)^2}$$

Here, $\frac{\partial y}{\partial x_i}$ represents the sensitivity coefficient, indicating how sensitive the concentration is to changes in the $i$-th input. In titrimetry, concentration is typically defined as $C_{analyte} = \frac{C_{titrant} \cdot V_{titrant}}{V_{sample}}$. Consequently, the sensitivity coefficients for the primary variables are:

For titrant concentration ($C_{titrant}$): $\frac{V_{titrant}}{V_{sample}}$
For titrant volume ($V_{titrant}$): $\frac{C_{titrant}}{V_{sample}}$
For sample volume ($V_{sample}$): $-\frac{C_{titrant} \cdot V_{titrant}}{V_{sample}^2}$

Simplification via Relative Uncertainty

While direct calculation of partial derivatives can be computationally intensive, practical applications in engineering and teaching often benefit from using relative uncertainty. Since concentration is calculated via multiplication or division of measured quantities, the relative combined standard uncertainty $u_r(y)$ simplifies the process significantly. It is defined as the ratio of the combined standard uncertainty to the measured value:

$$u_r(y) = \sqrt{ \left( u_r(C_{titrant}) \right)^2 + \left( u_r(V_{titrant}) \right)^2 + \left( u_r(V_{sample}) \right)^2 }$$

This equation reveals that the relative uncertainty of the concentration is the square root of the sum of the squares of the relative uncertainties of the input variables. This implies that if one input component possesses a significantly larger relative error, it will disproportionately affect the overall precision of the result. For instance, if a burette has a reading uncertainty of $\pm 0.05 \text{ mL}$ against a measured volume of $20.00 \text{ mL}$, the relative uncertainty is approximately $0.25%$. If the sample weighing introduces a larger relative error, it becomes the critical factor requiring immediate procedural improvement or higher-precision instrumentation.

Expanded Uncertainty and Confidence Intervals

The ultimate goal of uncertainty evaluation is to define a range within which the true value is expected to lie with a specified probability, known as the expanded uncertainty ($U$). This is achieved by multiplying the combined standard uncertainty by a coverage factor ($k$). For a 95% confidence level, $k$ is commonly set to 2:

$$U = k \cdot u_c(y)$$

When reporting analytical results, strict adherence to significant figure rules is essential. The last digit of the uncertainty must align with the last digit of the reported measurement value. For example, if the calculated concentration is $0.10234 \text{ mol/L}$ with an uncertainty of $0.00045 \text{ mol/L}$, the result should be rounded and reported as $0.1023 \pm 0.0005 \text{ mol/L}$.

Practical Demonstration

Consider a scenario involving the standardization of a hydrochloric acid solution using anhydrous sodium carbonate as a primary standard:

Mass of Primary Standard: $m = 0.1500 \text{ g}$ (Purity $99.9%$, Molar Mass $M = 105.99 \text{ g/mol}$). The relative uncertainty is $u_r(m) = 0.1%$.
Volume of Titrant: $V = 25.00 \text{ mL}$. The relative uncertainty of the burette is $u_r(V) = 0.08%$.
Calculated Concentration: $C = 0.1002 \text{ mol/L}$.

Applying the relative uncertainty formula:
$$u_r(C) = \sqrt{(0.1%)^2 + (0.08%)^2} \approx 0.13%$$

The standard uncertainty of the concentration is then:
$$u_c(C) = 0.1002 \times 0.13% \approx 0.00013 \text{ mol/L}$$

Using a coverage factor $k=2$, the expanded uncertainty is approximately $0.00026 \text{ mol/L}$. Following rounding rules, the final reported result is:
$$C(\text{HCl}) = 0.1002 \pm 0.0003 \text{ mol/L} (k=2)$$

Through this systematic evaluation and propagation of uncertainty, analysts can scientifically delineate the confidence interval of titrimetric results, providing a robust foundation for subsequent quantitative analyses.