816 TROUBLESHOOTING and outlet check valves seldom are interchangeable, and if they are not clearly marked, an identifying mark should be scribed on the check-valve body to indicate position and flow direction. Some check valves will come apart when inverted, so it is wise to check for this prior to sonication, or you may be surprised to find small parts in the sonicator after a cleaning attempt. We recommend placing each check valve in a separate beaker with enough methanol or isopropanol to cover the check valve; then sonicate for 5 to 10 minutes. In our experience, this will fix leaky check valves most of the time, presumably by removing unwanted contaminants from the ball and/or seat of the seal. If a check-valve comes apart, clean the parts in alcohol, then carefully reassemble them using forceps; avoid contacting the internal valve parts with paper, cloth, or fingers, because a small bit of fiber or a fingerprint can cause the valve to leak. See Section 17.4.2.2 for additional discussion on the possible causes of and corrections for check-valve sticking. 17.2.5.5 Leak Detection Mobile-phase leaks may be obvious—or not. Drips, puddles, and leak alarms usually make location of the leak simple. When buffers are used, a white, crystalline deposit may show up on a slowly leaking fitting—even where no liquid is obvious. A simple leak detector for hard-to-find leaks can be made from a piece of thermal-printer paper (e.g., a credit-card receipt). Cut a narrow, pointed strip of thermal paper (e.g., 1 × 5 cm, pointed at one end) and use the pointed end to probe suspected fittings, seals, or other possible leak sources. The paper will turn black when it contacts organic solvent; this can be useful for locating leaks that are hard to detect by other means. 17.2.5.6 Repairing Fitting Leaks Correcting a leaky fitting may be as simple as tightening the fitting one-quarter turn to see if a leak can be stopped. If this does not fix the problem, disassemble the fitting, rinse it, and reassemble it or replace the ferrule with a new one. Do not overtighten a stainless steel fitting, because the internal parts can distort enough that the ferrule will ‘‘mushroom’’ out beyond the fitting threads, making the connection impossible to disassemble. For stubborn fitting-leaks, PEEK (poly-ether-ether-ketone) ferrules often are superior to stainless steel because they deform sufficiently to seal with otherwise imperfect surfaces. When a leak is encountered with PEEK fittings and tubing, it is best to shut off the mobile phase flow, loosen the nut, reseat the tubing in the fitting body, and re-tighten the nut. Sometimes when a PEEK fitting is tightened with the flow on, the tube end can slip in the fitting, creating a small cavity at the tip of the tube, which in turn can cause unwanted peak broadening (extra-column effects, Section 17.4.5.3). 17.2.5.7 Cleaning Glassware Organic residues on ‘‘clean’’ glassware can be the source of ghost peaks or baseline drift in blank gradients, so a thorough cleaning of the glassware is essential. Various techniques to clean glassware have been recommended [6]. Extra rinsing (10 rinses with tap water followed by 10 rinses with deionized water) of glassware that had been washed with laboratory dishwashing detergent was found to be satisfactory 17.2 PREVENTION OF PROBLEMS 817 in one study [7]. Other workers [8] avoid detergents altogether, preferring to rinse glassware (including reservoirs, pipettes, and graduated cylinders) used only for HPLC mobile phases with water and then a clean organic solvent [6]. To avoid inadvertent contamination of the mobile phase by glassware, use only the cleanest possible glassware and take care that the cleaning process does not add contaminants to the glass surfaces. 17.2.5.8 For Best Results with TFA Trifluoroacetic acid (TFA) is a widely used additive for gradient-elution mobile phases. TFA is readily miscible with both water and organic solvents, provides a low-pH mobile phase (0.1% TFA ≈ pH-1.9), acts as an ion-pairing reagent with biomolecules (Section 13.4.1.2), can be used at wavelengths of <220 nm, and is sufficiently volatile to use with mass spectrometric or evaporative light-scattering detectors. TFA is available in a highly purified form suitable for HPLC use, but it degrades rapidly upon exposure to air. For best results purchase HPLC-grade TFA (or equivalent spectral grade) in 1-mL ampoules and use the entire ampoule at one time. TFA is available in larger containers (e.g., 25 mL) at a much lower cost per mL, but—in the experience of one of the authors—it is impossible to prevent rapid degradation of the reagent once the bottle is opened, even when working in an inert atmosphere and carefully resealing the bottle. However, once mixed with water, resulting TFA solutions are fairly stable (e.g., 1 week). With UV detection at <220 nm and gradient elution, some drift with TFA-acetonitrile mobile phases may be observed, because the absorbance of TFA depends on the %-ACN in the mobile phase [9]. To minimize baseline drift, add the same amount of TFA (e.g., 0.1%) to both the A- and B-solvents and work at 215 nm, where the absorbance is fairly constant for TFA/ACN mixtures [9]. An alternative to preparing TFA-containing solvents in the laboratory is to buy HPLC-grade solvents with TFA (or other additives) already added from one of the HPLC-grade solvent suppliers (e.g., Burdick and Jackson, J.T. Baker). 17.2.5.9 Improved Water Purity In the examples of Section 17.4.5.2 and accompanying Figures 17.11 and 17.12, the problem of mobile-phase contamination by impure water or additives is discussed. It is important to use the highest quality reagents in order to avoid unnecessary background peaks. This usually means purchasing HPLC-grade reagents for all salts, buffers, and organic solvents. Most laboratories prepare their own HPLC-grade water with a water purification system, such as the Milli-Q system (Millipore). Such water purifiers combine physical filtration (≤0.2 μm), ion exchange, and adsorption on carbon to remove organic contaminants. Ultraviolet photo oxidation may also be carried out in order to kill bacteria and oxidize organic contaminants [6]. A further cleanup of the water may be required for maximal removal of back- ground peaks, for example, for stability-indicating pharmaceutical assays—where peaks with areas ≥0.05% of the parent-peak area must be quantified. Passing the water through a C 18 HPLC column works [10], but this is inconvenient. For high-pressure-mixing systems, another option is to install a C 18 guard column between the A-pump and the mixer [10–12]. This ‘‘scrubber’’ column will trap organic materials before they reach the mixer and prevent them from entering the 818 TROUBLESHOOTING analytical column and producing background peaks. Other devices have been sug- gested for a final polishing of water prior to use. These include passing the water through a C 18 SPE cartridge prior to use, or using a low-back-pressure in-line filter before the mixer (low-pressure mixing) or A-pump (high-pressure mixing) [6, 13]. 17.2.5.10 Isolating Carryover Problems ‘‘Carryover’’ describes the repeated appearance of a peak in later chromatograms, when the sample is only injected in an initial run. That is, remnants of the sample remain after the first run and are somehow introduced into one or more subsequent blank runs. There are three main types of carryover: • volumetric carryover • adsorptive carryover • incomplete elution Each of these carryover problems is described below, with some tips for distinguish- ing among them and correcting the problem. In volumetric carryover a small amount of sample is trapped in the sample-injection system and is unintentionally injected the next time the injector cycles. It is characterized by a constant percent-area of the carryover peak relative to the previous peak. For example, if an injection gave a peak of 100,000 area-counts and 1% carryover was seen, a blank injection following a normal injection would have a peak 1% as large as the original (1000 area counts) with the same retention time. An additional blank injection would have 1% of the 1% peak, or 10 area counts. This constant serial dilution in subsequent blank injections characterizes volumetric carryover. Because of the dilution effect it is rare to have any carryover peak after 2 or 3 blank injections. If volumetric carryover is suspected, look for small unintentional volumes where sample might get trapped and then diluted out, such as poorly assembled fittings that only contact the mobile phase during sample injection. Sometimes if a system overpressure occurs, tubing held by PEEK ferrules may slip slightly and create small gaps within the fitting, loosen the fitting and reseat the tube end as described in Section 17.2.5.6. Ineffective autosampler flushing between samples also can be a problem source, make sure that the autosampler wash mechanism is working properly. A worn injection valve rotor can contribute to volumetric carryover; rotor replacement should correct this (see Section 17.4.1.4 for additional autosampler-related discussion). When column-inlet frits become blocked, sample may pool in a stagnant channel and become purged only when an injection occurs. Back-flushing or replacing the column (see split or distorted peaks in Section 17.4.5.3) should solve this problem. Adsorptive carryover may appear to resemble volumetric carryover, but it does not disappear as rapidly in subsequent injections. For example, the first blank may have 1% carryover from the original peak, but the second blank may also have 1% of the original peak or maybe a bit less, not 1% of the 1% peak. The source of such carryover is sample adsorption on surfaces within the system or column. Sample adsorption on the internal polymeric surfaces of the sample injector, the autosampler loop, and the inside of the injector needle are common sources of adsorption. Injection of a very hydrophobic sample dissolved in a polar solvent 17.3 PROBLEM-ISOLATION STRATEGIES 819 (e.g., water) is one common cause of adsorptive carryover. Addition of a few percent of an organic solvent to the injection solvent often will correct the problem. Also it may be useful to increase the strength and volume of the autosampler wash solvent or change the nature of the surfaces (e.g., replace a stainless-steel loop with a PEEK one). For a detailed discussion of sample adsorption in the injection valve, see [14] Sometimes volumetric carryover and adsorptive carryover can occur together. In such cases a constant-fraction drop-off in peak size would be seen in the first and perhaps second blank injection, as the true carryover peak disappeared, but the adsorptive peak would persist for later injections. When all the solutions mentioned above have been tried, and carryover still is unacceptable, you may need to adjust the sample injection sequence so that all high-concentration samples are followed by a blank injection (i.e., ignored) or that low-concentration samples are never injected directly after high-concentration samples. When approximate sample concentrations are not known in advance, the method instructions can be written to ignore a sample with a small peak following a large one; the sample then is re-injected following a blank to get an accurate peak size. Incomplete elution is highly unlikely in gradient elution. With isocratic sep- aration, if an analyte does not leave the column during the run, it may elute in a following injection as an unexpectedly wide peak. See Section 17.4.4.1 for a further discussion of late-eluted peaks. 17.3 PROBLEM-ISOLATION STRATEGIES The ability to isolate problems is a skill that is honed with practice and depends, in part, on a personal aptitude for such activities. In this section we describe 6 rules of thumb that are recommended components of a problem-isolation strategy. Experienced users will likely have at least some of these rules incorporated into their personal approach to isolating problems. Remember to keep safety in mind whenever you are working with an HPLC system. Eye protection should be worn at all times—a few drops of mobile phase seldom will cause a problem on your skin, but in your eyes it may result in irritation or a more serious injury. The high pressure of an HPLC system might at first seem dangerous, but the mobile-phase flow rate is low, so a broken piece of tubing or a leak may cause a mess, but that rarely is a physical danger. However, be careful to avoid trying to stop a leak by pressing against the leak with your finger or thumb—pressures are sufficient to inject mobile phase through your skin, which can cause serious tissue damage. Normal laboratory safety precautions usually are sufficient for HPLC troubleshooting and maintenance (e.g., eye protection, a lab coat, and in some cases solvent-resistant gloves). 17.3.1 Divide and Conquer This is an essential part of troubleshooting. Make changes that allow you to eliminate potential problems—the more sources eliminated with each change, the better. A typical example is to run a new-column test to determine if a problem is related to the analytical method (including the original column) or the hardware. Just install 820 TROUBLESHOOTING a new column and repeat the manufacturer’s column-performance test (Section 3.10.1.3, Table 3.5). If you get the same results (e.g., plate number and retention time within ≈10%) as the column manufacturer, you know the HPLC system is working satisfactorily and the method (and/or the original column) is more likely the problem source. The column-performance test checks isocratic performance; you may want to supplement this with a gradient linearity or gradient-step test (Section 3.10.1.2). 17.3.2 Easy versus Powerful It is important to balance which tests are done first, so as to make the best use of time. For example, if the problem takes longer than normal retention time, and there is more than one peak in the chromatogram, determine the retention-time ratio for each peak in the original and current chromatograms. If this ratio is constant for each peak, a decrease in flow rate is likely. This can be confirmed by a flow-rate check, which is easy and fast. Although it may not be as likely an answer to the problem to make up a new batch of mobile phase, a flow-rate check can be chosen first for convenience and speed. Of course, common sense should lead you to focus on the more common problem areas, even if they are not as easy to troubleshoot. 17.3.3 Change One Thing at a Time Also called The Rule of One, this reminds us to use the scientific method during troubleshooting. Make a change and evaluate the result. Sometimes it is faster to make several changes at a time, but this offers little insight into the real source of the problem, a knowledge of which can be used to solve the problem and design preventive maintenance procedures for the future. 17.3.4 Address Reproducible Problems Also called The Rule of Two, make sure the problem happens at least twice. Chromatographic problems that are not reproducible are difficult to troubleshoot, and it is even more difficult to know that they have been corrected. Make sure that the problem you are trying to solve is sufficiently reproducible that you can be confident you have corrected it. For example, if you have an extra peak or ‘‘spike’’ in a chromatogram, but do not see it with a reinjection of the same sample or in other samples, how will you know if you fixed the problem by making some change in the system? 17.3.5 Module Substitution Replacing a suspect part with a known good part, whether it is a column, check valve, circuit board, detector, or other part, is one of the easiest and most powerful ways to isolate a problem. This strategy constitutes a good argument to have multiple installations of a given brand and model of HPLC system in a laboratory—so that there are more equivalent parts to interchange. Always keep plenty of consumable items on hand, such as filters, frits, guard columns, columns, tubing, and fittings, so that they are available for substitution. 17.4 COMMON SYMPTOMS OF HPLC PROBLEMS 821 17.3.6 Put It Back This goes hand in hand with module substitution and reminds us that if we have substituted a known good part for a suspect part, and it does not correct the problem, we should re-install the original part. This avoids accumulating used parts of questionable quality. Of course, use common sense—it does not make sense to put the old seal back if replacing a pump seal did not solve the problem. 17.4 COMMON SYMPTOMS OF HPLC PROBLEMS One of the most effective ways to isolate and correct an HPLC problem is always to be on the lookout for the common problem symptoms listed below. A good habit that will help identify problems early—before they cause real trouble—is to carefully examine the system every day. For example, during system equilibration at the beginning of the day, trace the flow path through the system, looking for unusual conditions. Start with the reservoir (enough solvent, no visible contamination), trace the transfer tubing to the pump (no leaks, steady pressure), then the autosampler (no puddles, enough wash solvent), column oven (no leaks), detector (no leaks), and waste bottle (sufficient capacity). By following this procedure regularly, you may find problems that can be corrected prior to running samples; you will also get to know how the system works when it is working properly, so that you can recognize when something is wrong. For example, it is unlikely that you will record the sound of the system, but you will get to know the normal clicks and hums that occur when the system is working well. You are then more likely to recognize that ‘‘something doesn’t sound right,’’ so you can investigate further. The list of common symptoms in this section is not exhaustive, but it should cover the most common HPLC problems. For additional help, consult one of the references [1–5] listed in Section 17.1. The text in the present section is conveniently used with the data of Tables 17.2 through 17.11, where each symptom is listed with potential sources of the problem and a cross-reference to the discussion below. Note that these tables are grouped for convenient cross-referencing in Section 17.5 at the end of the chapter. If you get beyond your level of expertise (see Section 3.10.3.1), do not hesitate to ask for help from a more experienced worker, or request a repair visit from a service technician. The first time you perform a mechanical procedure, such as pump-seal replacement, it may be useful to supplement the manufacturer’s diagram(s) with digital photographs or a sketch of your own so that you can confidently reassemble all the parts properly. Problems can exhibit more than one symptom. For example, a loose fitting is likely to exhibit both a leak and low pressure. To avoid repetition, a detailed discussion of a problem will not be repeated for each symptom linked to that problem. Finally, the symptoms and problems in this section are based on the assumption that the system and/or method was working properly prior to observing the symptom. For this reason it is a good idea to consider what changes have been made since the system last worked acceptably—this may narrow down the possible problem sources. 822 TROUBLESHOOTING 17.4.1 Leaks Mobile-phase leaks are one of the most common sources of HPLC problems, and one of the easiest to locate and identify. Many HPLC systems contain leak sensors in various places throughout the system where leaks are likely to occur—as in the column oven. Usually the sensor comprises a pair of electrical contacts that, when bridged by a small amount of liquid, will activate an alarm to alert the user to the presence of a leak. These sensors are located at a low point in the module, so that any solvent will collect at this point and trigger the sensor. Some leaks are obvious, such as those that are clearly visible or identified by a leak sensor. Slow leaks may be more difficult to locate because the liquid evaporates before a drip is apparent (however, slow leaks may result in a visible deposit of buffer crystals at the point of leakage). To locate such micro leaks, a homemade leak detector can be useful (Section 17.2.5.5). A short summary of the causes of leaks in the HPLC system is contained in Table 17.3 and cross-referenced to the following sections. 17.4.1.1 Pre-pump Leaks Leaks before the pump are in the low-pressure portion of the HPLC system (Fig. 3.1), and most commonly are due to a loose low-pressure fitting (Section 3.4.2.1). Whenever parts that are made of two different materials are threaded together, they may tend to loosen over time because of differences in thermal expansion coefficients. Loosening is also accelerated by vibrations, as occur on the mobile-phase proportioning manifold for low-pressure mixing systems. If vibration-loosening is a recurring problem, several manufacturers sell low-pressure fittings with lock nuts to hold the low-pressure nuts more securely. If a loose fitting is found, tighten it (e.g., 1/4–1/2 turn). If the fitting still leaks after additional tightening, take the fitting apart and examine it for damage or contamination. If in doubt, replace the fitting. Nearly all low-pressure fittings are interchangeable (as long as the thread-pitch matches, i.e., English or metric threads with the same diameter and number of threads per unit length), so it is common to switch to another brand of fittings when factory-supplied low-pressure fittings fail. With high-pressure mixing systems (Fig. 3.14), the only low-pressure connec- tions are at the pump where the mobile-phase inlet tubing is connected to an inlet check-valve, solvent-selection device, or a splitter to supply solvent to two pump heads. The source of the leak should be obvious in this case. With low-pressure-mixing systems (Fig 3.15), in addition to the low-pressure connections at the pump, a proportioning manifold is used to blend the proper proportions of solvents to form the mobile phase. The fittings at this manifold are another possible source of leaks. The proportioning valves can also leak, although this occurs rarely. If fluid is present at the base of a proportioning valve, gently tighten the screws holding the proportioning valve(s) in place. These may be metal screws threaded into a plastic block, so be careful—they are easy to overtighten and strip the threads. If the proportioning valves continue to leak, it is best to exchange the entire proportioning manifold with a new one—the parts are carefully matched at the factory for best performance. It is also possible to have an air leak on the low-pressure side of the pump. Air can leak in through a gap too small to allow liquid to leak out. Air leaking in will 17.4 COMMON SYMPTOMS OF HPLC PROBLEMS 823 cause the pump pressure to be lower than normal (see Section 17.4.2.2 for more details). 17.4.1.2 Pump Leaks There are many parts of the pump that can leak, and leaks often are accompanied by low- or variable-pressure symptoms (Sections 17.4.2.2, 17.4.2.3). Low-pressure fittings generally are used on the pump inlet (Section 17.4.1.1) and high-pressure fittings on the outlet (Section 17.4.1.3); consult these Sections for more details. The remaining possible sources of pump leakage are the check valves, the pump seals, and auxiliary components, such as pulse dampeners. External leaks of fluid from check valves result from a check valve that is loose, cross-threaded, or contains a damaged seat. A loose check valve is the most common source of leakage. Tighten the check valve (e.g., 1/8–1/4 turn) with a wrench to correct this problem. Be sure to hold the pump firmly so that the whole pump is not twisted when you tighten the check valve. If this does not correct the problem, make sure that the check valve is properly threaded into the pump head. Turn off the pump and remove the check valve. Check valves should turn freely with your fingers once they are initially loosened. If a check valve requires a wrench to remove it or insert it most of the way, the threads may be damaged. Check the threads for damage and replace a damaged check valve—in some cases the threads in the pump head may be damaged, requiring a pump-head replacement. When the check valve is removed, examine the seat where the check valve contacts the pump head at the bottom of the threaded portion. This seat often is made of a hard plastic and can crack—if the seat is damaged or cracked, replace it. Consult the pump manual for details about your specific pump. Pump seals are a common source of pump leaks, if they are not replaced regularly according to a preventive maintenance program (Section 17.2.2). The function of the pump seals is described in Section 3.5.1. Because the pump seal continuously rubs on the moving piston during pump operation, it is the most wear-prone part of the HPLC system. As the seal wears, it will become less effective, with resulting leaks and/or pressure fluctuations. If pump seals are replaced every 6 to 12 months, it is rare that pump-seal failure will be encountered; for this reason we strongly recommend periodic pump-seal replacement as part of preventive maintenance. In addition to leaks, pump-seal failure usually results in the generation of small particles that work themselves downstream and block tubing, frits, or columns. If the seal leaks, usually liquid will appear below the pump head between the inlet check valve and the pump body—most pumps have a small drain hole in this position. Before replacing the pump seal, consult the pump operator’s manual for specific recommendations for your pump. A general description of seal replacement follows: • Turn off the pump and disconnect the inlet and outlet tubing at the inlet and outlet check valves. • Remove the pump-head retaining screws; this usually requires a hexagonal (‘‘Allen’’) wrench. Successively loosen each screw by a small amount until the pump head is free (so that the pump head is not twisted in the process). Pull the pump head straight off the pump so that no sideways pressure is 824 TROUBLESHOOTING placed on the (fragile) piston; alternatively, with some pumps, the head and piston come out as a unit. • Remove the pump seal. If the pump came with a seal-removal tool, use it. Alternatively, a brass wood-screw can be used like a cork screw to remove the seal. If you elect to pry the seal out of the pump head, use a plastic tool, not a screwdriver or other metal tool, to avoid damaging the pump head. Do not use the piston to pry out the seal because the piston is likely to break. As you remove the seal, be sure to note how the seal is installed—the open side of the seal, where the spring is visible, should be facing the high-pressure section of the pump. • Clean the piston. Usually rinsing with alcohol and wiping the piston with a disposable laboratory wipe is sufficient to clean the piston. If a residue remains, sometimes polishing the piston with a small amount of toothpaste will remove the residue (be careful to avoid fluoride-containing toothpaste if the system is used for ion chromatography). The toothpaste is sufficiently abrasive to remove buffer or other deposits, without scratching the sapphire piston. • Inspect the piston for any scratches or defects. A simple way to highlight any problems is to hold a laser pointer up to the end of the piston, which will cause it to act as a light pipe and glow brightly. Any imperfections in the piston surface are either scratches or residues. Examine the end of the piston to be sure it is not broken or rough. If further cleaning does not remove these imperfections, replace the piston. • Install the new seal. This usually can be done by pressing the new seal into the pump head with a fingertip. Be sure to install it in the correct orientation. Most seals come with a flange that prevents improper installation, but some seal designs can be installed backward. • Lubricate the seal and piston with a squirt of alcohol from a wash bottle so that it slides together easily. • Slide the pump head back on the piston, taking care not to twist the head and break the piston. • Successively tighten each retaining screw a small amount so that the pump head is not twisted, possibly breaking the piston. If a torque specification is given in the manual, tighten the screws as directed; otherwise, tighten the screws firmly. • Reconnect any tubing connections, and purge the pump (without a column connected, to allow particles to be flushed out); the pump is now ready for use. Leakage can occur at auxiliary pump components such as pulse dampeners, mixers, or pressure transducers. If the leak is at a fitting, follow the normal procedure for correcting leaks at high-pressure fittings (Section 17.4.1.3)—first tighten the fitting, and if this does not fix the leak, clean or replace the fitting. If the leak is in a pulse dampener or mixer, and the cause of the leak is not obvious or cannot be corrected easily, the component may need to be replaced. 17.4 COMMON SYMPTOMS OF HPLC PROBLEMS 825 17.4.1.3 High-Pressure Leaks High-pressure fittings are the most common location of leaks in the HPLC system. Such leaks become obvious when a leak sensor triggers an alarm or a puddle is discovered. More subtly, the pressure may be low or fluctuating (Sections 17.4.2.2, 17.4.2.3), or a white deposit of buffer crystals may appear where the nut threads into the fitting body. Sometimes an elusive leak can be identified using a piece of thermal paper as a probe (Section 17.2.5.5). (If you are not familiar with the parts and proper assembly of high-pressure fittings, read Section 3.4.2.2 and consult Figs. 3.8 and 3.9.) In most cases a leaky fitting can be fixed by tightening the fitting 1/4 to 1/2-turn. Note that with PEEK fittings the pump should be turned off, the nut loosened, the tubing pushed firmly into the fitting, and then the nut should be re-tightened; otherwise, the tube end can slip in the fitting, causing a void volume in the fitting (see Fig. 3.09b). If this additional tightening does not stop the leak, disassemble the connection, rinse it out, and reseal it. If it still leaks, replace the ferrule. PEEK ferrules can be used to overcome minor imperfections in the surface of the seat. While PEEK fittings are suitable for traditional HPLC system pressures (up to 6000 psi or 400 bar), note that higher pressure systems usually require stainless-steel fittings for secure, leak-free connections. 17.4.1.4 Autosampler Leaks Nearly all HPLC systems are operated with autosamplers today. But, if your system uses a manual injector, many of the same corrective measures apply, because the injection valves are very similar for both manual injectors and autosamplers. Injection-valve and autosampler design and operation are discussed in Sections 3.6.1 and 3.6.2, respectively. Leaks at the low- or high-pressure fittings in the autosampler can be addressed in the same way, respectively, as leaks before the pump (Section 17.4.1.1) or with other high-pressure fittings (Section 17.4.1.3). Other leaks can be divided into two categories: those associated with the injection valve, and those external to the valve. Leaks associated with the injection valve occur either at one of the fittings or connections, or as a result of injector rotor-seal leakage. Push-to-fill autosamplers (Section 3.6.2.2, Fig. 3.21) and most manual injectors make the connection between the needle and the injector valve with a low-pressure seal, shown in a schematic in Figure 17.2. A nut (n) and ferrule (f) are used to hold a polymeric sleeve (s) in place in the valve body (b). At the ferrule the sleeve is crimped slightly (c) so that it seals against the needle. This seal can wear or deform over time so that fluid leaks out at the top of the sleeve when the sample is transferred into the loop. A slight tightening of the nut, often will tighten the crimp slightly and correct the leakage problem. In other cases the sleeve may need to be replaced. Consult the autosampler manual for specific instructions for your autosampler. Needle-in-loop autosamplers (Section 3.6.2.3, Fig. 3.22) require a high-pressure seal between the needle and the injection valve. This seal is made of a hard plastic, such as PEEK or other hard material, that can wear, become distorted, or crack over time. Sometimes tightening the nut that holds the high-pressure seal in place will stop a leak, but usually a leak at the high-pressure seal will require installation of a new seal. . screw to remove the seal. If you elect to pry the seal out of the pump head, use a plastic tool, not a screwdriver or other metal tool, to avoid damaging the pump head. Do not use the piston to. sapphire piston. • Inspect the piston for any scratches or defects. A simple way to highlight any problems is to hold a laser pointer up to the end of the piston, which will cause it to act as. Clean the piston. Usually rinsing with alcohol and wiping the piston with a disposable laboratory wipe is sufficient to clean the piston. If a residue remains, sometimes polishing the piston with