Qualitative Errors in HPLC
in this part error detection modes and test data are discussed, which also work for GC.

There are several qualitative error sources in HPLC:

1. Poor separation power - which means a small Trennzahl TZ.
2. Some special substances from the samples can be adsorbed strongly in and on the stationary phase. This changes the overall chromatography behavior of the HPLC column stepwise from run to run. The elution of such strongly adsorbed material pushes the chromatographer to use from time to time mobile phases of strongest elution power but in many cases the adsorbed materials resist any elution. It costs time and material until the chromatography characteristics return to a stable and constant start condition The best way to check for such impurities is the use of a PLC plate with equal stationary phase - the chemisorbed or otherwise strongly adsorbed sample parts remain on the sample start line on the PLC plate. 
3. Selectivity changes of the stationary phase can also result from highly sorptive impurities in the mobile phase. No mobile phase is chemically clean down to the ppb level and many mobile phases are oxygen / light sensitive which let grow the concentration of unwanted mobile phase impurities.
4. The column temperature has an often much stronger influence in liquid chromatography than believed changing selectivity and separation power. This influences the separation of critical substance pairs and is an error source.
5. A too strong believe in the identification power of spectrographical HPLC detectors lowers the check and control activity, which is a further error source.
6. The total numbers of chemical substances qualified for a chromatographic analysis is very much larger in HPLC than in gas chromatography. But many HPLC columns can just only separate 50 substances in one run. The latest developments of comprehensive GC however reached a separation capacity of several ten thousand coexisting substances per run. This was found not only in coal tar or raw mineral oil but in important environmental samples. This should make HPLC-only analysts very positive for error detection concepts by using cut techniques, multi column systems, multi detector systems, direct coupling with GC and PLC or using HPLC elution cuts injected into GC instruments or sprayed onto PLC plates. This all will enlarge the safety of complete separations because it reduces drastically the total range of coexisting substances. Of course it costs time and money, but the costs of wrong analytical data are incomparably higher for sure.

The only recently found sample complexity when using comprehensive chromatography in samples like bio active natural extracts, maritime organic trace substances and many others makes clear, that a HPLC analysis with a maximum separation capacity of only 50 differing substances carries a high risk level for systematic qualitative errors even when used for years for the same type of samples just because regulation does not allow to check for errors.. Let me give you an example for industrial routine analysis done for years with one and the same tool and mode: in a technically important raw material in which the presence of impurities had an utmost importance the routine lab found regularly seven traces. All known impurities were known by name. The author introduced quantitative capillary GC because he disliked the visible quality control chromatograms. As a result the seven changed into seventeen. This meant: nearla all names were wrong, the quantitative data were completely false and the years old decision concept based on analytical data had to be corrected full scale. Please understand that at first the new method and the people behind had some trouble...

The following is only a minimum of reasons which should animate using error control techniques in qualitative HPLC despite the trouble just mentioned in the last sentences:

    a) Check the separation power of the chromatographic equipment - not only the column but including the injection system and all connection tubes. IfC proposes to check for the Trennzahl “TZ” and the “b0” value in order to see if TZ is at its maximum in the range of application, “b0” is at its minimum possible and the mobile phase flow speed is optimal. Optimal is not the one existing at the  questionable value for “HETP minimum”, see last figure of this part of the site.

    b) The use of two columns either parallel or in series under conditions of a physical and chemical selectivity optimization is better than the believe that MS and MS/MS alone will do it. The team of the specific detectors for substance purity checks together with an optimized selectivity at a large enough Trennzahl is very much stronger.

    c) Mobile phase optimization is not too easy especially when it becomes clear, that statements like “a third added specific trace to the mobile phase which already got two others has no effect” are simply caused by missing experiments. We found those effects in circular multiphase programmed planar chromatography, where the mobile phases were mixed in a wide range of concentration automatically with a fourth special add-on phase during the run. Only this fourth one “MADE it”.

    d) Just this last remark may bring the idea to practice: planar chromatography in connection with HPLC is an unbeatable team for much safer quality information in liquid chromatography.

    e) Check these statements: changed stationary phases bring faster success when the change to further mobile phases has reached a certain economical border. PLC for instance works with many more mobile phases than HPLC. Some mobile phases with most helpful chromatography characteristics can never be used by HPLC but are in easy reach for PLC.

    f) Economical testing is done by the direct coupling of chemically differing stationary phases. It is much cheaper to test the HPLC separation by direct coupling with a PLC plate for a possible peak overlapping. Perfect overlapping cannot be detected by HPLC detectors like an MS system especially in case of very similar structures. If for instance optical activity exists a trace of a d-structured substance under the main peak with l-structure is much safer, easier and much more economically found this way instead of using an optically active HPLC column just for testing purposes.

    g) Testing - testing - testing! The HPLC sample as such is sprayed in one corner of a PLC plate for a following two dimensional PLC run. After drying and photo quantitation the same plate can again be given to a next two dimensional run with other mobile phases. If under all conditions one spot remains at the spray on position, these are substances which will stay in the HPLC column. It has a too large k-value (see under k-value) for any elution. These substances will deteriorate the stationary phase in the HPLC column. If under all conditions substances run with the mobile phase front in PLC, these are substances with a k-value of zero. At k = zero there is no separation power independent of the Trennzahl or the plate number of the HPLC column. There exists peak overlapping for sure. Here even substance mixtures may overlap to a single sharp peak.

    h) HPLC columns cannot have enough separation power as the probability of a very complex substance mixture is larger in HPLC than in GC. Thus we advise to check for the real separation power. IfC has left already since many years the classical modes of measuring the separation power in HPLC, GC and PLC. In all three modes the concept of the “theoretical plate number n, theoretical plate height HETP”was good only for the column development, not for analytical optimization. The many follow up numbers based on the classical concepts do NOT help reaching the best possible separation power for a given column / capillary. Often enough we need a “big number”. Separation capacity, information productivity, correct enough results found in shortest time counts. Therefore in the following we discuss the ABT concept as it is easy to use, very helpful in practical application, and elegant for a better understanding how to use chromatography optimal.

ABT - concept, equally valid for HPLC and for GC.

The ABT concept uses retention time and peak width values of homologues separated isocratically. At least four homologues are used. The time and peak width data must be taken very precise and the isocratic conditions must remain constant during the complete run time. As no peak overloading is acceptable - which falsifies peak width and retention time data - the homologue sample is diluted in a solvent eluting later than the homologues. From this run the retention time and peak width data allow to calculate most accurate and helpful control values. They describe the quality of a column/capillary and the overall qualification of the complete analytical system. We name such a carefully taken homologue chromatogram an ABT chromatogram. We need at least five, better up to nine ABT chromatograms taken at differing mobile phase speed values. The series of runs start at a quite high separation speed, which then is lowered stepwise to a quite low speed. This produces very fast and finally quite slow chromatograms. Analyzing all control data over the speed axis we find the best possible separation capacity at the most economical chromatogram speed.
Without the “ABT-software” such tests are work intensive, and are useful only if they support the final analytical task with the real samples in question. That means in HPLC: the best possible stationary and mobile phases at a well selected separation temperature are taken, the ABT chromatograms are measured under these conditions. In GC the best possible stationary phase and a good temperature compromise are selected. This of course determines also the type of homologues qualified for the ABT chromatograms. In HPLC this may be a selection of four consecutive members from the series of methyl- to C18 esters of di nitro benzoic acids (easy to synthesize, or to buy). In GC this means to select four consecutive n-alkane homologues within the range from n-pentane up to C20H42. ABT data are in easy reach using proper software, but they can also be found by pencil, liner and paper. However accuracy and precision is necessary and then the paper page must be in the size of one meter side by side. If the retention time and peak width data are transferred automatically by the integration program into EXPORT data files, here only the used homologue values must be written manually into the EXPORT data file which is a question of seconds only and can be done by WORD software or an editor program. In HPLC the homologue numbers have independent of the taken ester homologues the index or homologue numbers 100, 200, 300 and 400. In GC these are the injected hydrocarbons with their carbon numbers times 100, that are for instance the homologue numbers 800, 900, 1000 and 1100 for C8H18 to C11H24.

ABT calculates the following basic values if the four peak width and retention time values of the test homologues are correctly and precisely measured in seconds:

1. The dead time “tm” [sec], which is the residence time in the mobile phase along the whole path of the substances from the injection point to the detection sensor. “tm” has nothing to do with chromatography, but all with the geometry of the whole instrument. tm is a perfect measure of the mobile phase speed and this is the fundamental value for the physical separation optimization. Chemical optimization has been done already by the optimal selection of the mobile and stationary phases and the temperature. Knowing the tm value all non adjusted retention time values “tms” can be changed into the adjusted net retention time values “ts”. “ts” is the individual residence time in the stationary phase. This is valid in HPLC and GC. Knowing “ts” values we can calculate in isocratic HPLC the important value “k”.
“k = ts / tm” . “k” is correlated with the chromatographic partition coefficient K. Using k as chromatography unit we can build a chromatography scale in k - units. The chromatographic process - but not yet the separation - starts at k = -1. Any substance or mixture which is insoluble or not sorbing on and in in the stationary phase leaves the instrument at k = 0. Here is a “blind time area”. Nothing at k = 0 can be separated. We may have overlapping peaks with the peak width in half height of “b0” [sec]. That means: The peak width value “b0” has nothing to do with chromatography but with the sampling process, all mechanical connection parts and their dead volumes, with the mobile phase volume, with the volume of connection tubes and the phase flow speed. b0 is an unwanted non chromatography value acting completely against separation. We must do everything practically to reduce b0 as much as possible.
The “tm” value stands for T in ABT. See figure 1 below.
The chromatographic time axis is unlimited long in elution techniques. But each chromatogram will be halted after the chromatographer believes, all substances are eluted and detected. This is NOT the case for all substances which irreversibly adsorb or chemisorb in or on the stationary phase or any solid wall within the instrument. These substances have the virtual k-value k = unlimited. They falsify for sure the quantitative results in HPLC and GC. But they are visible and even quantitized by PLC. Thus the use of PLC is a very important technique of error detection in HPLC and GC.

If one does not know about b0 and its size, one cannot do something against. Therefore the ABT calculation is necessary.

2. The ABT concept finds out the value of “b0” in seconds. This peak width part adds to all correct peak width values of all substances in isocratic HPLC and GC and even in programmed chromatograms. “b0” deteriorates the separation efficiency. “b0” stands for the B in ABT. b0 is found by a simple mathematical procedure in extrapolating a linear increase of peak width data over tm-time units.

3. A separation is only possible if substances differ in their residence time “ts” in the stationary phase.  But this is no residing in the sense of “sitting” unmoved. Substances which sorb in the stationary phase and leave it again with a differing speed of movement through the complete length of the stationary phase “rod” do this in very small time units. The “ts” value is an integral of such time units consisting of diffusion from the moving mobile phase into a non moving thin layer of locally staying mobile phase part. This covers the stationary phase. Only now specific separation starts due to the physico chemical sorption depending on structure elements. The way back is controlled by probability and costs again time Only in the second the always moving substance molecule reaches flowing mobile phase again, it can be moved forward towards the outlet of the column / capillary. The forward - backward movement through flowing and staying mobile phase parts as well as onto / into and out of the stationary phase with even porous structure costs time. The often porous structure of the stationary phase is also filled with non or less fast moving mobile phase. Under isocratic conditions this peak width increase is unbelievably constant over the time. The ABT concept uses this fact which for long time was questioned by some theoreticians. Their main argument was, that small molecules like methane or helium produce broader peaks due to their higher mobility. However the ABT procedure avoids using methane, helium or other smaller molecules like butane. We start with heptane or C12 n-alkane or the methyl ester of 3-5 di nitro benzoic acid. IfC has evaluated thousands of isocratic “ABT” runs in GC and HPLC. The retention times and peak width data have been measured very accurate and highly precise. The peak width growth with time is an important measure of the separation quality of a column/capillary. It is linear with utmost determination factors - see the figures 1 and 3 below. As only the chromatography related peak width growth has to do with the quality of the column/capillary, ABT calculates the function “peak width growth based on k”. Because of the strictly linear growth ABT calculates the function peak width = a * k + b0 and takes the constant value “a” [seconds] as column/capillary quality number. “a” stands for A in ABT. b0 stands for B in ABT and quantitizes the non chromatographic peak width enlargement. b0 is a most negative value which must be minimized by flow speed optimization, instrumental improvements, optimized sampling.

All three ABT data: “a”, “b0”, “tm” in seconds change drastically with the mobile phase flow speed “U”. Practically this flow speed is given in seconds per mm and can easily be calculated if one knows the length ”L”  [in mm] of a column / capillary. As the mobile phase needs “tm” seconds to reach the detector after a sample injection, it follows
                                                                as flow speed U = tm / L [sec/mm]

The classical theory concepts of chromatography uses other “quality numbers” than given by ABT, but all of them are based on the fundamental data “a”, “b0” and “tm” . Thus we only need the transfer formulas or can let calculate all of them automatically by a piece of software which takes four tms and four b0.5 values from four isocratically separated homologues.

In addition the ABT values allow to interpolate and within some limits to extrapolate any other time and peak width data in isocratic chromatography, because we have such a perfect (and from some colleagues experimentally never found) linearity of peak width growth over time. In addition we have a strict linearity of the log(ts) growth for the homologues. Within limits we can even extrapolate the analysis time of any substance from which we know the retention index. We can pre calculate the end of a chromatogram if we know about or ask for a “last compound”. This helps to decide if we should take a chair to sit waiting or better do not inject the sample without changing conditions.
To avoid any trouble caused by not eluted substances with too long retention times EVERY GC or HPLC instrument MUST have “backflush” installations.

ABT data  “a”, “b0”, “tm”:

See the localization of these three fundamental isocratic chromatography values in figure 1 and 3.
The k-scale is a relative time scale but in chromatography units. Remember: k = (tms - tm) / tm.


Figure 1: ABT data from a GC capillary. NOTE: this is valid for just only one mobile phase flow speed, which here is obviously too small. Depending on the capillary dimensions (inner diameter) and the carrier gas He, N2, H2 ... there is one best carrier gas speed for a maximum TZ, but this may not be the optimum for the routine use. There one may want a much higher separation speed as given at a Trennzahl maximum, see the tables below and figures.

Practical application of  “a”, “b0”, “tm”:

Any not adjusted retention time value tms can be transferred into k-values:
    k = (tms - tm) / tm  [no unit], but the value for tm must be calculated, it cannot be measured easily.
2. The peak width b05 of any peak with known tms value can be pre calculated:
    b05 (at tms) = b0 + a * (tms - tm) / tm  =  b0 + a * k.
3. The Trennzahl  for any time region can be pre calculated, see figure 2:
    TZ = [tms homol.(N) - tms homol.(N-1)] / [b05 homol.(N) + b05 homol.(N-1)] + 1
4. The peak capacity “PC”, which is the largest possible number of base line separated peaks within a selected analysis time can be pre calculated. Practically useful are “PC15” values which is the maximum number of baseline separated peaks within 15 minutes analysis time. The upper limit of an acceptable mass analysis time are 45 minutes and the corresponding peak capacity is “PC45”. In modern micro capillary chromatography the one minute PC = “PC1” is a useful value. Why can a “PC” value help ? The question of a non analyst to his chromatographer may be “give me data for all olefines  of this heptane cut. I need them to day”. If in this heptane cut (for certain reasons) only 150 olefines exiat, then “PC30” must be > 150. If however the standard “PC45” value for the available analytical instrumentation is just only 50, don’t inject samples. You may get very many overlappings and the non analyst will not get any correct results in the time he selected.

The classical theory works with the theoretical plate height “h” or “HETP” respective the theoretical plate number n to describe the separation capacity of a column packing or the separation power or packing quality of the column as such.

n = 5.545 * (tms / b05)2 = 5.545 * ((ts + tm) / (b0 + a*(ts / tm))2
HETP = L / n [mm]
L = column length [mm]

The formula for n and HETP above has the following problem: The factor “5.545” (often given as 5.54) is correct only for symmetrical GAUSS shaped peaks. Those do not exist in practice. The variance of practical HPLC peaks for instance differs easily by about 40 % from a real GAUSS shaped peak. We had after many weeks critical discussion with Prof. M. Golay based on hundreds of precisely measured analytical facts this agreement: “The peaks are truly Gaussian shaped, if we inject nothing into a column of unlimited length”. The next problem with the theoretical plate number n and HETP is: it depends very strongly on the non chromatographic value b0 and the k-value of the selected single test substance. With other words: n and HETP are substance dependent. Absolutely wrong values are found early in the chromatogram because of the then relatively strong influence of b0. Only n and HETP data at the end of a chromatogram have some usability. The third serious problem is: separation exists only if there is more than one substance in a sample. So n or HETP tell nothing about the practical separation quality a column or capillary in GC or HPLC has.
More realistic but still weak because of the single substance with non Gaussian peak shape is the “real plate number” Nreal = 5.545 * (tm / b0)2 .
These mentioned chromatography quality numbers follow now as function of the mobile phase flow
speed, see figure 2 for GC and figure 4 for HPLC:


Figure 2: TZ data from a GC capillary. NOTE: this is now valid for five mobile phase flow speed values. Depending on the capillary dimensions (inner diameter) and the carrier gas He, N2, H2 ... there is one best carrier gas speed for a maximum TZ, but this may not be the optimum for the routine use. There one may want a much higher separation speed as given at a Trennzahl maximum, see the tables below. The flow speed levels are seen in tm values.

HPLC ABT- and TZ-values:

It looks like simple repetition but the ABT concept as used for GC shown in figure 1 is now applied for HPLC. Isocratic runs with a 250 mm long HPLC column of 3 mm inner diameter, stationary phase Eurosphere 100 / C18, mobile phase methanol / water 85 /15 v/v, room temperature;
inlet pressure ranges at 8 levels from 255 bar (fastest flow speed) to 39 bar (longest analysis time).


Figure 3: ABT data from the HPLC column mentioned above under conditions near the TZ maximum. The TZ values as function of the mobile phase flow speed (between 0.3 and 2.0 mm/sec) are seen in figure 4 below, see all detailed values in the tables.  Compare figure 3, valid for HPLC, with figure 1, valid for capillary GC. Compare especially the b0 values in GC and HPLC, but see the last figure of this page about a wide ranging mobile phase flow speed to understand what is a flow speed optimum. Optimal for what ?


Figure 4: TZ data from the HPLC column mentioned above.
NOTE: this is now valid for eight mobile phase flow speed values. Depending on the column packing (particle size and range), the mobile phase viscosity and the column temperature,  there is one best mobile phase flow speed for a maximum TZ, but this may not be the optimum for the routine use. There one may want a much higher separation speed as given at a Trennzahl maximum, see the tables below. The flow speed levels are seen in tm values and in the table data, especially also in the last figure of this page..

Complete separation quality data and optimization values for a selected pair of mobile and stationary phases.

Only a good enough separation is basis of smallest possible quantitative errors and of error free qualitative data.
It is wasting of time & material if long base line regions have no any peak. It is a critical error situation, if the finally coming peaks (under isocratic conditions) have peak width values larger than
b05 = a * (tms-tm) / tm + b0 = a * k + b0

Optimization therefore pays back. The ABT data a, b0, tm can HELP.

But NOTE: Any classical value to describe physically the separation system like the theoretical plate number “n”, or the height of a theoretical plate “HETP”, depend on the mobile phase flow speed value.
In addition they can be taken ONLY under isocratic conditions and what is the most critical limitation, they depend on the “k” value.
Remember:  k = tms (which is measured) minus tm divided by tm. Again: tm can be calculated only exactly enough especially in HPLC, while in GC one can make detectors especially He sensitive and can use the Helium retention time tms as tm, because there is nearly NO residence time for He in most of the stationary GC phases.

IfC used the ISOPT software, which calculates a huge set of data (some of them shown in the following tables) and which analyzes quantitatively all column quality values over the mobile phase flow speed. The last figure shows how all ABT data as well as “n”, “HETP” and all other classical column/packing quality values depend on the mobile phase flow speed.
Optimal separation conditions depend in addition on the physical length of the column/capillary, the particle size and size distribution, the inner porosity and porosity distribution, the inner diameter of a HPLC column or a GC capillary. There is a strong dependence on the packing uniformity and packing density. With respect to the chemical part of the packing the uniformity of a phase distribution on and inside the particles is a next most strongly acting factor. In GC packed columns the situation is more or less equal as with HPLC column packings but in capillary GC the film thickness and thickness uniformity are strong factors. The ABT concept gave us enough precise and correct “b0” values. Their correlation with the inlet and outlet physics, the detector volume size and geometry and all technical details inside the sampling systems was a great help for optimization.
It must be mentioned, that the chemical and thermal part for separation system analysis and optimization is of strong influence.
Here we learned about the importance of the working range each separation system has: it is not only the sample load capacity but in GC the column/capillary temperature, the level of the temperature programming speed and the upper and lower temperature limits in connection with the stationary phase film thickness. Quite wide is the range within which GC works well with the mobile phase flow speed, but in HPLC columns there is soon a limit given because of the mechanical forces which destroy porous packing particles. The flow resistance heat produces not only strong polarity changes inside the packing but temperature gradients from hot (packing center) to cooler (at wall side) in quite thick layers of the packing near the column wall.

Thus we have a huge number of factors to be optimized as well as wide ranges for each factor to check for an optimum in practice under the conditions of sample numbers per time, economy, and needed data precision.

THAT MEANS: to classify a column or capillary in HPLC and GC as “O.K.” is the one side of the story and to concentrate this many factors to a few quality numbers like the “HETP” or the “plate number “n” is an other story.
It is also important to avoid such classical quality numbers if they are calculated using non adjusted retention time values. Only net-retention time values are acceptable, thus we reject ntheoretical numbers but use neffective numbers are based on ts values.

This all is valid ONLY under isocratic conditions. Flow, pressure, temperature, and mobile phase composition MUST remain strictly constant.

But from a good optimum (found by a lot of compromises) we can translate to non isocratic conditions, which are used by a majority of chromatographers.

The “Trennzahl” over the k-range needed for a given analytical task is then - under programmed conditions - the only separation system quality value we know.

Some values and their precision as example. NOTE: isocratic conditions ONLY !

separation quality numbers

in GC (see *1)

in HPLC (see *2)

at the Trennzahl maximum



ABT value a (in seconds)

0.43 +- 0.001 sec

7.43 +- 0.02 sec

ABT value b0 (in seconds)

0.43 +- 0.05 sec

16.74 +- 0.04 sec

ABT value tm (in seconds) (see *3)

38.27 sec; 0.9999

442.04 sec; 0.9999

TZ maximum at speed:

24.2 at 390 mm/sec

12.3 at 0.5 mm/sec

n real at k=10



h real at k=10

347 micro meter

12.7 micro meter

n theoretical in the meas.t-range

42400 - 43150

6700 - 13250

n effective in the meas. k-range


1260 - 9040

PC real (from k=0 to 10) (see *4)

105 peaks baseline

49 peaks baseline

PC15 (0-15 min separation time)

139 peaks

10 peaks

PC45 (0-45 min separation  time)

188 peaks

34 peaks

measured within the k-range

from 0.8 till 4.7

from 5.6 to 32.5

* 1: GC capillary 15 m long, 0.28 mm i.d., hydrogen 0.45 bar = 390 mm/sec phase speed;
       0.25 micro meter film thickness, 200 degr.C isothermal and isobar.
* 2: HPLC: 250 mm long, 4 mm. i.d. packed with Eurospher 100 C18; 75 Vol % methanol in water 80 bar         = 0.3 ml/min flow at 0.5 mm/sec speed; 20 degr.C isothermal and isobar.
* 3: tm is calculated by computer iteration until the correlation log tms over retention index is linear.
       34   steps needed. The retention index values in GC: 1100, 1200, 1300, 1400, the retention index           in HPLC from 100 to 500 (= methyl- to pentyl-dinitrobenzoic acid ester) - see EXPORT data block.
* 4: PC real is the peak capacity in the chromatography range from k=0 to k=10, that is from 0 x tm
        up to 10 x tm. PC means the number of baseline separated peaks.
PC15 = number of base line separated peaks in the chromatography range from k=0 to k=10,
        separated within 15 minutes working time starting at injection.
PC45 = as PC15 but working time = 45 min as the upper acceptable time limit in mass routine analysis
        under worst conditions.

NOTE: to days micro chromatography systems which separate at the upper limit of optimal speed need some seconds to 3 minutes separation time. Multidimensional systems separate in a few minutes
10,000 to 30,000 peaks but no classical chromatogram registration is possible and quantitation is still a problem. (2005).

All the HPLC data could be calculated by IfC ISOOPT software using the EXPORT data set given below.
This EXPORT data set or block is automatically printed as last job of the integration program. We used Beringer software. This is the only integration software package which calculates peak width data in high precision and accuracy. As for all the separation quality data we need only tms, b05 and INDEX values, the peak height and peak area values are not necessary. They are set to zero.

       retention time      peak width             peak area                 peak height             index
       tms (seconds)      b05 (seconds)        not stored = zero     not stored = zero    



























Figure 5:
 Main separation quality numbers for GC and HPLC - in this figure capillary GC data.

blue: separation speed given in peaks per second. The maximum is at about 680 mm/sec mob.phase speed.
violett: Trennzahl TZ over speed. Maximum (TZ=30) is at 280 mm/sec.
dark violett: classical theoretical plate number, maximum is at 250 mm/sec.
brown: peak capacity in numbers of peaks separated from k=0 to k=10. Maximum is at 390 mm/sec.
green/black/red: a, bo and tm decrease all with growing speed.

Why are the quality numbers so poorly qualified as seen in the graphics ? The measurements have been done in a public experimental run with about 24 guests in a hurry. ABT data need isocratic conditions. The change of the mobile phase flow speed needs several minutes until stability is reached Thus there is not yet an isocratic condition. We have not done any stability test. Even slight flow speed changes result in errors for the ABT values. As all other separation quality numbers depend on ABT, they also show instability. But the  ABT data errors are small enough so we see maxima as shown in figure 5. Optimal speed for mass analysis is at 400 mm/sec reaching aaximum separation capacity. For a publication showing highest separation speed we would use 680 mm/sec; to show how wonderful is our GC separation capillary one would take a speed between 280 and 390 mm/sec.
At this speed b0 is at 0.4 seconds, which is quite small. Therefor we realize that the capillary installation, the used connection tubes, the sampling speed etc. are perfect for a 0.3 mm i.d. capillary, 25 m length, hydrogen as carrier gas, 150 degr.C, isothermal, n-alkanes C10 to C14.


[Home] [We can help] [Systematic C-Errors] [Statistics] [Error Detector "sf4"] [Sampling/Calibration] [Qual.Error GC] [Quant.Error GC] [Qual.Error HPLC] [Quant.Error HPLC] [Qual.Error PLC] [Quant.Error PLC] [Integration] [Chrom. Combination] [µPLC Micro Planar LC] [Altern.Chrom.Theory] [Contact IfC] [About the Author]
[Home] [We can help] [Systematic C-Errors] [Statistics] [Error Detector "sf4"] [Sampling/Calibration] [Qual.Error GC] [Quant.Error GC] [Qual.Error HPLC] [Quant.Error HPLC] [Qual.Error PLC] [Quant.Error PLC] [Integration] [Chrom. Combination] [µPLC Micro Planar LC] [Altern.Chrom.Theory] [Contact IfC] [About the Author]