1. An Efficient Approach to Column Selection in HPLC Method Development
Craig S. Young and Raymond J. Weigand
Alltech Associates, Inc.
2051 Waukegan Road • Deerfield, IL 60015
Phone: ALLTECH • Web Site:

2. Introduction Common Mistakes in Method Development:
• Inadequate Formulation of Method Goals
• Little Knowledge of Chemistry of Analyte Mixture
• Use of the First Reversed Phase C18 Column Available
• Trial and Error with Different Columns and Mobile Phases
These Mistakes Result In:
• Laborious, Time-consuming Development Projects
• Methods that Fail to Meet the Needs of the Analyst
Method development in HPLC can be laborious and time consuming. Chromatographers may spend many hours trying to optimize a separation on a column that is not optimal to accomplish the goals. Even among reversed phase columns there is astonishing diversity, owing to differences in both base silica and bonded phase characteristics. Many of these show unique selectivity.
What is needed is a more informed decision making process for column selection that may be used before the chromatographer enters the laboratory.
The method of column selection presented here involves a minimal investment in time initially, with the potential of saving many hours in the laboratory.

3. HPLC Method Development - A Proposed Procedure
At Your Desk
• Define your knowledge of the sample
• Define your goals for the separation method
• Choose the columns to be considered
In the Laboratory
• Choose the initial mobile phase chemistry
• Choose the detector type and starting parameters
• Evaluate the potential columns for the sample
• Optimize the separation conditions (isocratic or gradient) for the chosen column
• Validate the method for release to routine laboratories
HPLC Method Development - A Proposed Overview
At Your Desk
• Define your knowledge of the sample
• Define your goals for the separation method
• Choose the columns to be considered
In the Laboratory
• Choose the initial mobile phase chemistry
• Choose the detector type and starting parameters
• Evaluate the potential columns for the sample
• Optimize the separation conditions (isocratic or gradient) for the chosen column
• Validate the method for release to routine laboratories

4. Choosing the Appropriate HPLC Column Should Be Based Both Upon Knowledge of the Sample and Goals for the Separation
Benefits of this Approach Include:
• Small initial time investment
• Big time savings in the HPLC laboratory
• More “informed” approach to column selection
• More efficient than “trial and error” approach
Choosing the Appropriate HPLC Column Should Be Based Both Upon Knowledge of the Sample and Goals for the Separation
Benefits of this Approach Include:
• Small initial time investment
• Big time savings in the HPLC laboratory
• More “informed” approach to column selection
• More efficient than “trial and error” approach

5. Knowledge of the Sample Influences the Choice of Column Bonded Phase Characteristics}
Knowledge of the Sample Influences the Choice of Column Bonded Phase Characteristics
Knowledge of the Sample
•Number of compounds present?
•Sample matrix?
•Structure of sample components? Column Chemistry
•pKa values of sample components? (bonded phase, bonding type,
•UV spectral information? endcapping, carbon load)
•Concentration range?
•Molecular weight range?
•Solubility?
•Other pertinent data?
Knowledge of the Sample
• Structure of sample components?
• Number of compounds present?
• Sample matrix?
• pKa values of sample components?
• Concentration range?
• Molecular weight range?
• Solubility?
• Other pertinent data?
Column Chemistry
(bonded phase, bonding type, endcapping, carbon load)

6. Goals for the Separation Influence the Choice of Column Particle Physical Characteristics}
Goals for the Separation Influence the Choice of Column Particle Physical Characteristics
Goals for the Separation
•Max. resolution of all components?
•Partial resolution?
•Fast analysis? Column Physics
•Economy (low solvent usage)? (particle bed dimensions,
•Column stability and lifetime? particle shape, particle size,
•Preparative method? surface area, pore size)
•High sensitivity?
•Other goals?
Goals for the Separation
• Max. resolution of all components?
• Partial resolution?
• Fast analysis?
• Economy (low solvent usage)?
• Column stability and lifetime?
• Preparative method?
• High sensitivity?
• Other goals?
Column Physics
(particle bed dimensions, particle shape, particle size, surface area, pore size)

7. Column Selection Chart
Method Goals
Default Column
(Good for most
Applications)
High Sample Loadability
Suitable for MW >2000
Low Mobile Phase Consumption
Stability at pH Extremes
High Efficiency
High Capacity
Low Backpressure
High Resolution
High Stability
High Sensitivity
Fast Analysis
Fast Eqilibration
Particle Size
small (3µm) • •
medium (5µm) •
large (10µm) •
Column Length
short (30mm) • • • • •
medium (150mm) •
long (300mm) •
Column ID
narrow (2.1mm) • • medium (4.6mm) •
wide (22.5mm) •
Surface Area
low (200m2/g) • • •
high (300m2/g) • • •
Pore Size
small (60Å) • •
medium (100Å) •
large (300Å) •
Carbon Load
low (3%) •
medium (10%) •
high (20%) • • •
Bonding Type
monomeric • •
polymeric • • • •
Particle Shape
spherical • • • •
irregular •
The Column Selection Chart aids us in relating specific method goals to particle bed physical parameters. We start with a default column (150 x 4.6mm, 5µm, 200m2/g, 100Å, 10% C, monomeric, spherical) and change any specific parameters as necessary to meet the method goals.

8. Choosing the Bonded Phase
Draw the molecular structures for all known components of the mixture. Identify the two compounds whose structures are the most similar.
e.g.:
Choosing the Bonded Phase
Draw the molecular structures for all known components of the mixture. Identify the two compounds whose structures are the most similar.
Prednisolone
Prednisone

9. Choosing the Bonded Phase
For these two molecules, circle the structural features that differ. It is these differences that should be exploited to optimize the separation.
e.g.:
Choosing the Bonded Phase
For these two molecules, circle the structural features that differ. It is these differences that should be exploited to optimize the separation.
Prednisolone
Prednisone

10. Choosing the Bonded Phase
Use the results of the structural comparison to select a bonded phase showing optimal selectivity for these two molecules. In this case consider using a silica column (no bonded phase) for its ability to retain polar solutes through hydrogen bonding.
Use the results of the structure comparison to select a bonded phase whose selectivity is optimal for separating molecules according to the key structural differences previously circled. Keep the sample solubility in mind. In the example below, the acetonitrile-soluble steroids are equivalent in the 4-membered fused-ring skeleton, but differ in the nature of the polar functional groups (a-hydroxyketone vs. a,a'-dihydroxyketone).
Accordingly, we may want to consider an Alltima™ CN or Platinum™ CN bonded phase for its ability to retain polar solutes by dipole-dipole and
CN ••• C=0 (p- p) interactions.
Prednisolone
Prednisone

11. Functional Group Polarity Comparisons
Polarity Functional Group Structure Bonding Types Intermolecular Forces Displayed
Low Methylene s London
Phenyl s , p London
Halide s London, Dipole-Dipole
Ether s London, Dipole-Dipole, H-bonding
Nitro s , p London, Dipole-Dipole, H-bonding
Ester s , p London, Dipole-Dipole, H-bonding
Aldehyde s , p London, Dipole-Dipole, H-bonding
Ketone s , p London, Dipole-Dipole, H-bonding
Amino s , p London, Dipole-Dipole, H-bonding, Acid-base chemistry
Hydroxyl s London, Dipole-Dipole, H-bonding
High Carboxylic Acid s , p London, Dipole-Dipole, H-bonding, Acid-base chemistry
This table is meant as an aid to functional group chemistry comparisons.

12. Choosing the Bonded Phase
Examples of bonded phases used for HPLC packing media:
C8 or Octyl bonded phases are also common. Like C18, they are non-polar, but not as hydrophobic. Therefore, retention times for hydrophobic compounds are typically shorter. Also, they may show somewhat different selectivity then C18 due to increased base silica exposure.
C18 or Octadecylsilane (ODS)
Very nonpolar - Retention is based on London (dispersion) interactions with hydrophobic compounds.
Example Alltech Phase: Alltima™ C18

13. Choosing the Bonded Phase
Phenyl
Nonpolar - Retention is a mixed mechanism of hydrophobic and p - p interactions.
Example Alltech Phase: Platinum™ Phenyl
Unique selectivity results in p interaction of the bonded phase with electron deficient functional groups of solute molecules.

14. Choosing the Bonded Phase
Cyanopropyl
Intermediate polarity - Retention is a mixed mechanism of hydrophobic, dipole-dipole, and p - p interactions.
Example Alltech Phase: Alltima™ CN
Retention is a mixed mechanism, resulting from both hydrophobic interactions and dipole interactions of the bonded phase CN group with solute amino groups or p - p interactions with sites of unsaturation. It is best for polar organic compounds and is versatile enough for use in both normal and reversed phase modes.

15. Choosing the Bonded Phase
Each bonded phase has unique selectivity for certain sample types.
As a practical example, to separate toluene and ethyl benzene:
• Note a difference of one -CH2- unit
• Choose a C18 bonded phase for retention by hydrophobicity
• Maximize hydrophobic selectivity with a high silica surface area, high carbon load material like Alltima C18
Each bonded phase has unique selectivity for certain sample types. For example: to separate toluene and ethyl benzene (different by only one -CH2- unit), we would choose a C18 bonded phase. Further, we would want to narrow the decision to a particular packing material that shows good or excellent retention of such hydrophobic compounds (i.e. high % carbon load) to be able to maximize this particular separation. Alltech’s Alltima™ C18 would be a good choice here, as it uses high surface area silica and has a high % carbon load (16%). The effects of surface area and carbon load are discussed in the next section. The stationary phase must be able to "hold on" to the two compounds long enough to resolve them by differential migration.
Toluene
Ethyl Benzene

16. Choosing the Particle Physical Characteristics
Use the Column Selection Chart
• Use “default” column as starting point
• Match up method goals with individual particle physical characteristics
• Change only those particle parameters that affect the method goals
• Recognize the “optimum” column as a possible compromise
Example:
Sample Type: hydrophobic compounds
Method Goal: highest resolution
Use the "default" column as a starting point. Next, match up your method goals with the individual particle physical characteristics down the left edge of the chart. The matches will guide you in how to deviate from any individual parameters. For example, the method goal, highest possible resolution of hydrophobic compounds, may lead to the following "optimum" column.
Alltima C18™, 5µm, 250 x 4.6mm Part No
We chose this Alltima column for its long length, its high surface area, and its high carbon load.
Note that the choice may represent a compromise. Here, the “optimum” column for resolution sacrifices speed.
(If speed is important a 3um 150 x 4.6mm column is preferred)

17. Choosing the Particle Physical Characteristics
Example:
Sample Type: hydrophobic compounds
Method Goal: highest resolution
Column Selection Chart
Default Column Optimum Column†
Column Bed Dimensions 150 x 4.6mm 250 x 4.6mm
Particle Size 5µm 3* or 5µm
Surface Area 200m2/g >200m2/g
Pore Size 100Å 100Å
Carbon Load 10% %
Bonding Type Monomeric Mono- or Polymeric
Base Material Silica Silica
Particle Shape Spherical Spherical
* mobile phase backpressure may be excessive
† Optimum Column: Alltima C18™, 5µm, 250 x 4.6mm (Part No )
*Note that the choice may represent a compromise. Here, the “optimum” column for resolution sacrifices speed.

18. Choosing the Particle Physical Characteristics
Column Dimensions
• Length and internal diameter of packing bed
Particle Shape
• Spherical or irregular
Particle Size
• The average particle diameter, typically 3-20µm
Surface Area
• Sum of particle outer surface and interior pore surface, in m2/gram
Column Dimensions
This refers to the length and internal diameter of the packing media bed within the column tube.
Particle Shape
Most modern chromatographic packings have spherical particles, but some are irregular in shape.
Particle Size
This refers to the average diameter of the packing media particles. Standard particle sizes range from 3µm (high efficiency) to 15-20µm (preparative). Recently, Alltech introduced 1.5µm particles in Platinum™ C18 and Platinum™ EPS C18 phases for use in short, high efficiency columns. A 5µm particle size offers a good compromise between efficiency and back pressure.
Surface Area
Expressed in m2/gram, the total surface area of a particle is the sum of the outer particle surface and the interior pore surface.

19. Choosing the Particle Physical Characteristics
Pore Size
• Average size of pores or cavities in particles, ranging from ,000Å
Bonding Type
• Monomeric - single-point attachment of bonded phase molecule
• Polymeric - multi-point attachment of bonded phase molecule
Carbon Load
• Amount of bonded phase attached to base material, expressed as %C
Endcapping
• “Capping” of exposed silanols with short hydrocarbon chains after the primary bonding step
Pore Size
This refers to the average size of the pores or cavities present in porous packing particles. Pore sizes range from 60Å on the low end to greater than 10,000Å on the high end.
Bonding Type
This refers to how the bonded phase is attached to the base material. Monomeric bonding uses single-point attachment of each bonded phase molecule to the base material. Polymeric bonding uses multi-point attachment of each bonded phase molecule to the base material.
Carbon Load
Carbon load refers to the amount of bonded phase attached to the base material. For C18, C8 and phenyl packings, the carbon load is a good indicator of hydrophobic retention.
Endcapping
Endcapping applies only to reversed phase chromatography and is the process of bonding short hydrocarbon chains to free silanols remaining after the primary bonded phase has been added to the silica base.

20. Column Dimensions Effect on chromatography Column Dimension
• Short (30-50mm) - short run times, low backpressure
• Long ( mm) - higher resolution, long run times
• Narrow ( 2.1mm) - higher detector sensitivity
• Wide (10-22mm) - high sample loading
Short columns (30-50mm) offer short run times, fast equilibration, low backpressure and high sensitivity. Long columns ( mm) provide higher resolving power, but create more backpressure, lengthen analysis times and use more solvent. Narrow column (2.1mm and smaller) beds inhibit sample diffusion and produce narrower, taller peaks and a lower limit of detection. They may require instrument modification to minimize distortion of the chromatography. Wider columns (10-22mm) offer the ability to load more sample.

21. Particle Shape Effect on chromatography
Spherical particles offer reduced back pressures and longer column life when using viscous mobile phases like 50:50 MeOH:H2O.
Spherical particles offer reduced back pressures and longer column life when using viscous mobile phases like 50:50 MeOH:H2O.

22. Particle Size Effect on chromatography
Smaller particles offer higher efficiency, but also cause higher backpressure. Choose 3µm particles for resolving complex, multi-component samples. Otherwise, choose 5 or 10µm packings.
Smaller particles pack into columns with a higher density, allowing less diffusion of sample bands between particles and causing narrower, sharper peaks. However, smaller particles also cause higher solvent back pressures. As a rule of thumb, choose 1.5 or 3µm particle sizes for resolving complex, multi-component samples. Otherwise, choose 5 or 10µm packings.

23. Surface Area Effect on chromatography
High surface area generally provides greater retention, capacity and resolution for separating complex, multi-component samples. Low surface area packings generally equilibrate quickly, especially important in gradient analyses.
High surface area silicas are used in Alltech’s Alltima™, Adsorbospherel® HS, and Adsorbosphere® UHS packings. Low surface area silicas are used in Alltech’s Platinum™, Econosphere™, and Brava™ packings.
Solute retention is greater on packings that have a high surface area. High surface areas generally provide longer retention, greater capacity and higher resolution. As a rule of thumb, choose a base material with maximum surface area for resolving complex, multi-component samples. High surface area silicas are used in Alltech’s Alltima™, Adsorbosil®, Adsorbosphere® HS, and Adsorbosphere® UHS packings.

24. Pore Size Effect on chromatography
Larger pores allow larger solute molecules to be retained longer through maximum exposure to the surface area of the particles. Choose a pore size of 150Å or less for sample MW  Choose a pore size of 300Å or greater for sample MW > 2000.
Larger pores allow larger solute molecules to be retained longer through maximum exposure to the surface area of the particles. Choose a pore size of 150Å or less for sample MW  Choose a pore size of 300Å or greater for sample MW > 2000.

25. Bonding Type Effect on chromatography
Monomeric bonding offers increased mass transfer rates, higher column efficiency, and faster column equilibration.
Polymeric bonding offers increased column stability, particularly when highly aqueous mobile phases are used. Polymeric bonding also enables the column to accept higher sample loading.
Polymeric bonding offers increased column stability, particularly when highly aqueous mobile phases are used. Polymeric bonding also enables the column to accept higher sample loading.

26. Carbon Load Effect on chromatography
Higher carbon loads generally offer greater resolution and longer run times. Low carbon loads shorten run times and many show a different selectivity, as in Alltech’s Platinum line of packings.
Higher carbon loads generally give higher column capacities, greater resolution and longer run times. Conversely, low carbon loads shorten run times and may show different selectivity because of greater exposure of the base material. Alltech’s Platinum™ line of packings are specifically designed to offer unique selectivity through increased exposure of the base material, while still preserving peak shape. Choose high carbon loads for complex samples which require the maximum degree of separation. Choose low carbon loads to give shorter analysis times for simpler sample mixtures and for samples which require a high water content for solubility or stability.

27. Endcapping Effect on chromatography
Endcapping reduces peak-tailing of polar solutes that interact excessively with the otherwise exposed, mostly acidic silanols. Non-endcapped packings provide a different selectivity than do endcapped packings, especially for such polar samples.
Alltech’s Platinum™ EPS packings are non-endcapped to offer enhanced polar selectivity.
Endcapping reduces peak-tailing of polar solutes that interact excessively with the otherwise exposed, most acidic silanols. Non-endcapped packings provide a different selectivity than do endcapped packings, especially for such polar samples. Alltech’s Plantinum™ EPS packings are specifically non-endcapped to provide “enhanced polar selectivity”.

28. Conclusion
In this approach to HPLC column selection, the bonded phase chemistry of the column is chosen on the basis of an analysis of the sample component structures. The physics of the column is chosen according to an analysis of the goals for the separation method. This approach succeeds in predicting unique, optimum bonded phase chemistries and particle bed physical characteristics that are likely to meet the goals for the separation method.

29. Column Selection Example #1
What goals do I have for the method?
Maximum resolution of all components? Best Peak Shape for difficult samples? 
Fast analysis? 
Economy (low solvent consumption)? 
Column stability-long lifetime?
Purify one or more unknown components for characterization?
High sample loadability?
High sensitivity? …Other (High Sample Throughput--Quick Equilibration) 
Number of compounds present
Sample matrix
pKa values of compounds?
UV spectral information about compounds? UV -254
Concentration range of compounds
Molecular weight range of compounds
What do I know about the sample?

30. Column Selection Example #1
Structures of Compounds

31. Column Selection Example #1
Which two sample components have the most similar structures?
Draw them, then circle the structural differences between them.
Normal phase silica NH2 CN
Reversed phase C18 C8 Ph CN
Note: The structural difference between these two compounds is the hydrophobic hexyl side chain. This suggests a non-polar C18 or C8 column would interact with this area of difference to help provide separation of these two compounds.
Anthracene
3-Hexylanthracene
Recommended bonded phase (silica based materials only) – mark one

32. Column Selection Example #1
Column physical characteristics – use Column Selection Chart and Method Goals
Default Column Ideal Column Column bed dimensions (mm) 150 x x 2.1
Particle Size (µm)
Surface area (m2/g) <200
Pore Size (Å)
Carbon Load (%)
Bonding type Monomeric Monomeric
Particle shape spherical spherical

33. Column Selection Example #1
Available packing alternatives meeting the above criteria:
Packing Base Particle Particle Carbon Pore Surface Bonding End-
Material Shape Size Load Size Area Type cap’d
(µm) (%) (Å) (m2/g)
silica Sph. 3, 5, Mono. Yes
silica Sph. 3, Mono. Yes
silica Sph. 3, 5, Mono. Yes
silica Sph. 3, 5, Mono. Yes
Column of choice: Brava BDS C18, 100x2.1, 5µm (Spherical , 185m2/g, monomeric)
Allsphere ODS-2
Brava BDS C18
Econosphere C18
Platinum C18
Increased Sensitivity,
Low Solvent Consumption,
Fast Analysis
Best Peak Shape
Quick Equilibration
Reduced
backpressure
Good balance of efficiency
& backpressure

34. Column Selection Example #2
What goals do I have for the method?
Maximum resolution of all components?  Partial resolution, resolving only select components?
Fast analysis?
Economy (low solvent consumption)?
Column stability-long lifetime? 
Purify one or more unknown components for characterization?
High sample loadability?
High sensitivity?  …Other
Number of compounds present
Sample matrix
pKa values of compounds?
UV spectral information about compounds? UV -254
Concentration range of compounds
Molecular weight range of compounds
What do I know about the sample?

35. Column Selection Example #2
Structures of Compounds

36. Column Selection Example #2
Which two sample components have the most similar structures?
Draw them, then circle the structural differences between them.
Notes: both structures very polar, with amine and pi bond functions--a RP CN column may give good separation by mixed mode retention of hydrophobic, CN---H---NR2 hydrogen bonding and - interactions with double bonds.
Normal phase silica NH2 CN
Reversed phase C18 C8 Ph CN
Recommended bonded phase (silica based materials only) – mark one

37. Column Selection Example #2
Column physical characteristics – use Column Selection Chart and Method Goals
Default Column Ideal Column Column bed dimensions (mm) 150 x x 2.1 Particle Size (µm)
Surface area (m2/g)
Pore Size (Å) Not critical
Carbon Load (%)
Bonding type Monomeric Polymeric
Particle shape spherical spherical

38. Good balance of efficiency
Column Selection Example #2
Available packing alternatives meeting the above criteria:
Packing Base Particle Particle Carbon Pore Surface Bonding End-
Material Shape Size Load Size Area Type cap’d
(µm) (%) (Å) (m2/g)
Adsorbosil CN silica Irreg. 5, Poly. Yes
Alltima CN silica Sph. 3, Poly. Yes
Allsphere CN silica Sph. 3, 5, Mono. No
Platinum CN silica Sph. 3, 5, Mono. No
Column of choice: Alltima CN, 250 x 2.1 , 5µm ( Spherical , 350 m2/g , polymeric)
High resolution,
High sensitivity
High res.
Good balance of efficiency
& backpressure
Reduced
backpressure
Robust