Chromatogram Integration is fundamental.
Only intelligent integration software produces data worth to be evaluated

The following list of conditions for optimal chromatogram integration is the result of many years BETA testing of commercial integration software done by the author and his coworkers.
The integration software should:

1. Analyze the characteristics of the electronic signal coming from the detector and signal amplifier.
2. Check for earlier information about the to be expected peak width, peak height, run time per chromatogram, mode of start signal (manual / automatic), mode of mobile phase action (isothermal, isobaric, programmed, soft steps, steep steps), mode of analog signal display, sample naming, raw data storage address, mode of reduced raw data treatment, or ask the chromatographer for all these data if known.

By the way: if the chromatographer knows nothing and the integration software has no own information about how to do the job a clever auto-optimizing pre run is helpful to finally have acceptable conditions from on the second run. This situation is reality in case of non routine research chromatography.

3. Now adjust for the optimal rate when taking raw data.
   Give a READY signal or start the run by auto-injection.

The following figure and formula is basis of chromatogram integration and shows the details which are prone to systematic errors.

The quantitation formula

There are many highly complex formulas to describe the quantitation basics of GC and HPLC detectors, but things can be made very easy and still top correct.
In the figure 1 above the peak area is the integral of the peak height h over the time Sp to Ep as long as h differs from the base line B position.
Let us first assume, the detector is a mass measuring one, thus any h x dt value represents a certain substance mass given in gram.

The chromatographer wants detectors, which are very sensitive, that is, which produce a largest possible signal for a given mass of substance.

In the following t means only time. The substance amount in gram which represents the total peak area is named “E”. During the chromatographic run the flow of substance passing the detector changes from the peak start Sp in figure 1 over a maximum down to the peak end at Ep. In the same way the detector signal measured in ampere named “I” changes from no signal above the baseline B through a maximum down to no further signal above the baseline B again. Thus we can describe the quantitative process by formula [1] as

                 I [ampere] = q  x E / t [gram x seconds]                                                         [1]

where q is a factor describing the sensing quality of the detector. The units are given below:

                             q =  (I x t) / E   with the unit [ampere x seconds per gram]

which are “coulombs per gram”. This tells us, a very sensitive detector produces a large amount of coulombs per gram substance.
However it also tells, that probably not every chemical compound will produce the same amount of coulombs per gram, that is “q” will be substance specific and q will not be a universal constant.
Who looks critically to formula 1 will realize, that E = (I x t) / q which means, if we express the peak integral in amperes times seconds we only need to know the substance specific and detector depending value q and we know how much grams is an integrated specific peak. Who also thinks immediately in an analytical way will understand, that it might be quite helpful to check the correctness of a trace analysis this way:

We take one gram fruit sample, extract specifically and with a well known yield one toxic substance “S” in a solution of let say 10 ml, inject from this volume an accurately taken volume of let say 10 microliter, make the chromatogram, find through formula [1] 2 nanogram toxic substance “S” and can calculate:
From 1 gram fruit we got 10 ml extract and injected 10 microliter. This means we in fact injected the “S” extract from 1/1000 gram fruit which is 10^-3 gram absolute.
This corresponds to the found 2 x 10^-9 gram “S”.
Thus the fruit contains 2 x 10^-6 grams “S” per gram or 2 x 10^-4 weight %.

The trace analyst will realize, that we made an absolute quantitation without any trace level concentration We avoided the unknown specific adsorption loss which may be disastrous in the trace concentration level. We analyzed an extract which (hopefully) has the substance “S” high enough enriched and we repeated the analysis to check for any trend. We however MUST know how correct worked our enrichment - which means the yield of extracted substance. This can be checked by many well known procedures.

Who knows his quantitative detectors quantitatively (through the knowledge of q for substance “S”) can check quantitative results not only the regulated company specific way but by “absolute quantitation”. This is a most powerful error control technique. This is described in many details in IfC “teachware”, see under Teachware in this site.
The majority of detectors are not mass flow sensitive but concentration sensitive. It is easy to find out which type a special detector is: mass flow or concentration measuring. One just stops the mobile phase flow when a peak elutes. As the flow decreases now, flow measuring systems loose their signal. Concentration sensitive detectors keep it. Amperes means electricity flows. Volts means it stays. Thus concentration sensitive detectors work according to formula [2]:

                         U [volt] = q x E / v  [gram x ml]                                                       [2]

                 q =  U/E/v with the unit [volt per gram / ml]

q is substance specific and a non constant factor. In fact the detector “quality” value q depends also from the detector non linearity and from the working range level. For error free quantitation we need a non linearity correction factor if we use “absolute quantitation” concepts. This will be discussed shortly here but in all details in the IfC TEACHWARE. The knowledge of q-data is a very helpful way to detect and reduce quantitative systematic errors. q-data are available through calibration procedures. They also help to select detectors depending on the analytical task to do. Trace analyses should be done with detectors working with large q-values, but if the working range is not wide, careful non linear calibration is needed.