Introduction to Modern Liquid Chromatography, Third Edition part 88 ppsx

17.4.1.6 Detector Leaks The pressure at the column outlet is lower than at the inlet by an order-of-magnitude or more, so detectors are subject to much lower pressures than the preceding

b

s

needle

n

c

Figure17.2 Schematic of low-pressure needle-seal Valve body (b), nut (n), ferrule (f), poly-meric sleeve/seal (s) and distorted portion of sleeve (crimp) that forms a seal around the nee-dle (c)

Failure of the injection rotor-seal can result in leaks at the injection valve The standard 6-port injection valve uses a rotating seal, the rotor, which turns against

a stationary stator, to which the tube connections are made The rotor typically is

made of a hard polymer, such as PEEK, Vespel, or one of the fluoropolymers, and contains tiny passages that connect the fluid passages in the stator The stator may consist of the stainless-steel valve body, or it may be an insert in the valve made of ceramic or another hard, smooth surface In the schematic of Figure 3.17, the rotor comprises the portion of the valve inside the circle, whereas the stator is the part immediately outside the circle In the most popular injection valves, the rotor is a flat disk with three kidney-shaped grooves in the surface, similar to that shown in the

sketch of Figure 17.3a These grooves line up with the ends of the tube connections

on the stator, shown as the dashed circles in Figure 17.3b In one position (e.g.,

load), the flow channels that result connect ports 1–2, 3–4, and 5–6, as seen in

Figure 17.3b When the rotor is rotated 60◦ (e.g., to the inject position), a different set of ports is connected, in this example, 2–3, 4–5, and 6–1

If a small piece of hard material, such as a particle of column-packing or a bit of stainless steel from a poorly cut tube, gets caught in one of the passages of the injector, it can scratch the rotor This can form a connecting passage between

two of the grooves, as shown between ports 4 and 5 in Figure 17.3c (arrow) The result is cross-port leakage, where fluid from one hydraulic portion of the system

leaks into another This can show up as fluid leaking out the injection or waste port, as a problem of precision due to liquid leaking into or out of the sample loop,

or sometimes as a carryover problem (Section 17.2.5.10) Rotor-seal (and possibly stator) replacement will be required to correct this problem Also be sure to clean the remainder of the valve thoroughly to remove any particulate matter

More commonly the rotor seal will fail as a result of normal frictional wear The rotor seals are designed with a service lifetime of>100,000 cycles For many

laboratories, this will mean several years of operation before failure, so routine replacement of the rotor seal does not make much sense—unless the system has a

3

4

5 6

2

(a)

(c)

(b)

Figure17.3 Schematic of injector rotor-seal (a) View showing kidney-shaped grooves in sur-face of polymeric seal; (b) as in (a), showing connecting ports (numbered, dashed circles) for tubing; (c) scratch between ports 5 and 6 of b, causing cross-port leakage (d) Cross section of kidney-shaped groove (normal condition); (e) worn seal resulting in burr of seal material at

edge of groove (arrow)

counter to automatically record the number of injection cycles, so as to allow the remaining rotor seal life to be estimated When viewed in cross section, the grooves

of the rotor shown in Figure 17.3a are U-shaped as in Figure 17.3d As the surface

is worn during normal operation, the rotor will become slightly thinner and often a small burr of rotor material will form at the open edges of the groove, as seen at the

arrow in Figure 17.3e These burrs can break off and shed particles that can block

tubing or frits As the surface or the rotor wears, the contact pressure between the rotor and the stator is reduced, and eventually the seal will leak Replacement of the rotor should solve this problem; thoroughly clean the valve before reassembly The pressure limit of the seal between the valve rotor and stator is related

to how tightly the two surfaces are held together Higher pressure-limits require that the surfaces be held together more tightly, but this also increases the friction between the surfaces and the effort to rotate the valve Normally the injection valves are adjusted to withstand 6000 psi (400 bar) for traditional HPLC applications For U-HPLC use (>6000 psi, Section 3.5.4.3) the two surfaces must be held together

more tightly, so rotor lifetimes are expected to be shorter; alternative injector designs may overcome this problem When an injection valve is disassembled for servicing,

the rotor-to-stator sealing pressure may need to be adjusted; consult the service manual for additional instructions

A major exception to the general use of rotary injection valves is found in some Waters-brand autosamplers, which typically use a ‘‘seal-pack’’ design that incorporates slider valves and high-pressure seals instead of rotary valves Some of these parts are user-serviceable, and some require replacement of a subassembly with a new or rebuilt unit Consult the user manual for more information on troubleshooting and repair

Other points of leakage in autosamplers will vary from one design to another and often are unique to one model If tightening or replacing a connecting fitting does not correct the problem, consult the autosampler manual for more information 17.4.1.5 Column Leaks

Leaks at the column will be associated with the fittings Leaks at the tube connections are treated as outlined in Section 17.4.1.3 If the column end-fitting itself is leaking,

it may be possible to stop the leak by tightening the nut 1/4-turn These larger fittings will take more effort to tighten than fittings for the 1/16-in o.d connecting tubing If the fitting continues to leak, it may be best to discard the column, because disassembly of the column end-fitting can result in permanent damage to the column

In the past, it was common to remove the column-inlet fitting to replace the frit if

a blockage was suspected, but with today’s column-packing techniques, removing the end-fitting may allow column packing to ooze out and permanently damage the

column For this reason removal of the end-fitting for examination or repair is no

longer recommended.

Cartridge-type columns comprise a disposable column that is held in a reusable holder If the end-fitting on a cartridge column leaks, try tightening it to correct the problem If it still leaks, disassemble the holder, rinse the fitting, and reassemble In some cases the polymeric seal between the column and the holder may need to be replaced to stop a leak

17.4.1.6 Detector Leaks

The pressure at the column outlet is lower than at the inlet by an order-of-magnitude

or more, so detectors are subject to much lower pressures than the preceding high-pressure components Because leak-related problems correlate strongly with the local system pressure, leaks at the detector are much less common than in other parts of the HPLC system

Fittings at the detector inlet usually are the same type of high-pressure com-pression fittings used in other high-pressure parts of the system; leaks at these tubing connections should be treated as described in Section 17.4.1.3 Some detectors use 1/32-in o.d tubing instead of the standard 1/16-in tubing used elsewhere in the system This smaller diameter tubing is easily twisted and kinked, so take extra care when working with it Some detectors operate at sufficiently low pressure on the inlet side that low-pressure plastic fittings can be used; many detectors use low-pressure fittings on the detector outlet because the pressure is quite low Correction of leaks

in low-pressure fittings is described in Section 17.4.1.1

UV detectors (Section 4.4) are the most popular HPLC detectors A generic version of the detector cell is shown in Figure 17.4 The cell typically comprises a

end cap quartz window

gasket

inlet

outlet cell body

mobile phase light

in

light out

10 mm

Figure17.4 Schematic of typical UV-detector flow cell

stainless-steel block (e.g., 10-mm long) with a hole (e.g., 1-mm diameter) drilled through it; quartz windows are held on the ends by an end-cap with O-ring seals The tube connections to the cell body at the inlet and outlet may be compression-type fittings, which can be adjusted in the normal way (Section 17.4.1.3) if leaks occur

In other cases the tubing is soldered or welded to the flow-cell body, in which case factory repair or replacement will be required The inlet tubing connected to the cell usually is thin-walled, narrow-diameter tubing (e.g., 1/32-in o.d.) that also functions as a heat-exchanger to stabilize the temperature in the flow cell Be careful not to twist or kink this thin tubing when tightening fittings

Because air bubbles create noise in the cell, the UV detector often is operated with an after-market back-pressure restrictor on the cell outlet (Section 4.2.1) This creates enough pressure (e.g., 50–100 psi) to keep the bubbles in solution, but not

so much pressure as to cause the window seals to leak Spring-loaded back-pressure restrictors work well to accomplish this An alternative is to use a narrow-bore waste tube (e.g.,≤0.010-in i.d = ≤0.25 mm i.d.), but as the flow rate is increased,

the pressure also is increased With this type of restrictor, a high flow rate may create sufficient back-pressure to exceed the upper pressure-limit of the detector cell (e.g., 150 psi), causing leakage at the window seals Once a detector cell leaks, it may continue to leak if the window seals become distorted Some detector cells are user-serviceable, whereas others will require service by a factory-trained technician,

or complete replacement—consult the detector manual for more details on your model

Other types of detectors have different flow-path and detector-cell designs For any flow-through detector (fluorescence, electrochemical, conductivity, etc.), leaks can occur at the detector flow cell Evaporative detectors, such as the evaporative light scattering, corona discharge, or mass spectrometric detector will have fluid leaks only at the interface, so leak isolation may be a little easier Consult the appropriate section of Chapter 4 for a general discussion and generic design of

specific detectors This information, plus details from the detector manual, will help you locate and correct leaks in specific detector models

17.4.2 Abnormal Pressure

Abnormal pressure is always a symptom of some other problem in the HPLC system Normal system pressure will be different for different HPLC systems and applications, so it is a good idea to record the system pressure on a regular basis For example, if you use a batch record or run sheet, record the pressure as part of the system suitability records; then you will be able to compare a questionable high- or low-pressure symptom to the historical record (Section 17.2.4) Conventional HPLC systems are capable of operation up to 6000 psi (400 bar), but most applications operate at pressures of 1500 to 3000 psi (100–200 bar) U-HPLC systems are designed to operate at pressures>6000 psi—as much as 15,000 psi (1000 bar) or

more However, as of this writing, these systems typically are operated in the 8000

to 10,000 psi (550– 650 bar) range

Ideally the pressure will be constant during an isocratic separation; however, small pressure fluctuations in the range of 1–2% of the operating pressure (e.g., 10–20 psi, 1–2 bar) are normal for many applications Pressure also will vary with

mobile-phase composition, as illustrated in Table 17.1, so pressure changes during

gradient elution are normal Methanol (MeOH) is more viscous than acetonitrile

(ACN), and blends of MeOH and water are considerably more viscous than either MeOH or water alone On the other hand, the pressure generated by ACN/water mixtures decreases in a fairly linear fashion as the mixture is changed from 100% water to 80% ACN More detailed information on solvent viscosity can be found in Table I.3 of Appendix I

Table 17.1

Example Pressures for Various Solvent Blends

B-Solventa

aA-solvent is water.

bApproximate (calculated) pressure for 250× 4.6-mm, 5-μm particle column operated at 35◦C and

2 mL/min (Eq 2.13a).

Most HPLC pumps have pressure-limit settings that serve to shut off the pump

if the limits are exceeded The upper pressure-limit protects the system against damage or leaks if a blockage occurs causing excessive pressure Most workers will set the upper limit well above the normal maximum operating pressure for

all methods For example, if normal method pressures are 2000 to 3000 psi, the upper limit might be set in the 4000 to 5000 psi region This will provide some protection, yet is not low enough to result in frequent pump shutoff The lower pressure-limit shuts the pump off if the pressure drops below the set point, such as when a mobile-phase reservoir runs dry There is not much danger of pump damage

if the solvent supply is exhausted, and the pump will not pump air, so the column will not dry out Nevertheless, it is a good idea to set a lower pressure-limit in the 20

to 50 psi region so that the pump stops if the solvent supply is interrupted Pressure problems are discussed below and summarized in Table 17.4

17.4.2.1 Pressure Too High

Higher-than-normal pressure is a symptom of a blockage (assuming that the system settings are correct) Most commonly the pressure will rise gradually with successive batches of samples as debris collects on the in-line filter, guard column, or frit at the head of the column This is normal with methods for the analysis of samples that may not be completely free of particulate matter Remember, the pressure can increase by as much as 60% over the starting pressure during a gradient (Section 17.4.2, Table 17.1), so set the upper pressure-limit to accommodate this normal pressure fluctuation It is also important to remember that column pressure is

related to particle size, column length, and column diameter (Eq 2.13a) Nominally

equivalent columns can also differ in their pressure drop (e.g., two different brands of

150× 4.6-mm i.d., 5-μm particle C18columns) Finally, HPLC systems designed for higher pressure operation (e.g., U-HPLC) often have very small orifices and reduced tubing diameters (e.g., ≤0.005-in or ≤0.12-mm i.d.), resulting in >1000 psi of

system pressure without a column installed, so allowances for normal background pressure must be taken into account during troubleshooting

Occasionally a sudden pressure increase will occur, and likely trigger the upper pressure-limit with a resulting pump shutdown This can happen upon injection of a very dirty sample, such as untreated plasma, if sufficient particulate matter is present

to completely block a frit or a piece of connecting tubing Blockage can also take place when buffer and organic solvent are mixed on-line under conditions where the buffer solubility is poor and precipitation occurs, such as blending phosphate buffer and acetonitrile in some HPLC equipment

Problem isolation for excessive-pressure problems is quite simple Just work your way upstream from the column outlet, loosening the tubing connections as you go, with the pump operating (If you strongly suspect a blockage at a particular point in the system, such as the in-line filter, start there and save time.) When a fitting is loosened and a sudden pressure drop results, the blockage is immediately downstream from that fitting Remember, there normally is >1000 psi pressure

drop across the column (depending on the flow rate), so much of the pressure drop observed when loosening the column-inlet connection is normal With conventional HPLC systems designed for operation≤6000 psi, once the column is removed, the pressure should be very low (e.g., ≤100 psi) As noted above, systems designed for higher pressures (Section 3.5.4.3) and sub−3-μm particles may have a system pressure of 1000 psi or more with no column attached For reference purposes,

it may be useful to go through the blockage isolation procedure with a normally operating system, to see what the normal system pressure is at each connection (for some fixed flow rate)

Once the problem location is isolated, appropriate corrective action should be taken For example, if a piece of tubing is blocked, the tubing should be replaced If the in-line filter is blocked, replace its frit If the column is blocked, column reversal may help (the procedure is included in the discussion of split and distorted peaks in Section 17.4.5.3); otherwise, it may be necessary to replace the column

Before putting the system back into service, consider if steps need to be taken

to prevent the problem from recurring—or reduce its impact the next time For example, if the problem is frit blockage from sample particulates, you may want

to institute a filtration or centrifugation step during sample preparation, and use

an in-line filter (Section 3.4.2.3) If buffer precipitation blocked a tube in the pump

or mixer, consider reducing the buffer concentration or pre-mixing the buffer and organic solvent

17.4.2.2 Pressure Too Low

System pressure that is too low is a sign of a leak or a pump problem (assuming all the system settings are correct) If the pressure is cycling, the problem is more likely

at the pump (see the further discussion in Section 17.4.2.3), whereas if the pressure

is steady and low, a leak is more likely Because low-pressure and cycling-pressure problems are closely related and sometimes hard to tell apart, consult Section 17.4.2.3 for additional information

If the pump has shut off due to the lower pressure-limit, and the cause is not obvious (e.g., an empty mobile-phase reservoir), restart the system and see what happens Sometimes the lower pressure-limit sensor is too sensitive and will stop the pump if a momentary drop in pressure occurs, such as with the passage of a bubble through the pump If this is a regular occurrence, it may be best to disable the lower pressure-limit

Mobile-phase leaks can also cause low pressure; identify and correct the source

of the leak using the instructions of Section 17.4.1 and Table 17.3 Although mobile-phase leaks are the most common type of leak, it is possible for air to leak into the system through a loose low-pressure fitting (Section 17.4.1.1) If efforts

to find a liquid leak are fruitless, tighten each of the low-pressure fittings to see if this corrects the problem If a proportioning valve is not sealed properly, it may be possible to pull air into the system through an unused solvent supply tube

A sticking inlet check valve can prevent the pump from building sufficient pressure Check-valve sticking is particularly problematic with ball-type check valves

(Fig 3.12a) when used with acetonitrile (ACN) When ACN is used as a solvent,

the machined surface of the sapphire valve seat can catalyze the polymerization

of minor contaminants in the ACN (aliphatic amines) [15, 16] This polymer then results in a smoothing of the contact surface, so the ruby ball sticks to the seat via increased surface tension Sonicating the check valve in methanol seems to correct this problem, at least temporarily There is also speculation that sonication in dilute nitric acid might serve to remove this polymeric buildup [17], but this had not been confirmed at the time this book was completed HPLC pumps that use active check

valves (Fig 3.12b,c) are not subject to this sticking problem; unfortunately, pumps

with ball-type check valves cannot be retrofitted with active check valves

A pump that is starved for mobile phase will not be able to generate the expected pressure Check that sufficient mobile phase is available at the pump, by

carrying out the siphon test described in Section 3.2.1 Impediments to free-flowing mobile phase include blocked inlet-line frits, faulty proportioning valves, and pinched

or blocked tubing

Insufficient mobile-phase degassing or insufficient pump purging can leave enough air in the mobile phase that the pump loses prime, thereby lowering the pressure Ensure that the mobile phase is properly degassed (Section 3.3), and purge the pump by opening the purge valve, then pumping 10 to 20 mL of degassed mobile phase to waste (a high flow rate will sometimes displace bubbles in the pump) Frequent bubble problems can point to a faulty degassing module If this is suspected, try an alternate degassing method (Section 3.3) to isolate the problem

A worn pump seal can cause leaks (Section 17.4.1.2), but before leaks are apparent, the seal problem may prevent the pump from being able to provide the expected pressure Carefully check for seal leaks at the drain hole on the bottom

of the pump head (between the inlet check valve and the pump body), and check the system logbook (Section 17.2.4) to see if the system is due for a scheduled seal replacement Seal replacement (see Section 17.4.1.2) should correct any problems When the pump head is removed, check to see if there is any piston damage—a scratched or broken piston also can cause the pump to underperform

17.4.2.3 Pressure Too Variable

As mentioned above, the HPLC system pressure in the normally fluctuates Typically this fluctuation is 1–2% (e.g., 10–20 psi, 1–2 bar) of the operating pressure, but this will vary between systems and applications It is therefore a good idea to make

a note of the normal pressure variation as part of the records kept for each batch of samples (Section 17.2.4) Also keep in mind the normal pressure cycle during each run for gradient elution

A summary of variable-pressure symptoms and solutions is given in Table 17.4 The most common sources of these problems are bubbles in the pump, sticky check valves, worn pump seals, broken pump pistons, and an inadequate mobile-phase sup-ply to the pump The identification and correction of these problems is almost exactly the same as for low-pressure problems (Section 17.4.2.2), so further instructions are not needed The main difference between low-pressure and variable-pressure problems is that the latter may be limited to one pump head of a dual-piston pump,

or one pump of a two-pump system Thus part of the system may be working normally (higher pressure) while part of the system is not delivering enough mobile phase (lower pressure) The two most likely problem sources are a bubble in a pump head or a sticky inlet check valve The simplest initial approach to correcting the problem is first to purge the pump to see if this fixes the problem If it does not, sonicate the check valves (Section 17.4.2.2) in methanol Additional information is found in Table 17.4

17.4.3 Variation in Retention Time

Retention times for analytes should be constant within a sample batch (e.g., same day, same batch of mobile phase) if all chromatographic conditions are held constant

If a change in retention is observed, it indicates that at least one condition has changed Because it is impossible to hold all variables exactly constant, there is a normal variation in retention time for every method Typically this is in the range of

±0.5% of the retention time or ±0.0.02 to 0.05 minute; this normal variation can be determined from historic sample-batch records When the retention-time variation exceeds the normal variation, steps should be taken to assess the cause and take corrective action

A tool to implement the initial divide-and-conquer approach (Section 17.3.1)

is to calculate the retention factor k (Eq 2.6), and then compare changes in k and retention time t R with the help of Table 17.5 Table 17.5 will guide you to one of the following sections for more information about possible changes in the mobile phase, column, or column temperature As discussed in more detail below, mobile-phase changes tend to occur in a stepwise fashion when some intentional change is made, column changes usually take place over a period of weeks or months, and temperature changes tend to cycle during the day These general patterns can greatly aid the identification of the problem source

An alternate way to classify retention-time problems is by the observed change

in t R—increased, decreased, or variable If you want to approach the problem in this manner, consult Table 17.6 and Section 17.4.3.6 first The following sections (17.4.3.1– 17.4.3.5) cover each of these symptoms It may be appropriate at this point to run the system reference test (Sections 3.10.1.3, 17.2.1) to determine if the problem is related primarily to the equipment or the method Finally, there is much overlap among the various causes of variable retention, so it is a good idea to read all of Sections 17.4.3.1 through 17.4.3.5 in order to gather as many ideas as possible, if the solution to the problem is not quickly reached

17.4.3.1 Flow-Rate Problems

A change in the flow rate will change t R but not k, because k is independent of flow rate, while t Rvaries in inverse proportion to flow rate (an exception can occur for pressures>5000 psi, because of a slight dependence of k on pressure) Never

underestimate the power of operator error—it is a good idea to verify that the proper flow rate is selected Once a setting error has been eliminated, the only possible cause of a higher than normal flow rate is a problem with the system controller, which will require the skills of a trained technician to fix

Flow rates that are lower than normal cause retention times to be too long, a problem that can be caused by bubbles in the pump, pump starvation, faulty check valves or pump seals, or leaks Sections 17.4.1 and 17.4.2.2 describe corrective actions for these problems As discussed in Section 17.4.2.3 for pressure, the causes

of variable pressures or flow rates often are the same as low-pressure or low-flow problems

17.4.3.2 Column-Size Problems

A rare, but possible cause of a change in t R but not k is installation of the wrong

column size, but correct type (e.g., Symmetry C18) Of course, the column size cannot change without operator intervention The obvious fix is to look at the column label and then install the proper column

17.4.3.3 Mobile-Phase Problems

A change in the mobile phase (e.g., %B) can result in changes in both t R and k.

When a new mobile-phase batch is prepared incorrectly, any changes in retention

will be noticed at that time—whether the mobile phase is pre-mixed (‘‘off-line’’) or prepared by on-line mixing When on-line mixing systems are used, small variations

in mobile-phase composition may occur due to problems with the proportioning system Alternatively, small, continuous changes in the mobile phase (and sample retention) can occur over time, although this is much less common For example, a volatile buffer, such as ammonium carbonate, may evaporate so as to change the pH

of the mobile phase If continuous helium sparging is used for degassing, a volatile organic component of the mobile phase could be selectively evaporated

Errors in formulating the mobile phase are a likely cause of shifts in retention (with increased retention for a reduction in %B, and vice versa) The rule of 2.5 (Section 2.5.1) indicates that a 10% change in %B will change the retention factor

by approximately a factor of 2.5 times; a 1% error in mobile phase %B can account for≈10% change in k.

An error in mobile-phase pH can have a much larger effect on the retention

of acidic or basic solutes (Section 7.3.4.1) than neutral analytes The concentration

of mobile-phase additives, such as ion-pairing reagents (Section 7.4.1.2), also can affect retention During method development the robustness of the separation to small changes in mobile-phase composition should have been examined (Section 12.2.6) The results from robustness testing can be useful in determining the specific mobile-phase error that was made From a practical standpoint, however, the most direct solution is to make up a new batch of mobile phase and determine whether the problem has been corrected

17.4.3.4 Stationary-Phase Problems

With continued use of the column, changes in retention and selectivity are common,

but t0 is unaffected; consequently values of both k and t R will change Retention shifts due to a change in the stationary phase rarely are the only symptom observed

Usually the plate number N will also drop, peak tailing will increase, and the column

pressure will rise Past records of method use (Section 17.2.4) in combination with

recent data on the performance of a column (values of N, pressure, etc.) can be used

to avoid its (highly undesirable!) failure during the assay of a series of samples As a rule, a column lifetime of 500 to 2000 sample injections should be expected for most applications (and will account for<1% of the total cost of analysis) Some methods

can degrade columns more quickly, while other methods may allow a longer use of the column Expect shorter column lifetimes when the column is operated outside the

2< pH<8 region or at temperatures >50◦C The use of in-line filters (Section 3.4.2.3) and guard columns generally will extend column life In any event, the column should

be considered a consumable item that will wear out (hopefully gradually) over time

The easiest way to check for stationary-phase related problems is to replace the column (module substitution, Section 17.3.5) If a guard column is in use, first replace or remove the guard column to see if the problem is resolved If the column repeatedly fails prematurely, check to be sure that it is operated within its recommended limits (consult the column care-and-use instructions for specific guidelines) In some cases it may be appropriate to find and use an equivalent column (Section 5.4.2) that is more stable If the failure is due to the injection of dirty samples, additional sample-preparation steps (Chapter 16) may be necessary