TROUBLESHOOTING

• FLOWCELL LEAKS
• HIGH PRESSURE
• LOW PRESSURE
• PRESSURE FLUCTUATIONS
• HIGH BACKGROUND
• LOW BACKGROUND
• REGULAR BASELINE NOISE
• IRREGULAR BASELINE NOISE
• BASELINE SPIKES
• BASELINE DRIFT
• SMALL PEAKS
• NO PEAKS
• CONTAMINANT PEAKS
• LATE ELUTERS
• SPLIT PEAKS
• BROAD PEAKS
• RETENTION TIME CHANGES

FLOWCELL LEAKS

A small amount of salts accumulating around the juncture between the auxiliary and working electrodes is normal. But any visible liquid or dripping should be dealt with. The most common reasons for this are:

• Torn or scratched gasket. Replace.

• Scratched surface on working electrode or auxiliary electrode. Polish or replace.

• Improperly assembled flowcell. Pay particular attention that all parts are mating properly and are in the correct order, and that the two alignment pins on the auxiliary electrode DO NOT sit in the same holes as the two on the backing plate. Review assembly instructions.

• Excessive pressure in the waste line. This could be caused by a combination of clogging, crushing, too narrow a bore, too long a length, and a high flow rate. We suggest using no more than 12" of 0.010" ID tubing. Try switching to 0.020" ID tubing.

HIGH PRESSURE

Pressures as high as 4000 psi are not necessarily a problem. However, a rise in pressure above what is typical for a given column and system may indicate a problem such as clogged tubing or an occluded column. If a high-pressure problem develops, look at the following:

• Is the injection valve fully turned to the 'inject' or 'load' position? In between these positions the passages may be partially to fully blocked.

• Is the column, precolumn, or in-line filter clogged? Disconnect these sequentially from the detector towards the pump (allow pressure to drop to zero before disconnecting the column). Run the pump and note the pressure after each removal. A return to normal pressure indicates that the clogged component has just been removed (knowledge of typical pressures for each component is needed for this troubleshooting.) The clogged component should be cleaned or replaced as necessary.

• Is the tubing clogged? If there is excessive pressure with the column, precolumn, and in-line filter removed, there may be a clog in the lines. Begin removing tubing and work back towards the pump until the section with high pressure is identified.

• Has the mobile phase changed? In particular, methanol will give much higher pressures than acetonitrile.

• Is there something wrong with the column? Check with the manufacturer of the column, or compare with a known good column. The type of packing may have changed, or certain lots may have higher pressures than others.

• Does the pressure gauge work? This is harder to check, but worth keeping in mind.

• Has the column temperature gone down? The lower the temperature, the higher the column backpressure.

LOW PRESSURE

As with high pressure, low pressure is only a problem if it is unusual for a given system and column. A drop in pressure can indicate a problem with the pump, or a massive leak in the system. Examine the following items if low pressure develops:

• Are there any obvious leaks? Look for dripping. Salt accumulations indicate small leaks that should be corrected, but they are unlikely to cause low pressure.

• Is the flow rate within specifications? Measure the flow from the column with a graduated cylinder. If it is lower than expected there might be a large leak, the pump motor or gearbox may need repair, or a check valve might not be seating properly. Although rare, a piston may be broken.

• Are the check valves working? Low pressure, especially in conjunction with pressure fluctuations, can be caused by check valves that do not seal properly because of an air bubble or particle. Check valves can be sonicated or replaced to diagnose and correct the problem.

• Is there a problem with the column? Substitute a known good column, or contact the manufacturer to inquire about problems with the lot.

PRESSURE FLUCTUATION

Pressure fluctuations are a consequence of the alternating pulsations of the piston pump. They are minimized by good system design, such as the size and stroke length of pistons and the use of a pulse damper. Leaks and air bubbles are the major causes of excessive pressure fluctuation. Check the following items:

• Is a pulse damper being used? Sensitive gains require pulse-dampened flow.

• Are any fittings leaking, even slightly? Leaks allow pressure to drop as fluid is forced out in response to each piston stroke.

• Is there air in the system? Air can lodge almost anywhere, but check valves and pulse damper are likely locations. Purge the system at a high flow rate with freshly degassed, warmed mobile phase. If air bubbles recur consider their source: inadequate degassing of the mobile phase, restricted flow in the mobile-phase intake line, a loose connection in the intake line that allows air to be sucked in, or on-line mixing of aqueous and organic phases.

• Are the check valves working properly? Pressure drops that occur during the stroke of only one piston of a dual-piston pump suggest an improperly seating check valve. The source of the problem could be either the inlet check valve on the side with the pressure drop, or the outlet check valve on the opposite side. The check valves could be damaged or merely dirty. Soncicate the suspected check valves to clean them. If this doesn't help, replace them with known good ones.

• Are the seals leaking? Pressure drops that occur during the pressure stroke of only one piston of a dual-piston pump suggest a leaking plunger seal. Replace and break in (if necessary) in accordance with the manufacturer's directions.

• Are the pistons scratched? Scratches where the piston engages the seal will cause leaks. An indicator of scratched pistons is the rapid return of leaks after seal replacement. Replace any scratched pistons.

• Is the pump itself malfunctioning? Worn or broken gear teeth, cams, bearings, etc. could cause pressure fluctuations.

HIGH BACKGROUND

The background reading represents the electrochemical activity of the mobile phase, including all its components and contaminants. It is measured as the detector output when all offsets (zeroing) have been turned off. Every combination of mobile phase, electrode material, and applied potential has a characteristic background. The typical background current for your application is of diagnostic value — it should be noted each time the system is used and when mobile phase or hardware is changed. Backgrounds increase with temperature, applied potential, and contamination, and can decrease if the working electrode becomes coated or the reference electrode becomes depleted.

Many chromatographers don't understand the importance of a low background. They think that rezeroing gives them a zero background. But zeroing simply moves the frame of focus — the high background is still there, and it contributes to noise and distortion. This is true for both electrochemical and optical detectors.

Whether it be UV or EC detection, a high background makes it difficult to distinguish small peaks, as they represent only a tiny fraction of the total signal. When the background is low, however, these same small peaks are easier to distinguish because they represent a greater proportion of the total signal. High backgrounds can cause increased baseline noise, and are an indication that something is wrong with the system. Check the following if high backgrounds become a problem:

• Is the working electrode dirty? Electroactive material may have built up on the electrode. Try wiping it with methanol or acetonitrile, and if this doesn't work, polish it.

• Is the mobile phase old? Contaminants or microbial metabolites may be accumulating. Try a fresh batch of mobile phase. If this doesn't help, clean the system with solvents and try again with fresh mobile phase.

• Is the mobile phase new? If the problem began with a change of mobile phase, the mobile phase may have been improperly made, or a contaminant could have been introduced. Does the deionized water have > 15 MOhms resistivity? Was a different bottle of one component used? Was only clean glassware used?

• Is the column dirty? Try bypassing it. If the background goes down, the column should be cleaned.

• Is the mobile-phase uptake frit dirty? Remove it to see if the background goes down.

• Is Fe²⁺ being oxidized to Fe³⁺at the electrode? Try adding a metal chelator (e.g., 1 mM ethylenediaminetetraacetic acid) to the mobile phase.

• Is the potential set correctly? Higher potentials will produce higher backgrounds.

• Is the working electrode damaged? Try substituting a known good electrode.

LOW BACKGROUND

Low backgrounds may seem desirable, but atypically low ones may indicate a problem. Be suspicious of a low background, especially when coupled with an unnaturally quiet baseline and small peaks. Examine the following items if backgrounds are low:

• Is the mobile phase new? The concentration of buffer may be too low.

• Is the working electrode coated with something? Try wiping it with methanol or acetonitrile, and if this doesn't help, try polishing.

• Is the reference electrode depleted? Substitute a known good one.

• Is the potential set correctly? Lower potentials give lower backgrounds.

REGULAR BASELINE NOISE

Regular baseline noise (fluctuation) has a constant period that usually can be traced to a system component, especially if the period matches that of the pump. Air bubbles (and leaks) will produce baseline noise as they compress (or leak) at each pump stroke. To test whether the noise is flow related, change the pump speed. The period of the baseline noise should change proportionately. Consider the following items when tracking down baseline noise:

• Is there air in the flowcell? Purge the air as follows: turn off the detector, remove the reference electrode, allow fresh mobile phase to fill up the reference well, then reassemble.

• Is there enough backpressure on the flowcell? Backpressure reduces outgassing in the cell. Try using two feet of 0.010" tubing as the exit line from the cell. Alternatively, use a commercial backpressure regulator that provides 100 psi.

• Is the flowcell leaking? Clean and dry the mating surfaces, and use a fresh cell gasket.

• Is the reference electrode cracked? Replace it.

• Is there fluid leaking around the reference electrode? Replace the o-ring.

• Is there air in the check valves, pulse damper, or tubing? Purge at high flow rate with freshly degassed warm mobile phase.

• Are you mixing buffer and organic solvent on line? Incomplete mixing can cause a regular noise pattern. Try mixing and filtering the mobile phase before putting it on the LC system.

• Are there excessive pressure fluctuations? See this section.

• Is there a ground loop? Ground loops can cause regular or irregular noise. Check that the EC detector is properly grounded.

• Is the period very long (many minutes or hours)? See the section on baseline drift, especially temperature fluctuations.

IRREGULAR BASELINE NOISE

Irregular baseline noise (fluctuation) can be difficult to track down. Sources can be internal or external to the LC system. Consider the following items:

• Is there a ground loop? Ground loops can cause regular or irregular noise. Check that the EC detector is properly grounded.

• Is there electrical interference from other equipment? Try running a heavy-duty extension cord from another location. If the baseline improves, consider a dedicated power line for the LCEC system.

• Is radio interference from pagers causing intermittent problems? Consider shielding the detector, or moving the system to another location.

• Does the baseline change if you touch or stand in front of the system? Check the grounding, and make sure the flowcell is within an enclosure that acts as a Faraday cage.

• Is the noise associated with high background? Solving the background problem may reduce the noise.

• Is the column dirty? Continual elution of small peaks may look like noise. Clean the column or substitute a known good one.

• Are the electrodes damaged? Substitute known good reference and working electrodes.

• Is the period very long (many minutes or hours)? See the section on baseline drift, especially temperature fluctuations.

BASELINE SPIKES

Spikes are fast deflections in the baseline, either positive, negative, or both. They usually occur at irregular intervals, but depending on their source can be regular as well. Evaluate the following possible causes:

• Is there a ground loop, or no ground? Ground loops can cause regular or irregular noise. Check that the EC detector is properly grounded.

• Are all electrical connections clean and tight? Check cell leads, ground wire, and connections to chart recorder or data system.

• Is there electrical interference from other equipment? Try running a heavy-duty extension cord from another location. If the baseline improves, consider a dedicated power line for the LCEC system.

• Is radio interference from pagers causing intermittent problems? Consider shielding the detector, or moving the system to another location.

• Does the baseline change if you touch or stand in front of the system? Check the grounding, and make sure the flowcell is within an enclosure that acts as a Faraday cage.

• Are there bubbles passing through the detector? Submerge the exit line from the cell in a beaker of water and watch for bubbles emerging from the end. If found, track down the source and correct it. Sources include inadequate degassing of the mobile phase, restricted flow in the mobile-phase intake line, a loose connection in the intake line that allows air to be sucked in, or on-line mixing of aqueous and organic phases.

BASELINE DRIFT

Baseline drift is a change in background over a long period of time, usually hours. A decrease in background after initially turning on the cell is normal, as the detector equilibrates. For other causes, consider the following items:

• Has the composition of the mobile phase been changed? Be sure to purge the system to remove the old mobile phase. The column will need time to equilibrate with the new mobile phase.

• Does the baseline continually drift down? Are peak heights also decreasing? The working electrode may be becoming coated as the mobile phase passes over it. The coating can be removed by wiping the electrode with methanol or acetonitrile at regular intervals. Higher-quality mobile phase components, or better sample pretreatment, could eliminate the problem.

• Does the baseline drift up or down after a new bottle of mobile phase is put on the system? The temperature of the fluid could be changing.

• Does the baseline continually drift up? This could signify a buildup of contaminants from the mobile phase or samples. Do not recycle the mobile phase into the solvent reservoir while injecting samples. Change the mobile phase every three days if it is acetate- or phosphate-based, as these buffers are good growth media for bacteria. The electrode can be accumulating an electroactive coating — this can be removed by wiping with methanol or acetonitrile at regular intervals.

• Does the baseline drift up and down over a period of hours? This could be a response to the cycling of room temperature. Try operating the flowcell 5 °C above ambient temperature to isolate it from fluctuations.

• Does the baseline increase during the day and decrease at night? This pattern, or its inverse, could be a response do different day and night temperatures in the building. Try operating the flowcell 5 °C above ambient temperature to isolate it from fluctuations.

• Does the baseline drift up during gradient runs? A certain amount of drift is unavoidable, but a larger amount may be caused by old or impure components of the mobile phase. Try using the best grades available (at least AR grade, but preferably HPLC grade).

• Does the baseline drift up only during the first gradient of the day? Impurities in the initial mobile phase may be collecting on the column during overnight recycling. Always run the gradient once to clean the column before injecting samples.

SMALL PEAKS

If peaks have normal shape and retention times but are smaller than expected, check the following:

• Have the standards been made/diluted properly? Have they degraded?

• Has the mobile phase been changed? Was it made properly?

• Is the chart recorder or data system set properly?

• Are the detector range and potential set properly?

• Is the working electrode coated? Try wiping with methanol or acetonitrile, or polishing.

• Is the reference electrode depleted? Try substituting a known good electrode,

• Is the electrochemical detector functioning properly? Perform the self test if available for this detector.

NO PEAKS

Lack of peaks can be either a separation or a detection problem. Consider the following:

• Have the standards been made/diluted properly? Have they degraded?

• Is the detector on? Are the range and potential set properly?

• Is the detector responding to anything that is injected? A normal response to the injection, and normal baseline noise, suggest a separation or injection problem rather than a detector problem.

• Is the chart recorder or data system set properly?

• Is the working electrode coated? Try wiping with methanol or acetonitrile, or polishing.

• Is the reference electrode depleted? Try substituting a known good electrode,

• Is the electrochemical detector functioning properly? Perform the self test if available for this detector.

• Is there something wrong with the mobile phase? If the problem developed after the mobile phase was changed, try remaking it.

• Are the peaks eluting in the void? This is common when developing a new method. Compare the void produced by an injection of mobile phase with that produced by standards dissolved in mobile phase. A larger void in the latter case suggests the presence of unretained analytes. Try using less solvent in the mobile phase to increase their retention.

• Are the peaks eluting very late? Leave the chart recorder or data system running and go to lunch. Once late peaks are located, the solvent strength of the mobile phase can be increased to elute them earlier.

CONTAMINANT PEAKS

Contaminants are not necessarily a problem, unless they interfere with an analyte peak. If a contaminant does cause a problem, its source must be known before it can be dealt with. Common sources include the sample, chemicals used in extraction or purification, and mobile-phase components. Consider the following when tracking down a contaminant problem:

• Are the contaminants present in the system? Do they appear when the injector is thrown to the 'inject' position with only mobile phase in the loop? If so, try cleaning the column and the loop. The contaminants may have been introduced with the mobile phase. Consider the quality of each mobile-phase component. Ion-pair reagents in particular have been known to cause these problems. Try another source for these reagents if you suspect a problem.

• Are the contaminants in the syringe? If no peaks eluted when the injector was thrown to the 'inject' position, fill the loop with mobile phase using the syringe, then inject. If the contaminants appear now, the syringe may be dirty. Alternatively, the waste line of the injector may be clogged, allowing contaminants to wash back into the loop.

• Are the contaminants in the standards? Inject freshly diluted standards.

• Are the contaminants in the extraction procedure? Extract a blank sample. If contaminant peaks appear now, change the extraction components one by one until the problem is found. Old buffers are likely suspects. In catecholamine assays, alumina that is old or has not been stored in a desiccator may have adsorbed contaminants that elute with the analytes.

• Are the contaminants in the samples? Only a better cleanup or extraction procedure will solve this.

LATE ELUTERS

These are normal sample components that elute much later (even hours later) than the analytes of interest. Consider the following:

• Can the sample preparation scheme be modified to remove these? Solid-phase extraction may be useful here.

• Can the injection schedule be timed so that the late eluters from the previous sample elute in a clear area of the current sample? Timing is critical here.

• Can a certain number of injections be performed, followed by a quick wash of the column to remove the late eluters? Alternatively, can a wash be programmed after each injection?

SPLIT PEAKS

A split peak is one peak that appears to be two poorly separated peaks. Consider the following:

• Is the peak truly split, or are there really two compounds? If only one peak in the chromatogram appears to be split and the others appear normal, it is likely there really are two closely eluting peaks. Try injecting a fresh dilution of standards.

• Is there a poor connection at the top of the column? Try remaking this connection so there is no dead space.

• Has the column bed developed a void or channels at the top? If this occurs, some of the analyte travels through the "open" spaces, thus eluting slightly before the rest of the analyte that travels in the bed. All peaks in the sample will be split. Try reversing the column. A temporary improvement suggests that the column bed has developed problems. Replace the column.

• Is the injection fluid too strong (organic strength or pH)? Try injecting a standard dissolved in mobile phase. If this helps, adjust the solvent strength or pH of the injection fluid, or reduce its volume. If pH is a problem try increasing the buffering capacity of the mobile phase.

BROAD PEAKS

The lowest detection limit is achieved when an analyte is in as compact a band as possible within the flow stream. We see this compactness as a narrow peak on the chromatogram. In isocratic elution peaks broaden as retention time increases, so it's best to elute the analytes as early as possible, consistent with good separation. Unusually broad peaks suggest the following:

• Has the loop been changed? Larger injection volumes lead to more dispersion of the sample, hence broader peaks.

• Has new tubing been installed somewhere between the injector, column, and detector? tubing with an internal diameter larger than 0.010" can cause sample dispersion.

• Is the column losing active sites or developing voids? Compare to a known good column.

• Is the flowcell volume large compared to the volume of the peak? Dead volume in the flowcell causes peak dispersion. Microbore systems require lower flowcell volume than do standard-bore systems.

• Is the pH of the mobile phase near the pK_a of the analyte? This could produce two species of the analyte that differ in hydrophobicity, and hence in retention time.

RETENTION TIME CHANGES

Retention times can vary randomly, or steadily increase or decrease. Consider the following:

• Is the column thoroughly equilibrated with the mobile phase? Mobile phases containing ion-pair reagents require longer time (greater volume) to reach equilibrium.

• Is the pump malfunctioning? Check the flow rate several times to make sure it is consistent.

• Is the composition of the mobile phase changing? Evaporation of organic solvent will cause retention times to progressively increase. Compare with fresh mobile phase.

• Is the column losing bonded phase? Column deterioration will cause peaks to elute sooner. Compare to a known good column.

• Is ambient temperature fluctuating? Do day temperatures differ from night temperatures? Retention varies inversely with temperature. Keep the column in a thermostatted oven at least 5 °C above ambient.

• Is the sample being injected at the proper time? Injection and the start of data collection must be simultaneous.

• Is this a gradient separation? Slight differences in piston position at the start of a run can cause variations in gradient composition from run to run, hence retention times vary randomly. Use a pump that synchronizes piston position for the start of each run.

• CONTENTS