The CHROMacademy Essential Guideto HPLC Troubleshooting - Eluents and Solvent Delivery Systems
The Essential Guide from LCGC’s CHROMacademy presents the first in our series of webcasts on HPLC Troubleshooting. In this session, Scott Fletcher (Technical Manager , Crawford Scientific) and Tony Taylor (Technical Director, Crawford Scientific), present practical troubleshooting and maintenance information around HPLC eluents and solvent delivery systems (pumps). This session considers important practical concepts such as good practice for eluent preparation, filtration and degassing as well as common related issues such as retention time drift and selectivity changes. We also explain the working principles of HPLC solvent delivery systems and highlight the components which are most susceptible to problems. The webcast concludes by considering various trohbleshooting tests and strategies as well as common maintenance operations for solvent delivery systems. A must see for anyone who uses HPLC equipment and who wants to improve their troubleshooting and instrument maintenance skills.
Good practice for eluent preparation
Troubleshooting retention time and selectivity issues
Degassing – do’s and don’ts
Anatomy of a mobile phase delivery system
Components that go wrong and why
Symptoms of a poorly pump
Troubleshooting and tests
Recommended maintenance operations
Who Should Attend:
Anyone who uses HPLC equipment and who wants to improve their troubleshooting and instrument maintenance skills.
Key Learning Objectives:
Identify good practices for preparation of HPLC eluents
Appreciate the requirements for successful filtration and degassing of eluents
Explore the links between HPLC eluent chemistry and chromatogram retention time / selectivity issues
Appreciate the various designs of HPLC Solvent Delivery System (pump)
Identify the most susceptible components of the pump
Use tests and troubleshooting strategies to correctly identify pump issues
Develop a maintenance schedule for your HPLC Solvent, Degasser and Pump
High quality (HPLC Grade) solvents are important for use in HPLC, especially when performing gradient analysis.
Impure solvents can give rise to noisy baselines, peaks in blank gradient runs, high background (low sensitivity) analysis and, if improperly filtered, blockages in various filters throughout the system. Solvents used for gradient HPLC analysis over a wide range of %B, at low UV wavelength (<220nm) and where high sensitivity require the highest purity solvents and all potential sources of contamination should be avoided. Water purity is especially important and can be a significant source of contamination. Ultra pure de-ionised water of 18MΩ resistivity or better is recommended.
Many instances of ‘spurious’ or artifact peaks in the chromatographic baseline of gradient HPLC analyses can be attributed to low quality solvent usage or solvent contamination. The extent to which this problem can affect the chromatographic baseline is shown in Figure 1 below.
Figure 1. Spurious peaks generated from contaminated solvents (30 min equilibration top trace / 10 min equilibration lower trace). Conditions: Intersil ODS 3 150 x 4.6mm 5μm, 30oC, 1.5ml/min, 0-100%B over 15 mins. with a 5 min. hold at 100%, solvent A 5:95 (v/v) acetonitrile – pH 7.0 potassium phosphate buffer (10mM), solvent B 80:20 (v/v) acetonitrile – pH 7.0 potassium phosphate buffer (10mM) (Reproduced with permission LCGC magazine Volume 16 Number 11)
The spurious peaks in Figure 1 are caused primarily by impurities in the solvents accumulating on the head of the analytical column during re-equilibration, and are eluted as the eluotropic strength of the mobile phase increases. This is borne out by the relative increase in peak height / area as the gradient equilibration time is increased in Figure 1.
Potential sources of contamination in the typical laboratory include:
Cross contamination from glassware (ensure proper washing and rinsing)
Contamination from soap solution sued to wash glassware (rinsing of up to 10 times in clean deionised water may be necessary to remove soap solution, even from glassware washed in a professional dishwasher)
Calibration and storage solutions used with pH meters all have the potential for cross contamination
The salt solution used to fill the pH meter electrode may transfer contaminants into the solvents if left in contact for extended periods of time (check pH using aliquots extracted from the bulk solution to minimize the potential for this contamination)
Figure 2. Chromatograms recorded from various potential sources of contamination.
a) glassware contact
b) solvent filtration via a 0.45μm filter
c) pH meter probe
d) degassing (via Helium Sparge)
(See Figure 1 for conditions)
(Reproduced with permission LCGC magazine Volume 16 Number 11)
Figure 2 shows the results from a gradient analysis using gradient analysis of solutions brought into contact with various potential sources of contamination. Soap extracted from glassware used to prepare the mobile phase and the pH electrode salt solution (extended contact >10 mins) provided the most significant background peaks.
All buffers should be prepared freshly on the day required. This practice ensures that the buffer pH is unaffected by prolonged storage and that there is no microbial growth present.
Buffer reagents can contain a stabilizing agent, for example, sodium metabisulphite. These stabilizing agents often affect the optical and chromatographic behaviour of buffer solutions, so it is often worth buying reagents that contain no stabilizer. Containers of solid reagent are easily contaminated by repeated use. For this reason, we recommend that regents be purchased in low container weights. If buffer solutions are stored, be aware that they have a finite lifetime. Refer to pharmacopoeia monographs or similar for further guidance on buffer shelf life.
Differences in mobile phase preparation technique often account for difficulties in reproducing or transferring an analysis. Retention time and selectivity can both be affected when the mobile phase is incorrectly prepared. Liquids should be prepared volumetrically and solids gravimetrically.
Mobile Phase preparation involves:
Measuring the appropriate volume of each solvent
Adding buffer salts or liquids and adjusting the pH of aqueous component
Adding other additives to the aqueous component of the mobile phase (ion pairing reagents for example)
* Filtering and degassing steps may be carried out in tandem.
Mobile phase pH (or more accurately the hydrogen ion activity) of a mobile phase can affect both analyte retention and separation selectivity when ionisable analytes are present. In some separations, small changes in pH (<0.1pH units) can have drastic effects on the separation and so accurate and precise pH measurement and adjustment is critical.
The upper chromatogram in Figure 3 shows the separation of six weakly acidic analytes at a mobile phase pH of 5.10. At this pH all six components are adequately resolved, with the critical peak pair (1, 2) exhibiting a resolution value of 1.50 − the minimum recommended peak separation for quantitative purposes.
However, the chromatographic separation of the same six analytes at a pH of 5.20 would be as represented in the lower chromatogram of Figure 3. At a pH of 5.20 the six analytes are now not adequately resolved, with the critical peak pair (5, 6) separated with a resolution of only 0.58. This is appreciably less than the recommended 1.50, and consequently their individual peak areas may not be adequately integrated for quantitative purposes.
Figure 3. Separation of 6 weakly acidic compounds at two different pH values, pH 5.10 top and pH 5.20 bottom. Conditions: C8 150 x 4.6mm 5μm, 25oC, 1.0ml/min, ammonium acetate buffer (10mM various pH values)
Whilst we are concerned with the pH of our aqueous:organic mobile phase mixtures to control analyte retention and selectivity, it is the hydrogen ion activity (aH) which actually controls these parameters. Calculation or measurement of the hydrogen ion activity in the mixed solvent system is not simple, and so we tend to use the measured pH of the aqueous component of the mobile phase which we know will result in a suitable hydrogen ion activity in the mixed mobile phase to give the retention and selectivity that we require (i.e. this has been empirically determined sometime in the past).
Whilst measuring pH using a probe is a suitable way to prepare buffered mobile phases, volumetric and gravimetric means can be both quicker and more precise.
The excellent buffer calculator at this link http://www.liv.ac.uk/buffers/buffercalc.html will give you recipes to make buffers by pH adjustment or by volumetric / gravimetric means, the latter of which also take into account the hydrogen ion activity correction. For more information on this subject see reference 1.
The important point here is that the correct buffer is chosen and pH is adjusted in the correct way. The buffer is typically chosen so that the target pH is within +/- 1pH of the buffer pKa for optimum buffer capacity. The buffer strength (buffering capacity) will have been empirically determined previously but one should note that higher buffer capacity is required when the sample diluent and mobile phase hydrogen ion activity (pH or buffer strength) differ greatly.
One should be sure that the pH measurement is made accurately and consistently according the pH meter manufacturers instructions and local practices. This may seem a somewhat rudimentary discussion, however, one should stop to consider the following questions before assuming that you know how to correctly measure pH:
Is a two or three point calibration recommended for your meter?
What calibration solutions are required to ‘bracket’ the required pH?
Does one need to uncover a filling hole in the pH electrode?
Are you going to stir (using a magnetic stirrer) the calibration and/or test solutions
Does one rinse with deionised water, tap water or the calibration / sample solution
Are you testing the bulk aqueous component or small aliquots removed from the bulk?
How long does one wait for the pH to ‘settle’ post addition of acidic or basic solution?
Most mobile phases are a mixture of at least two solvents. The volume of each solvent should be measured independently prior to combining the two solutions.
This avoids problems associated with volume change on mixing certain solvents – a 60:40 mix of water:methanol may be incorrect by up to 10% due to latent heat of mixing affecting the overall volume of the mixed solution. A 1% error in mobile phase composition can result in as much as 5% change in retention .
Mixing is often carried out on-line by the HPC system – this tends to overcome the problems when pre-mixing mobile phase components – which one has to do when using isocratic pumping systems.
When using buffers in gradient analysis make sure they are entirely soluble in the full range of expected mobile phase compositions. Less soluble buffers may precipitate as the organic strength of the mobile phase is increased – mobile phases containing more than 60% acetonitrile are particularly renowned for this phenomenon.
Figure 4. Apparatus commonly used in mobile phase preparation.
To protect the HPLC and your column from particulate matter, manufacturers recommend filtering of mobile phases prior to use. This removes any particulate matter that may cause blockages. After filtration, the solvents should be stored in a covered reservoir to prevent contamination with dust etc. Filtering HPLC solvents will benefit both your chromatography and the wear and tear of the HPLC system. Pump plungers, seals and check valves will perform better and lifetimes will be maximized.
This task can be accomplished with simple vacuum filtration devices purchased through HPLC supply catalogs. Add all buffers and modifiers prior to filtration. A vacuum is applied to pull the solvent through a filter –which also acts to simultaneously degass the mobile phase. One should take great care to avoid loss of the more volatile component in pre-mixed mobile phases under vacuum conditions. Volatile buffers /ion pairing reagents (such TFA) and the organic solvent (typically methanol or acetonitrile) may be lost under the reduced vapour pressure conditions in the vacuum (receiving) flask – thus altering the mobile phase composition and potentially affecting analyte retention time or separation selectivity.
Handle filters with tweezers and make certain the filtration apparatus is clean at all times. Nylon 66 is a good filter material for aqueous mobile phases, while PTFE is an excellent filter for most organic solvents. Inorganic membranes are resistant to chemical degradation from a wide range of HPLC solvents.
Be aware that Teflon filters cannot be used with water due to the material’s nonpolar characteristics. HPLC mobile phase filters have a pore size of around 0.45 microns.
One needs to ensure that filter fibers do not find their way into the degassed solvent as problems with blocked / sticking pump check valves will occur.
Most HPLC systems will incorporate an eluent reservoir filter which will filter the mobile phase as it is drawn into the system using a frit mesh size of 0.2 to 0.45μm filter.
Figure 5. Inline mobile phase filters – made from both ceramic and metal materials are used as sinkers in the eluent reservoir.
Eluent reservoir filters should be regularly cleaned. Soaking on 0.1N nitric acid for 30 minutes is recommended. Do not sonicate ceramic filters or they will be irreparably damaged.
Mobile phase degassing is used to remove dissolved gases from the mobile phase, which can cause a variety of problems:
Pumping reliability decreases. Air in the pumping system can cause irreproducible system back-pressure, retention time drift and poor solvent mixing.
Air bubbles can get trapped in the check valves –causing a distinctive pressure ripple with associated baseline appearance.
An air bubble in the piston chamber can cause a loss of pumping efficiency, as shown in Figure 5. A sinusoidal base line is seen due the lower compressibility of air in the pump head. In extreme cases it is possible that air will accumulate in the pump head which will stop the mobile phase flow.
Figure 6. Air bubbles trapped in the pump head can cause a variety of issues including a sinusoidal baseline profile, variable system back-pressure and irreproducible retention times.
Methods of Degassing
There are four commonly used methods for degassing mobile phases:
The relative efficiency of some of these methods is shown in Figure 7 – the figure is based on the percentage of oxygen removed from a set volume of methanol.
Broadly the degassing techniques are spilt into online and offline techniques – with online techniques being preferred. Offline techniques might be highly effective – but once the degassing stops and the eluent bottles are placed onto the HPLC system –ingress of air will begin almost immediately unless the solvent is held under a blanket of inert gas.
Ultrasonic degassing is perhaps the least favored of the off-line techniques. Not only is it relatively inefficient, it also risks loss of volatile mobile phase components (when the phase has been pre-mixed) through solvent heating in the bath. This can lead to irreproducible retention times.
Figure 7. Relative efficiency of various online and offline degassing techniques for removing dissolved oxygen from methanol (reproduced from reference 3 with permission).
Refluxing is perhaps the most effective degassing technique –however it is arduous and not without risk when trying to degass large amounts of solvent at any one time. For these reasons refluxing is rarely used.
Helium sparging removes approximately 80% of the dissolved gases by solubilising dissolved oxygen and nitrogen from the mobile phase into itself. Helium however, is not particularly liquid soluble and will agglomerate and out-gas, carrying the other dissolved gases with it.
Helium spargers can be simply constructed by connecting a length of PTFE tubing with a frit at one end to a helium cylinder. Immerse the frit into the mobile phase and adjust the regulator to release a gentle stream of bubbles. It is important that the helium is released at the bottom of the eluent reservoir and permeates throughout the whole of the eluent.
Generally speaking 1 litre of mobile phase will be fully degassed with 1 litre of helium at a rate of 100-200 ml/min.
Excessive degassing with helium can cause the loss of more volatile mobile phase components and may result in a poorly degassed phase – especially if the surface of the eluent liquid can be seen to be ‘bubbling’, whereupon more air may entrained back into the phase than is released at the surface in the helium bubbles. Best practice is to sparge at 100 – 200 L/min for a few minutes and then reduce the sparge flow to a trickle.
Further, one should consider the relatively high cost and shortage of Helium when considering mobile phase degassing options.
Figure 8. Helium sparger (fashioned from PTFE tubing with a fritted metal sparger). One should avoid excessive helium flow as efficiency of degassing will be considerably reduced.
The mobile phase enters the gas permeable tubing and passes in to the vacuum chamber. Any dissolved gas will be drawn out from the mobile phase, through the tubing walls, into the lower pressure of the vacuum chamber.
Pressure will increase in the vacuum chamber and the sensor monitors this change. Once a pre-defined specification pressure is reached, the control circuit opens the solenoid valve, turns the pump on and the gas is removed from the vacuum chamber and the correct level of vacuum re-established.
This method of degassing is perhaps the most popular for modern instruments and most manufacturers will provide an online degassing unit such as that shown in Figure 9. Whilst online vacuum degassing is not the most efficient technique –it is both quick and convenient –leading to it’s widespread adoption. Modern on-line vacuum degassers are capable of producing an eluent stream that satisfies all the stringent performance characteristics of modern HPLC.
Figure 9. Principles of online vacuum degassing.
Mobile Phase Storage
Changes may occur to the mobile phase on standing. Volatile mobile phase components may evaporate from the surface causing gradual changes to the retention time, when using pre-mixed mobile phases. Typically retention time will increase as it is the more volatile organic (more highly eluotropic) solvent which is lost in preference.
Selectivity of the separation may change if volatile ion pairing reagents (such as TFA) are lost through evaporation. This will alter the pH of the mobile phase which in turn may also lead to selectivity changes.
On standing, CO2 may also ingress into the mobile phase, which being acidic, will also lead to a reduction in pH of the mobile phase.
Always ensure that the mobile phase is capped, but has a mechanism to prevent a partial vacuum forming in the mobile phase reservoir. This will reduce the amount of solvent evaporation and pH change due to ingress of CO2. If a separation is particularly susceptible to changes in eluotropic strength or pH, then it is possible to provide a inert blanket gas such as nitrogen over the eluent solution to increase the vapor pressure and prevent evaporation. This becomes increasingly important as the eluent level drops and the headspace volume above the eluent liquid increases.
Before starting an analysis, calculate the total volume of mobile phase required.
Prepare all the phases at the same time and place in reservoirs that are large enough to accommodate them. Insufficient mobile phase may cause the system to pump dry, which is
undesirable because it will fill the system with air. If this does happen, purge the solvent lines and pump with fresh solvent then allow the system to pump solvent until it is equilibrated.
Mobile Phase Properties
Do not use highly acidic or basic solvents unless your HPLC system and column have been engineered to accommodate them. Seals, plungers etc. can be damaged by extreme pH conditions.
Highly aqueous mobile phases are ideal breeding grounds for microbes. Ensure that an organic solvent is flushed through the HPLC system and column at least once every 48 hours to kill unwanted microbial growth. Alternatively add a small amount of sodium azide to the aqueous solvent to inhibit growth.
Note: Never allow a HPLC column or system to stand with water or buffer in it for an extended period of time (for example, over a holiday break). Always flush with a solvent mix that contains a minimum of 20% organic in water.
Figure 10 below shows a schematic representation of the single piston pump. Most modern HPLC pumping devices are based around this simple concept of a pumping chamber with two check valves to regulate the liquid flow.
The component parts are:
Eccentric cam: As the eccentric cam rotates it moves the piston into and out of the piston chamber –causing mobile phase to enter or leave the pump head chamber
Spring mounted piston: The piston moves forward to force the liquid out of the chamber and through the check valve. With the piston at the end of its stroke the spring is fully compressed. The spring pushes the piston back and liquid is drawn into the piston chamber.
Liquid filled chamber
Check (ball) valves: to regulate flow direction
Outlet Check Valve (Ball Valve): This ball and seat valve allows unidirectional flow of liquid out of the piston chamber. As the piston moves into the chamber the ball (usually made from ruby) lifts from the seat (usually made from sapphire), to allow liquid to flow out of the pump head. As the piston moves out of the chamber, the ball is forced down against the seat under the vacuum created in the pump head. This stops back flow of mobile phase into the pump head.
Inlet Check Valve (Ball Valve): This ball and seat valve allows unidirectional flow of liquid into the piston chamber. As the piston moves out of the chamber the ball (usually made from ruby) lifts from the seat (usually made from sapphire), to allow liquid to flow into the pump head. As the piston moves into the chamber, the ball is forced down against the seat under the pressure created in the pump head. This stops flow of mobile phase out of the pump head.
Figure 10. Operating principles of a single piston pump. NOTE: all modern HPLC systems use a combination of pumps in what is known as the ‘reciprocating’ design (see Figure 11).
An electric motor is used to drive an eccentric mounted cam and spring mounted piston or ball drive piston which moves into (the compression or delivery stroke) and out of (the filling stoke) a chamber drilled into a pump head.
The filling stroke causes mobile phase to fill the chamber and the liquid flows through the inlet check valve as the ball is lifted from the check valve seat. No mobile phase flows from the pump during this phase of pumping cycle.
The delivery stroke compresses the liquid within the chamber and forces the liquid out through the outlet check valve and to the HPLC system. This produces a sinusoidal pressure / flow profile.
To create a continuous flow of mobile phase (at constant pressure), a reciprocating design is typically used which is shown in figure 11, in which two pumps operate at 180o out of phase.
Figure 11. Operating principles of a dual piston reciprocating pump.
Solvent Delivery Systems for Online Mixing
In Gradient HPLC –a more strongly eluting (usually organic) solvent is gradually introduced into the mobile phase to progressively elute more strongly elute analytes within a reasonable timeframe. To meet the requirements for gradient HPLC there is a need to employ a more sophisticated pump design that is capable of pumping and mixing exact proportions of more than one liquid simultaneously and reproducibly. There are two ways of achieving a gradient profile using HPLC pumping equipment:
Two pumps working in unison but each delivering a different volume fraction of the total flow – this is called a Binary Gradient Pump
One pump fitted with a proportioning value to allow exact volumes of liquids to be mixed prior to the pump –this design is called a Ternary or Quaternary Gradient pump
Binary Pumps consist of two ‘channels’ or pumps – each channel has an identical reciprocating pump, both channels being connected to a low volume mixing chamber and to a single pulse damper, which in turn may be connected to a second mixing chamber. Computing firmware within the pump, or an external computer, controls the pumping speed (or volume) of each reciprocating pump so that the solvent volume fraction from each channel will be ‘proportioned’ to give the desired total flow. That is, if a gradient composition of 50% solvent A and 50% solvent B is required the pumps will each pump at an equal flow rate. If the total desired pump output is 1mL/min, each of the reciprocating pumps will deliver solvent at 0.5 mL/min.
Figure 12. Operating principles of a binary (high pressure mixing) pump.
Low Volume Mixing Chamber: Primary mixing of the mobile phase from the two pump heads prior to pulse damping
Pulse Damper: Removes pressure fluctuations in the mobile phase flow caused by the operation of the two pump heads
Second Mixing Chamber: A larger volume mixing component which also acts to further reduce pressure ripple to ensure the most stable baselines and lowest pressure ripple
The quaternary pump will deliver up to four different solvents simultaneously via a mixing device located prior to the pump(s). Rather than use four pumps, which would be prohibitively expensive, the quaternary pump uses one dual-piston reciprocating pump and a solenoid controlled portioning valve located in-line between the solvent degasser and the pump head.
The design of the quaternary pump dictates that all mixing of solvents is done prior to liquid compression – these pumps are therefore known as low pressure mixing devices.
The solenoid controllers of the proportioning valve open to control the volume fraction of each mobile phase component at any instant. The time cycle of the proportioning valve are usually based on the pump ‘duty cycle’. If the pump duty cycle (one full pump cycle) is around one second – then a 25:25:25:25 A:B:C:D mobile phase composition would be achieved by opening each solenoid valve for one quarter of a secnd each. The proportioning valve is usually controlled by firmware resident with the pump module which is in turn programmed by a computer data acquisition system. The use of a PC make this whole process relatively simple and the end users needs only concern themselves with entering the gradient conditions correctly in the pump set-up screen of the operator software.
Figure 13. Operating principles of a quaternary (low pressure mixing) pump.
Proportioning Valve: Each solenoid valve will be open for a proportion of what is called the ‘duty cycle’ of the pump (often based on the piston stroke time, typically around one second) – i.e. when mixing four solvents at equal proportion, each solenoid will be open for one quarter of the duty cycle (0.25 seconds)
Inlet Valve: Provides unidirectional flow into the piston chamber
Outlet Valve: Provides unidirectional flow out of the piston chamber
Damper: Removes pressure fluctuations in the mobile phase flow caused by the cyclical operation of the two pistons
The ‘check’ or ‘ball and seat’ valve may be positioned in–line before and / or after the piston chamber. The valve is designed to allow flow of mobile phase in one direction only.
Constructed from inert materials such as ceramic or ruby, the ball sits in a small cylinder or seat made of similar material (sapphire being another popular material for seat manufacture). Depending on the stroke of the piston the ball is either positioned against the cylindrical seat, thus stopping mobile phase flow, or in the main body of the valve where it offers little resistance to the flowing liquid.
Typical ‘cartridge’ type mechanical ball and seat valve.
Typical electronic ‘solenoid’ two way check valve
(typically for low pressure use only).
Figure 14. Check Valve operating principle and problems with contamination (top), typical check valves.
Check valves are designed to give uninterrupted flow of liquid over many thousands of hours. However, check valves and their constituent components are classed as consumable items and will need to be replaced after extended use. The use of highly corrosive (mineral acids or acid halides), un-filtered, or high buffer concentration mobile phases may considerably shorten the lifetime of the check valve.
Check valve lifetime will be shortened if particulate matter is present in the mobile phase, either through poor mobile phase filtration – or by the precipitation of solid buffers
A failing (sticking) check valve will give a noisy baseline appearance, sometimes, but not always, with a discernable pattern and a systematically erratic flow or pressure
A failed (stuck open) check valve will give symptoms similar to that of air bubbles in the system with a periodically erratic flow / pressure
A failed (stuck closed) check valve typically result in a system over pressure
Figure 15. Typical baseline profiles associated with failing check valves compared against the sinusoidal profile obtained with air entrained in the pump head.
If a failing check valve is suspected then investigate which of the check valves is at fault by comparing the flow / pressure output with the pump cycle and isolate the problematic component.
Mechanical check valves can be cleaned in a sonic bath using solvent (typically methanol or isopropyl alcohol) or using a dilute solution of nitric acid (0.1N) for 30 minutes. Note – some manufacturers advise against cleaning of cartridge type valves and electronically actuated valves (solenoids) should never be cleaned by sonication or solvent soaking. Please check with your manufacturer for the correct cleaning procedure.
Pistons and Piston Seals
The piston is used to compress the mobile phase in the ‘analytical volume’ – i.e. the piston chamber within the pump head. It should be physically robust and chemically inert to withstand the rigors of pumping potentially corrosive solvents at high pressures (up to approx. 5000psi). Typically pistons are constructed from sapphire rods with rounded ends –mounted onto a passivated metal holder. Depending upon the pump design the piston may fit exactly into the piston chamber, however, it is much more usual for the piston have a small tolerance between itself and the inner surface of the piston chamber and a small amount of solvent will flow between these surfaces for purposes of lubrication.
The piston needs to move through a seal, which will isolate the liquid in the piston chamber from the pump mechanism. Piston seals are typically constructed using a metallic spring encased in a solvent resistant plastic material -PEEK and CALRES are commonly used plastics. The seal must fit the piston plunger tightly enough to avoid leaks at high pressure but must also avoid excessive wear on the piston. Both the piston and the seals are consumable items, however it is reasonable to expect a longer lifetime from the piston if the pump is properly maintained. Extended periods of inactivity with buffered mobile phases, or poorly filtered mobile phases are likely to reduce the lifetime of both the piston and piston seal.
Figure 16. Typical piston and piston seal (note the seal is not to scale and is designed to fit tightly around the body of the piston plunger).
The piston is commonly made of sapphire, which offers an extremely hard, durable and inert surface to the flowing mobile phase. Precise alignment of the piston should enable thousands of hours of continuous use. Problems will arise generally from poor use or maintenance of the pump and crystallization of buffer solutions or particulate material in unfiltered mobile phases is the most common cause of damage to the piston. If the mobile phase within the pump head is allowed to evaporate on standing – the buffer salts will crystallise. When the pump is restarted the buffer crystals will score the piston and seal – causing a leak. Similarly, if solvent system are used which result in precipitation of buffers at high concentration (i.e. high organic phases containing acetonitrile for example) the piston and pump seal will also be scored – again causing a leak.
Initially any leaks from the piston seal may be so small that they do not appear to affect the delivery of the pump. However left unattended the result will be gradual deterioration of the performance of the pump. Small leaks can allow small amounts of air to be introduced to the system – resulting in noisy baselines, sometimes with a suggestion of a cyclical pattern.
Ultimately the piston seal will fail - causing a cyclical baseline and corresponding pressure ripple. Catastrophic failure of the pump seal will result in loss of system pressure.
Figure 17. Typical baseline profiles associated with minor and major leaks of the pump seal.
To extend the lifetime of the pistons avoid particulate matter in the mobile phase, involatile buffers at high concentration and employ a flush method at the end of a sequence of injections.
Figure 18. Recommended flushing procedure for an HPLC pump (system).
If buffers are to be used at high concentration then it is recommended that a piston seal wash is utilized to help extend the lifetime of the pistons.
Whenever the seals are replaced the piston should be examined for scratches and /or deposits of crystallized buffer. Any deposits can be removed with a lint free cloth dipped in alcohol. Always clean from the base of the piston forwards. Any signs of scratching or wear on the piston will require the piston to be replaced. Several manufacturers recommend that the piston and seal are always changed as a ‘pair’. Seal replacement must be meticulously carried out – following manufacturers guidelines, which typically involve pre-soaking / softening the seal in alcohol / solvent followed by careful insertion using a special tool. Only careful insertion will prevent issues with premature seal wear.
A simple preventative maintenance (PM) schedule should be implemented which includes changing the piston seals. The duration between PM’s will be determined by the user and will vary depending on the instrument usage.
Typically either 3, 6, 9 or12 months between PM’s is normal. With a good PM strategy it is common not to require any major maintenance of the instruments outside these pre-determined dates.
Even when using reciprocating piston pumps a slight pulse can be detected in the baseline – especially at higher attenuation. In order to achieve the lowest possible pressure flow / ripple characteristics, modern HPLC pumps use pulse damping units. The damping unit is filled with a compressible liquid, and is separated from the flowing mobile phase via a flexible membrane. The mobile phase is allowed to enter the pulse damper after leaving the liquid filled chamber of the pump head. The pulse damper absorbs energy fluctuations of the pulsed flow and can mechanically smooth the flow output from the pump. Typically a greater than 98% reduction on the observable pulse of a reciprocating piston pump will be observed when using a pulse damping unit.
It is usual for the flexible membrane within the pulse damping unit to be attached to a tension or strain gauge - this is used to measure and report pressure and pressure ripple via the software control system.
Pulse dampening units require no maintenance and rarely fail. Particulate matter is always a possible cause of a blockage and care should be taken when using and changing mobile phase composition.
Figure 19. Various pulse damping devices – diaphragm type (top) are used on most modern HPLC systems, coil type (bottom).
Diaphragm Type - the most usual damper type is based on a membrane or diaphragm, usually having a very low internal volume (< 0.5 mL). The compressibility of the filling liquid is enough to compensate for the pulsations of the dual piston pump with piston volume up to around 100 µL. Pressure ripple should be reduced to less than 2% of the system pressure with this type of unit. Greater pressure ripple indicates problems with the piston / piston seal combination.
Coil Type - as the pump strokes, the coil flexes, absorbing the energy of the pulsations. This type of pulse damper holds a large amount of liquid which must be purged during solvent changes or when performing gradient elution.
The purge valve allows solvent to be primed (drawn) into the pump head. Without this – the mere pumping action of the pump head itself might not be enough to draw solvent from the eluent reservoir, through any online degassing equipment that might be present and into the pump. Purge valves can be used to prime the HPLC pump in two separate ways:
The purge valve is opened and a higher than usual pump flow selected (~5 – 10 mL/min. typically) in order to effectively ‘draw’ the solvent into the pump head
The draw off valve has a port to which a syringe might be attached – once the valve is open the user may manually draw solvent from the reservoir into the pump head
The purge valve is located after the liquid filled chamber of the pump outlet and before the injection valve (autosampler) in the HPLC system and can be either fully open or fully closed. When open the mobile phase flow bypasses the injector, column and detector and flows to waste – a useful feature when work on the system is required but when stopping the mobile phase flow would be inconvenient – whilst inserting a column into the system for example. In routine use it is common to open the purge valve only when changing solvents or priming the system. When priming the system it should be noted that at least 2x the system volume up to the pump head should be drawn through to eliminate the previous phase, avoid precipitation issues and reduce column / mobile phase incompatibility issues. Some systems may have up to 20mL system volume prior to the pump head if online degassers are included in the configuration.
Figure 20. In-line / Purge Valve filters - integral filter cartridge design (top) (Reproduced with permission from Agilent Technologies, Santa Clara, California), simple syringe draw off valve (bottom).
Depending upon the instrument manufacturer, the instrument may have one or more in-line filters between the eluent reservoir and the pump head. Indeed there may also be filters between the pump and the autosampler.
In general these filters are used to trap particulate materials which have not been filtered out of the eluent or precipitated buffer salts in the instance where involatile buffers have been used.
The purge valve should operate without maintenance for many thousands of hours. The two common problems associated with purge valves are particulate matter accumulating from worn pump seals and particulate material in the mobile phase or by overenthusiastic closing of the valve.
If particulate matter blocks the Purge Valve Frit a significant (> 5 Bar) back pressure will be observed when the purge valve is open. In such cases the frit should be replaced.
The symptoms of a damaged purge valve are a pressure drop or leak from the valve.
High back-pressure - results in the pump working under greater resistance – usually due to a blockage in the system. High back pressure will lead to an increase in the need for maintenance: pump seals and pistons will need to be replaced more regularly and the lifetime of the pump will be reduced. There are a number of potential contributing factors to high back-pressure in an HPLC system and most are due to blockage in the fluid (hydraulic) path.
It is important to know (and perhaps record) the normal working back-pressure for the column, flow and mobile phase conditions you are using in order to correctly identify high system back-pressure. Some data systems are capable of recording the pump back pressure profile across a chromatographic run or period of time.
The pump may be investigated as the cause of blockage by isolating it from the rest of the system. This might be achieved by disconnecting from the autosampler or by opening the purge (priming) valve. Pressure higher than 1-2 bar (5-15 psi) after isolation indicates the blockage is within the pump unit. Figure 21 highlights the areas on the binary pump that are susceptible to blockage.
Figure 21. Possible sites of blockage in a binary HPLC Solvent Delivery System.
Partially blocked inlet filter or tubing from the solvent reservoir – soak the filter in 0.1N Nitric acid for 30 mins, rinse with water / methanol and replace. Alternatively replace the filter
Partial blockage in the inlet check valve (high back pressure will be accompanied by a noisy / cycling baseline) –remove a sonicate in 50:50 Methanol:Water for 15 minutes. Shake to ensure the ball is free within the valve. If required soak in 0.1N Nitric acid for 30 minutes rinse with water / methanol and replace. Alternatively replace the filter
Partial blockage in the outlet check valve (high back pressure will be accompanied by a noisy / cycling baseline) – remove a sonicate in 50:50 Methanol:Water for 15 minutes. Shake to ensure the ball is free within the valve. If required soak in 0.1N Nitric acid for 30 minutes rinse with water / methanol and replace. Alternatively replace the filter
Partially blocked purge valve frit – if possible change the purge valve filter or sonicate the frit in 50:50 Methanol:Water for 15 minutes. Alternatively replace the valve
Blockage in the damping unit / defective damper – flush (at least 10 ml) with 0.1N Nitric acid followed by isopropanol. Flush (at least 10 mL) with 50:50 Methanol:Water. Alternatively replace the damping unit (may require engineer intervention)
Variable back pressure - is usually the result of trapped air within one of the pump heads or a failing check valve (see also Figure 15 and the earlier section on Check Valves for more details).
As the mobile phase is de-pressurized within the pump head (during the fill stroke), any gas dissolved within the solvent will tend to ‘outgass’ and agglomerate to form a single air bubble within the pump head. As the viscosity of the air is low, it is not swept out of the pump head at the typical pressure and flow encountered in an HPLC pump and becomes trapped. Out-gassing is of particular importance in reversed phase gradient HPLC analysis as the solubility of air in the separate solvents is often higher than in the combined mobile phase. Out-gassing is worse with low pressure mixing systems (quaternary pumps) as the mixing is carried out at low pressure and prior to the pump heads, rather than after which is the case with high pressure mixing systems.
Air trapped in the pump will lead to a sinusoidal pressure / flow output as the compressibility of air within the pump head is significantly different to that of the mobile phase. As the air is compressed in the pump head there will be no, or much reduced, mobile phase flow and so system pressure and flow will significantly reduce as shown in Figure 22.
Figure 22. Sinusoidal flow / pressure output from an HPLC pump with as air bubble trapped in the pump head.
Trapped air can easily be removed from the system via the use of the purge valve on most modern HPLC systems. The valve is manually or automatically opened and the pump run at a high flow rate (5 – 10 ml/min. typical) which forces the air from the pump head. On older systems, or if this operation is unsuccessful, it may be necessary to loosen an outlet check valve fitting to remove all air bubbles whilst the pump is running under pressure. Typically the fitting is not removed, but loosened and then re-tightened after several seconds.
Ensure that well degassed mobile phase is used to prevent problems with out gassing in the pump and other system components.
It is possible for fittings on the low pressure side of the system to be loose enough to allow ingress of air, without presenting a leak. If out gassing is an on-going issue even after mobile phase degassing, ensure the viability of all connection on the low pressure side of the pump, especially those emerging from the on-line degasser where relevant.
When multiple solvents are mixed together at different proportions during gradient analysis, the measured flow rate can vary from the set flow rate due to the compressibility of the liquids involved, which even though small in absolute terms, can be significant. This flow rate accuracy issue can be compensated for using the built-in solvent compressibility compensation software which is found in most modern HPLC systems. Many of these systems will allow you to manually enter the actual liquid compressibility values for each solvent (pump channel) used. This can result in better baseline stability and less pump noise. Note how Water has a compressibility value of ~ 46, but a very common HPLC solvent such as Methanol has a value of 120.
Most pumps are pre-set with a compressibility value of '100'. A 50/50 mixture of Water: Methanol run isocratically would have a predicted compressibility value of 83 [(46 + 120) / 2 = 83)]. This is a best guess value as the best compressibility value for a mixture of liquids must be determined through actual experiments and one should choose the value which results in the lowest pump pressure ripple and/or noise.
Solvent Compressibility (10-6 per bar)
Ethyl Acetate 113
Carbon Tetrachloride 106
Typical Solvent Compressibility Values
The values shown opposite are approximate and assumed to be accurate at 20oC . Various grades/purity of solvent may have different compressibility values so please verify the values of your own solvents before use.
If the flow and pressure output remain variable after addressing all of the points above, then one might suspect the Pulse Damper unit of the HPLC system. If the damper diagram splits or suffers mechanical failure (depending upon instrument design) then it will no longer function to reduce ripple within the pump output. Consult your equipment supplier for further advice on problem isolation and potential component replacement.
Figure 23. Possible sites / components to consider when troubleshooting variable pump pressure / flow output.
Ensure all mobile phases are properly degassed and check all fittings on the low pressure side of the pump if pump purging and mobile phase degassing do not overcome the problem
Partial blockage in the inlet check valve (high back pressure will be accompanied by a noisy / cycling baseline) –remove a sonicate in 50:50 Methanol:Water for 15 minutes. Shake to ensure the ball is free within the valve. If required soak in 0.1N Nitric acid for 30 minutes rinse with water / methanol and replace. Alternatively replace the filter
Air trapped in the pump head. Prime the pump by opening the priming / purge valve and run at 5-10ml/min. for 30 seconds. Take care to reduce system flow before closing the purge valve
Mechanical failure of the pulse damper – refer to manufacturer
Solvent compressibility settings – especially during gradient operation. Ensure the correct compressibility settings are entered into the system software
Gradient Formation Issues (Irreproducible retention times or selectivity in gradient analysis)
Mixing solvents on-line using a gradient pump can sometimes cause problems in HPLC. The different approaches to on-line solvent mixing (high pressure mixing using a binary pump, or low pressure mixing using a quaternary pump), and problems with the hardware used can lead to variability in retention time, sometimes accompanied by spurious peak shapes. Crucially –when performing gradient analysis, the selectivity of the separation may be altered if the gradient composition is irreproducible between subsequent injections. If variability in retention time from injection to injection is encountered –the following potential causes should be investigated:
Binary Pump Systems
The flow rate from one or other of the pump heads is inaccurate (check using a flow meter running one pump head at a time, at a flow of 1ml/min. using a calibrated electronic flow meter)
There is a blockage between one of the pump heads and the solvent mixing device –check as above
Occasionally the mixing units can block or fail. Symptoms are usually a high pressure, a noisy baseline or a combination of both.
Figure 24. Potential causes of issues with gradient formation in a binary solvent delivery system.
Blockages in the inlet filter may cause cavitation on the solvent lines leading to the pump system. This may cause irreproducible flows from either of the pump heads
Retention times may vary and the selectivity of gradient analyses may alter from one injection to the next
Problems with the inlet or outlet check valves of each of the pumps heads in a binary system can lead to problems with the flow rate – and lead to inaccurate and/or irreproducible retention times due to poor gradient composition reproducibility
The flow controller may be causing problems –check the flow from each pump head using an electronic flow meter
Blockages between either of the pumps and the gradient mixing device will lead to an irreproducible gradient composition
This might be investigated by disconnecting the tubing and measuring the flow output directly from each pump head
Blockages in the pulse damping device or any subsequent mixers may lead to irreproducible flow rates –however they are unlikely to give rise to irreproducible gradient composition
Gradient analyses are likely to retain the same selectivity even in retention times changes
Quaternary Pump Systems
The gradient former or active inlet is not delivering an accurate mixture of solvents. This can be investigated by setting a 50:50 gradient composition and lifting (momentarily) the eluent lines out of the solvents. If the air which you introduce chases through the tubing at approximately the same rate – then the composition of the solvent should be fairly accurate (don’t forget to purge the air before starting your analysis!)
For more accurate diagnosis, measurement of UV absorbance of a 50:50 mix of Water : 0.1% Acetone (aq) should give a constant UV absorbance at 265nm (see Gradient Performance Checks) for more details
Occasionally the mixing units can block or fail. Symptoms are usually a high pressure, a noisy baseline or a combination of both.
Figure 25. Potential causes of issues with gradient formation in a quaternary solvent delivery system.
Most low-pressure mixing valves in quaternary pump systems use solenoid valves to proportion the solvents in the on-line mix
If these valves fail or are sticking – retention times will be irreproducible and gradient analyses may show changes in selectivity
Check using an air bubble introduced into the solvent line or by using an UV eluent at a fixed wavelength (see Gradient Performance Checks for details)
Gradient Performance Checks - Low Pressure Mixing Systems
There are series of simple tests which can be carried out on both low and high pressure mixing systems which help to accurately diagnose the system performance in terms of gradient formation and delivery.
For low pressure systems (quaternary pumps) channels A and B are placed in HPLC water and channels C and D in HPLC grade water doped with 0.1% acetone. Detection is via UV at 265nm. Stepped gradients are run from 50:50 A-B to 10:90 A-C and all other valve combinations – see Figure 25. Eluent flow rate is set to 1 ml/min. The column is replaced with a back pressure restrictor coil (1m x 0.12mm i.d. tubing is satisfactory)
Figure 26 shows the results of a typical low pressure mixing (gradient proportioning valve) test.
Figure 26. Results from a low pressure gradient proportioning valve test (valve / solvent combinations as shown), gradient step time 2 mins.
Plateau Response (mAU)
The allowable limit of variation in response between the maximum and minimum plateau values and the response average is 5%. In the case above the variation is large (12.4%) and indicates that there may be an issue with the proportioning valve or the inlet reservoir filter of one or more solvent line.
In the case above, the active valve ‘A’ was replaced and the test was passed successfully.
Gradient Performance Checks - High and Low Pressure Mixing Systems
Further high and low pressure mixing system checks are also required in which channel A is water and B is water doped with 0.1% acetone. All other conditions are the same as the low pressure mixing test above. The following tests are carried out:
A series of 4 min steps is run in 10% increments (0% B, 10% B, 20% B… 100% B) including extra steps at 45% B and 55% B
A 15 min linear gradient of 0–100% B
Figure 27 shows the results of typical gradient.
Figure 27. Results from gradient performance checks (see text for details): a) 10% step check b) 1% step check c) linearity check
Figure 27a looks fine and the responses of each step compare within 1% to the theoretical response change – so this test passes. However, the linear gradient response in Figure 26c clearly shows an issue with linearity at the mobile phase compositions highlighted with an arrow. Closer inspection of the middle problematic region (Figure 27b) shows a step between 50 and 51% B where the response (and therefore the gradient composition) changes by around half as much as it should – clearly indicating a proportioning issue.
Performance characteristics of the HPLC pump output can be used to monitor and diagnose faults with the system. Monitoring the pressure, the flow rate stability and even the sound of the pump will alert the user to potential problems.
Computer control has improved the level of feedback available, and modern HPLC pumps can often alert the user of potential problems in advance of any failure. To minimise the system down time due to pump failure a calibration and maintenance regime should be employed. Certain components within the HPLC pump (such as pistons, seals, inlet filters, purge valve filters, check valve filters/ frits, check valve cartridges, active (solenoid)valves) should be replaced according to a schedule to ensure maintenance is pro-active rather than reactive post failure. A typical HPLC preventative maintenance schedule is shown below.
Clean in-line filters (soak in methanol and or 0.1N nitric acid)
Replace pump seals
Every 6 months to 1 year depending upon use
Check Valve filters (if fitted)
3-6 months depending upon use
Purge valve or in-line filter
Monthly or as required depending upon use
Performance Verification (PV) is commonly used to calibrate and ‘assure’ an HPLC system. Depending on the working environment PV may be required by external regulators. In such situations high levels of traceability of the calibration, including time and date verification, will be required.
In externally regulated environments such as pharmaceutical manufacturing or environmental analysis there may be clear guidelines to follow regarding the calibration and testing of HPLC pumps and the information on calibration / verification test details should be sought from the appropriate regulatory body.
As a member of CHROMacademy, you will also get access to LCGC magazine articles from your favourite authors - John Dolan, John Hinshaw, Mike Balough, and Ron Majors; over 300 news items refreshed daily; vendor application notes, webcasts and podcasts; electronic laboratory tools and calculators; and access to the growing Essential Guides archive, recorded tutorials by industry experts.
Subscribe now for $399
and get instant access to all CHROMacademy Essential Guide Webcasts
The following subjects are covered in CHROMacademy.com
The Theory Of HPLC
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
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)