HPLC

Troubleshooting
Guide
1 of 86
What are the differents elements of a HPLC system : Isocratic or gradient

2 of 86
To help you identify normal operation conditions

Record a map of your LC system

Keep a log
Run test chromatogram regularly

A log is especially important in a lab in which the system is shared by multiple

operators

3 of 86
all the listed items in the standard HPLC system could be a potential for
chromatographic problems

A troubleshooting strategy includes five primary process:

Identifying the symptoms
Understanding the possible causes
Isolating the exact possible causes following a logical
sequence
Do you have the expertise to repair
Resolving the problem or contact the manufacturer

4 of 86
Isolating the Problem
Start troubleshooting by determining if the problem is caused by the hardware,
that is the instrumentation, or the chemistry. Overall, chemistry includes the
column, guard column, the mobile phase or solvent and the sample itself. It is not
always easy to categorize the problems as either chemistry or hardware related.
Your suppliers’ service or support department is always a good place to ask for
help to isolate the problem.

5 of 86
6 of 86
7 of 86
Reservoirs - glass or plastic?
Keep covered
Don’t “top off”
Tubing - No kinks or twists
Effects of cold flow
Filter/sinker - 10m, not a substitute for solvent prep,
Keeps tube at bottom of reservoir.
Draw down
No effective way to clean them, replace if plugged/corroded
Sparge - Necessary when blending solvents in line (gradient, “dial-a-mix”)
Degasser - Necessary when blending solvents in line (gradient, “dial-a-mix”)
UV Cutoff - Wavelength at which solvent has 1AU
Contaminants - Particulate and chemical
Solvent Prep - Filtering
pH measurement, volume measurement for pre-mix

8 of 86
Explanation of the characteristics of a solvent

9 of 86
10 of 86
11 of 86
Effect of undegassed solvent on the reproducibility of the retention time

12 of 86
13 of 86
14 of 86
15 of 86
Filtration sur membrane :
• 2 actions : élimination des insolubles et des gaz dissous.

• Compatibilité de la membrane avec les solvants utilisés. Exemple :

HA, FH, HV ou LCR

• Diamètre des pores : 0,22 ou 0,45 µm.

Elimination des insolubles ou stérilisation des solutions.

• Probléme si mélange de composés volatiles.

16 of 86
17 of 86
18 of 86
19 of 86
Lorsqu ’on réamorce ou purge une voie,il faut connaitre le type de chambre que l’on
possède pour être sur d ’avoir purger correctement les chambres

20 of 86
21 of 86
22
23
UV absorption of 2, 4 DMP in 55:45 méthanol:phosphate buffer is used to monitor
the ionization of the 2, 4 DMP.
Series 1 shows the result of adjusting the pH of the water and then adding the
methanol.
Series two shows the effect of adding the methanol and adjusting the pH of the
methanol: buffer mixture.
Yes it does make a difference.
[D.V. McCalley /J. Chromatogr. A 664 (1994) 139-147]

24 of 86
Back Pressure and Solvent Viscosity
It is important to know that not all solvents give the same system pressure. The
chart indicates that mixtures of water (or buffer) and organic solvents have
varying viscosities depending on the volume percent of each component. When
you go from a mixture of pure water to a mixture of pure methanol, the viscosity,
and hence the pressure, increases to reach a maximum around 50 % methanol,
and then decreases again. This viscosity and pressure change is not as
pronounced for mixtures of water and acetonitrile as it is for water and ethanol
mixtures.

25 of 86
26 of 86
27 of 86
UV absorption of 2, 4 DMP in 55:45 méthanol:phosphate buffer is used to monitor
the ionization of the 2, 4 DMP.
Series 1 shows the result of adjusting the pH of the water and then adding the
methanol.
Series two shows the effect of adding the methanol and adjusting the pH of the
methanol: buffer mixture.
Yes it does make a difference.
[D.V. McCalley /J. Chromatogr. A 664 (1994) 139-147]

28 of 86
29 of 86
30 of 86
31 of 86
Il faut noter aussi que le ph peut aussi jouer sur le forme des pics

32 of 86
33 of 86
34 of 86
35 of 86
36 of 86
1: 37 of 40
In the original experiment, a pH 11.1 solution of 0.5% NH4OH in water was
prepared and placed in a four liter amber glass bottle (from J.T. Baker). This bottle
is manufactured of a certain composition of borosilicate glass. This 0.5% NH4OH
solution leeched silicates from the amber glass bottle . These silicates traveled
throughout the plumbing of the Alliance 2695 Separations module. When these
silicates in the mobile phase were introduced into the pumping cylinders, they
caused a “sand paper-like” action as the plunger passed through the plunger seal.
This in effect caused the seal to be slowly scraped into tiny particles and over time,
deposited on the inline filter of the system.
A number of inline filters containing the deposits were submitted to the analytical
group for evaluation. The analytical group did find two types of deposits on the
inline filter, an orange deposit (polyethylene material from the plunger seal) and a
white deposit (silicates from the mobile phase). These results were confirmed by
FTIR.

38 of 86
39 of 86
Crépine en inox:10um

40 of 86
Too large an internal diameter tubing between the injector and column or column
and the detector will cause broad peaks.
Too narrow an internal diameter tubing anywhere in the system can cause high
pressure and may affect the separation in a gradient separation.

41 of 86
Il n ’y a pas de standardisation dans les raccords

attention quand on utilise plusieurs fabricants de colonne

42 of 86
Setting the ferrule the proper distance from the end of the tube is necessary to
minimize bad spreading. This can be a problem with the tubing fittings on the inlet
and outlet of the column if columns from multiple manufacturers are routinely
exchanged in a chromatographic system.
if ferrule is not seated may result in
leaks
'mixing chamber'
dead volume
broader peaks
Fitting Differences

43 of 86
An improperly installed ferrule can cause band spreading, especially on fittings
from the injector to the column and from the column to the detector
.

44 of 86
Le PEEK se dissout dans le chlorure de methylene. Il s ’assouplit dans le THF et il
est recommandé de ne pas l’utiliser avec des diamétres internes supérieur à 30
millieme

45 of 86
46
As a means of illustrating the effect that tubing has, let’s compare tubing internal
diameter and the resulting volume,

It is important to note that the achievable linear velocity will vary as tubing size
changes. For smaller diameters laminar flow characteristics become increasing
important because the difference in volume of the tubing and the volume of the peak
reduces and thus tubing effects are more significant.

47
48 of 86
Causes of Extra-column Band Spreading
Using the incorrect tubing diameter is a common cause for increased band
spreading. This graphic shows some of the commonly used internal diameter
tubing. The 0.009” diameter is recommended for all connections after the injector.
The two other dimensions can be used in the pump and from the pump to the
injector.

49 of 86
50 of 86
How to know that the spreading is becoming too high : Suitabillity

51 of 86
20 second gradient run on 1.0 x 50mm, 1.7µm ACQUITY UPLC BEH C18
Sample: 1µL injection of System Qualification Test Mix

52
53
 Troubleshooting, Tips and Tricks

In-line filter installation problem

In this example the in-line filter was not installed correctly, resulting in a tailing
peak with lower plate count. When the filter unit was replaced with a correct one the
peak shape was back to normal.

Rev 17310 Chapter 8: Page 54 of 57

 Troubleshooting, Tips and Tricks

Distorted peak shapes

In above example the reusable fitting was not installed correctly, resulting in tailing
peaks with lower plate count. When fitting was correctly assembled with a new
ferrule, better peak shape was observed.

Rev 17310 Chapter 9: Page 55 of 57

 Same injection volume,flow rate etc… the only change is the ID tubing

Diamètre interne
inches cm volume (microl/cm)
0.005 0.013 0.13
0.009 0.023 0.41
0.020 0.05 2.03
0.040 0.100 8.11

56 of 86
Let us start with the dead-volume. It is also called the extra-column volume, which
is somewhat clearer. It comprises the volume of an HPLC-system between the point
of injection and the point of detection, but excluding the part of the column that
contains the packing.
Therefore it includes the injection volume, the volume of the injector, the volume of
the connection tubing before and after the column, the volume in the endfittings of
the column, including the frits, and the detector volume. Actually, to be precise, it
includes half the injection volume and half of the detector volume.
We are concerned about the extra-column volume, because it causes extra-column
bandspreading. Bandspreading means that the peaks become broader as they flow
through the extra-column volume. This is undesirable since it may destroy some of
the separation achieved in the column. We would like to keep the extra-column
bandspreading as small as possible.

57 of 86
Approximate volume per inch of typical HPLC tubing.
0.009” OD 1 mL/in
0.020” OD 5 mL/in
0.040” OD 20 mL/in
The chart above shows that the inside diameter (ID) of a tube has a greater effect on
bandspreading than the length of the tubing in a chromatographic system. Also that a
change in system volume does not necessarily produce a comparable change in
system bandspread.

58 of 86
59 of 86
60 of 86
These are things which are unavoidable parts of the chromatographic process.
The efficiency of an HPLC system tells how well it will separate (resolve) peaks.
The narrower the peaks, the higher the efficiency and the better the resolution.

Time can be used instead of volume in these calculations, however all

measurements must be in time units.
Vtot becomes RT (retention time)
Peak Width can be taken from System Suitability calculations, be sure to use the
tangent (4s) result.

62 of 86
For those not familiar with testing systems, here is a recap for using a data system to
perform a bandspread test.

63 of 86
measurement of system band spreading

64 of 86
65 of 86
66 of 86
Impact of System Band Spreading
The impact of system band spreading on column performance (plate count) is
dramatic. The plate count can easily drop 20% if the band spreading volume is
increased from, for example, 70 µL to 120 µL.

67 of 86
68 of 86
69 of 86
Retention time precision is directly related to the precision of solvent flow.
Peak area can also be affected by erratic pump flow. For example, 0.995 ml/min to
1.005 ml/min yields a 3-4% area variation.

Visible leaks - Buffer trails out weep holes on head supports (510,600 pumps)

Strange Noises - Stop the pump immediately! Serious damage can result.

Ramp test - See Reference section.

GPV Test - See Reference section

70 of 86
71 of 86
72 of 86
73 of 86
74 of 86
The figure below illustrates the relationship between retention times and their slopes (vs. % methanol)
and composition for a separation of propyl and butyl paraben in water methanol mobile phases on a C18
reverse phase column. This figure indicates that at moderate values of k’, the retention time varies
strongly with composition. (k’ for propyl paraben at 50% methanol is 1.24; for butyl, k’ is 2.62)
This figure indicates that the retention time of a small molecule eluted at k’  3 is a sensitive function of
the composition of the mobile phase delivered. Consequently, the functioning of the gradient
proportioning valve can be verified by comparing the retention times of a moderately retained peak for
various combinations of valve pairs which result in the same nominal concentration. The slope of the
butyl paraben retention time at 50 percent methanol is –1.21 minutes per % methanol; i.e., a 1 percent
change in methanol concentration will result in a 1.21 minute change in retention time. A 0.1% change
in methanol composition will result in a retention time change of 0.12 minutes or 7.2 seconds which is 4
to 6 times greater than the typical flow rate precision specification of a modern hplc pump. A careful
selection of mobile phase, column dimensions and retention probe compound allows the demonstration
of suitable gpv functioning without contamination of the chromatograph with a strong UV absorber and
a miniumum of user intervention. This measurement does combine flow rate and compositional errors
for the retained (k’  3) peak, but the void peak will experience only the flow rate errors

75 of 86
76 of 86
77 of 86
 As a rule of thumb, if you make an error of 1% in the amount of organic solvent,
the retention time can change by between 5% and 15%, typically by about 10%.
This means that you have to measure the amount of solvent very carefully. The best
approach is to prepare the mobile phase gravimetrically rather than volumetrically.

78 of 86
To demonstrate how HPLC system design affects baseline performance, a simple gradient of
increasing Acetonitrile concentration was used employing a mobile phase modifier (0.1%
trifluoroacetic acid) which absorbed at low UV wavelengths. Figure 1 clearly demonstrates how
baseline stability is affected by HPLC system design and how Waters Alliance Technology yields
superior results to that of Brand A (Agilent) in this example. (Note: The increase in baseline from 0
to 26 min is normal since 214nm absorbance of TFA increases with increased concentrations of
acetonitrile.)

79 of 86
80 of 86
81 of 86
82 of 86
83 of 86
• Serial flow design.
• Two strain gauge transducers allow pressure monitoring each head independently.
• Solvent is pulled into the prime/vent valve only.

84 of 86
85 of 86
The effect of the delay volume is shown here. The original method was developed
on a 4.6 mm i.d. column, using a system with a large delay volume. When the
gradient was scaled down to a 2.1 mm column on the same system, a significant
delay of the separation was observed due to the large system delay volume. The
reason for this is the fact that at the lower flow rate it takes longer to purge the delay
volume. We then reduced the system delay volume by removing the gradient mixers
and using 9/100 tubing instead of 40/1000 tubing in the connections. This
significantly reduced the difference between the separation observed on the large
diameter column and the small diameter column.
By the way: to scale a gradient from one column to another, the correct way of
doing this is to keep the ratio of the gradient volume to the column volume constant.
This is very simple.

86 of 86
Waters Service Clients

# 87
Waters Service Clients

# 88
89 of 86
90 of 86
91 of 86
As we have mentioned before, the system delay volume or gradient delay volume
becomes important in gradient chromatography.The delay volume is the volume
between the point where the gradient is formed and the inlet of the column. I
comprises the volume of the gradient mixers, the pump, the injector and all the
connection tubing up to the column inlet.
Two methods for the measurement of the delay volume are shown here. Use
methanol as your A solvent, and methanol with a small amount of an UV absorber
as your B solvent. Then you can either run a linear gradient from 0 to 100 % B in 10
minutes (method 1) or a step gradient (method 2). In the first case, you draw a line
through the linear section of the gradient. Its intercept with the baseline gives you
the delay volume. In the step gradient method (method 2), the point where 50% of
the height of the step is reached is the gradient delay volume. The new Waters
Alliance solvent management system has an exceptionally low delay volume.
When you develop a gradient method that needs to be transferred to other HPLC
systems, it is important to measure the delay volume and report it along with the
method, just as any other experimental parameter.

92 of 86
Delay volume is the volume from the Gradient Proportioning Valve (GPV) through
the pump and injector to the head of the column.
Delay volume affects gradient separation. You cannot transfer a gradient method
developed on a 600 directly to a 2690, or the other way around. You will need to
modify the gradient table to get the same separation

As we have mentioned before, the system delay volume or gradient delay volume
becomes important in gradient chromatography.The delay volume is the volume
between the point where the gradient is formed and the inlet of the column. I
comprises the volume of the gradient mixers, the pump, the injector and all the
connection tubing up to the column inlet.
Two methods for the measurement of the delay volume are shown here. Use
methanol as your A solvent, and methanol with a small amount of an UV absorber
as your B solvent. Then you can either run a linear gradient from 0 to 100 % B in 10
minutes (method 1) or a step gradient (method 2). In the first case, you draw a line
through the linear section of the gradient. Its intercept with the baseline gives you
the delay volume. In the step gradient method (method 2), the point where 50% of
the height of the step is reached is the gradient delay volume. The new Waters
Alliance solvent management system has an exceptionally low delay volume.
When you develop a gradient method that needs to be transferred to other HPLC
systems, it is important to measure the delay volume and report it along with the
method, just as any other experimental parameter.

93 of 86
94 of 86
 Gradient Elution

Curve Profiles
Some HPLC systems allow different curve profiles. The most common curve type
is the linear one, where the change in the percentage of the eluent B as a function of
time follows a linear relationship. The convex type curves (2 to 5) shown here
change the eluting strength faster in the beginning and slower in the end. You can
achieve the opposite effect by choosing one of the concave curves (7 to 10).

Rev. 14109 Chapter 5: 95 of 25

Above is an example of correct chromatography and flowless flowing pressure
traces.

96 of 86
97 of 86
In a properly functioning solvent delivery system, the crossover from one plunger
delivering a mobile phase to the other plunger delivering mobile phase is not visible
much below 4000psi.
The tail on this chromatogram is shorter than the previous one. This indicates a
slightly lower flow rate.
The ramp pressure decay indicates there is a leak in the system at high pressure.
From the chromatogram it appears the leak does not affect the separation. This
system should be monitored and the leak corrected when convenient.

98 of 86
Notice that the retention time is twice what it was with the proper pump operation.
This indicates that one head is working perfectly and the other is not. This will
happen due to air in the affected head, a failed inlet check valve, a massive failure
of the plunger or the plunger seal.

99 of 86
The only indication of a problem (left) is the retention times have slightly increased.
The ramp test (right) exhibits the classic symptom of a seal leak. In this example,
the left plunger pressure stroke indicates a pronounced crossover, a different rate of
pressure increase and less pressure per stroke. It is important to remember that a
ramp test can rapidly demonstrate a seal leak, but seal performance is not indicated
by the decay rate. When the over pressure limit is reached and the pump stops, the
outlet check valves close and isolate the seals from the transducer.

100 of 86
The flowing pressure trace as well as the late retention times (left) indicate erratic
flow and a reduction in flow rate. The ramp test (right) also indicates that a fluidic
problem exists. Starting at 0 psi , the left plunger is first to deliver its pressure or
displacement stroke. A loss of pressure is observed as the right plunger displaces the
mobile phase. Upon first observation, the right plunger would appear to be the
problem, but upon closer examination, it is discovered to be the left head. The left
plunger cannot displace the mobile phase and build up pressure if either its inlet
check valve does not close, or if the outlet check valve on the right side does not
seal. Knowing that these check valves are functional, only the right inlet and left
outlet remain suspect. The conclusion is the the left outlet check valve failed. Inlet
check valve failures look different as we shall see later.

101 of 86
102 of 86
103 of 86
104 of 86
• Bubble detection is automatic and can not be switched on or off.
No more with Version 2.02
• Automatically activated above 650 psi and a C/D ratio of 1.8+

105 of 86
No Peaks - Does needle go into vial?
Do valves open and close?
Was sample withdrawn?

Poor Shape - Tailing, shoulders, doublets indicate valve problem

Determine system bandspread

Poor Area - Poor peak area reproducibility indicates syringe or valve

problem

Ghost peaks - Residue from previous injection

Needle wash solvent vs. mobile phase/sample
solvent
Contamination in injector

A system bandspreading test can be performed to check the efficiency of the

injector.

106 of 86
107 of 86
 Maintenance and Troubleshooting

Isolating Peak Shape Problems

First, you should notice is if all peaks are affected (have a too broad, strange peak
shape), or if only one or a few peaks are affected. If all peaks are affected, this is
nearly always related to a mechanical problem. This could be incorrect tubing or
fittings, a voided column, a contaminated in-line filter or a problem with the
injector. One chemical problem that can also affect all peaks is using an incorrect
sample solvent relative to the mobile phase. In all these instances, all peaks will be
affected. If only one or a few peaks are affected, the reason is probably “chemical”,
and can be caused by interaction of the specific analytes with the stationary phase
surface, or an overload of the column. A co-eluting compound that might come from
a previous injection, can also cause a distorted peak.

Rev. 14109 Chapter 9: 108 of 77

Waters Service Clients

# 109
110 of 86
.
Down Count (77 – 90)* ½ Seal W idth + stream to top of
61 – 95 idéal:83 lower seal

Up Count (310 – 337)* Lower sensor to stream

257 – 605

Seal W idth (72 – 82)* Top of lower seal to bottom of

57 – 100 lower seal

* Résultats typiques

111 of 86
112 of 86
113 of 86
114 of 86
The normal valve configuration is
Valve 1 (inject) open,
Valve 2 (syringe) closed,
Valve 3 (vent) closed,
Valve 4 (Needle Wash) closed.

Because of the restrictor loop the flow splits 95% flows through the sample loop,
5% flows through the restrictor loop.

115 of 86
116 of 86
117 of 86
118 of 86
Waters Service Clients

# 119
Look at this chromatogram for example.

In this chromatogram peak 5 represents a contamination.

We don’t know where it comes from. It could be sample carry over due to problems
with the injector, it could be the mobile phase or . . .
. . . . it could be from the vial.

120 of 86
 Rev. 30012012
Troubleshooting, Tips & Tricks

Definitions
Contamination is a very general term that describes the presence of any unwanted
substance in a chromatographic system that appears as either peaks or high
background noise. The chromatographic data shown here is an example of system
contamination which resulted from a contaminated lot of mobile phase additive and
appeared in the gradient baseline without an injection being made on the system.
Contamination can be difficult to pinpoint but by performing a few different types
of blank runs, its source can often be pinpointed.

Chapter Nine: Page 121 of 67

 Rev. 30012012
Troubleshooting, Tips & Tricks

Carryover
Carryover is defined as sample material that remains in the system after an injection
and appears as peaks in subsequent injections and compromises quantification.
Ideally we want to eliminate carryover entirely from our analysis, however it is
usually sufficient to minimize it to an acceptable level for the separation (which
requires it to be a set amount below the limits of quantification). If we do find that
we have carryover for a particular analysis, the type of carryover will determine
how we approach improving it. There are two different types of carryover. The first
is dilution carryover. It appears in the injection immediately after the sample
injection and is washed away with multiple blank injections or by increasing the
wash volumes or times. Adsorptive carryover results when the sample adsorbs
somewhere within the system, such as on the column or the materials inside the
wetted flow path of the injector or system and elutes much slower. To reduce
adsorptive carryover requires changing the composition of the wash solvent, the
materials within the fluidic path, or depending on where the adsorption is occurring,
the method.

Chapter Nine: Page 122 of 67

 Rev. 30012012
Troubleshooting, Tips & Tricks

Carryover example
This is an example of a typical carryover scenario. The blank injection prior to the
sample analysis is blank, however the blank injection immediately following the
sample injection has the peak of interest in the baseline. However, the type of blank
injection we perform can be very telling of what the predominant source of
carryover in the system is. Let’s look at some of the different types of blanks we
can run to help troubleshoot the source of carryover.

Chapter Nine: Page 123 of 67

 Rev. 30012012
Troubleshooting, Tips & Tricks

Measuring carryover
When we think of a blank injection, most of us will run the sample diluent or the
matrix without the analyte. The carryover is then expressed as the response from
the analyte of interest that appears in the blank divided by the response of the peak
in the previous sample injection and is expressed as a percentage. Instrument
manufacturers strive to create an injector that yields a very low carryover spec, for
example the I-Class specification is < 0.001%. However, simply observing that
there is carryover does not give us enough information on how to improve carryover
performance. By injecting multiple blanks after our sample injection we can assess
if the carryover is a dilution effect (easily removed by subsequent injections) or an
adsorption effect (not easily removed by subsequent injections) or maybe a
contamination issue. Additionally, by injection mobile phase blanks versus matrix
blank can help us assess whether the carryover is exasperated by the matrix.

Chapter Nine: Page 124 of 67

 Rev. 30012012
Troubleshooting, Tips & Tricks

Measuring carryover
Besides a traditional blank injection, it is advantageous to run a zero volume blank
injection. This can pinpoint any sources of contamination within the vial or the
blank. Even more telling is being able to run the gradient without making an
injection. This will give us information on whether the observed carryover is
associated with the injector or another component within the system. If the
carryover peak is present without making an injection we know that our main source
of carryover is not injector but either contamination in the system or the solvents or
it could be related to method carryover. We will discuss sources of method
carryover later in this chapter.

Chapter Nine: Page 125 of 67

 Rev. 30012012
Troubleshooting, Tips & Tricks

Measuring carryover
Sometimes when we are performing highly sensitive analyses, contamination from
the mobile phase can appear as carryover. Sometimes this appears as an interfering
analyte that elutes at the same time as our peak of interest and appears to be
carryover, or in some instances glassware and laboratories can become
contaminated with our analyte of interest and be transferred into the mobile phase
and it is eluted off the column during the gradient in the same manner as the
injected sample. To troubleshoot this type of contamination we can implement a
series of blank injections which have different re-equilibration times of the initial
conditions. If the peak area of the ‘carryover’ peak increases with increasing
equilibration times, then the source of carryover/contamination is the mobile phase.

Chapter Nine: Page 126 of 67

127 of 86
128 of 86
129 of 86
Vials are also the least expensive component of your HPLC system . . .
. . . But they could be a big source for problems you encounter

These problems can be of mechanical or chemical nature.

Mechanical problems are:

- Reproducibility issues with injection volume
- Septum of the vial was falling off or part of the septum was cut by the
injection needle
- Injector damage or Robotic malfunctioning
- Too much residual volume (often they want to take ALL sample from the vial)
Chemical problems are:
- Evaporation of sample or solvent
- Extra peaks
- Degradation of sample or analyte

130 of 86
Let’s go back to the troubleshooting and into more detail now.

The majority of problems with vials are coming from choosing the wrong septum.
These problems could be :

Contamination - which result in extra peaks - due to incompatibility of septum with sample
or eluent.

Evaporative loss . . . from the fact that the septum is not closing tight enough.

Reproducibility issues . . . especially for repetitive injections.

Septum coring . . . the injector needle cutting off part of the septum.

Needle damage . . . when using very hard and thick septa over a period of time.

Finally Septum dislodging . . . Septum falling out of the cap . . . during transportation or
during injection process.

131 of 86
How to choose the right septum is shown in this table

Start from the bottom !

The most recommended septum today might be the PTFE/Silicone pre-slit version.
It re-seals after each injection (reasonably good)
No issues with septum coring
No reproducibility issues as it allows adequate venting.

Only if the sample or solvent is very volatile, you may to go to non-pre-slit.

Finally, if Silicone is not compatible with the sample or the solvent you may go to
PTFE.

132 of 86
Knowing the right septum, we have several choices for the cap design.

The crimp cap is still the most widely used.

It is very cheap and it closes well . . . . BUT you need a crimping tool and often you
may have difficulties to re-open the vial again.

The snap cap is the closest to the crimp cap. It even fits on some crimp vials.
It is extremely easy to use and does NOT require any tools.

The universal solution is the screw cap.

With the development of our new screw caps, they become more and more
attractive.
The next slide shows you why.

133 of 86
134 of 86
Waters Service Clients

# 135
Waters Service Clients

# 136
137 of 86
Another problem may appear when you do several injections out of ONE vial.
This problem does NOT appear if you do ONLY ONE SINGLE injection.

What you may see is the following:

First injection (and only the first one) gives lower peak area than all the following injections.
Question :
Is this an injector (or needle) issue or does it come from the vial ? Could be one or the other.

What could happen is that you have a vial with closed (and probably thick) septum.
When the injection needle passes such septum, the septum seals around the needle and forms
a vacuum.
The vacuum draws some sample back out of the needle. So, the volume the needle takes is
less what we wanted.

This effect will be even stronger when the vial is filled up to the top.

Troubleshooting Tips :
First, never OVERFILL a vial.

Second, if you see different peak areas, make the following test :
remove cap and septum from the vial, perform several injections out of the same vial and
measure the peak area.
If they are now all the same, it’s the septum you need to change.
138 of 86
Not only the septum can cause a problem in terms of compatibility with the sample and
solvents.
Even the vial body - either colourless glass, amber glass, or certain plastic - can contribute to
interaction with the sample or the solvent. Some compounds may show degradation or
others can react with the surface itself.

For clear glass vials you can have 2 different types of glass.
One is called Type 1 Class A which is borosilicate of high inertness.
The so called Type 1 class B borosilicate is a more alkaline glass.
ALL OUR NEW VIALS are Type 1 Class A !

Amber glass vials are always Type 1 class B and are used for light sensitive compounds.

Where samples like certain Fluoro-compounds are aggressive to glass, a nonionic, non
reactive polypropylene might be the most appropriate solution.

And the ultimate solution - for example for some sensitive protein or peptide studies - are
silanized glass vials which will be available beginning 2002 from Waters .

139 of 86
Waters Service Clients

# 140
Waters Service Clients

# 141
Waters Service Clients

# 142
Waters Service Clients

# 143
Waters Service Clients

Selecting a Needle-Wash Solvent

Use a needle-wash solvent based on the sample and mobile phase chemistries, making sure all
solutions/buffers are miscible and soluble. A gradient separation typically involves the initial use of a
weak solvent combination, whereas final conditions usually have a combination of stronger solvents.
In general, it is best to choose a needle-wash solvent similar to final gradient conditions. For
example, in a separation with an initial gradient of 80% water:20% methanol and a final gradient of
10% water:90% methanol, an effective needle-wash solvent would be 20% water:80% methanol.
The choice of a needle-wash solvent similar, but not equivalent, to final gradient composition
ensures solubility of early eluting peaks. When sample solubility in the mobile phase is pH-
dependant, choose a needle-wash solution that is similarly pH- adjusted. If acids are used, avoid
long-term exposure of the needle-wash system to halide-containing acids.
Using samples containing strongly hydropholic components may require experimentation to
determine the best needle-wash solvent. In these cases, the use of a needle-wash solvent such as 20%
water:80% methanol with tetrahydrofuran (THF) or dimethylsulfoxide (DMSO) [1-5 mL THF or
DMSO/99-95 mL needle-wash solvent] may be superior.
Table 3-5 lists some needle-wash solvents recommended for use with certain mobile phase
conditions. High sample concentrations may require other needle-wash solvents.

# 144
Waters Service Clients

# 145
Waters Service Clients

Priming the Needle-Wash Pump

The needle-wash pump flushes the needle in the sample management system, preventing carryover
of sample between injections. The needle-wash also extends the life of the injector seals by removing
buffered mobile phase and sample from the needle.
To prime the needle-wash pump:
1. Be sure the needle-wash fluidic lines are properly installed, as described in Section 2.6.6,
Installing Needle Wash and Plunger Seal Wash Reservoirs.
2. Press the Diag screen key in the Main screen. The Diagnostics screen appears (Figure 3-14).
3. Press the Prime NdlWash screen key. The 30-second needle-wash procedure begins. If solvent
does not flow out of the waste tube within 30 seconds, press the Start Again screen key.

# 146
Waters Service Clients

Priming the Needle-Wash Pump

The needle-wash pump flushes the needle in the sample management system, preventing carryover
of sample between injections. The needle-wash also extends the life of the injector seals by removing
buffered mobile phase and sample from the needle.
To prime the needle-wash pump:
1. Be sure the needle-wash fluidic lines are properly installed, as described in Section 2.6.6,
Installing Needle Wash and Plunger Seal Wash Reservoirs.
2. Press the Diag screen key in the Main screen. The Diagnostics screen appears (Figure 3-14).
3. Press the Prime NdlWash screen key. The 30-second needle-wash procedure begins. If solvent
does not flow out of the waste tube within 30 seconds, press the Start Again screen key.

# 147
Waters Service Clients

– The operator has the ability to choose one of three options for needle wash duration as
shown below.

– “Normal”, the default value, was the standard, non-programmable value in previous
versions of software. The needle wash portion of the injection cycle operates for
approximately 13 seconds when “Normal” is selected.
– In cases when excess carryover is seen, you may choose either “Double (2X)” or
“Extended (4X)” to increase the length of time the needle is washed.

– The needle wash portion of the injection cycle is increased either by an additional ~13
seconds, or by an additional ~39 seconds, respectively if “Double (2X)” or “Extended
(4X)” is chosen as compared to “Normal”.

# 148
Waters Service Clients

Replace frit?
• The lower needle wash frit is intended to be replaced by users on regular basis
• Soak or sonicate the needle wash frit in the wash solvent prior to use to ensure
proper “wetting”.

Seal Pack Rebuild Kit W/Needle (WAT271019)

The Seal Pack Rebuild Kit includes all of the parts necessary to replace seals in
the seal body, including the high pressure seals. Replacement of these seals
requires use of HPMV Tool Kit (WAT045427).
The Rebuild Kit includes: one Teflon washer, three needlewash filters, one
upper spacer, one lower spacer, one alignment tube, one wash tube seal, one
needle, two high pressure seals, one compression screw, and one ferrule.

# 149
Waters Service Clients

A Bottom Port (conventional) design needle

B Waters traditional needle used on 717+ and 2690/2695
C Waters Alliance HT needle – designed to not core septa/cap-mats samples near
bottom of container

– The PerformanacePlus needle (p/n 700001247) was made to

provide a solution to the destruction of seal packs by needles that
have been corroded by acids or bases. It is made of a corrosion
resistant nickel/cobalt alloy (MP-35N).

– These new needles are much less prone to etching by acids;

therefore it may adsorb much less protein and yield lower
carryover than the standard stainless steel needle for samples in
acidic solutions.

– This new needle is not standard equipment and is only

available as a spare part
• For very “sticky” applications (e.g. biologicals), consider using the P+ needle,
p/n 700001247

# 150
151 of 86
Pour le 717
Vitesse et taille sont mis dans CONFIG .

Pour le 2690
Taille de la seringue est Config
La vitesse est dans la méthode instrumentale page 2

152 of 86
This slide shows the effect when the sample is injected using the extraction solvent.
Note that the mobile phase contains only a relatively small concentration of organic
solvents. Under these circumstances, the extraction solvent is a strongly eluting
solvent, therefore peak broadening and peak distortion can occur at large injection
volumes. Due to the increased concentration of organic solvent in the mobile phase
at the column top, the equilibrium at the column top is disturbed and the sample
does not bind uniformly. This results in the distortion of the peaks. The problem
goes away, when the sample is evaporated and reconstituted in HPLC mobile phase.
(to the speaker: the final complete HPLC conditions for this analysis are listed in a
later slide)

153 of 86
The differences are due to the amount of time/volume for the gradient to reach the
column.

These differences can be even more pronounced when the method was developed on
one type of pumping system/injector and transferred to a different pumping
system/injector. (Delay volume differences)

When doing in-line mixing, isocratic separations are not generally affected by delay
volume differences. The exception is when the volume is reduced so much that
complete mixing does not occur.

154 of 86
155 of 86
Inline Filter - Usually 2m pore size placed between the injector and the column.
Prevents particulate from reaching the column.

Guard Column - A short section of column placed before the analytical column to
scavenge damaging sample components from the analytical
column. Need to consider the packing material.

A Guard column is not a substitute for sample preparation.

156 of 86
3: 157 de 45
158 of 86
159 of 86
 Troubleshooting, Tips and Tricks

Effect of blocked in-line filter

The limited channels in a plugged filter act as a restrictor to increase pressure and
will cause band spreading and/or poor peak shapes, or even as seen here split peaks.
In this case changing the inlet filter in the column restored column performance
completely.

Rev 17310 Chapter 9: Page 160 of 57

Waters Service Clients

# 161
High pressure - particle size, solvent composition
Temperature - retention time can vary with changes in temperature
Retention time shifts - dirty column (chemical)
Ghost peaks - clean column with strong solvent
Poor peak shapes - voids, dirty, packing degrading
Loss of efficiency - voids, dirty, packing degrading
Direction of flow - reversing flow direction can cause bed to shift
Pressure shocks - high to low or low to high pressure almost instantly
pH compatibility - 3 - 7 for silica based columns
Compatible end fittings - ferrule set on tubing
Baseline drift - long retained material bleeding off column
Column storage - wet in compatible solvent, temperature, mechanical shock

162 of 86
163 of 86
164 of 86
Perhaps the most common problem is tailing peaks.
Column overload - too much mass, too large a volume of sample
Competing mechanisms - adsorption on C18 and silanol sites.
Poorly packed column - results in voids
Voids - can be induced from pressure shocks, mechanical shocks, drying out
Viscosity too high - sample viscosity
Bad guard column - sample bleeds onto analytical column
Need ion-suppression or PIC - application dependant
Injector bleeding sample - not a column problem, but keep this possibility in mind

165 of 86
166 of 86
Here we see mass overload. Indicated by a rapid detector response along the y axis.
How to remedy the problem.
We dilute the sample by 50% then reinject.
One can observe better resolution and slightly rounded leading edge of the peak.
We then dilute the sample by 50% again until we observe a much more pronounced
leading edge of the peak of interest.

A good approach to a loading study is to increase the amount injected by a factor of 2.

This is fine enough to find the point of maximum load but course enough to cover a large
concentration quickly. It is often not necessary to perform measurements on the
chromatogram. A visual inspection is often sufficient. One can see the loss of resolution in
front of the peak as the amount of sample is increased beyond 2.4mg for the scaling
column. One can see that the peaks behind the major peak are barely affected by the
overload of the major compound, even up to a load ofover 19mg

167 of 86
168 of 86
(Course Name)Method Transfer from HPLC to
UPLC®

COMPANY CONFIDENTIAL 169

(Course Name)Method Transfer from HPLC to
UPLC®

COMPANY CONFIDENTIAL 170

(Course Name)Method Transfer from HPLC to
UPLC®

COMPANY CONFIDENTIAL 171

(Course Name)Method Transfer from HPLC to
UPLC®

COMPANY CONFIDENTIAL 172

(Course Name)Method Transfer from HPLC to
UPLC®

COMPANY CONFIDENTIAL 173

174 of 86
This chromatogram of caffeine, aspartame, benzoic acid and sorbic acid shows the
repeated shoulders that suggest a column void or injector failure as probable causes.

175 of 86
Weak acids and bases that are run in a mobile phase that is near their pK may show
tailing even on highly functionalized columns. This is typically explained as an
ionization problem rather than a column interaction problem.

The pKa for Aspirin is 3.5

176 of 86
177 of 86
178 of 86
179 of 86
180 of 86
181 of 86
182 of 86
183 of 86
184 of 86
185 of 86
Drift - Change in baseline over one hour
Noise - Changes in baseline with a duration less than one minute
Wander - Changes in baseline with a duration between 1 minute and one hour
Bubbles - Out gassing before or inside detector cell
Light source - Lamp or LED, includes power supply in this category
Flowcells - Be very careful if you clean the flow cell
Time constant - “Averages” signal for a specified time, used to reduce noise

186 of 86
187 of 86
188 of 86
189 of 86
190 of 86
191 of 86
192 of 86
193 of 86
194 of 86
195 of 86
196 of 86
197 of 86
198 of 86
199 of 86
200 of 86
•Important Considerations when setting the wavelength range:
• HPLC grade water has no significant absorbance in the UV.
• HPLC grade solvents like methanol and acetonitrile have UV cutoffs, I.e., at a certain
wavelength they have strong absorbance.
• Buffers which are used in HPLC have UV cutoffs as well.
• The final mobile phase, for example 40% Water and 60% Methanol, has some UV cutoff.
• Freshness and degassing of solvents does make a difference!

201 of 86
202 of 86
203 of 86
•Effect of Background Absorbance:
• Background absorbance reduces the working range.
• Most mobile phases have absorbance.
• Depending on the UV cutoff of the mobile phase and the selected wavelength which the user
is monitoring there can be varying degrees of background absorbance.

204 of 86
205 of 86
206 of 86
207 of 86
208 of 86
Waters Service Clients

# 209
210
Note the apparent loss of resolution between peaks 6 through 9 by lengthening the
time constant.

211 of 86
3: 212 de 45
3: 213 de 45
Waters Service Clients

# 214
215 of 86
216 of 86
Waters Service Clients

# 217
Waters Service Clients

# 218
219 of 86
Troubleshooting Diagram
You can use this diagram to troubleshoot chromatographic problems. For
example, if you see problems with the reproducibility of the retention times, the
likely reason is either a problem with the pump or the mobile phase composition.

220 of 86
221 of 86
222 of 86
223 of 86
224 of 86
225 of 86
226 of 86
227 of 86
228 of 86
229 of 86
230 of 86
231 of 86
232 of 86
233 of 86
234 of 86
235 of 86
236 of 86
237 of 86
238 of 86
Noise can be introduced to a system by a loose analog signal cable, or by a bad A/D
device.

Peak generators create a chromatogram electronically so the identical signal can be

input to any number of systems. This will allow you to determine if the problem
with a chromatogram is caused by a problem in the “wet” system or in the data
system.

239 of 86
240 of 86
241 of 86
242 of 86
243 of 86
244 of 86
245 of 86
246 of 86
247 of 86
248 of 86
249 of 86
250 of 86
251 of 86
What to do?
Inject acetone solution to make a peak.

252 of 86
253 of 86
254 of 86
Common sources include switching valves, compressors, muffle
furnaces, fraction collectors, lighting fixtures, “dirty” power source,
power conditioners.

255 of 86
256 of 86
257 of 86
258 of 86
259 of 86
260 of 86
261 of 86
262 of 86