The Essentials of Peak Integration: What, When, and How to Optimize Your Results

Integration of chromatographic peaks (determination of height, area, and retention time) is the first and most important step in the data analysis of HPLC and GC methods. This information is used for all subsequent calculations, including construction of calibration curves and calculation of unknown concentrations.

This first installment will define what chromatographic peak integration is and the primary chromatography data system (CDS) settings which are required.

What is Integration?

Chromatographic peak integration defines an operation in which the area under the chromatographic peak is measured. The measurement is based on the integral technique of splitting the peak into a large number of rectangles, which are then summed to provide an estimate of the total area under the peak (Figure 1).

Two events need to be defined in order for the data system to carry out the calculation; these are the peak start and end which are determined using threshold and peak width settings in the data system. The method of determination has to be reproducible for rugged integration. The baseline is then drawn between the peak start and end points created by the data system (Figure 2).

Several other integration events exist to give ruggedness and flexibility to peak integration, however, they can be manufacturer specific, and hence, familiarity with the particular chromatography data system (CDS) in use is essential.

It should be noted that peak height measures the distance between the peak apex and the intersection with the integrated peak baseline; incorrectly integrated peaks may provide incorrect peak height measurements.

The two most important parameters for peak integration are peak threshold and sampling rate.

Peak threshold allows the integrator to discriminate the start and end point of the peak from the baseline (Figure 3). This is achieved based on the rate of change when the detector signal rises above (peak start) or falls below (peak end) the value(s) set in the CDS - front and back slope settings can be applied separately to detect the peak start and end points. Careful setting is required so as not to falsely identify baseline noise as peaks (threshold too low) or to miss small peaks (threshold too high). At the end of the peak if the threshold is too low, too much of the peak tailing is counted. Conversely, if this setting is too high the peak will finish early which will result in an underestimation of the area.

The peak end point is defined as the point at which the signal returns to where the baseline was prior to the start of the peak, however, elution of several peaks or a rising baseline mean that this value will not be reached. A baseline tolerance value can be set to determine how far the baseline can drift from the initial pre-peak value. This is different from peak threshold, as peak threshold determines how fast the baseline changes in order to detect a peak.

The sampling rate (or peak width) determines how often the detector output is sampled; this directly impacts on the accuracy of peak area measurement. If the sampling rate is too slow it may result in fast eluting peaks being missed or a valley between two peaks. At low sampling rates, peak apices can be missed and numerical integration will be inaccurate. Low data sampling rates alone do not lead to peak broadening (lower resolution), rather a combination of sampling rate and filtering electronics. When data sampling is too fast, acquisition rate can be managed by smoothing (bunching) in the CDS. Peak should be sampled at least 20-30 times across their width for accurate quantitation. Conventional HPLC typically requires a sampling rate of 0.5-1 Hz, faster capillary GC applications require sampling rates in the range 5-20 Hz, and UHPLC needs a sampling rate of 20-50 Hz (with most CDS being capable of up to 100 Hz).

There is also the issue of how the detector acquires the data and reports to the data system (this can differ between manufacturers). For example, in one system, if the detector is set to acquire data at 40 Hz lots of data points will be detected across the peak and the data will exhibit fine structure (Figure 4). Reducing the sampling rate to 1.5 Hz results in one out of every 15 points, relative to the first case, being detected resulting in a loss of some of the fine structure, for example, the peak apex. In this case, however, the baseline is less noisy. An alternative system may exhibit peak broadening when the detector sampling rate is reduced - this can probably be attributed to how the data is handled in the analogue or digital domain inside the instrument module or software.

Some good rules of thumb to follow when integrating chromatographic peaks are:

🔹Determine the best integration technique by analyzing known samples and collect data for both peak height and area. Calculate the results using both techniques and use the method that gives the most accurate and precise results

🔹Always tailor the integration parameters to each individual method

🔹Always use the same integration method for a particular analysis

🔹Integrating well resolved peaks is always the easiest and produces the most accurate and reproducible data, therefore, develop robust methods that exhibit the best possible resolution for the analytes

🔹Understand how to use the CDS and ensure that each chromatogram is examined to check the integration has been carried out correctly

Each CDS will have its own unique set of parameters which can be used to integrate chromatographic peaks and all users should become familiar with these. Table 1 details some of the common settings which can be found.

Incorrect peak integration ultimately results in poor quantitative results (Figure 5). However, when faced with peaks that are always going to be difficult to integrate (i.e. those which are poorly resolved or of very different sizes/ratios) some decisions and sacrifices may have to made based on the correct knowledge of peak integration. The following installments of this series will deal with strategies for integrating problematic peaks by using different integration methods.