No thanks! I would like to know more about CHROMacademy

 Over 3000 E-Learning topics / 5000 Articles & Applications
The CHROMacademy Essential Guide to Developing Better Methods for Reversed Phase HPLC (Part 2)
Analytical Goals to Consider jump to section »
Selecting the right HPLC Column jump to section »
Where to begin?
jump to section »
Use of pH jump to section »
Analytical Goals to Consider    

Often the main difficulty in developing a HPLC method is knowing:
What are the method aims and goals?  l  When have you arrived at your goal?

  1. How many samples will you have , do you need a high throughput method – this will help with choices between isocratic and gradient methods as well as the mobile phase composition and column geometry
  2. How many analytes of interest will be present – this helps to make choices on efficiency and resolution of your method
  3. What will be the sample matrix – will you need to separate the analytes from many matrix components and are special sample preparations steps (such as Solid Phase Extraction) required
  4. Is the method qualitative or quantitative?
  5. What is the concentration of the analyte (note you may need to deal with a range of analyte concentrations) – do you need high sensitivity separation and detection techniques
  6. Analyte / Sample matrix solubility & stability – you may need to exclude certain HPLC techniques when dealing with unstable or sparingly soluble compounds

Perhaps one of the most difficult aspects of method development is knowing when you have achieved a suitable set of conditions and chromatographic outcomes, and it is useful to consider what the completed might ‘look like’ prior to embarking on method development.  You may want to use some of the following criteria to define your success criteria:



Retention factor of all analytes of interest has a retention between 2 and 10 (2 < k < 10)rentention-factor
Retention factors of above 1 are sometimes acceptable however care must be taken to ensure that all early eluting peaks are separated from other analytes and from non-retained or poorly retained matrix components. 

In general, when retention factor (k) exceeds a value of 10, increasing retention as a means of improving resolution is not effective, as decreases in efficiency due to longitudinal molecular diffusion negate improvement in peak spacing.

Retention factors up to 20 are sometimes necessary, especially for very complex samples, however longer retention sometimes leads to broader peaks and potentially reduced resolution.

Examples of low and higher values for retention factor and their effect on chromtaographic separation.  

Tailing factor of all peaks should be less than 1.5 (Tf ≤ 1.5)

This will be essential for good peak integration for quantitative measurement and to ensure satisfactory resolution and essentially describes the symmetry of the

chromatographic peak.  Some laboratories and regulatory agencies stipulate that a tailing factor of less than 2 is sufficient.

Efficiency should be close to the optimum based on column and system geometry
Efficiency (or plate count, N) is used as a measure of system and column performance and gives an indication of how 'narrow' the peaks in the chromatogram will be, which is important in obtaining good peak resolution.  Different column and packing geometries and system extra column volume will give different achievable efficiency however some reference values (based on real world rather than optimized test compounds), however for reference a 150 x 4.6mm x 5mm or 100 x 4.6mm x 3mm HPLC column should generate a plate count (N) of 10,000  - 12,000.  Note that the traditional equations used to measure column efficiency are only meaningful for isocratic separation conditions and a different approach needs to be taken for gradient HPLC.

One useful system test is to measure the efficiency values of peaks within an isocratic separation, which should give increasingly greater plate counts from the first to last peak.  If the plate count does not increase in this fashion, the HPLC system, rather than the analytical column, is having the greatest influence on peak dispersion and system extra column volume should be reduced.


It is useful to have an 'efficiency target' based upon the column and support particle geomtery used for your analysis.  The table shown here gives some useful 'real world' target efficiency values for various column geometries.

Resolution between analyte peaks of interest or analyte and interferent peaks should be greater than 2 for robust methods (Rs> 2)
Resolution is determined (as is shown in equation X), by the retention factor, efficiency and selectivity of the separation and a resolution of 2 or greater between each critical peak pair will ensure that quantitative measurement is possible and baseline resolution is achieved for all peaks.  Of course – this may be unrealistic with ‘real world’ samples and one should decide, based on quantitative or qualitative requirements, on the actual ‘required’ resolution for the method.  Method development goals in one famous HPLC laboratory are defined as ‘Obtain a separation which results in the required resolution between all peaks in the minimum analysis time on a consistent basis’ resolution

These criteria should be used to decide upon the suitability of the method you have developed and can be used to form the basis of a system suitability test or target validation criteria should these be required.  Remember that with experience, the end point for method development can very much become a matter of considering the ‘look and feel’ of the chromatogram as well as incorporating the quantitative measures outlined above.

Selecting the right HPLC Column   back to top »

Selecting the right column is perhaps the most fundamental choice to be made when developing HPLC methods.  Prior to making a specific column choice, we should consider the ‘mode’ of chromatography that will be used.  For example, isomeric compounds may require normal phase conditions, highly polar compounds may need aqueous normal phase, mixed mode or HILIC conditions, ions are best analysed using ion-pair or ion exchange chromatography, chrial compounds require a chiral column and so forth.

The method of choosing an appropriate HPLC for a particular separation varies widely from using a ‘favorite’ column from the column store, the recommended column from an application or similar method, to a rigorous screening of columns chosen for their orthogonality based on a study of analyte and/or column classification characteristics.

Perhaps the most challenging method development situation involves a sample containing analytes whose chemistry or physicochemical chemical parameters (Log P, Log D, Molecular Weight, pKa etc. etc.) are not well known.  In this case, the best approach involves performing a ‘screen’ typically using 4 to 6 columns known to produce orthogonal selectivity (separation factor or peak spacing) using a shallow and steep gradient at three different pH values – the data from which can be entered into one of the popular chromatography optimization programs. If no optimization software is available, one might perform a standard scouting gradient at two pH levels (see later).

Whichever approach is taken, there is no doubt that an improved knowledge of column chemistry is invaluable in choosing a suitable column, or understanding the results of various method development separations.

Stationary phase design and characterisation was the subject of a recent CHROMacademy Essential Guide
and as such, only a brief guide is given below. » view Resolver Isssue 2 here

Reversed phase HPLC columns are available with a wide range of bonded phases, primarily to vary the hydrophobicity of the stationary phase.  By varying the type and degree of interaction between the various analytes and the stationary phase surface, this offers a very powerful tool to alter the selectivity of a separation – and hence attain one of our crucial method development goals.

Typical stationary phase types include:

  1. Hydrophobic alkly species – C18, C8 etc.
  2. Shorter alky species with relatively polar end groups – cyano propyl
  3. Alkyl species with polar embedded groups near the modification layer – C12 with imide or carbamate at the C3-4 position
  4. Phenyl Bonded phases with Π electron systems – phenyl, phenyl hexyl
  5. Polar bonded phases – diol, cyano
  6. Bare silica
  7. Mixed mode phases – C8 plus a weak or strong ion exchanger

Whilst being nominally identical, C18 alkyl bonded columns from different manufacturers can produce separations with widely different selectivity,
and there are a number of other factors to consider besides the nature of the bonded phase including;

  1. Nature of the base silica (Type I or Type II silica), metal ion content and several other factors
  2. Nature and degree of end-capping
  3. Ligand density to surface area ratio (as a better measure of surface coverage than simple % Carbon loading)
  4. Silica or silica inorganic hybrid
  5. Particle size, pore size and latterly, fully or superficially porous
  6. Use of monolithic silica or carbon

In modern HPLC method development it is perhaps more useful to ‘characterise’ stationary phase materials in terms of a fixed set of criteria using probes which are designed to better describe the phase performance.  The three popular approaches to column characterization use slightly different test probes and this topic will be the subject of a future Resolver and Essential Guide Webcast.  A full overview of one of the more straightforward approaches to column characterisation and how this approach might be used to select columns according to analyte properties or based on experience with other phase types can be found at the following links:



Where to begin?   back to top »

Significant Method Development Variables
It is important to understand, for the mode of chromatography chosen, what variables are available to change the selectivity of the separation.

For reversed phase HPLC separations, the table below outlines some of the available variables alongside other important considerations:

Variable Impact on Selectivity (α) Easy to Change Robust /Easy to Control?

Stationary Phase
Organic Modifier
% Organic Modifier
Eluent pH
Eluent Ionic Strength
Gradient Slope
Use of an Ion Pair reagent

Large 1
Medium 2
Medium 2

Yes 3


1 – for ionisable analytes 2 – large when analyzing ionisable species
3 – when using quaternary pumping systems


Optimising Mobile Phase Composition in the Chromacademy Essential Guide » view webcast here


Analyte Assessment
Prior to commencing method development it is important to make some fundamental decisions based on analyte information, if this is available.  Listed below are just some of the important factors that need to be considered prior to any experimentation:

Analyte LogP – if this is known then the relative hydrophobicity of the analyte can be assessed. This gives an indication of the eluent strength required to elute the analyte (high positive values require higher organic for elution).

Analyte pKa – gives an indication of the pH at which the analyte changes it’s state of ionization which affects its relative ‘polarity’ and therefore it’s extent of retention under reversed phase HPLC conditions.  The analyte may have more than one ionisable functional group which may render the method development process more complex.  When dealing with ionisable analytes the first approach is usually to completely suppress or promote  ionization in order to make optimization of the eluent organic composition more straightforward.  A common modern approach is to employ 0.1% Trifluoroacetic acid as the aqueous component of the eluent system – which results in a pH of approximately 2.1.  At this pH, most organic acid functional groups will be fully ion suppressed and most basic functional groups will be fully ionised.  These two extremes can then be handled using a combination of column chemistry and mobile phase eluotropic strength.  Whilst this approach lacks finesse, it does simply initial method development work and will give important pointers as to whether, for example, pH optimization is necessary to obtain a satisfactory separation.

First Steps – The Gradient Scan
For most method development exercises involving Reversed Phase HPLC, the % organic modifier in the eluent will be one of the first parameters to optimize.  This can be easily automated using one of the simple method optimization programs available or can be collected and interpreted manually.  Depending upon the column type selected and the nature of the analytes the pH may be adjusted to a higher or lower value (2.1 and 10 are typical values) to control extent of analyte ionization and control this potentially confounding variable.  This is especially important where little or no information about the analyte exists.


Manual Gradient Scan
This technique uses a fairly standard gradient to ‘investigate’ or ‘scout’ the elution behavior (effectively the hydrophobicity) of the analytes within the sample.  We would typically use a 5 – 100% gradient over around 20 minutes and examine the results to make some ‘rule of thumb’ predictions regarding the optimum mobile phase composition.  Typical gradient conditions are shown on the chromatogram.  For neutral species its not strictly necessary to control pH however at lower mobile phase pH, peak shape for polar compounds will improve.

The use of 0.1% TFA as the aqueous eluent component results in an aqueous solution pH of around 2.1.  This will render most organic acids fully ion supressed and most bases fully ionised (see further information on the use of pH control below). This helps us to rationalise the results of the chromatogram obtained interms of analyte retention based on hydrophobicity and degree of ionisation.

Figure 1
So let's suppose we did the scouting exercise and got the results shown.  The first rule of thumb states that if the compounds within the sample elute (even if they are not fully separated) over a time which is less than one quarter of the gradient time (5 minutes in this case), then isocratic analysis can be considered.



The isocratic ‘best estimate’ isocratic conditions are obtained by calculating the %B at the average retention time of the components within the sample – you can see how we calculated that on the slide here. The composition at 8.46 mins (the middle of the retention range) is around 39%B. Remember that one should always take into account the system dwell volume as we have done above.

Predicted isocratic separation at 40%B

System dwell volume is the time taken from  the gradient composition being mixed in to the pump to the point at which this composition enters the column (and therefore affects analyte retention).  You can find a more in depth discussion on how this is measured within CHROMacademy, however for this system the value is 1.2 minutes.  We can simply subtract this value from the average elution time to adjust for the actual gradient composition in column at this theoretical ‘average’ retention time, as you can see above.

The composition we derive by this method may need to be further optimized in order to get the exact separation we need – however, it represents a great way to get into the ‘ball park’ without several repeated injections and adjustments of the mobile phase composition.

Gradient HPLC
If we find that the elution profile of our analytes spreads over more than 25% or the scouting gradient, then gradient analysis is recommended.  This may well happen if we have samples which contain analytes with widely differing hydrophobicity, or mixtures of acids and bases or zwitterionic compounds).

We will cover the theoretical and practical aspects of gradient HPLC in a future issue of resolver and an accompanying ‘Essential Guide’ webcast. 
To find out more in the meantime there are a host of resources within CHROMacademy - LC / HPLC Channel / Theory of HPLC /Gradient HPLC

Gradients in HPLC are defined by three values:  Initial %B, Final %B, Gradient time (or steepness).

Developing a  gradient profile using these three values is relatively straightforward.  The elution composition of the first peak within the scouting gradient can be used to estimate the initial %B.  In our example this would correspond to:

Retention time of first peak (Ti) – 5.98 mins
System dwell time (VD) – 1.2 mins
Retention time of first peak (Ti), adjusted for System Dwell – 4.78 mins.

Estimated initial %B -
95%B/20 = 4.75%B / min
(4.78 x 4.75) + 5 (the starting gradient composition) = 27.7% B

We can estimate the final gradient composition by calculating the elution strength at the time at which the final peak within the chromatogram elutes during the scouting gradient.

Retention time of final peak (Tf) – 10.94 mins
System dwell time (VD) – 1.2 mins
Retention time of first peak (Ti), adjusted for System Dwell – 9.74 mins.

Estimated final %B -
95%B/20 = 4.75%B / min
(9.74 x 4.75) + 5 (the starting gradient composition) = 51.3% B

We now need to decide upon the gradient steepness and the ending eluent composition.

Gradient steepness is controlled by the mobile phase starting and ending composition and the gradient time. The steepness of the mobile phase gradient can have a significant effect on the separation.

The equation for the gradient retention factor (k) shown here can be found in many textbooks. Gradient retention factor is difficult to visualise as it differs from its isocratic counterpart and resembles more the profile of the gradient elution. It is effectively defined as the retention factor for an analyte that has migrated half way down the HPLC column and is often defined using the notation k*.


Figure 2
Estimating the gradient slope from our initial scouting experiment is possible:

ΔΦ = 0.241
(51.3 – 27.7 %B )
S = 4
F = 2
tg = ?
Vm = 1.7ml
k* = 5

tg = k x (1.15 x S x DF x Vm) / F

tg = 5 x (1.15 x 4 x 0.241 x 1.7) / 2

tg = 4.72 mins.


It is best practice to include an isocratic portion at the beginning of the gradient to allow transfer of our method to other HPLC systems.
Therefore – our initial gradient for method development will be

Time %B

1.2 (VD)
5.92 (tg+VD)

27.7 (initial %B)
51.3 (final %B)

We may need to adjust the initial and final %B to obtain the required selectivity and resolution, however
this method can be usefully used to estimate a useful set of starting conditions.

The predicted and final (partially optimised) gradient chromatograms are shown above.  The final organic composition and the steepness (rate of change of organic modifier), have both been lowered to achieve the final result.  It is these two parameters which are often altered to affect a separation of the required quality. The final gradient profile is shown here.

Method Development with Ionisable Compounds
Acid, basic or zwitterionic compounds usually require some form of eluent pH control in order to control their retention under reversed phase conditions.  As discussed above, it is fashionable to attempt to eliminate pH as a variable by having eluent pH either high (~pH 8 – 9) or low (~pH 2-3).  For example, when using 0.1% TFA the eluent system pH will be approximately 2.1, rendering most acidic analytes in the neutral form and most basic analytes in the fully protonated form.  The retention of both species is then controlled using a combination of stationary phase choice, organic solvent and mobile phase eluotropic strength as described above.  This low pH approach has the added benefit that peak shapes are improved.

However there are situations in which this approach does not work.  Ionised compounds may elute too early (k<1) or the required selectivity may not be achieved using the available variables.  In this case, using eluent pH to control analyte retention can be a useful additional method development variable.


Analyte pKa
The pKa or ionisation constant is defined as the negative logarithm of the equilibrium coefficient of the neutral and charged forms of a compound. This allows the proportion of neutral and charged species at any pH to be calculated, as well as the basic or acidic properties of the compound to be defined.

As the eluent pH changes, the extent of analyste ionization changes and therefore, because it’s relative ‘polarity’ is changing, the analyte retention time alters.  This is outlined below

Thermodynamic Ionisation constants


HA = H+ + A-
Ka = {H+}{A-} / {HA} ({} = activity in mole L-1)
pKa = -log10(Ka)


BH+ = H+ + B
Ka = {H+}{B} / {BH+} ({} = activity in mole L-1)
pKa = -log10(Ka)

  • At around pH 9.5 the analyte is in it’s 50% ionised form – this is equal to the pKa value of the analyte molecule
  • At around pH 6.5 the analyte is fully ionised (i.e. is fully protonated) – lowering the pH will not significantly affect the degree of ionisation
    (and hence the retention behaviour)
  • At around pH 11.5 the analyte is fully non-ionised – again raising the pH will not significantly affect the degree of ionisation (and hence the retention behaviour)
  • It should be noted that the extent of ionisation behaviour is exactly the opposite to that encountered with acidic species.
The pH corresponding to the point at which the two forms of the analyte (ionised and non-ionised) are present in equal concentrations (i.e. the analyte is 50% ionised) is the called the pKa value. Each ionisable functional group on the analyte molecule will have it’s own pKa value.

The partial acid (or base) dissociation (ionisation) constant is defined as the negative logarithm of the equilibrium coefficient of the neutral and charged forms of a compound. This allows the proportion of neutral and charged species at any pH to be calculated, as well as the basic or acidic properties of the compound to be defined. It is very important to note the extent of ionisation at different pH values around the analyte pKa.

The 2pH Rule   back to top »

The extent of ionisation of an analyte molecule against pH has been demonstrated for both acids and bases.

Perhaps you noticed that the change in degree of ionisation happens over a limited pH range.  In fact because pH and pKa scales are logarithmic, it can be shown that at 1pH unit away from the analyte pKa, the change in extent of ionisation is approximately 90%.  At 2 pH units away from the pKa the change in extent of ionisation is approximately 99%, at 3 pH units 99.9% etc.  Therefore – a rule a thumb known as the ‘2 pH rule’ is useful in predicting extent of ionisation.  The 2pH rule for weak acids and bases is shown here:


Schematic Representation of the 2pH Rule describing the effects
of mobile phase pH on degree of analyte ionisation in solution


Using pH to Control Selectivity
It is important to realise that the two forms of ionisable analyte molecules give different retention characteristics. The ionised form is much more polar, and its retention in reverse phase HPLC is much lower (shorter retention time (tR), smaller retention factor (k’)). This behaviour is expected, as the more polar analyte has a higher affinity for the mobile phase and moves more quickly through the column. The converse is true of the non-ionised form as it is much more hydrophobic, relative to the ionised form.

This is nicely demonstrated with the separation of Ketoprofen and Indoprofen.

The plot of capacity factor against eluent pH is shown and this is an important tool in method development.  As can be seen from the chromatogram, at pH 2 when both analytes are non-ionised (because the eluent pH is well below the pKa value of the acidic analyte), retention is good and they are well separated due to their differences in hydrophobicity, which is exploited by the stationary phase interaction.  As the pH is raised and the analytes become increasing ionised, their relative polarity also increases and their retention in the reversed phase mode decreases significantly.  At around pH 8 when both analytes are fully ionised they both elute within the void volume, indicating little or retention and there is no separation of the analytes.

In Reversed Phase method development with ionisable compounds, plotting retention against pH for the various analytes in the separation can be a very enlightening.  Plots of the separation at pH 3.5, 5.5. and 7.5 can often help to decide the optimum pH for the separation.  This can help even with fairly complex analytes as is shown in the very famous example below:

There are several observations from the pH versus retention volume (retention time x eluent flow rate) plot that highlights the usefulness of this approach to optimizing separation selectivity.

  • Whilst we have no indication of peak efficiency, it can be assumed that as lines become closer together the separation (and the hence the resolution) between analytes reduces
  • Crossing lines – indicate co-elution of the analytes, and therefore the pH would not suitable for the separation
  • At point 1  (pH 5) all analytes are separated, although the peaks are fairly close and we would need to assume good peak efficiency to ensure the required resolution between all peak pairs. The retention time is reasonable (circa. 10mins. if we assume a flow rate of 1ml/min.).  More importantly, the response curves for all analytes are fairly flat – we will see why this is important in the next two points.
  • At point 2 (pH 7) the peaks are all separated (in terms of selectivity), however retention time is slightly longer.  However the response curves of compounds  4 and 5 are changing rapidly with pH which of great concern for the robustness of the method.  Small changes in mobile phase pH will lead to large changes in analyte retention and therefore potential loss of resolution between analytes – which is to be avoided at all costs in method development.
  • At point 3 (pH 8), whilst the peaks are best separated, the retention time has increased significantly and peak 5 is at this point very susceptible to changes in pH – again not recommended.
  • Conclusion – attempt separation at pH 3

Using Software to Optimise pH
It is possible to use software to help with pH optimization.  Typically this will involve making two or three injections at different pH values and entering the peak retention data into the program which will then ‘model’ the separation at all pH values and most will also give a predicted chromatogram under these conditions.

For the separation of 8 benzoic acids retention data was entered into DryLab (TM) software for separations at pH 2.9, 3.5 and 4.  The prediction program can be used to map the separation for best resolution and robustness and can help save time in finding the optimum conditions, especially where multiple ionisable analytes or mxtures of ionisable and non-ionisable analytes are involved.  This is one of a number of computer based chromatographic modelling and prediction programs available.


The following subjects are covered in full multi-media at

The Theory Of HPLC
Introduction (1.5hrs)
Chromatographic Parameters (3hrs)
Band Broadening (3hrs)
Column chemistry (4hrs)
Reverse phase (partition) chromatography (6hrs)
Ion-Pair Chromatography (3hrs)
Normal phase (absorption) chromatography (3hrs)
Gradient HPLC (3hrs)
Quantitative and Qualitative HPLC (3hrs)
FAST HPLC (4.5hrs)

Theory and Instrumentation of GC
Introduction (1.5hrs)
Chromatographic Parameters (3hrs)
Band Broadening (3hrs)
Gas Supply and Pressure Control (2hrs)
Sampling Techniques (4.5hrs)
Sample Introduction (5hrs)
GC Columns (5.5hrs)
GC Temperature Programming (3hrs)
GC Detectors (2.5hrs)

Instrumentation of HPLC
Mobile Phase Considerations (3.5hrs)
Solvent Pumping Systems (4hrs)
Autosamplers (4.5hrs)
Detectors (4.5hrs)

Solid Phase Extraction
Molecular Properties (4hrs)
SPE Overview (3.5hrs)
SPE Mechanisms (4.5hrs)
Method Development (5.0hrs)
Primary Sample Preparation Techniques (2hrs)
Liquid / Liquid Extraction Techniques (1.5hrs) Approaches to Automation for SPE (1.5hrs)

Fundamental GC-MS
Introduction (1.5hrs)
GC Considerations (4.5hrs)
GC -MS Interfaces (2.5hrs)

Fundamental LC-MS
Introduction (1.5hrs)
Electrospray Ionisation Theory (6.5hrs)
Electrospray Ionisation Instrumentation (4hrs)
Mass Analyzers (9.5hrs)
Atmospheric Pressure Chemical Ionisation (3.5hrs)
Atmospheric Pressure Photoionisation (3hrs)
Solvents, Buffers and Additives (3.5hrs)
Vacuum Systems (3hrs)
Flow Rates and Flow Splitting (3hrs)

MS Interpretation
General Interpretation Strategies (11hrs)

loading data
loading data
loading data
loading data
loading data
Home | About UsContact Us | SubscribeTerms and Conditions | Advertise | Privacy Policy |

loading data

loading data

loading data


loading data

loading data