Everyone dealing with HPLC equipment knows that they have to cope with disturbances and defects. Some of these can be attributed to the equipment itself (e.g. pump disturbances, a malfunctioning thermostat, etc.) while other problems arise from less direct factors (e.g. blockages). These problems are not solely dependent on the HPLC‑technique, but can be of a more general character. 
HPLC equipment is diverse and high standards are set on the electronics, machinery, and materials. The lifetime of these parts is limited and, consequently, wear and tear can eventually cause instability or failure. 

Scope of the topic
In this Topic circle the most common problems occurring in the day to day operation of an HPLC system will be considered. We shall not deal with:

• Disturbances of an instrument‑specific nature and problems that need specialist attention, such as electronics problems.
• Defects that are not specific to HPLC (such as problems of a chemical nature like instability, oxidation or reduction sensitivity, acid‑base reactivity, etc.)

Any disturbance or defect/failure should be dealt with systematically.  If, after a brief check, the problem is not solved, then a systematic search should be carried out to pinpoint the cause of the disturbance. The entire disturbance‑sensitive area must be defined, after which it can be subdivided into different areas, each of which is related to one part or parts of the HPLC equipment. 

Each of these sub-areas should be checked as the possible source of the problem. In this way one, or possibly more than one cause may be indicated. Of course, every solution has to be checked as to whether it is actually the source of the problem. On the face of it, this technique may seem time‑consuming and roundabout, but it results in a problem‑solving way of thinking that, in time, becomes automatic and intuitive.

Before disassembling the instrument, a simple preliminary check should be made. When, within a laboratory, several people work with the same instruments and parts are being interchanged, it is useful to check on the power, connecting cables, voltages (especially in relation to data systems), etc.

No matter how trivial it may sound, to resolve a problem, one must detect it first.  If an error or malfunction is detected, one can then identify, analyse, and remedy the problem. Though many problems are fairly obvious, there are many more that are not so evident.  Of course, the latter are more common. A malfunction in the LC is generally recognized in the long run, but other errors and sources of error in an analysis may go undetected for long periods of time.  The development of the QC process within the lab is meant to catch even the latter, silent errors with its more comprehensive approach.

Errors are generally spoken of as either random errors or systematic errors:

1. Random errors are deviations of single measurements from the expected average value due to the expected statistical variation of results.
2. Systematic errors are more difficult to detect because it looks as if there is no error. Systematic errors are generally detected by analyzing samples or standards of known composition. 

Obviously, any inability to recognize disturbances in results is a serious problem. Thus, it is important to develop effective QA/QC strategies within a laboratory. Quality systems and manuals detailing the steps for analyzing disturbances can be of great use. When documenting errors and troubleshooting analyses, the analyst must work methodically, generally paying close attention to his actions, documenting the steps taken and subsequent results, and editing the list of possible causes frequently.  Once again, the whole process is simplified if the lab has a well-documented strategy for troubleshooting analyses. 

The following are generally key parts of such a strategy:

1. Visual inspection of the chromatogram. Every chromatogram should be verified. Retention values, integration marks, and the calculated baseline should all be inspected since peak integration and data processing are ready sources of error.   
2. Daily routine tests of the chromatographic system: The operating parameters/variables (flow,  eluent composition, background signal of the detector etc.) should be noted.  The proper functioning of “smaller” items and the proper condition of instrument control switches should also be verified. 
3. Daily tests on the system if it operates well: Test standards, test samples and test blanks provide information on the function of the chromatographic system.
1. Is the number of peaks in agreement with the expectation?  
2. Are all retention times as they should be?
3. What is the shape of each peak?
4. Is the baseline normal? 
5. Is the eluent consumption normal?
6. What is the noise level of the detector?
4. It is important to test everything influencing the quality of the analysis daily. Obviously, matters that do not influence anything need not be tested. In general, the end use of the analytical results dictates the severity of the QA/QC process.  
5. Weekly (or otherwise regular) tests and inspections: All less variable and/or critical analytical and instrumental matters should be tested regularly: syringes, filters, calibration standards, blanks etc.
6. Preventive maintenance: Changing septa, cleaning the injector, cleaning of the detector, etc. should be done rather than leaving these items until they actually result in a problem. Each chromatograph should have an associated plan indicating which parts require maintenance, as well as the maintenance interval. 
7. Annual maintenance: As with a car, a chromatograph requires serious maintenance on a periodic basis. This is best left to a maintenance contract or a competent service department.  In most cases the equipment supplier can provide such maintenance.

Any deviation leads rapidly to poor chromatographic results. Every HPLC‑user should therefore know how to localize and solve disturbances and sources of trouble. Without this knowledge a lot of time will be wasted in unsystematic and arbitrary searches for possible causes of the problem.

The golden rules of trouble shooting:

            1. Try it once again once more

            2. Install it again

            3. Throw it away

            4. Write down everything you do

            5. Remove the cause

To remove a source of disturbance does not always mean that one has located the cause of the disturbance. Disturbances can have a reason. When you do not find it, the problem will probably return and the disturbance has not been removed after all.

When the disturbance is supposed to arise from a certain part of the instrument, this part should be replaced with one that is known to be working properly, followed by a test of the system. If the change resolves the problem, remove the malfunctioning part at once. Do not keep it in the laboratory, running the risk that it may be used again by an unsuspecting operator.

Document!
It is important that all troubleshooting operations be well-documented, especially in the case of complex disturbances and problems. In such situations, troubleshooting and testing can extend over several days. If one does not record the results well, there is a risk of going around in circles, repeating measurements that have already been completed. 

Make a written record of each action in the troubleshooting process in which the symptoms, the tests, the possible solutions, the final solution and all relevant information are recorded. This type of information is also profitable for colleagues (exchanging experiences) and for the future. Nothing is more frustrating than remembering that the current problem has occurred before, but not remembering how it was resolved: keep in mind that problems may recur. Along these lines, remember that every piece of equipment has its own manual, and that the manual should specify the expected lifetime of the equipment.

An effective remedy at the analysis of complex problems is the Problem-Analysis-Scheme with which a certain problem can be wholly or partly elucidated by means of the symptoms viewed. Systematic approach of the symptoms structure the manner of working and result in a solution of the problem.

This remedy also requires some discipline and practice of the operator, because it looks initially academic and time-consuming. Practice proves the opposite.

Using the Problem-Analysis-Scheme is particularly useful for:

1. People without excellent practical experience
2. For complicated problems or disturbances and
3. When more people work at one problem.
4. It is particularly useful in the first stage of the process of solving problems when the symptoms are described and analysed.

The Problem-Analysis-Scheme consists of a table with four columns: symptoms, possible causes, possible solutions and counter-arguments.

Problem:

The scheme is filled in from left to right. When the first symptom is mentioned, the first line is filled in before completing the second symptom. Symptom for symptom is systematically completed in this way. You can stop when all symptoms have been mentioned and considered.

The most critical element when using this scheme is the definition of the problem. Observations should be mentioned clearly and without any form of interpretation or subjectivity. If this is not done there is a danger of limiting the prospective with the risk of arriving at a dead end. At that moment the use of such a scheme deteriorates and can be turned against the operator.

Any deviation in chromatographic behaviour that can be observed in the chromatogram could be due to changes in flow rate. Check whether the pump is operating at the correct flow rate and whether the output is constant.  Although the nominal flow rate reading given by the pump should be checked as well, it is necessary to measure the actual flow rate accurately (e.g. with the aid of a calibrated burette). This is most easily done after the detector.  It is rare to find too high a flow rate as long as the pump is adjusted correctly.  Too low a value can indicate blockages, leakage, or a malfunctioning pump.  Changes in the column (e.g.  pulverisation, sagging, etc.) can also result in changes in the flow rate.

In contrast to gas chromatography, leaks and blockages in HPLC can usually be detected fairly easily. Leaks become apparent if the operator notices the smell of solvent or observes a drip, while blockages will generally cause the pump to shut itself off automatically in order to prevent overpressure.  In the case of large leaks, eluent will leak out of the system at certain places. If a leak occurs in the instrument itself (e.g. detector), it may not be visible immediately, but if you look under the instrument there may be a pool of liquid. Large leaks are generally associated with a lower than usual pressure (though it may be constantly low).  In many cases droplets will not be visible as the liquid is only present as a very thin layer along the tubing and connections. To detect small leaks, follow the flow path with a finger or a piece of tissue. This can give an immediate indication of the source of the problem. Also, the pressure may not change significantly. Since volatile organic solvents are often used in HPLC, it is essential to check for leaks frequently.

A proper read‑out of the pressure in the system can prevent many potential problems. A pressure increase, for example, indicates blockages, injection of polluted samples, sagging of the column packing, etc.  Conversely, a leak-induced drop in pressure will also be detected, as well as the pressure fluctuations caused by a malfunctioning pump. In the event of a pump malfunction, the pump head responsible can even be singled out.  Pressure fluctuations during and after rotating the injection valve are an indication of blockages in the valve or the column. Strangely, many instruments are actually equipped with fairly poor pressure read-out facilities.   

In the event of a blockage (indicated by the pressure gauge), the first step (after turning off the pump and allowing the pressure to fall) is to disconnect the column from the injector and to turn the pump back on. If the pressure drop is virtually non‑existent, the blockage must be in the column or the detector. In the event that the pressure remains high, there might be a blockage somewhere between the pump and the column. Examine this section of the system by disconnecting all fittings, starting at the injection valve and working towards the pump. 

The flow through the injection valve should be checked both in the "load" and "inject" positions to ensure that all the ports of the valve and the sample loop are flushed. It is possible to clear blocked capillaries or sample loops with the HPLC‑pump or an ultrasonic bath.  However, to avoid a repetition of this particular problem, one should replace the capillary completely. Keep in mind that the capillary ends must be square and free of burrs.  A poor injection may be due to a leaking valve, but it is also often caused by an accidental interchange of the "load" and "inject" position. Valco valves are especially prone to this problem as the two positions are not marked on the valve.

The peak shape is often indicative of chromatographic problems.  If all the peaks are broadened in a chromatogram, especially the early eluting peaks, it is clear that the separation efficiency of the chromatographic system is diminishing.  As the total efficiency is determined by the entire chromatographic system, the cause of the excessive peak broadening has to be sought not only in the column, but also in the pre‑columns and the dead volume in the injector, detector, and connecting tubing.

Tailing peaks are the result of adsorption in the column or of dead volumes between the injector and detector. In the latter case, unretained components show peak tailing. 

Fronting is almost always the result of overloading the column. Too large an injection volume or injection of samples dissolved in a solvent stronger than the mobile phase can cause leading or broad peaks. Shoulders and double peaks are the result of dead volumes or small channels in the packing of the column. In the event of channels or dead volume, note that all peaks will show such shouldering. 
 
Negative peaks usually near the dead time t₀, are the result of refractive index phenomena which are caused by mixing the solvent sample with the mobile phase. Methanol/water mixtures are particularly prone to this effect depending on concentrations and the geometry of the detector cell.  When using a refractive index detector, negative peaks are quite common and are to be expected.
 
Nowadays, more and more analyses are carried out using a UV‑absorbing component in the mobile phase. This enables the detection of non‑UV‑adsorbing components (so-called indirect UV‑detection) in which a rise in UV‑transmission is registered. In fact negative peaks are recorded, but by reversing the polarity of the detector or the recorder, the signal is registered positively. In such a system, a UV‑absorbing component will register as a negative peak.
 
Undesired, interfering peaks can arise from several causes e.g.: 

• pulsating pump flow
• electronic disturbances (spikes) 
• late‑eluting components from previous injections
• gradient elution
• signals from an integrator/computer
• air bubbles

It is a well-known fact that any person using analytical equipment is a source of errors. These errors can be quite accidental, subject to motivation, self control, day, hour, the weather and the like, but they can also be systematic, such as a certain injection technique, measuring the peak areas by hand etc. When these errors are minor it does not matter, because they fall within the total variance of a measurement result. It is possible, however, that an acquired method can be so annoying that it cannot be tolerated. Improper operations and also wrong interpretations can lead to entirely wrong results. The usual method to track down errors or to eliminate them is to have the analysis done two or three times.

Systematic personal errors can be discovered when someone else is performing the analysis. This should not be regarded, although some people do, as a "motion of no-confidence", but as a method to find errors and disturbances which after elimination provides better results.

The management of a laboratory has the important task to see to it that the co-workers are sufficiently educated and motivated. There should be a personal development plan for each co-worker in which the skills and knowledge are mentioned (gained during internal and external courses and training) as well as the knowledge to be gained as yet.

It is essential to have knowledge and experience when dealing with trouble shooting. Experience can be acquired in time, but also as a result of conversations with the support of experts, like a maintenance man or a chromatography specialist.

If a first check did not come up with a solution for the problems, one should search further. The systematic of trouble shooting has already been discussed. Locating the areas in which the problem occurred is an essential part of it. Work from large to small, or in other words: first divide the entire analysis roughly into sub areas and then, if the problem has been located in a certain area, into smaller areas which have to be examined step by step.

Important aids are spare parts when the trouble spot has to be isolated. Replace in that case a component of the entire instrumentation by something of the same quality and watch its effect.

In order to shorten the time in which errors are located one might change temporarily, for example, a printer or a column. In addition there should of course be the necessary smaller devices such as syringes, fittings, tubing, filters, spare parts for the pump, reference column, test standards, etc.

Suitable standards are irreplaceable parts of trouble shooting. By standards we mean homogeneous and stable mixtures of components of which the compositions and qualities are known. Think of test samples, test standards, blanks, calibration standards etc. Quality systems dictate that an analytical method used, a gas chromatographic analysis and the equipment employed, materials and means should be tested to guarantee the reliability of the analysis results. The nature, the composition, the accuracy, the tenability and the use of standards should be known. Accurate documentation is a part of a quality system. Such standards are particularly useful in trouble shooting.

Test and calibration mixtures have a certain hierarchy with regard to accuracy and reliability. At the top are the certified test samples and the certified test solutions. The difference is the absence or presence of a certain matrix. These certified standards are primarily used to validate a certain total analytical method. In addition they can be used in ring tests between laboratories. Certified test mixtures are commercially available in a limited extent. For that reason when looking for method errors often use is made of laboratory test mixtures. These mixtures should also have a higher degree of guarantee with regard to the applicability and use.

Calibration standards and test standards consist in general in a solution of specific components in a properly chosen solvent. In case of diluted standards it is recommended to depart from a concentrated initial solution from which the diluted solutions are generated. Concentrated solutions are stable during a longer time, without the occurrence of deterioration. In diluted solutions the deterioration effect is much stronger and the standard is then useless. Diluted standards have a shorter storage life. Deterioration is primarily caused by three reasons: decomposition, evaporation and adsorption to for example glass walls. The stronger a solution is diluted, the more strongly these sort of effects occur.

In addition to deterioration the solution can also improve. Evaporation of the relative volatile solvent is than the main cause.

It is therefore recommended to depart from fresh solutions in both the diluted standards and the standards with volatile components and/or solvents.

Check in all cases the purity of the solvent used. A blank HPLC-analysis gives a definite answer.

Standards should be lightly shaken after they have been made and then stored coolly in a well-closed glassware.

Characterisation parameters of standards

• Name
• Identity
• Serial number and date of production
• Purity
• Concentration
• Homogeneity
• Stability
• Other relevant characteristics and particulars.

In addition to calibration standards there should also be test standards present in the laboratory: measuring solutions to test the equipment and to judge functioning.

The following test standards are recommended:

1. test mixtures to determine the instrumental detection borders. This value can be tested by means of the signal/noise ratio. See to it that this is carried out under the same circumstances as with the manufacturer: column temperature, gas velocity, injection technique and the like.  In gas chromatography the signal/noise ratio is determined dynamically or statically. In the first instance the signal/noise ratio is measured by means of the chromatogram obtained. In the second case there is a correction for the chromatographic band broadening. The manufacturer can give the answer which method should be used.
2. test mixtures to check the status of the analytical column: efficiency, inertness and separating power. This can be a test mixture which is given by the column manufacturer; it can also be an test mixture made. A good test mixture contains a number of inert components and critical components: whether or not with inert (polar) components that are difficult to chromatograph. It is obvious that the critical components have a certain relation with the relevant analysis. Also the concentration in which the components are present in the test mixture should correspond with the concentration reach of the relevant analysis. It is not inconceivable that components give symmetrical peaks in a higher concentration level, while they give distorted peaks (tailing) at lower concentrations.

Problems in relation to a quantitative condition often have a difficult nature. It already starts with signalling the problem: the analysis looks all right, only the measuring values are not correct. The latter cannot be observed from the measuring values.
The approach of quantitative problems requires a logical, structured and systematic approach, in particular with the first part of the trouble shooting process: the identification of the problem. The flow diagram mentioned before is therefore an excellent help.
It is a good thing to make a distinction between accidental (static) errors and systematic errors. Accidental errors can always be exposed by repetition which is not the case with regard to systematic errors. These are connected with an incorrect (repeatable) action, equipment, a laboratory or a method. The degree of difficulty is then extended. When a certain analyst provides systematically higher measuring values that is a personal systematic error which can simply be discovered.
 
That is more difficult in a systematic error in a method. Even with a ring test such a failure need not be exposed; if all laboratories participating operate according to the same method, a systematic error remains hidden. Even when using different methods, the interpretation of the deviations are not simple to detect. When eight laboratories provide a comparing measuring value which deviates from the measuring values of two other laboratories, we wrongly assume to point an accusing finger at those two labs. Systematic method failures can only be located by means of certified reference materials. If the analytical values of such a standard deviate repeatedly and significantly from the actual value, we have to deal with a systematic method deviation. A ring test where various methods are evaluated by a certified standard, usually provide data on the value of every analytical method.
 
Below in this section we do not deal with the general failure determination within an analytical method; we restrict ourselves to the quantitative part of a gas chromatographic analysis.
 
 
Poor reproducibility
 
If a quantitative analysis gives problems regarding a poor reproducibility, one of the first matters which should be investigated whether the deviating results have a cause which is subject to trend or only a large spreading where the average of a large number of observations is correct. In the latter case the error is most probably with the injection: a poorly functioning autosampler, a poor syringe or a poor controllable manual injection, whether or not in combination with a certain type of sample. For these kinds of problems, we refer to the section "Syringes and injection".
A deviation subject to trend is the category "sample chemistry". In all cases one should bear in mind that the problem is not caused by the injector or the integration. This should be checked in advance. Integration-problems occur in particular with peaks being asymmetric (tailing) or when peaks are separated inadequately. Starting point of a peak, final point and the baseline should be chosen in such a way that it cannot cause any problems. If it does cause problems, alter the integrator parameters. When in doubt, a manual peak height or peak area measurement is recommended.
 
When a deviation coincides with a large spreading, a test with a chemical inert standard would be useful. Good reproducible results can indicate a poor wetting of the syringe which occurs at highly viscous liquids. In case it enters the matrix of the sample, the problem is solved. Of course the standard peak should give a proper separation with the sample components and with the matrix peaks.
 
 
Traces of contamination in the solvent can also give problems: they can interact chemically with the sample-components. To test this one can try to use several solvents: solvents of various manufacturers and with a different purity.
Matrix-components that have been deposited or condensed in the injector or in the first part of the column, can form active places to which sample-components can adsorb or decompose. Regular maintenance or replacement of the injector insert solves the problem on the injector side. As to the column a retention gap is recommended which in its turn should be regularly maintained or replaced.
 
An increase in the peak surface can be caused by the evaporation of the (volatile) solvent from the sample during the storage. An internal standard which is added in time, compensates this effect.
In trace analyses an increase in the peaks are possibly the result of self-activation of the analytical system during the course of analysis series. Active places are covered by sample-components until there is a balance. A pre-injection of a relatively highly concentrated sample will solve the problem. The adsorption balance has already been created prior to the analysis series.
 
Most problems around "sample chemistry" occur at trace analyses, thus in strongly diluted samples (ppb-reach). In addition to poor reproducibility of the peak signal, sample chemistry can also be viewed by peak distortion, especially tailing.
 
Injection
 
Systematic errors can always occur with any injection technique and sample preparation. This is often attributed to various matrix effects which could provide differences between the "dirty" sample and the "clean" standards. Addition of a known amount of the components to be determined to the sample and the analysis, can give information about the effect of the matrix. See to it that the amount added does not tamper with the sample.
Consider that a so-called spiked solution does not need to be the same as the real sample. The matrix can have a different effect on the spiked components and the components present.
 
Calibration
 
Co-elution of contamination from a standard is a very important reason for a systematic error. Certainly with multi-component standards one should be aware of this. The gas chromatographic purity of the components and the solvent should be known in advance to exclude this problem. The analysis of all compounds in question under the measuring circumstances give the desired information.
 
Another important source of systematic errors is the extrapolation of calibration values to a higher or lower measuring range. It is assumed here that the measuring range is completely linear. Practice teaches that this is a wrong assumption. This is known in the high measuring range but not in the low measuring range. In multi-calibration the measuring range is within the measuring points, unless the linearity is verified outside this region.
One-point calibration is the major culprit. One supposes a completely linearity range in which the calibration line is also going through the origin. Practice proved that one should take this assumption very carefully.

A multi-point calibration of at least five measuring points on a large number of decades is preferred. If it is established that the calibration line is not linear (e.g. in the low concentration range or in the use of an Electro Chemical Detector) another calibration in the linear part is inevitable.
See to it that the injection of the calibration standards takes place in an arbitrary order. If the standards are injected each time in a certain order (e.g. from low concentration to high concentration or the reverse) the problems in the region of "sample chemistry" or cross-contamination can be masked. Random injection with blanks in between prevents this type of systematic problems.
 
Finally a general remark: 

If nothing works, read the manual!! 

In many cases this document contain valuable information how to run and maintain the instrument and how to trace and solve problems with the system and the HPLC application.