Making your LC Method Compatible

with Mass Spectrometry

Technical Overview

Author
Edgar Naegele
Agilent Technologies, Inc.
Waldbronn, Germany

Introduction
The power of traditional high-performance liquid chromatography with UV detection
(HPLC-UV, DAD) can be readily extended by simply coupling a mass spectrometer
(MS) to an existing system. In the early years of liquid chromatographic mass spec-
trometry (LC-MS), this coupling was considered exotic and complex. After more than
15 years of refinement, LC-MS systems are robust and easy to use, and provide speci-
ficity unattainable by any other detection scheme. With the increased analytical capa-
bility, challenges may be tackled from several different and complementary directions.
This Technical Overview provides an overview to the integration of MS into your exist-
ing HPLC system, showing the analytical strengths that MS brings and the ease with
which it can be added.
Part 1 presents a basic introduction to the technique, applications that highlight its
power, and some considerations on sample preparation methods. Part 2 describes the
theory of electrospray ionization (ESI), which is the most common technique. Part 3
describes less common alternative techniques like atmospheric pressure chemical
ionization (APCI) and atmospheric pressure photoionization (APPI) which are quite
useful for various applications.
Table of Contents
Introduction 1

HPLC/UV/MS: Part 1, Introduction and Applications 3

Introduction 3
The Basics of Atmospheric Pressure Ionization (API) Mass Spectrometry 3
Adapting Existing LC Methods to MS 4
Applications Using Electropsray Ionization (ESI) 4
Impurity Identification by LC-MS Coming from a Non-volatile Buffer LC
Method and Identification of Unknown Impurities 4
Identification of Impurites of a Topical Steroid Formulation in Oil at Low Levels 6
Considerations for LC Performance with MS 8
Conclusion 8
HPLC/UV/MS: Part 2, Theory of Electrospray Ionization Mass Spectrometry
and its Coupling to HPLC-UV 9

Introduction 9
Electrospray Ionization 9
ESI Related to Chromatography Conditions 10
Mobile Phase Buffer Selection 11
Considerations for LC Performance with MS 12
Conclusion 12
HPLC/UV/MS: Part 3, APCI (Atmospheric Pressure Chemical Ionization)
and APPI (Atmospheric Pressure Photoionization) as Alternative Ionization
Techniques for MS 13

Introduction 13
Atmospheric Pressure Chemical Ionization (APCI) 13
APCI Related to Chromatography Conditions 15
Atmospheric Pressure Photoionization (APPI) 16
The Multimode Source: Combined ESI and APCI 17
Conclusion 18

2
HPLC/UV/MS: Part 1.
Introduction and Applications
HPLC-UV capabilities enhanced by the addition of MS as demonstrated by
accelerated stability samples of both a pharmaceutical active and a formulation

Introduction The Basics of Atmospheric Pressure Ionization

(API) Mass Spectrometry
HPLC is the mainstay analytical technique of today’s chemical,
pharmaceutical, and bioanalytical industries. MS is a powerful Mass spectrometry detects ionized analytes from the LC eluent
tool for increasing quantitative capability, providing peak identi- after UV detection. The sample is desolvated, ionized, analyzed
fication, and elucidating the structure of unknowns. The addi- by mass/charge ratio, and detected. Based on the mass-to-
tion of MS to HPLC with UV detection to form an LC-UV-MS charge ratio (m/z), the mass spectrometer measures unique
system (see Figure 1) adds a rich additional data source with- masses from individual analytes that are used to confirm
out compromising existing data collection. This allows the ana- known compounds and identify unknowns. Analyses are easily
lyst to monitor masses relevant only to the target analytes, and performed in both the positive ion mode and the negative ion
the resulting increased specificity provides multiple advan- mode. In positive ion mode, protonation occurs on a basic site
tages: of the molecule. In negative ion mode, deprotonation produces
a negatively charged molecule.
• Improves sensitivity, resolution, throughput, and productivity
MS detection for samples in liquid phase (HPLC) is typically
• Integrates easily with HPLC-UV (DAD) accomplished by atmospheric pressure ionization. This method-
• Enables multiplexed experimental optimization (since the ology is compatible with a broad range of compounds, has fem-
two detection methods are entirely different mechanisms, togram-to-picogram sensitivity, and delivers both qualitative
they combine to yield a single, powerful, orthogonal and quantitative information. There are three typical atmos-
approach) pheric pressure ionization methods, with the method deployed
• Facilitates problem-solving, as will be described largely dependent on the polarity of the analytes: ESI, APCI, and
APPI. Of these three methods, ESI is the most common. See
following chapters for more detailed discussion on the theory
of electrospray as well as additional atmospheric pressure ion-
ization techniques.

+ =

HPLC MS Highly resolved, mass specifications

Figure 1
Multiplexing of HPLC with MS yields a powerful tool.

3
Adapting Existing LC Methods to MS spectrometry to HPLC-UV. Accelerated stability studies are rou-
tinely conducted to estimate product shelf-life in the pharma-
LC-UV mobile phases typically contain non-volatile buffers
ceutical and biopharmaceutical industries. Impurities or degra-
which build up in the MS. However, adapting to a volatile buffer
dates detected by HPLC-UV-MS are quickly and easily identified
at the same pH is generally a simple matter. For example, if
by molecular weight confirmation. Also, the data show that
working at pH 3 with phosphate, using formate will also bring
when an active is tested in formulation, the source of the impu-
the pH of the mobile phase to 3 and interface well with the
rity can be determined as coming from the active or the formu-
mass spectrometer.
lation. These results can then be used to make improvements
Since MS detection requires the formation of ions, the mobile in both products and processes.
phase should be used to create charged analytes. This means
that mobile phase pH and sample pKa information are critical. Impurity Identification by LC-MS Coming from a
Selecting a mobile phase pH that will give positively charged or Non-volatile Buffer LC Method and Identification
negatively charged analytes will increase sensitivity. Low pH of Unknown Impurities
will generally ionize basic compounds. Acidic compounds
require more careful mobile phase pH selection because they In the example given in Figure 2, an HPLC-UV method contain-
are more likely to have pKa values in the pH 2-to-5 region. ing phosphoric acid was developed for studying the stability of
Neutral compounds can also be analyzed when they associate an active pharmaceutical ingredient. During the analysis of
with ions such as acetate or ammonium that are in the buffer. accelerated stability samples, new unknown peaks were
Applying a voltage to the electrospray probe will induce ion for- observed. An HPLC-UV-MS compatible method with compara-
mation, but good mobile-phase selection to provide preionized ble performance to the phosphate method was needed to
molecules dramatically increases ion formation, and therefore determine the molecular weights of the impurities.
sensitivity. Volatile formic acid replaced the phosphoric acid/triethylamine
components used in the original LC method. The resulting
Applications Using ESI LC-MS analysis showed a retention-time shift, which was
The analyses of accelerated stability samples of both a phar- acceptable for this particular study, but the chromatography
maceutical active and a topical steroid in formulation were was improved with less fronting, yielding excellent detection of
chosen to demonstrate the benefits of adding mass the active pharmaceutical compound (APC).

DAD1 - B:Sig=280,10 Ref=550 Agilent 1200 Series LC System with Diode Array
API LC system:
100 Detector
MS system: Agilent 6410 Triple Quadrupole Mass Spectrometer
75
Injection volume: 10–40 µL
mAU

50 Column: Agilent ZORBAX SB C18 4.6 mm × 150 mm Stablebond,

5 µm particies
25 Column temp.: 40 °C
0 Flow rate: 1.5 mL/minute
2 4 6 8 10 12 14 16 18 20 22 24 Detector l: Diode array, 230 nm, 280 nm individual signals.
Acquisition time (min) 220–400 nm stored
Ionization mode: Electrospray ionization, positive and negative ion
LC-UV method MS mode: Scanning, m/z 110–1100
Solvent A: 0.03% Triethylamine in water, pH adjusted to 3.0 with Gas temp.: 300 °C
phosphoric acid. Gas flow: 11 L/min
Solvent B: Acetonitrile Gradient program
LC-UV-MS method Time %A %B (ACN)
0 90 10
Solvent A: 0.1% Formic acid in water, pH = 2.7
7 90 10
Solvent B: Acetonitrile 17 73 27
20 40 60
26 40 60
26.1 90 10
30.1 90 10

Figure 2
Method comparison for Active Pharmaceutical Ingredient analysis (API).

4
The LC-UV-MS analysis of an actual accelerated stability
API*
sample shown in Figure 3 indicates a match of the five impuri-
+ TIC Scan
ties between the MS total ion chromatogram (TIC) and the UV 40

Response (%)
30
trace with adequate resolution of known impurities. MS
20
After a degradation period of two months, four known impuri- 10
0 Changes to chromatography
ties were identified in the degraded samples. More importantly, G conditions have not affected
A C DE
two unknown impurities were also present above the 0.1% 46 separation or resolution of
known impurities
threshold. These unknown impurities labeled by the observed 44
*API- active pharmaceutical
molecular weights 358 and 497 are shown in Figure 4. 42 DAD-UV
ingredient
Note that impurity B is not seen in the TIC from the MS data. 40
3 6 9 12 15 18 21 24
The absence of the peak in the TIC does not mean impurity B Acquisition Time (min)
was not detected, although differences in intensities are to be
Figure 3
expected when comparing the UV and MS data. The TIC scans
Comparing LC-UV and LC-MS data.
the entire range of molecular weights selected by the user and
within the instrument’s range. Knowing that the impurity is
present from the UV (DAD) data and knowing the identity, one
can readily find the impurity in the mass spectrometry data by + TIC Scan Unknown 358
Unknown 497
searching for impurity B’s molecular weight. This is easily done 70 Two new unknown
impurities need to
by extracting the molecular weight in an extracted ion chro- 50 be identified
matogram (chromatogram of only specific molecular weight 30 MW: 358
Response (%)

ions). 10 MW: 497

DAD1 - B:Sig=280,10
This demonstrates how easy it is to confirm impurities when 1.2
Impurity B
Impurity A Although you can see
Impurity D them in UV, you cannot
mass spectrometry and UV data are complementary, as in 0
Impurity G
get information regarding
Figure 4, and reference material is available, as in Figure 5. In _1.2
structure without
running a standard
Figure 5, the mass spectral data of impurity B match that of the
reference standard in both the positive ion mode and the nega- 3 6 9 12 15 18 21 24
tive ion mode. The molecular weight, 165, is confirmed by the Acquisition time (min)

protonated molecular ion in the positive ion mode (M+H

Figure 4
m/z 166) and the deprotonated ion in the negative ion mode
Acquiring more information with mass spectrometry: Analyzing impurities
(M-H m/z 164). Also, ion fragments from the molecular ion pro- in two-month stability samples.
duce a spectral match between the sample and the reference
standard (m/z 138 and m/z 136).
Collecting and
evaluating positive Reviewing MS spectral
and negative MS based on DAD peak data Compare sample
data to standard to
verify impurity
+ Scan (3.94-4.16 min)
×10 2 138.0 166.0
Sample

MS+
+ Scan (3.89-4.13 min)
×10 3
111.0 138.0
166.0
Standard
120.0 204.0
148.0 187.9 219.8 260.9
100 140 180 220 260

×10 2 _ Scan (4.02-4.19 min)

112.6
136.0 Sample
163.9

MS_ _ Scan (3.82-4.17 min)

×10 3 136.0 Standard
113.0 164.0

100 120 140 160 180

Counts vs. Mass-to-Charge (m/z)

Figure 5
The power of tandem UV and MS: Matching of known impurities.

5
Another difference between the UV and MS scans is the pres- 6 ×10 7 Coeluting peaks not of
ence of new and unknown impurities detected at 18.8 min. An interest (See EIC for 653 TIC
below)
expanded view of the TIC, the UV (DAD) scan, and two
extracted ion chromatograms (EIC) at m/z 653 and 498 are
0
compared in Figure 6. The TIC shows two co-eluting peaks Peak of interest DAD

mAU
_
0.5
while in the UV (DAD) chromatogram, we see one main, _
1
unidentified peak of interest. By extracting specific ions on EIC,
compounds of molecular weights 653 and 498 are shown to 1.5 ×10 6
Not of interest +EIC 653
elute at similar retention times. The mass peak at m/z 498 cor-
responds to the peak of interest in the UV trace.
6 ×10 6
The positive and negative mass spectra for the peak of interest Peak of interest +EIC 498

at 18.8 min are shown in Figure 7. The molecular weight is con-

firmed to be 497 from both the positive and negative ion MS
18.2 18.4 18.6 18.8 19 19.2 19.4 19.6 19.8 20
traces. Acquisition time (min)

Identification of Impurities of a Topical Steroid Figure 6

Finding the unknown impurities using UV and MS.
Formulated in Oil at Low Levels
New challenges arise from accelerated stability samples of
active ingredients in formulation. The source of new impurity 100% + Scan
498.2 Spectral data from
peaks becomes ambiguous as it is not always clear if the impu- positive and
negative ESI used
rities come from the formulation matrix or the API. For the fol- for confirmation
lowing example, a MS compatible HPLC-UV method was devel-
oped for formulation analysis. Many new peaks were observed 260.3
when analyzing the accelerated stability sample. Although
%BPI

100 200 300 400 500 600 700 800 900 1000
some peaks were identified based on retention times, others Mass-to-change (m/z)
70% _ Scan
remained unknown. The steroid, formulated at 0.1%, required 496.0
identification of impurities at low ppm levels (~0.1% of the
API). When MS detection is added to the LC-UV method, impu-
rity identification becomes more complete and accurate.
A six-month accelerated stability sample of a topical steroid in
100 200 300 400 500 600 700 800 900 1000
oil at low levels showed significant impurities by LC-UV as Mass-to-change (m/z)
seen in Figure 8. Three impurities had similar retention times to
peaks in the placebo, but two were present in much greater Figure 7
amounts and had retention times similar to known API impuri- Mass spectral data identifies the peak of interest: Molecular weight of 497
confirmed.
ties. Initially, the peak at 21.1 min was assigned to the formula-
tion matrix, the peaks at 27.6 and 39.3 were assigned to known
API impurities, and the peak at 48.7 was unknown. These iden-
DAD1 A, Sig=238,4 (PLACEBO)
tifications were made by LC-UV, but were they correct? DAD1 A, Sig=238,4 (6 month SAMPLE)

Based on the UV response, it is very difficult to determine if the

23.3

10
48.7

API
larger impurity peaks in the six-month sample are due to matrix
or the API. When MS detection was added, unequivocal deter- 8
39.3

mination of the source of the impurities was possible.

27.6

6
mAU

25.804

Based on UV data (RT Comparison)

4 RT 21.1 min - Matrix affect
21.1

RT 27.6 min - Degraded API

2 RT 39.3 min - Degraded API

Are these peaks matrix or

0
degraded API?

0 10 20 30 40 50
Acquisition time (min)

Figure 8
Comparing samples: Changes in chromatography with six-month sample.

6
The peak at 27.6 minutes was initially assigned as an API *MSD1, ES-API, Pos, Sample MSD1 427, ES-API, Pos, Sample
27.2

468.2 M+H+ACN
impurity by LC-UV. Figure 9 shows the mass spectra associated 427.2
M+H
Max: 4637
80 4000
with the impurity found in the sample. The positive ion spec-
trum (top left) shows the protonated molecular ion at M+H 40 2000

500.2

677.2
m/z 427 for the sample impurity. The negative ion spectrum 0 0
200 400 600 m/z
(bottom left) does not show the deprotonated molecular ion 0 10 20 30 40 50 min

Max: 488
m/z 425, only the trifluoroacetic acid (TFA) adduct m/z 539. As *MSD2, ES-API, Neg, Sample
539.2
MSD1 427, ES-API, Pos, Placebo
0.04

488.2
80 M+TFA
such, the positive ion mode was chosen to determine the

529.2
471.2
0

423.6
source of the impurity. The positive ion mass chromatograms 40

355.0
_0.04
(top and bottom right) show a strong molecular ion peak for the 0
sample, but no peak in the formulation placebo. The MS results 400 450 500 550 m/z 0 10 20 30 40 50 min

definitely show that the impurity is not a matrix product of the Results:
With MS data - Compound in sample has MW = 426
formulation; thus it is a degradation product of the API. With MW 426 not observed in placebo.
the inclusion of mass spectrometry, the impurity peak at 27.6 is Therefore, it is degraded API and original assignment was correct.

confirmed to be due to API degradation.

Figure 9
A second impurity peak (at 39.3 minutes) was originally identi- Identifying peak at 27.2 min with mass spectrometry.
fied as another API degradation peak by LC UV. The reference
standard for the tentatively identified impurity, impurity
data conclusively show that the impurity peak was incorrectly
number 12, shows a protonated molecular ion at M+H m/z 495
identified as the degraded impurity no. 12 by LC-UV and might
(Figure 10, top left) in the positive ion mode and a M+TFA
be an increased amount of a matrix peak.
m/z 607 adduct ion in the negative mode. Mass spectra of the
sample peak do not show either of these ions. Instead a The peak at 39.3 minutes was incorrectly assigned based on
m/z 227 ion was observed for the 39.3 minute peak (top center LC-UV retention time alone. Based on MS, the impurity peaks
and top right) in the positive ion mode and a m/z 339 ion at 21.1, 39.3, and 48.7 minutes were all due to matrix
(bottom center and bottom right) in the negative mode. These degradation.

ES-API, Pos, Impurity No. 12 ES-API, Pos, Sample ES Pos 227, Sample, Placebo
100 Max:46156
Max: 1739 2000 39.4
495.2 100
227.2

80 M+H
80 1500
60
60
1000
538.4

40
40
536.2 500 Peak was initially identified known
20
M+H+ACN 20 impurity number 12 based on UV
0 0 retention time alone. Mass Spectra
0
500 550 600 m/z 100 300 500 m/z 0 10 20 30 40 50 min shows that it is actually a higher
ES-API, Neg, Impurity No. 12
level of matrix peak. Assignment
ES-API, Neg, Sample ES Neg 339, Sample, Placebo
100 100 500
based on UV retention alone is incorrect.
339.2

Max:6079 607.2 Max: 274

80 M+TFA 80
249.0

400
60 60 300
539.2
553.2

40 40 200
513.2
375.2

20 20 100
0 0 0
400 500 600 m/z
200 400 600 800 m/z 10 20 30 40 50 min

Figure 10
Identifying peak at 39.3 min with mass spectrometry.

7
Considerations for LC Performance with MS Conclusion
As previously discussed, using the appropriate mobile phase
The addition of mass spectrometry to HPLC-UV provides capa-
buffers and pH is critical to robust and sensitive MS perfor-
bilities to solve problems and allow for definitive identifications
mance. Sample preparation is also important to the success of
not possible with HPLC-UV analyses alone. In the examples
HPLC-UV-MS analysis. Inadequate sample preparation can
given, the addition of MS identified the molecular weights of
result in ion signal suppression or interferences. The main
unknowns and differentiated the source of degradation as
items to consider prior to ESI-MS are:
being from either the active pharmaceutical ingredient or the
• Matrix components: Salt suppresses the ionization matrix of the formulation. All this was obtained while perform-
process, detergents interfere with the evaporation process, ing the conventional HPLC-UV analysis typically used for moni-
and high analyte concentrations can saturate the detector. toring and quantitation.
• Eliminate matrix/salt/detergent effects. Some typical tech- Chromatographic performance was maintained when switching
niques are ultrafiltration, solvent extraction/desalting, the solvent system to MS-compatible solvents. The addition of
liquid-liquid extraction, solid-phase extraction (SPE), a single quadrupole mass spectrometer used in these exam-
immunoaffinity, on column concentration, and column ples significantly enhanced the problem-solving capabilities of
switching (LC/LC). HPLC. The addition of a mass spectrometer to HPLC-UV can
• If salts and detergents cannot be avoided, remove them also accelerate method development by facilitating peak
using chromatography (a short column is sufficient) or a tracking.
cut-off filter. The ease of implementing MS along with the industry-proven
• Concentration issues: Dilute the sample in the solvent robustness of the technique has made HPLC-UV-MS an indis-
composition that exists at the start of the LC method. pensible tool for rapid and definitive analyte analysis and
characterization.

8
HPLC/UV/MS: Part 2.
Theory of Electrospray Ionization Mass Spectrometry and its
Coupling to HPLC-UV
Introduction cohesive forces leading to droplet explosions. This process is
repeated until analyte ions are ultimately desorbed or ejected
As described in Part 1, the analytical capability of high-perfor- into the gas phase. By applying another model, the electrospray
mance liquid chromatography is greatly enhanced by the addi- process, may be thought of as ionization followed by ion evapo-
tion of MS. MS provides analyte specificity that is structurally ration driven by strong electric fields on the surface of the
based, providing unparalleled analyte identification and confir- microdroplets (see Figure 12).
mation. The structure of unknown components can often be
rapidly elucidated, providing answers to problems during
routine analyses.
100,000
The most common approach to MS detection for liquid samples
is atmospheric pressure ionization. For atmospheric pressure API-Electrospray
ionization, the sample is desolvated, ionized, analyzed by
10,000
mass/charge ratio, and detected. The three typical atmospher- Molecular weight
ic pressure ionization methods are ESI, APCI, and APPI. The
appropriate method largely depends on the polarity of the ana- APPI
lytes; however, ESI is the most common and will be described 1000
below. APCI and APPI will be described in Part 3. APCI

Electrospray Ionization
ESI can be used for both high and low molecular weight com- 100
Nonpolar Very polar
pounds, making it suitable for a great diversity of samples.
Almost anything that can form an ion in solution and can be API-ESI = Atmospheric pressure electrospray ionization
APCI = Atmospheric presure chemical ionization
separated in a volatile mobile phase can be analyzed by ESI. APPI = Atmospheric pressure photo ionization
ESI has femtogram-to-picogram sensitivity, and can deliver
Figure 11
both qualitative and quantitative information. Applicability of LC/MS ionization techniques.
Several factors relating to ionization are obvious from the mole-
cular structure of the analyte. Samples that contain het-
eroatoms typically accept a charge on N, S, O atoms (for exam- Charged Droplets Analyte Ions

ple, carbamates, benzodiazepines, acids, or bases), and hence + -

+ +
+
analyze well by ESI. Along the same line, compounds that can + - +
- Solvent Ion Clusters
+
- Salts/Ion pairs
accept a charge by induction can also be analyzed by ESI. With
+
-
+ +
Neutrals
electrospray ionization, strongly non-polar samples should be
avoided as charge induction will be inefficient and not much Evaporation

signal will be produced. These compounds are better tackled by +

+ Solvent Ion Cluster
++ ++-- + +
APCI or APPI (see Figure 11).
+ +
Rayleigh - - - ++
+
- -
+
Limit + +
+ +
+
ESI uses an electric field to generate a fine spray of droplets. Reached + -- +
+
+
+ Analyte Ion
The charged droplets are attracted toward the MS inlet, pass- Coulomb Explosions
ing through a counter-flow of heated nitrogen drying gas. This
desolvation process shrinks the droplets and carries away Figure 12
Theory of electrospray ionization.
uncharged material. The droplets continue to shrink until the
repulsive electrostatic (coulombic) forces exceed the droplet

9
The formation of ions in ESI is highly dependent on the pH
Positive ion mode (ph <7)
value of the mobile phase and the pKa value of the analyte, as
electrospray requires preformed ions in solution. This follows R1 R1
the well-known principles of acid-base theory (shown in
-
Figure 13) to produce either positive or negative ions which are :N R2 + HA + HN R2 + A
detected in the positive or negative mode respectively. R3 R3

Typical examples of data obtained as an ESI-MS spectrum of a Base Acid Sample

[M + H]+
small molecule and large molecule are shown in Figure 14.
ESI-MS tends to be a soft ionization process leading to limited Negative ion mode (ph >7)
analyte breakdown or fragmentation (Figure 14a). In this figure, O O
the amine, phenylbutazone, picked up a hydrogen atom (H) -
R C OH + :B R C O + H:B+
under ESI conditions and provided a very simple M+H ion as
the base peak. Acid Base Sample _
[M _ H]
Most mass spectrometers have acceptable ranges for mass
detection of compounds typical of the pharmaceutical and bio- Figure 13
Solution chemistry.
pharmaceutical industries. With large molecules such as pro-
teins and peptides, typically carrying multiple ionizable sides,
multiply-charged ions are produced. Since the detector moni-
14a) Singly Charged Ions in ESI
tors the mass/charge, multiple charges allow these large mole-
cules to be detected in a MS such as a single quadrupole Abundance 309
LC/MS system with a mass range of (for example) 2000
46000 [M+H]+
Daltons even if their mass is in the range of 10,000 Daltons. N N
38000
An example of the ESI spectra of the protein myoglobin is O O
shown below in Figure 14b. The spectrum is spread out 30000
between 600 and 1200 amu, depicting the typical pattern of a 22000
CH3
mass spectral envelope of multiply-charged ions. With simple
algebra, one can extract the molecular ion from the location 14000

and spacing of these multiply-charged peaks. The process of 6000 MW = 308

de-convoluting the molecular ion is generally performed by
software and can also easily be performed manually. The bot- 100 200 300 400 m/z

tom panel shows the de-convoluted spectrum, with a molecular API-ES spectrum of Phenylbutazone
ion at 16,959 Daltons. 14b) Multiply-Charged Ions in Electrospray Mode
Neutral molecules, which have a propensity for hydrogen bond- Abundance
40000
ing, can form adduct ions with ammonium or alkali metal ions A+21 A+20
A+22 808.60
848.95 A+19
(add a buffer of ammonium acetate or sodium acetate to facili- 30000 A+23 771.90 893.55 A+18
738.45
tate ionization); typical examples are carbohydrates. 20000 A+24
943.15
A+17
707.65 998.55 A+16
A+25 1060.85 A+15
10000
ESI Related to Chromatography Conditions 679.45 1131.50 A+14
1212.30

Mass spectrometry has certain operating parameters that need 700 800 900 1000 1100 1200 m/z
to be considered when coupling with HPLC. Lower LC flow
Abundance
rates are generally preferred since the MS operates under 280000 A
vacuum and can only handle a limited gas load generated from 16959.09
200000 Assume adduct is a proton:
the LC eluent. MS works at elevated temperatures and requires
M1 = (A+n)/n
volatile mobile phase buffers as the mass analyzer separates 120000 M2 = (A+n+1)/(n+1)
and detects gas phase analyte ions based on a mass to charge Solve for A
40000
ratio (m/z).
16400 16800 17200 17600 m/z
As described above, ionization factors must be considered. The
API-ES spectrum of Myoglobin
consistent and stable ion formation is the foundation for repro-
ducible measurements and is driven by well-understood chem- Figure 14
istry. The ion formation in the liquid to gas interface (the MS ESI-MS of singly-charged and multiply-charged compounds.
source) is the key to success; thus, different types of ion

10
sources are available for interfacing HPLC to a mass spectrom- analysis (triethylamine or ammonium hydroxide [pH 10–11]).
eter. The rest of the MS instrument is robust, requiring little Ion pairing reagents can ionize and create high MS background
attention for routine analyses. and strong ion pairing with an analyte can prevent ionization of
the analyte. Also, some mobile phase additives will cause per-
When choosing chromatographic conditions with an electro-
sistent background problems (TEA interferes in the positive ion
spray ion source, one should use the HPLC mobile phase to
mode [m/z 102] and TFA interferes in the negative ion mode
create charged analytes. This means that mobile phase pH and
[m/z 113]).
sample pKa information is critical. One can quickly select a
mobile phase pH that will give positively charged or negatively A summary of the volatile mobile phase choices that will work
charged analytes and thereby increase sensitivity. Low pH will for LC/MS are shown in Figure 15. For positive ion analysis of
generally ionize basic compounds. Acidic compounds require basic analytes, the buffer choices will be acetate, propionic
more careful mobile phase pH selection because they are more acid, formate, and TFA. These buffers provide the most reliable
likely to have pKa values in the pH 2-to-5 region; finding a good and consistent chromatography.
buffer is therefore more difficult. Neutral compounds can also
be analyzed when they associate with ions such as acetate or Positive ion detection of basic Negative ion detection of acidic
ammonium that are in the buffer. Applying a voltage to the analytes analytes
electrospray probe will induce ion formation, but good mobile Buffer choices (10 mM or less) Buffer choices (10 mM or less)
phase selection to provide ionized molecules dramatically • Acetate pKa 4.8 • Ammonia pKa 9.2
increases ion formation and therefore sensitivity. • Propionic acid pKa 4.9 • Diethylamine pKa 10.5
• Formate pKa 3.8 • Triethylamine pKa 10.7
ESI-MS interfaces operate over a wide flow rate range, from • TFA highly acidic • Piperidine pKa 11.1
5 µL/min up to 2.0 mL/min, for ESI with thermal gradient Typical analytes – amines, amides, Typical analytes – acids, hydroxyls,
focusing (Agilent Jet Stream). Most often they are operated at antibiotics phosphates, sulfates
or below 0.5 mL/min, with optimum sensitivity achieved at Amphetamine Salicylic Acid
lower flow rates. Standard 2.1 mm id HPLC columns are com- COO-
patible with most HPLC instruments and are ideally operated
CH2CHCH3 OH
around 0.25 mL/min. This mobile-phase flow rate leads to good
sensitivity with ESI. NH3+

ESI is readily compatible with reversed-phase and normal Figure 15

phase solvents. Since heating is not required there is little Mobile phase polarity and buffer selection for ESI.
chance for flammability problems with normal-phase solvents.
LC/ESI-MS has some limitations from a chromatographic per- Compounds that protonate to form a positive charge tend to be
spective. ESI is particularly well suited for the analysis of small, basic; for example, amphetamine, where the amino functional
polar analytes which might be difficult with the reversed phase group can be ionized. Acidic compounds deprotonate and are
retention (when selecting a mobile phase to provide ionized negatively charged, for example salicylic acid where the car-
molecules). In addition, adduct ions are possible with some boxylic acid group can be ionized. Analysis can be performed by
analytes (in addition to or in place of the protonated molecular positive and negative ion ESI and one wants to choose buffers
ion, M+H). LC/ESI-MS also has poor compatibility with non- that create these charged analytes based on the sample pKa.
volatile modifiers and ion-pairing agents. For optimum MS performance, buffer concentration should be
less than 10 mM for better droplet evaporation. For negative
Mobile Phase Buffer Selection ion detection of acidic analytes at high pH, some good buffer
Volatile buffers are used to modify mobile-phase pH in mass choices are ammonia, diethylamine, TEA and piperidine.
spectrometry not only because of deposit build-up from non- Compromises are possible to work closer to pH 7 where com-
volatile buffers but also because metal ion buffers interfere pound classes like carboxylic acids will have a negative charge,
with ionization and surfactants interfere with droplet evapora- but the mobile phase will not compromise column life. In
tion. The volatile buffers may also be added to the mobile- general, to optimize the mobile phase:
phase eluent as a post-column addition. This technique pre- • adjust the pH to be 1 to 2 units away from the pKa of the
serves chromatographic separation while optimizing analyte analytes,
ionization.
• avoid using salts and detergents, and
In general acidic solutions favor positive ion mode analysis
(formic acid, 0.1–1.0%; acetic acid, 0.1–1.0%; or TFA • use solvents that enhance ion desorption (for example, sol-
[trifluoroacetic acid], 0.05–0.2%). Ammonium salts favor pro- vents with low surface tension and low heat of
duction of single ammonium adducts (ammonium acetate or vaporization).
ammonium formate). Basic solutions favor negative ion mode
11
Considerations for LC Performance with MS temperatures, flow rates, and voltages for good droplet desol-
vation and ion evaporation. Settings for a typical electrospray
Evaluating Robustness
ionization source are shown in Figure 16.
It is important to distinguish the mass spectrometer source
from the analyzer. The source is where the liquid-to-gas phase Electrospray ions
transition and analyte ionization occurs. The operator has sig-
nificant control over reactions taking place in this region. A cor- Nebulizer (gas
rect setup (mobile phases, flow rates, source parameters) leads shown in red)
Vcap
to robustness, while inferior setup can lead to problems. The Drying gas
analyzer region is considered everything behind the first vacu-
um region; this region is very robust and can be treated as a Solvent spray
black box for normal applications.
+⊕
+
The chromatography must resolve interfering compounds and ⊕⊕⊕⊕
⊕ ⊕ ⊕ ⊕ ⊕ ⊕ ⊕ ⊕ ⊕ ⊕⊕
MS
the analyte must ionize in the source. These two criteria are
essential; all other criteria are secondary.
When adapting a method from HPLC-UV to HPLC-UV-MS, one Neutral
should consider fitness for purpose. In general there are no Molecules
hard-and-fast rules about buffers and modifiers, but simply Clusters
Salts Fragmentor
guidelines. Choices outside those guidelines will not damage
the instrument, although frequent source cleaning may be nec- Nebulizer Pressure
• Higher liquid flow requires higher nebulizer pressure for
essary. If the sample set is small, method development time is efficient spray
limited. The use of an established method with a semi-volatile • 100 to 1000 µL/min = 15 to 60 psi
or non-volatile buffer might be the most pragmatic choice. A Vcap
• Just need enough to perform the electrospray. (set and forget)
simple cleaning of the source when the analysis is complete
Drying Gas Flow
will prevent problems with subsequent analyses. • High water needs higher flow
• High gas flow keeps the transfer capillary clean when analyzing
In short, the standard guidelines given above can be modified, dirty samples or using semi-volatile modifiers.
but one must be mindful of the price one has to pay (loss of • When in doubt, use excess (set high and forget unless sensitivity
sensitivity, more cleaning). is an issue)
Drying Gas Temperature
Adapting Existing LC Methods to LC/ESI-MS • Higher for low vapor pressure solvents
• Start with 300-350 °C
If an existing method does need to be adapted to LC/MS, some
Fragmentor
slight changes in the separation may occur, but that can be • Low value breaks up solvent clusters but transmits molecular
optimized. First, replace non-volatile buffers with volatile ions (typical)
• Normally set and forget
buffers at a concentration of < 10 mM for electrospray. For
example, replace phosphate buffers with acetate or formate
Figure 16
buffers. TFA may be used as well. If the chromatography Typical electrospray source settings.
changes too dramatically and a non-volatile buffer must be
used, then select a buffer where only the anionic or cationic Conclusion
part is non-volatile—for example, ammonium phosphate
instead of potassium phosphate. The column id and flow rate The placement of a MS detector in tandem after an LC-UV
should be kept low to increase sensitivity as much as possible (DAD) provides increased performance and capability. Such
and minimize the build up of non-volatiles. Maintain the pH as enhancements include increased specificity, resolution, sensi-
in the original separation if possible. If working at pH 3 with tivity, productivity, and problem solving capabilities. MS offers,
phosphate, replacement with formate is recommended first in addition to UV detection, a sensitive detection of both polar
since formate will also buffer at pH 3. If any type of ion-pair and non-polar compounds. A few considerations such as flow
reagent must be used to increase retention, then the use of a rate, mobile phase buffer selection, and MS source parameters
volatile ion pair reagent like heptaflurobutyric acid (HFBA) and are important for optimum HPLC-UV-MS performance.
tributylamine (TBA) is recommended. In general it is possible Chromatographic conditions have a significant impact on mass
to adapt an existing method using the same basic principles spectrometric analysis and can be optimized and adjusted
and paying attention to the original pH of the separation. easily. Whether analyzing small molecules or large peptides
and proteins, even a single quadrupole mass analyzer has suffi-
Electrospray Source Settings
cient mass-resolving power for excellent, highly sensitive
To achieve optimum sensitivity, the ESI source requires specific quantitative and qualitative mass spectral analysis.

12
HPLC/UV/MS: Part 3.
APCI and APPI as Alternative Ionization Techniques for MS

Introduction Atmospheric Pressure Chemical Ionization (APCI)

APCI is distinguished from ESI as it uses a corona discharge to
As described in Parts 1 and 2, the coupling of MS to HPLC-UV
ionize molecules. Unlike ESI, no preformed charged analytes
provides a robust analytical technique that provides a dimen-
are needed because of a charge transfer process that occurs in
sion of specificity with problem solving capabilities not possi-
APCI. APCI generates ions by first nebulizing the liquid analyte
ble with HPLC-UV alone. In Part 2, ESI was presented as the
into small droplets. This is followed by evaporating the droplets
most common approach to atmospheric pressure ionization. In
to produce gas phase solvent and analyte molecules, the sol-
this section, two other approaches, APCI (Atmospheric
vent molecules being ionized by the corona discharge. A
Pressure Chemical Ionization) and APPI (Atmospheric pressure
corona discharge needle serves as a charge source. The gas
photoionization), are presented.
phase analyte is then ionized by gas phase chemical ionization
Atmospheric pressure ionization is the most common sample (CI) via proton addition, proton abstraction, or by electron cap-
introduction for MS detectors in liquid phase separations ture processes. CI is a process where the solvent acts as a CI
(HPLC). For API, the sample is desolvated, ionized, analyzed by reagent gas to ionize the sample. The ionization process is pro-
mass/charge ratio, and detected. The three typical API meth- tonation (for example, H3O+) for bases and a charge exchange
ods are ESI, APCI, and APPI. The appropriate method largely deprotonation for acidic compounds. APCI is also an electron
depends on the polarity of the capture mechanism for halogens and aromatics. The evapora-
analytes. tion and ionization processes are shown in more detail in
Figure 17.
After the eluent passes through the UV detector, the mass
spectrometer detects ionized analytes. Based on the mass to
charge ratio (m/z), the mass spectrometer measures unique [Solv+H] + A & Solv + [A+H]
masses from individual analytes that are used to confirm Analyte containing
aerosol
known compounds as well as identify unknowns. Analyses are
easily performed in both the positive ion mode and negative ion +
+
+
+
+

mode. In the positive ion mode, protonation or the addition of +

+
+
+
+

Evaporation + +
Charge transferred
hydrogen occurs on a basic site of the molecule. In negative +
+
+ + +
+ +
+ to analyte
ion, deprotonation produces a negatively charged molecule. +
+
+

Vapor Charged reagent

As discussed in Parts 1 and 2 of this technical overview, ESI is +
gas formed +

the most common method used for LC/MS. ESI can be used for Analyte +
+
Corona ions +

both high and low molecular weight compounds, making it needle

suitable for a large number of diverse samples—almost any-
thing that can form an ion in solution and can be separated in a Figure 17
Theory of atmospheric pressure chemical ionization (APCI).
volatile mobile phase.
Not all samples or chromatography techniques are best ana- Just as in the API-electrospray design, the APCI inlet is posi-
lyzed by ESI. While electrospray may be thought of as ioniza- tioned orthogonally to the inlet of the capillary, uses the same
tion followed by ion evaporation, not all analytes can be ionized nebulizer design, and takes advantage of the drying gas heater
in solution or use chromatography conditions that are suitable design. All of this results in low noise, high signal, and maxi-
for ESI. In these cases, the alternative ionization methods of mum system uptime at high HPLC flow rates. A high probe
APCI or APPI may provide a solution (Figure 11). temperature is typically used to desolvate and vaporize the

13
sample, although temperatures that are too high can lead to
Nebulizer
sample decomposition. The externally removable corona dis- Pressure HPLC Flow Rate > 500 µL/min
charge needle can be easily replaced without venting the Nebulizer pressure
Heater •60 psig
vacuum system or opening the spray chamber (Figure 18). Corona
Drying Gas Temperature
current
•Start with 350 °C
APCI is a good technique for small molecules with molecular Drying gas flow
weights of less than 1500 Daltons. These small molecules can Vcap •4 L/min
be polar or somewhat non-polar (substituted PAHs and PCBs, Vaporizer temperature
•Optimize with
fatty acids, phthalates). APCI is not a good technique for bio- flow injection analysis (FIA)
molecules (peptides and proteins) because this form of ioniza- Vcap
•Optimize with FIA (2000-6000)
tion rarely results in multiply-charged species. Therefore, for Drying gas •Start with 2500 V
large molecules that ionize, the mass-to-charge ratio would Temperature Fragmentor Corona current
and Flow •Optimize with FIA
remain high and would generally not be in the range of the MS
•Start with 25 µA (neg) or 4 µA (pos)
instrument. In general, APCI should be thought of as a comple-
mentary technique to ESI, because in APCI, samples targeted Figure 18
are generally those not charged in solution. Samples that APCI spray chamber settings.
charge in solution will typically be detected with greater sensi-
tivity using ESI. APCI is also not the method of choice for ther-
mally unstable or photosensitive samples. These tend to frag- ESI APCI
ment completely, thus not producing the parent ion which is Ionization: Pre-formed analyte ions Ionization: Charge exchange of gas
desirable in LC/MS for molecular weight and compound identi- transferred to gas phase phase neutral analysis
fication. These types of molecules are better attempted by Mobile phase issues: Mobile phase issues:
atmospheric pressure photoionization (APPI) discussed later in
• Organic solvent: • Organic solvent:
this technical overview. little effect on ionization MeOH usually best
A general comparison of ESI and APCI is presented in • pH: key to pre-formed ions • neutral analytes
Table 1. For sensitivity, if a sample can be ionized by both tech- • Buffer concentration: < 25 mM • Buffer concentration: < 100 mM
niques, electrospray is generally more sensitive and has less
• Flow rate: < 0.5 mL/min • Flow rate: > 0.5 mL/min
background noise. However, ESI is more adversely affected by
sample and solvent matrix than APCI (for example, signal sup- Table 1
pression). Electrospray also requires a lower concentration of General comparison—ESI vs APCI.
volatile buffers relative to APCI. The choice of organic solvent
strongly affects ionization in APCI while ionization in ESI is APCI typically generates just singly charged ions; however, it is
largely dependent on the choice of mobile phase buffers. possible to get doubly charged ions where the charge sites are
Another significant difference is that ESI works best at low separated from each other (usually by a hydrophobic region).
flow rates (< 500 µL/min) while APCI has a broad flow rate For positive-mode APCI, the solvent must be capable of donat-
range up to 1.5 mL/min. APCI is more sensitive and has less ing a proton and the analyte must have higher proton affinity
noise than ESI at high flow rates (> 750 µL/min). ESI, like a UV than the reagent gas. For example, acetonitrile/water is a com-
detector, is a concentration sensitive isolation technique while monly used mobile phase for LC/ESI-MS, but acetonitrile is not
APCI is mass sensitive. For APCI, sample dilution is not a factor a protic solvent (that is, does not have a proton to donate) and
for sensitivity. water in the gas phase is a very strong base. Therefore,
There are advantages of APCI over ESI. APCI is less sensitive acetonitrile/water is not a good solvent choice for APCI posi-
to solution chemistry effects than ESI. APCI tolerates higher tive mode. In negative-mode APCI, the reagent gas must be
flow rates than ESI and accommodates some solvents that are able to abstract a proton or capture an electron. Where ESI is
not compatible with ESI. Most importantly, APCI may ionize thought of as ionization followed by evaporation, APCI can be
neutral or more non-polar compounds that cannot be ionized by thought of as evaporation followed by ionization.
ESI.

14
APCI Related to Chromatography Conditions reserpine. However, as the “volatile salt” concentration is
increased, it becomes more difficult for the analyte to effective-
Typical rigor used in LC applies to LC/APCI-MS. Only highly
ly volatilize. Caffeine, which is a weaker base than reserpine,
purified water and organic solvents should be used. Volatile
does not show the same signal increase at low buffer concen-
solvents are best for APCI, as the eluent needs to be in the gas
trations since ammonia competes with caffeine for protonation.
phase for ionization to occur. Where ESI requires low buffer
concentrations, APCI works in a wide range of buffer concen- Protic solvents like methanol should be used for the positive
trations. This is because the APCI process for ionization is ion mode. These solvents have a proton to transfer to the M+H
based on charge transfer and not the evaporation from the fine ion. Therefore, methanol is often the better organic solvent
particles to give charged analytes. Figure 19 and Figure 20 choice over acetonitrile. For the negative ion mode, choosing a
show the effects of volatile buffer concentrations on the ioniza- solvent that can readily capture an electron will make the ion-
tion of caffeine and reserpine. For ESI as shown in Figure 19, ization process more efficient. In general if an ammonium salt
the buffer ions compete with the analyte about the droplet sur- is used in the mobile phase, ammonium adducts are likely to
face and the analyte has increased difficulty in escaping the form. This is detrimental if monitoring [M+1] but may be benefi-
droplet as the buffer concentrations increases. Smaller mole- cial if monitoring for [M+18] as sensitivity can be increased
cules such as caffeine are desorbed earlier in the ionization with an ammonium adduct.
process. For APCI as shown in Figure 20, at low concentration
APCI can accommodate a very wide flow rate range, up to
the ammonium acetate in the buffer aids proton transfer for
about 1.5 mL/min, so column dimensions are not as restricted
as with some other techniques. Columns of 3.0 mm id are very
popular for APCI. These operate around 0.5 mL/min, which is
1.20E+07 ESI also a flow rate that can be used by ESI with very good sensi-
tivity. However, LC/APCI-MS is also well suited for 4.6 mm id
Reserpine Caffeine
columns which are popular in HPLC/UV methods. So adapting
Area

6.00E+06 an analytical method from HPLC/UV to HPLC/UV/APCI-MS

does not require changing column dimensions or adjusting flow
rates. Like LC/ESI-MS, reversed phase chromatography
0.00E+00 typically precedes APCI along with the use a buffered mobile
0 50 100 150 200 250 300 350
Concentration (mM)

2.E+07 APCI
O CH3
N
H3C N H3OC N Reserpine Caffeine
N H H 2.E+07
H
OCH 3
O N N H
Area

CH3 H3COOC
1.E+07
OOC OCH 3
OCH 3
OCH 3 5.E+06

Caffeine Reserpine 0.E+00

0 50 100 150 200 250 300 350
ESI: At higher buffer concentrations, the analyte has more Concentration (mM)
difficulty escaping the droplet. Smaller molecules such
as caffeine are desorbed earlier in the ionization process.
O CH3
LC Conditions N
H3C N H3OC N
Mobile phase: Ammonium acetate in 50:50 methanol:water. N H H
H
OCH 3
Flow rate: 0.6 mL/min. O N N H
Injection: 1µL of reserpine (84 ng) or caffeine (125 ng) CH3 H3COOC OOC OCH 3

MS Conditions: OCH 3
OCH 3
SIM: 195.2 and 609.3.
Caffeine Reserpine
Drying gas: ESI – 350 °C, 8L/min; APCI – 350 °C, 5 L/min.
Nebulizer: ESI – 30 psig; APCi – 60 psig. APCI: Ammonium acetate initially aids proton transfer for reserpine.
Vcap: 4000 V Increasing “volatile salt” concentration makes it more difficult to
Fragmentor: Ramped 70 V for 195.2; 120 V for 609.3 effectively volatilize the analyte. Caffeine is a weaker base than
reserpine and ammonia competes for protonation.
Vaporizer: 400 °C

Figure 19 Figure 20
Effect of volatile buffer concentration on ESI response. Effect of volatile buffer concentration on APCI response.

15
phase. APCI generally can accommodate a little more salt than
ESI, though a volatile buffer is still preferred. APCI is possible APPI
with selected normal phase solvents for normal phase HPLC, HPLC inlet
but highly flammable solvents should be avoided because of Nebulizer
HPLC Flow Rate >500 uL/min
the heating process or special caution has to be taken. Nebulizer pressure
• 35 psig
Drying Gas Temperature
Atmospheric Pressure Photoionization (APPI) • Start with 275 °C
Vaporizer Drying gas flow
Another API technique complementary to ESI is APPI. APPI is • 11 L/min
(heater) Drying gas
typically best for hydrophobic conjugated ring systems (corti- Vaporizer temperature
costeroids, PAHs) and may be less susceptible to ion suppres- • Optimize with FIA
hν + Vcap
sion than ESI. APPI is an ionization technique where the ana- +
+ + + + + • Optimize with FIA (2000-6000)
lyte and mobile phase is first evaporated then ionized using UV • Start with 2500 V
light. UV Lamp Capillary

The APPI interface was introduced by Agilent and Syagen at

ASMS in May 2001 (Figure 21). Like APCI, it is used for low to Figure 21
medium polarity analytes. The APPI process generates ions by APPI Sources for LC/MS.
first nebulizing the liquid analyte into small droplets, followed
by evaporating the droplets to produce gas phase analyte mole-
cules. The gas phase analyte is then ionized by photoionization
from a krypton lamp. Analyte containing +
aerosol +
+
The evaporation processes for APPI is similar to APCI and the +
+
analyte is not ionized until after evaporation. There are two
mechanisms for ionization in APPI (Figure 22). The first mecha- Photon ionizes + +
nism, direct photoionization, results from the analyte absorbing hυ analyte + +
Evaporation +
a photon of light from the krypton emission. In order to get this + +
+ + + Analyte ions
type of ionization, the analyte must have an ionization potential + + + +
hυ
that is less than the 10.6 eV of the lamp. Direct photoionization Vapor + + +
+ +
can only result in positive ion formation. In general, the analyte + + +
+
+ + + +
molecule needs to have only about seven carbon atoms in +

order to be photoionizable; but it needs to absorb the radiation Dopant is photoionized

and acts as reagent gas
energy. The second mechanism, as compared to APCI, can be
thought of as photo-induced chemical ionization. In this case, a
Figure 22
second reagent called a dopant is added to the mobile phase. Theory of atmospheric pressure photo ionization (APPI).
The dopant is photoionized and the charge is then transferred

16
from the dopant to the analyte. Typical dopants are acetone Nebulizer
HPLC inlet
and toluene, although other compounds can be used.
In order to do negative-mode APPI, there must be a source of
thermal electrons. While the light striking the metal of the
vaporizer barrel can be a source of some electrons, it is far ESI Zone
better to use a dopant for this. Acetone is an excellent dopant
for negative-mode APPI and does not interfere with positive-
APCI Zone
mode ionization. As with APCI, the LC mobile phase can inter- Thermal
fere with ionization if the solvent has more affinity for the container
proton than the analyte. Also as with APCI, acetonitrile can be
a problem for some analytes and it is recommended to try
Capillary
methanol first.
Corona
needle
The Multimode Source: Combined ESI and APCI Drying gas
The most versatile ion source for the single quadrupoles is cer-
tainly Agilent’s G1978A multimode source (Figure 23). While
ESI and APCI are essentially incompatible processes, each Parameters ESI APCI Mixed mode
needing its own conditions for aerosol drying and electrical Capillary voltage (Vcap)
fields, it is possible to form ions simultaneously from ESI and single ion polarity 2000 V 2000 V 2000 V
APCI if the two ionization regions are separated in space. The polarity switching 1000 V 1000 V 1000 V
HPLC effluent is nebulized using the same sprayer as for a ded- Charging electrode 2000 V 2000 V 2000 V
icated ESI source. The droplets are emitted into the “ESI zone,” Corona current 0 µA 4 µA 2 µA
where a high-voltage electrode performs the charging of the
droplets and induces the formation of ions. Ions formed in this Drying gas flow 5 L/min 5 L/min 5 L/min
region pass through the source and enter the capillary. Drying gas temperature 300 °C 300 °C 300 °C
Residual droplets are dried using two IR lamps (not shown) Nebulizer pressure 60 psig 30 to 60 psig 40 to 60 psig
emitting at the absorption frequency of water, and the vapor
and analyte(s) enter the APCI zone where they are ionized by Vaporizer temperature 150 °C 250 °C 200 °C
this process. Ions are then drawn into the capillary as in the Figure 23
case of the dedicated ESI and APCI sources. The multimode source: A combination of ESI and APCI (Starting conditions
for method development).
If the multimode source is operated as an ESI or APCI source
only, there is no loss in sensitivity. In fact, because the new
source uses infra-red lamps for droplet evaporation, the effi-
ciency of the evaporation process actually improves. Increased
APCI sensitivity using the multimode source has been docu-
mented. This drying technique has also helped in achieving
flow rates up to 2 mL/min. The standard APCI source, which
used convective heating, works well to flow rates less than
1.5 mL/min. If the multimode source is operated in ESI and
APCI simultaneously, sensitivity losses of up to a factor of two
for some compounds can occur. Therefore, one must weigh the
benefits of running analyses in both modes simultaneously
against losses in sensitivity. For most applications, a loss in
sensitivity of less than a factor of two is negligible.

17
Conclusion
The addition of MS to HPLC-UV provides an increase of analyti-
cal capability for the chromatographer. The most common
atmospheric pressure ionization interface to HPLC is ESI. Not
all samples or chromatographic techniques are well suited for
ESI. For these applications, two alternative atmospheric pres-
sure ionization techniques have been developed: APCI and
APPI.
APCI and APPI both ionize the sample after evaporation while
ESI evaporates the sample after ionization. As such, both APCI
and APPI are able to ionize the less polar molecules that can-
not be ionized by ESI. Since both APCI and APPI are mass
dependant analyzers, sample dilution does not reduce
sensitivity, so higher flow rates and larger LC column dimen-
sions can be used. APCI is good technique for small molecules
(< 1500 Daltons) which can be polar or somewhat non-polar
(substituted PAHs and PCBs, fatty acids, phthalates). APPI is a
good technique for hydrophobic conjugated ring systems (corti-
costeroids, PAHs) and compounds that are thermally labil
during APCI.
The development of a multi-mode source combining ESI and
APCI increases the range of polarity for analyzing samples.
When used in solely ESI or APCI modes, the convenience of
having both techniques in one source increases productivity.
The availability of ESI, APCI, and APPI for HPLC/MS analysis
provides options for diverse sample analysis as well as varying
HPLC techniques.

18
19
www.agilent.com/chem/lcms
Agilent shall not be liable for errors contained herein or
for incidental or consequential damages in connection
with the furnishing, performance, or use of this material.

Information, descriptions, and specifications in this

publication are subject to change without notice.

© Agilent Technologies, Inc., 2011

Printed in the USA
April 28, 2011
5990-7413EN