Chem. Rev.

2001, 101, 361375

361

Methods of Ion Generation

Marvin L. Vestal
PE Biosystems, Framingham, Massachusetts 01701
Received May 24, 2000

Contents
I. Introduction
II. Atomic Ions
A. Thermal Ionization
B. Spark Source
C. Plasma Sources
D. Glow Discharge
E. Inductively Coupled Plasma (ICP)
III. Molecular Ions from Volatile Samples.
A. Electron Ionization (EI)
B. Chemical Ionization (CI)
C. Photoionization (PI)
D. Field Ionization (FI)
IV. Molecular Ions from Nonvolatile Samples
A. Spray Techniques
B. Electrospray
C. Desorption from Surfaces
D. High-Energy Particle Impact
E. Low-Energy Particle Impact
F. Low-Energy Impact with Liquid Surfaces
G. Flow FAB
H. Laser IonizationMALDI
I. IR MALDI
J. Delayed Extraction
K. New Mass Analyzers
V. Present Status and Future Prospects
VI. References

361
362
362
362
362
362
363
364
364
365
367
367
367
367
367
369
369
370
371
371
371
373
373
373
374
374

I. Introduction
Mass spectrometry separates ions according to
their mass-to-charge ratio. Except for the quadrupole
ion trap, all conventional mass analyzers require ions
in the gas phase in a sufficiently good vacuum that
collisions with background gas are insignificant. With
few exceptions, such as flames and plasmas, samples
of interest are not charged and are present in
condensed phases or as gases at atmospheric pressure or above. The challenge in selecting an ionization method for mass spectrometry is to choose a
technique that preserves the properties of the sample
that are of interest while at the same time converting
it to ions which can be mass analyzed. This review
will concentrate on ionization methods of most practical utility at present but will also include a brief
summary of some that are of historical or fundamental importance.

Marvin L. Vestal received his B.S. and M.S. degrees, 1958 and 1960,
respectively, in Engineering Sciences from Purdue Univesity, Layfayette,
IN. In 1975 he received his Ph.D. degree in Chemical Physics from the
University of Utah, Salt Lake City. From 1958 to 1960 he was a Scientist
at Johnston Laboratories, Inc., in Layfayette, IN. From 1960 to 1967 he
became Senior Scientist at Johnston Laboratories, Inc., in Baltimore, MD.
From 1960 to 1962 he was a Graduate Student in the Department of
Physics at John Hopkins University. From 1967 to 1970 he was Vice
President at Scientific Research Instruments, Corp. in Baltimore, MD. From
1970 to 1975 he was a Graduate Student and Research Instructor at the
University of Utah, Salt Lake City. From 1976 to 1981 he became
Associate Professor of Chemistry at the University of Houston. From 1981
to 1987 he was Professor of Chemistry at the University of Houston.
From 1983 to 1993 he was President of Vestec Corp. in Houston, TX.
From 1993 to present, he has been Vice President, Biospectrometry, at
PerSeptive Biosystems, Framingham, MA.

Mass spectrometry traces its roots to the work of

J. J. Thomson1 beginning nearly a century ago.
During the first half of the century, the technique
was developed and practiced by physicists and the
applications were mainly limited to precise determination of the masses of the stable nuclides and
natural abundance of isotopes.2 During and immediately following World War II, analytical applications and fundamental physical measurements on
molecules began to emerge. The early work required
intense beams of atomic ions. The fact that most of
the elements of interest were naturally available as
components of molecules or crystalline solids was a
nuisance to be overcome by the choice of suitably
energetic means of producing ions such as gaseous
discharges, arcs, and sparks. During the last half of
the century, there has been a relatively continuous
shift toward applications involving molecules of
greater and greater complexity, and within the past
decade biological applications have become, by far,
the most important and fastest growing segment.

10.1021/cr990104w CCC: $36.00 2001 American Chemical Society

Published on Web 01/25/2001

362 Chemical Reviews, 2001, Vol. 101, No. 2

II. Atomic Ions

A. Thermal Ionization
This is one of the earliest techniques in mass
spectrometry and was originally used for producing
beams of positive and negative atomic ions for precise
measurements of masses of the nuclides and relative
intensities of the isotopes. This ionization source
produces intense, stable beams that can be maintained for long periods to allow precise measurements
but is limited to elements of relatively low ionization
potential. An excellent review of the theory of thermal ionization and early experimental results has
been given by Ivanov.3
In thermal ionization, one or more metal filaments
are heated, directly or indirectly, by passage of an
electrical current. The sample may be either in the
gas phase or deposited directly on the surface. A
widely used practical source is the triple-filament
design of Inghram and Chupka.4 In this arrangement
the sample is deposited on the outer two filaments
that are heated to produce a neutral vapor of the
sample and the inner, hotter filament causes the
ionization. The ratio of ions to neutrals at the surface
of the hotter filament is given by the well-known
Langmuir-Saha equation. As discussed in detail by
Ivanov,3 the effective work function is strongly dependent on the chemical and physical properties of
the surface and may not be simply related to those
for the clean, crystalline metal surface. This mode
accounts satisfactorily for ionization of alkali metal
and other easily ionized atoms on heated metal
surfaces. This source is still widely used for accurate
determinations of isotope ratios in geochemistry
applications5 and in nuclear technology.6

Vestal

developed plasma ionization sources but is still used

in many laboratories, particularly for determination
of trace impurities in metals.

C. Plasma Sources
The dominant techniques presently in use for
elemental analysis by mass spectrometry fall into the
general category of plasma sources. An excellent
review with an extensive bibliography has been
published by Hieftje.10 These may be separated into
those operating at reduced pressure, ca. 1 Torr,
producing a glow discharge (GD), and those operating
at atmospheric pressure using inductively coupled
plasma (ICP) generators. Glow discharge ionization
has a long history in mass spectrometry, beginning
with its early use in isotopic studies and as a means
of ionizing volatile organic compounds. These applications generally employ other techniques at
present, but glow discharges are still used extensively
in the analysis of trace elements in semiconductors,
metals, and especially high-purity metals.11

D. Glow Discharge
A schematic diagram of a glow discharge ion source
for a magnetic mass spectrometer is shown in Figure
1.12 The cathode is formed as a metal pin, ca. 2 mm
in diameter, mounted on a removable probe. The
stainless steel anode contains the ion exit aperture,
typically 0.5 mm in diameter. The anode is separated

B. Spark Source
The thermal ionization source is simple and wellbehaved but is highly selective. Elements with low
ionization potentials are efficiently ionized, but others
such as most transition metals are difficult or impossible to ionize thermally, and the ionization efficiency
may vary by many orders of magnitude. The vacuum
spark was originally developed by Dempster7 to
extend mass spectrometry to analysis of metals. In
this source an electrical spark is produced between
a rod electrode of the material to be analyzed and
the wall of an aperture in a tantalum disk. Both
Tesla-coil and pulsed-RF sources have been used for
excitation of the spark. Ions passing through the
aperture in the tantalum electrode are accelerated
to ca. 15 kV and analyzed in a double-focusing mass
spectrometer.8 The ions from the source have a very
wide energy spread (ca. 1000 eV), so double focusing
is required to obtain useful resolution. Another
problem with this source is that the spark is subject
to random fluctuations in ion intensity and produces
considerable RF noise which may interfere with
electrical detection. Despite these limitations, the
spark source provides a very sensitive method for the
analysis of trace impurities in solids that give approximately equal sensitivity for all elements.9 It has
been replaced for some applications by more recently

Figure 1. Schematic diagram of the glow discharge source

for an MS9 magnetic mass spectrometer: A, beam focusing
plates; B, ion exit (anode) plate; C, machined ceramic
insulator; D, source block; E, sample pin (cathode); F,
insulated cathode cap; G, stainless steel probe shaft; H,
quartz support plates. (Reprinted with permission from ref
12. Copyright 1989.)

Methods of Ion Generation

from the source block by a machined ceramic insulator. The anode to cathode gap is about 5 mm. The
cathode is insulated from the source block by a
shaped cap of either ceramic or vespel through which
the sample pin protrudes. The discharge gas, usually
high-purity argon, is supplied through a needle valve
and a length of fused silica capillary to maintain a
pressure of about 1 Torr in the discharge. A floating
power supply provides a voltage of about -1000 V to
the cathode, and a series resistor limits the discharge
current to 1-2 mA. The sample of interest is either
the cathode material itself, in the case of metals, or
a layer of material depositied on the surface of the
cathode. In this configuration the glow discharge is
limited to analysis of conductive solids, but a RFexcited version has also been developed and applied
to nonconductive solids.13
The processes occurring within a glow discharge
have been studied extensively14,15 For a given discharge voltage and gas pressure, the voltage drop
adjacent to the cathode (the cathode fall) occurs
mostly within the thin dark space near the cathode.
The plasma in the remaining negative glow is
almost electrically neutral overall and is close in
potential to the anode. Sputtering is caused by both
ionic and neutral argon bombardment of the probe
with an energy which is estimated to be about 1/4 of
the discharge voltage at a pressure of 1 Torr. Neutral
atoms or molecules are sputtered into the gas phase
from the surface of the cathode. Any cations formed
would not be able to escape the electric field of the
cathode fall. The neutrals are then ionized by electrons, excited atoms (Penning ionization), or by ionmolecule reactions (chemical ionization). The ions
sampled by the mass spectrometer are those formed
close to the ion exit that are swept out by the flowing
gas. The major disadvantage of the glow discharge
for elemental analysis is that it takes relatively long
times (ca. several minutes) for the extracted ion beam
to reach equilibrium with the elemental concentration of the original sample. Thus, sample throughput
may be low.

E. Inductively Coupled Plasma (ICP)

This plasma source was originally developed as a
superior excitation source for elemental analysis of
solutions by atomic emission spectroscopy (AES).16
Techniques for sampling flames and plasmas into
mass spectrometers were developed earlier for fundamental studies.17 Gray18 demonstrated that useful
mass spectra of elemental constituents in solution
could be obtained from a DC plasma, and Houk and
co-workers19 produced the first analytically useful
spectra from an ICP.
A typical example of an ICP source coupled to a
quadrupole mass spectrometer is shown in Figure 2.20
The excitation portion of the source is identical to
that used in AES, but instead of viewing the emission
with a spectrograph, the ions produced at atmospheric pressure are sampled into the vacuum of the
mass spectrometer. The plasma is sustained by RF
energy applied to the load coil. The electromagnetic
field transfers energy from the coil to the plasma.
Most of the energy is added to the outer, annular

Chemical Reviews, 2001, Vol. 101, No. 2 363

Figure 2. ICP ion source and sampling interface: A, torch

and load coil; B, induction region of ICP; C, solution aerolsol
being injected in axial channel; D, initial radiation zone;
E, normal analytical zone; F, nickel cone with sampling
orifice in tip; G, skimmer cone; H, boundary layer of ICP
gas deflected outside sampling orifice; I, expanding jet of
gas sampled from ICP; J, first element of ion lens.
(Reprinted with permission from ref 20. Copyright 1986
American Chemical Society).

portion of the plasma inside the load coil. The sample

traverses only the axial channel, which is heated by
energy transferred from the outer, induction region.
The physical separation of the sample from the region
where the energy is added has important analytical
implications. The sample composition can change
substantially and have little effect on the electrical
processes that sustain the plasma. Also, the absence
of electrodes in physical contact with the ICP prevents contamination by sputtered electrode material.
The challenge with ICP-MS is to efficiently transfer
ions from a plasma at atmospheric pressure and ca.
5000 K into the vacuum of the mass spectrometer.
As shown in Figure 2, the plasma flows around the
tip of a water-cooled metal cone. This cone has a
circular orifice of ca. 0.5-1 mm in diameter at its tip.
Gas from the ICP enters this orifice from a cross
section of the axial channel which is approximately
three times the orifice diameter; thus, a large fraction
of the gas in the axial channel is sampled. The central
orifice of a conical skimmer is located beyond the
sampler at an appropriate distance to transmit as
much of the sampled beam as possible into a second
vacuum chamber without overloading the second
vacuum system. The pressure here is low enough (ca.
10-5 Torr or less) so that ion lenses can be used to
collect, focus, and transmit the ions to the mass
analyzer with minimal losses by scattering.
The ICP is a very efficient source for producing
singly charged elemental ions. Positive ions have
been detected for all of the naturally occurring
elements except F, Ne, and He, all with ionization
potentials greater than Ar used to form the plasma.
On theoretical grounds, some 54 elements are expected to be ionized with efficiencies of 90% or more.
Even metalloid or nonmetallic elements such as P
and As are ionized with useful efficiency. Doubly
charged ions are produced from some elements having low second ionization potentials, such as Ba and

364 Chemical Reviews, 2001, Vol. 101, No. 2

several rare earths, but these are generally less than

10% of the singly charged ions of the same element.
Oxide ion peaks (MO+) are observed for some elements.21 In some cases, doubly charged ions or oxide
ions from abundant elements may interfere with
detection of ions from elements present at lower
abundance.
For some applications, helium may be used in place
of argon in the ICP.22 This eliminates the spectral
interferences due to polyatomic argon-containing
species and provides more efficient ionization of
elements with high ionization potential, such as
halogens.23 Helium microwave-induced plasma sources
operating at subatmospheric pressure have also been
developed, particularly for the detection of halogenated hydrocarbons.24

III. Molecular Ions from Volatile Samples.

A. Electron Ionization (EI)
The electron ionization source was first used by
Dempster,25 and most commercially available sources
are based on the design of Nier.26 A beam of electrons
is directed into molecular vapor at reduced pressure,
one or more secondary electrons are ejected, and
relatively unstable odd-electron or multiply charged
positive ions are produced. Negative ions may also
be produced when samples have high electron affinities.
A schematic diagram of a Nier-type ion source is
shown in Figure 3. Electrons are produced by heating
a metal filament, usually formed from a fine wire or
ribbon of tungsten or rhenium. These electrons are
accelerated by a potential difference between the
filament and the ions source box, pass through
entrance and exit apertures in the box, and are
collected by a trap electrode. Voltage is applied to a
repeller electrode within the box that accelerates ions
toward the ion exit aperture. A collimating magnetic
field is applied parallel to the electron beam axis, and
the field strength is chosen to provide high transmission of the electrons with minimal perturbation of
the ion beam. A field on the order of 100 G is typically
used. The electron current is stabilized by monitoring
either the total emission current or the current to
the trap with feedback to the filament heater power
supply to maintain a constant current. Samples are
introduced directly into the ion source box, and for
most analytical applications, the vapor input rate and
the orifices in the source box must be chosen so that
the probability of ion-molecule reactions is minimal.
In the chemical ionization (CI) source, discussed

Figure 3. Schematic diagram of electron impact ion source

adapted from the original design of Nier. (Adapted with
permission from ref 26. Copyright 1947.)

Vestal

below, the pressure in the ion source box is elevated

and EI is used to generate reagent ions that react
with the sample molecules to produce CI spectra. To
obtain pure EI spectra it is usually necessary to
maintain the ion source pressure below about 0.1
milliTorr. This is particularly important if the results
are to be used for compound identification by comparing spectra with those in a database.
The major applications of EI are the determination
of molecular weight and structure of molecules that
are sufficiently stable to survive vaporization without
decomposition. Most of the advances since Nier have
been in improved techniques for handling difficult
samples, either by chemical derivatization to increase
volatility and/or thermal stability or by novel sample
introduction methods which minimize unwanted pyrolysis. The cross section for ionization of molecules
is a strong function of the electron energy. Ions are
first detected at the appearance potential, and the
ion current increases with electron energy to a
maximum that typically occurs for singly charged
ions in the range of 50-100 eV. Measurements of
appearance potentials by EI for both parent and
fragment ions was a major source of data on ion
energetics in the past,27 but this has been supplanted
by photoionization techniques in more recent work.
For analytical applications, an electron energy of
70 eV has been adopted as the standard. This is a
reasonable, though arbitrary, choice in that it corresponds approximately to the maximum cross section
for most molecules and provides nearly the maximum
absolute intensity for molecular ions while providing
relatively intense fragment ions which carry structural information. Fragmentation of the odd-electron
molecular ions produced by EI is well understood in
principle using the quasi-equilibrium theory (QET)
of Wahrhaftig and co-workers,28 which is equivalent
to the more general RRKM theory of unimolecular
kinetics.29 While this theory has been used to explain
many of the features of ionic fragmentation processes,30 it is difficult to apply in practice without
detailed information on the energetics and structures
of the ions involved. There exists a vast literature
on the interpretation and prediction of EI mass
spectra,31 and not surprisingly, considering the complexity of the processes involved, the former is much
more successful than the latter.
Despite difficulties in interpreting and predicting
EI mass spectra, the spectra observed under standard
conditions are reproducible and largely independent
of the instrument used. Furthermore, the spectra are
highly correlated with molecular structure even
though rapid isomerization may, in some cases, occur
before fragmentation. Thus, fragment ions are often
observed which clearly must be formed following
rearrangement of the neutral molecule ionized. Nevertheless, the relative intensities of the observed
fragment ions may still reflect the isomeric structure
even though the connection between the observed
masses and the structure may not be obvious. For
these reasons, the spectra of compounds to be identified are often treated as molecular fingerprints that
are used in a library search to generate a hit list of
compounds arranged in order of similarity to the

Methods of Ion Generation

unknown. Extensive libraries of EI mass spectra have

been generated over the past 50 years and now
contain some 180 000 entries. Recent work has been
aimed at improving the reliability of this approach
by critical evaluation and correction of the existing
libraries32 and by developing better methods for
generating higher quality data by extracting pure
component mass spectra from complex GC/MS data
files.33
The major limitation of EI mass spectrometry is
the requirement for stable neutral molecules in the
gas phase prior to ionization. Many molecules are
simply not stable at the elevated temperatures
required for vaporization. Chemical derivatization
techniques have been highly developed for at least
partially overcoming this limitation, particularly for
molecules of biological interest,34 but these techniques are of limited use with trace amounts of
material present in complex mixtures. Another approach is based on the observation that intact molecular ions could be observed from relatively nonvolatile samples if they were vaporized rapidly from
weakly bonding surfaces such as fluorocarbons.35 This
approach, which became known as direct or desorption EI, is illustrated in Figure 4, where it is
compared with sample introduction by conventional
solids probe. In this approach samples are vaporized
from a heated wire, which is generally coated with
an inert substrate such as a polyimide, and the
filament is inserted directly into the path of the
electron beam.36 This approach extends the EI technique to somewhat less volatile molecules but is

Figure 4. Illustration of (a) conventional solids probe

method for introducing a solid sample into and ion source
and (b) the direct exposure method. The conventional
method requires vaporization of the sample prior to it entry
into the active region of the ion source, while the direct
method places the solid sample into the ionizing region.
(Adapted with permission from ref 36. Copyright 1980
American Chemical Society.)

Chemical Reviews, 2001, Vol. 101, No. 2 365

limited in utility for molecules of biological interest

since many of these appear not to yield useful
intensities of molecular ions even when vaporized as
intact molecules.

B. Chemical Ionization (CI)

The chemical ionization technique was developed
by Munson and Field37 and initially applied to
hydrocarbons and petrochemicals. The technique
uses an ion source similar to that for EI, as illustrated in Figure 3, except that it is designed to
work at higher ion source pressures of ca. 1 Torr. This
is accomplished by using small ion source apertures,
introducing a reaction gas directly into the ion source
box in addition to the sample, and by using a larger
vacuum pump on the ion source housing to maintain
adequate vacuum in the mass analyzer. The electron
energy is generally increased to 500-1000 eV so that
the electron beam penetrates into the center of the
ion source box. The essential requirements for the
reaction gas are that it produces a set of ions which
are essentially nonreactive with the reaction gas
itself but can undergo exoergic ion-molecule reactions with the samples of interest. The reaction gas
used initially was methane, which produces the
stable reagent ions CH5+ and C2H5+. These react by
either proton transfer or hydride abstraction to form
either MH+ or (M - H)+ ions, and the former
predominate for samples of higher proton affinity.
This new technique aroused some academic interest
but was not widely accepted as an analytical technique until the utility for biologically important
samples not yielding molecular ions by EI was
demonstrated by Fales and co-workers at NIH using
a commercial prototype CI source.38
While the ion source for CI is superficially quite
similar to the EI source, the nature and energetics
of the ions produced are quite different. In EI, a
relatively high-energy incoming particle is used (ca.
70 eV) and the energy lost in the ionizing collision is
shared between the ion and the secondary electron.
This leads to a rather broad internal energy distribution (typically 10-20 eV in nominal width), and
virtually all energetically allowed fragments are
observed. In CI, the collisions between reactant ion
and analyte molecule occur at thermal energies and
the energy available to drive fragmentation is limited
to the exoergicity of the reaction. For proton transfer
reactions, this is simply the difference in proton
affinity between neutral sample molecule and that
of the neutral reactant gas. Furthermore, ionization
occurs in a gas at ca. 1 Torr, where collisional
relaxation may compete with fragmentation. Also,
except for rare gas reactants where charge exchange
may be the dominant reaction, the ions formed are
even-electron species which are inherently much
more stable than the odd-electron molecular ions
produced by EI. Thus, it is not surprising that CI
yields molecular ions from many molecules for which
molecular ions are very weak or absent by EI.
The degree of fragmentation can be controlled by
the choice of reactant gas. This can be correlated with
the gas-phase acidity of the reactant. Thus, hydrogen
or methane may be used when fragmentation is

366 Chemical Reviews, 2001, Vol. 101, No. 2

Vestal

Figure 5. Simplified schematic diagram of a heated pneumatic nebulizer LC/MS interface combined with an atmospheric
pressure chemical ionization source. (Adapted with permission from ref 44. Copyright 1986 American Chemical Society.)

desired and a weaker acid such as ammonia when

predominately protonated molecular ions are the
goal. Of course, if the proton affinity of the sample
molecule is less than that of the reactant, proton
transfer will not occur although complex formation
with the reactant may be observed. This selectivity
is an important feature of CI when the goal is to
determine molecular weight but may be a distinct
disadvantage when the goal is determination of
molecular structure. The problem is that the fragmentation pattern is very dependent on the operating
conditions such as choice of reactant and the temperature and pressure in the ion source. Furthermore, it may be influenced by uncontrolled variables
such as trace impurities, which may drastically
change the energy transferred in the ionization
process. For example, traces of water vapor in
reactants such as methane or hydrogen will cause
H3O+ to be the major reagent ion reacting with
samples which will substantially reduce the energy
transferred and hence reduce the observed fragmentation. Thus, libraries of CI spectra have not been
generated for compound identification because this
lack of reproducibility has generally limited their
potential utility. Dual sources are available on most
commercial GC/MS systems with CI used to determine molecular weight and EI for identification by
searching the MS library. The direct vaporization
probe illustrated in Figure 4 is particularly useful
with CI since the combination of controlled vaporization and soft ionization often allows determination
of molecular weight on difficult samples.
The CI source has been particularly useful for
interfacing with liquid chromatography. Several approaches to interfacing LC with EI mass spectrometry have been described, including the original direct
coupling technique of Talroze39 and the particle beam
interface described by Willoughby and Browner.40
These systems have not been widely used mainly
because of the difficulty of vaporizing liquid and

efficiently transferring samples into the low-pressure

ion source required in EI. In contrast, the higher
pressures used with CI improves heat transfer and
allows larger amounts of solvent vapor to be vaporized in the ion source. Successful interfaces with CI
mass spectrometry include direct liquid introduction
(DLI)41 and thermospray.42
In recent years these earlier LC interfaces have
been largely replaced by atmospheric pressure ion
sources. These have the distinct advantage that the
liquid is vaporized and the sample is ionized at
atmospheric pressure and ideally only the ions in a
bath of relatively dry gas enter the mass spectrometer. This approach was pioneered by Horning and
co-workers43 and has become widely available commercially, typically on instruments that also provide
electrospray ionization. A schematic diagram of an
atmospheric pressure chemical ionization (APCI)
source with a heated pneumatic nebulizer is shown
in Figure 5.44 In this system the LC effluent passes
through the central tube while the nebulizer gas and
makeup gas are introduced coaxially into the heated
region. The combination of heat and gas flow desolvates the droplets to produce a dry vapor of solvent
and analyte molecules. Ionization of solvent molecules is initiated by corona discharge, and the
ionized solvent serves as the chemical ionization
reagent to ionize the samples by ion-molecule reactions. The ions are focused and declustered in a dry
nitrogen curtain gas and pass through as small (ca.
100 m) orifice into the high-vacuum region of the
mass spectrometer. This interface performs well for
samples that survive vaporization as neutral molecules without decomposition but is not suitable for
very nonvolatile or thermally labile samples. For
these electrospray is the method of choice for LC
interfacing, and since the interface with the mass
spectrometer is similar, both techniques are frequently installed on the same MS.

Methods of Ion Generation

C. Photoionization (PI)
The ion source for photoionization is similar to that
shown in Figure 3,45 with the beam of electrons
replaced by a beam of monoenergetic vacuum UV
photons. The photon beam may be produced by a
monochromator with either a conventional discharge
light source or by radiation from a synchrotron. More
recently, laser multiphoton ionization of vibrationally
and rotationally cooled beams of molecules has been
used extensively.46 These ionization methods have
been primarily used to obtain accurate values of
ionization and appearance thresholds for both positive and negative ions. The results provide much of
the basic data required in gas-phase ion and neutral
thermochemistry.47

D. Field Ionization (FI)

The earliest example of a soft ionization technique that produces molecular ions from gas-phase
molecules with little or no fragmentation is the field
ionization technique originally described by Inghram
and Gomer48 and further developed by Beckey and
co-workers.49 In this technique a sharp edge or small
diameter wire is held at high positive potential
relative to a grounded counter electrode. Chemical
etching or growing of whiskers on the metal surface
may be used to further increase the local field
strength by introducing points with very small radii.
At field strengths on the order of 1010 V/m, electrons
may be extracted from gas-phase molecules that
approach within a few angstroms of the metal surface
and transferred to the metal. Very little excess energy
is transferred to the molecules by the ionization
process, and very little fragmentation is observed.
With condensable samples such as water or methanol, protonated molecular ions were observed which
were shown to be produced by condensation on the
emitter tip followed by field desorption of protonated
ions, including under some conditions protonated
clusters. These observations lead to the development
of the field desorption technique for analyzing nonvolatile molecules discussed below.

IV. Molecular Ions from Nonvolatile Samples

By the mid-1970s, mass spectrometry with electron
ionization (EI) and chemical ionization (CI) was
widely recognized as a powerful analytical tool,
particularly in combination with on-line separation
of samples by gas chromatography and with dedicated minicomputer systems for acquiring and processing the data. Applications were limited to those
involving relatively volatile samples, and despite
some success with chemical derivatization to increase
volatility and direct introduction techniques to minimize pyrolysis, it was generally understood that mass
spectrometry was of little use in biology. Even for
small nonvolatile biological samples such as the
amino acid arginine, it was nearly impossible to
produce detectable intensities of molecular ions and,
except for small peptides containing volatile residues,
biological polymers were out of the question. It
appeared that mass spectrometry was reaching maturity, and while the applications of GC-MS were

Chemical Reviews, 2001, Vol. 101, No. 2 367

expanding rapidly, many felt that the exciting work

in developing analytical mass spectrometry was
finished. Then Macfarlane and co-workers50 made the
astounding discovery that intact molecular ions of
cysteine and arginine could be desorbed from thin
layers of these compounds excited by high-energy
fission fragments from 252Cf-radioactive decay. This
was followed by the observation of molecular ions
from difficult molecules such as tetrodotoxin,51
peptides with molecular weights in excess of 3000,52
and protected oligonucleotides up to m/z 12 500.53
These observations revitalized mass spectrometry for
biological applications, and many other techniques
were rapidly developed.

A. Spray Techniques
Static electrification processes were extensively
studied in the laboratory of Professor Leonard Loeb
before and after World War II.54 In one of those
studies, Chapman55 measured ion mobilities produced at atmospheric pressure by vaporization of
charged liquid droplets that were consistent with
molecular ions in the gas phase, although they could
also be explained as relatively small droplets with
an unusally high charge. Dole and co-workers56
developed an electrospray apparatus aimed at producing molecular ions from nonvolatile polymers.
Spraying solutions containing high molecular weight
polystyrene produced mobility spectra that were
consistent with molecular ions of these molecules, but
attempts to observe ions in a time-of-flight mass
spectrometer were unsuccessful. Iribarne and Thompson57 repeated the measurements of Chapman
with a higher resolution mobility apparatus and
concluded that desorption of small molecular ions did
indeed occur. Subsequently they showed that molecular ions from a wide variety of solute species could
be produced by spray electrification and that these
species could be detected and analyzed in a mass
spectrometer.58 Independently, in connection with
developing the thermospray interface between liquid chromatography and mass spectrometry, it was
found that ions were produced in the spray process
when the hot filament normally used for initiating
ionization failed and the ion beam persisted. Furthermore, intact molecular ions were produced by the
spray ionization process from molecules which were
not amenable to normal CI even with direct vaporization.59 Following these earlier observations, Fenn
and co-workers60 applied their expertise in supersonic
molecular beams to developing a new approach,
superficially similar to that described earlier by Dole
but with spectacularly different results.

B. Electrospray
The work by Fenn and co-workers in developing
electrospray ionization was largely ignored by the
mass spectrometry community until, before a small
audience at the ASMS meeting in 1988, they presented their first results on proteins.61 A small
sample of the early results is shown in Figure 6. The
fact that these highly charged molecular ions from
proteins as large as 40 000 molecular weight could

368 Chemical Reviews, 2001, Vol. 101, No. 2

Vestal

Figure 6. Examples of the first electrospray ionization spectra of proteins. Each spectrum is the result of a single 30 s
scan over the indicated range of m/z values and comprises a family of multiply charged peaks from the same nominal
mass. In this case the charge carrier is a proton, and the molecular weight of the molecule and the charge state for each
peak can be calculated from measured m/z values for two or more peaks. (Reprinted with permission from ref 61. Copyright
1988.)

be observed with no evidence of fragmentation was

truly astounding. The apparatus used in these experiments is shown schematically in Figure 7. This
system employs a glass capillary in place of the
simple nozzle used in an earlier configuration.62 This
allows the electrospray needle to be operated at
ground potential and is convenient for interfacing to
LC. For positive ions with the needle at ground,
typical applied voltages are as follows: cylindrical
electrode, -3500; metallized inlet end of the glass
capillary, -4500; exit end, +40; skimmer, +20; and
entrance lens in front of the quadrupole, ground.
Negative ions are produced by reversing the polarity
of the voltages; in addition, it is useful to add a small
stream of oxygen or other electron scavenger near
the needle to inhibit the onset of corona discharge,
which occurs at lower voltages in the negative-ion
mode. The mobility of ions in the capillary is sufficiently low that the gas flow through the capillary

Figure 7. Schematic diagram of and electrospray ion

source. (Reprinted with permission from ref 62. Copyright
1985 American Chemical Society.)

can drag the ions up the potential gradient and

deliver them to the exit. This approach has also been
used successfully for interfacing with magnetic mass
spectrometers where the ion source may be as much
as +15 kV relative to the needle at ground potential.63
Sample solution enters through the stainless steel
needle at flow rates usually between 1 and 40 L/
min. The field at the needle tip charges the surface
of the emerging liquid, and Coulomb forces disperse
it into a fine spray of charged droplets. The droplets
are driven by the electric field toward the inlet end
of the capillary through a countercurrent flow of bath
gas (dry nitrogen) typically at 800 Torr, entering
temperature of 320-350 K, and flow rate of about
100 mL/min. The solvent vapor from the evaporating
droplets along with any other uncharged material is
swept away from the capillary inlet by the bath gas.
As the highly charged droplets vaporize, they become
unstable, subdivide, and eventually produce molecular ions. If species are present as solutes which can
accommodate large numbers of elemental charges,
such as proteins, these are observed in the mass
spectrometer as shown in Figure 6. Some of these
ions are entrained in the flow of dry bath gas that
enters the glass capillary and emerge at the exit end,
carried along in a supersonic free jet into the first of
two vacuum chambers. A core portion of this free jet
passes through a skimmer into a second vacuum
chamber and delivers ions to the mass analyzer.
The details of the mechanisms by which molecular
ions are desorbed from evaporating charged liquid
droplets are still somewhat controversial, but the
major features are now generally accepted. The
droplets are initially formed with a charge near but

Methods of Ion Generation

somewhat below the Rayleigh limit,64 which corresponds to the droplet radius at which the forces of
surface tension holding the droplet together are equal
to the Coulomb repulsive forces between the charges.
As the droplet vaporizes, this limit is exceeded and
the droplet becomes unstable. This hydrodynamic
instability, at least for larger droplets, results in
release of a jet of small charged droplets. As the
initial droplet vaporizes further, this process is
repeated until the liquid is entirely dispersed. There
is general agreement about the first part of this
process, but the mechanism leading to molecular ions
from very small charged droplets has been the source
of some debate. Iribarne and Thomson58 proposed the
model of field-induced ion evaporation in which the
high electric field at the surface of a small charged
droplet was sufficient to make ion evaporation competitive with solvent vaporization. Roellgen65 challenged this model on the basis that such high
localized fields would cause the onset of hydrodynamic instability on the surface producing a jet
releasing a number of charged droplets. Roellgens
view is that the hydrodynamic disintegration continues until the droplets are so small that they
contain only a single solvated ion, and further desolvation leads to the ions observed in the mass
spectrometer. It is fairly clear that the observed
energetics for the ion evaporation model requires that
the emitted ions must be heavily solvated. If so, the
two models become almost indistinguishable and
mostly semantic since in the one case we have a
solvated ion and in the other a very small charged
droplet containing a single species which can compete
with the solvent for the charge as vaporization
proceeds. In either case, to efficiently produce ions
by electrospray it is desirable to produce small
charged droplets with as much charge as possible and
cause those droplets to vaporize in a bath of gas. If
the concentration of sample or nonvolatile salts is too
high, then the efficiency of ionization is low since the
droplets may dry to a solid particle without releasing
much of the sample into small droplets that can
produce molecular ions.
A number of variations on the design of electrospray ion sources have been described in the literature, and many of these are in use on commercial
instruments. These variations are mostly concerned
with the sampling of the electrospray ions into the
mass spectrometer to maximize sensitivity and improve reliability. In some of these, the countercurrent
flowing bath gas is not used and residual droplets
and some of the solvent vapor enters the nozzle
interfacing with the mass spectrometer. In these
systems vaporization is completed and resolvation
prevented by using either a heated capillary nozzle66
or a heated chamber downstream of the skimmer.67
In applications requiring continual sampling of the
effluent from an HPLC, a major concern has been
contamination of surfaces in the interface or mass
spectrometer by nonvolatile components which may
cause the system to fail and require it to be disassembled and cleaned frequently. Several ingenious
schemes have been described recently in an attempt
to efficiently sample the ions but reject any liquid

Chemical Reviews, 2001, Vol. 101, No. 2 369

droplets or macroscopic particles which might be

present.68,69
Electrospray is a concentration-dependent ionization method that works best at low flow rates. By
reducing the diameter of the electrospray needle,
stable sprays of very fine droplets can be produced
at flow rates of 1 nL/min or even less.70 Furthermore,
these nanospray needles work at very low voltages
(ca. 1000 V or less) because of the very high field
produced at the small tip, and they can be located
very close to the nozzle sampling ions into the mass
spectrometer since the flow of solvent is so low. This
approach allows very high sensitivity for small
amounts of sample, but because of the small orifices
involved, special care must be taken to avoid plugging.

C. Desorption from Surfaces

With the benefit of 30 years of hindsight, it is now
intuitively obvious that spray techniques should be
very soft ionization methods. The concepts of disrupting liquids to produce very fine droplets, charging the
droplets, and gently vaporizing the droplets to remove the neutral solvent leaving behind a multiply
charged protein ion in the gas phase now seems
perfectly reasonable. Most of the surface desorption
techniques do not. One exception is the field desorption technique (FD) developed primarily by Beckey
and co-workers.71 This might be considered a forerunner of the spray techniques in that it involves
application of a high electric field to a liquid surface
resulting in the production of intact molecular ions
from nonvolatile samples. The major difference is
that FD is done in the vacuum of the mass spectrometer and the spray techniques require a bath gas at
higher pressure. The mechanism of ion production
for both positive and negative ions by FD has been
extensively studied by Roellgen and co-workers.72
Even though this technique is not widely practiced
at present, these studies have provided much of the
fundamental understanding of the processes involved
in spray ionization.

D. High-Energy Particle Impact

The idea that Mev fission fragments impacting on
surfaces can produce molecular ions from nonvolatile
and thermally labile molecules is certainly not intuitive. This ionization method was discovered as a
byproduct of studies on beta decay by Macfarlane and
co-workers.50 These studies used time-of-flight techniques to measure the time between emission of the
electron and detection of the recoil partner at a
detector several meters away. Peaks were observed
in these spectra that were not correlated with beta
decay. The conclusion was that contaminants on the
surface were also being ionized and emitted. The beta
emitter was replaced by a 252Cf fission fragment
source to increase the energy deposition on the
surface, and the surface was doped with known
nonvolatile samples. In the earliest experiments,
protonated molecular ions were observed from arginine, a molecule generally considered impossible by
techniques then in use.

370 Chemical Reviews, 2001, Vol. 101, No. 2

Figure 8. Schematic of a

252Cf-PDMS

Vestal

experiment. (Reprinted with permission from ref 74. Copyright 1985.)

A schematic diagram of a plasma desorption system is shown in Figure 8.74 Ionization is initiated by
spontaneous fission of a 252Cf nucleus which produces
fission fragments with energies of about 100 Mev and
charges of +20 or so. Results have also been obtained
using similar beams of particles from an accelerator.73 The linear energy transfer is these ions in a
solid is so high (0.1-1 keV/) that it is important
that the source be thin. A typical source strength is
10 microcuries spread over a circular area 3-5 mm
in diameter. The thickness of the californium layer
is about 1 ng/cm2. The foils covering the source must
be thin enough to allow one of the fission fragments
to reach the start detector and the other fragment
to pass through the foil and ionize the sample
deposited on the surface of the foil. A grounded grid
is placed 2-5 mm in front of the sample surface, and
a high voltage (ca. 5-10 kV) is applied to the sample
to accelerate ions toward the detector.
Prior to electrospray and MALDI, this technique
was becoming a standard tool in mass spectrometry
laboratories focusing on biological samples74 and was
commercially available. At that time it was the only
technique available that could reliably produce molecular ions from proteins with molecular weight over
10 000. The spectrum of one of the larger proteins
measured is shown in Figure 9; this illustrates both
the power and the problems with the technique. The
rather low quality of the spectrum, by contemporary
standards, is now understood to be primarily due to
metastable decomposition of the ions in flight. The
plasma desorption technique has been largely supplanted by electrospray and MALDI but marked an
extremely important milestone in the development
of ionization techniques for mass spectrometry of
difficult samples. It both renewed interest in the
largely dormant field of time-of-flight mass spectrometry, particularly for large molecules, and showed for

Figure 9. Positive-ion 252Cf-PDMS of PSP (pancreatic

spasmolytic polypoptide). Each time channel is 32 ns wide;
counting time 8 h with a source producing 1450 fragments/
sec. (Reprinted with permission from ref 129. Copyright
1984.)

the first time that such incredible ions could exist in

the gas-phase long enough to be measured.

E. Low-Energy Particle Impact

Shortly after the publication of Macfarlanes work,
Benninghoven and co-workers presented the first
results on ionization of nonvolatile molecules by lowenergy ion impact.75 Mass spectrometric analysis of
ions produced by bombarding surfaces with keV ions
was already well established as a technique for
analysis of surfaces.76 The early applications were to
fundamental studies which were based on the fact
that secondary-ion mass spectrometry (SIMS) provided high absolute sensitivity for many surface
components and unlike other available techniques
could detect hydrogen and determine isotopic compositions of elements on the surface.77 Benninghoven
et al.78 showed that SIMS was a very sensitive
technique for the detection and identification of
organic compounds provided it is operated in the
static mode. The main problem in detecting intact

Methods of Ion Generation

molecules in SIMS can be the damaging effect of the

primary-ion beam. This effect can be minimized by
operating in the static mode in which the surface
is bombarded at very low primary current density.
Damage cross sections are typically on the order of
10-14 cm2; keeping the total dose of primary ions
below about 1012/cm2 ensures a small probability that
a surface area damaged by the impact of an ion is
struck by another primary ion during the measurement. The static SIMS technique has been applied
to a wide variety of organic compounds including
amino acids, peptides, nucleosides, nucleotides, vitamins, and many others.79 Almost all of the spectra
show high abundances of protonated or cationized
molecular ions. Secondary yields as high as 0.1
molecular ion per incident primary ion have been
measured, but the yields are strongly dependent on
the substrate and the properties of the primary beam.
Noble metal substrates generally produce the highest
yields.
Most of the early SIMS work employed either
magnetic deflection or quadrupole analyzers, and the
need to work in the static mode together with the
ion losses caused by the duty cycle of scanning
analyzers made the sensitivity lower than desired.
One solution to this problem was the use of the timeof-flight analyzer that allowed nearly every ion
desorbed by a pulse of primary ions to be detected.80
Also, the linear time-of-flight analyzer only requires
that the ions be stable for the time required to
traverse the accelerating field; thus, ions that would
go undetected in the scanning instruments may be
detected as the result of fragments arriving with
essentially the same time-of-flight as intact molecular
ions. Significant yields of molecular ions were obtained from a number of intractable compounds.
Spectra obtained by TOF-SIMS are very similar to
those from 252Cf TOF, as shown by detailed comparisons on the same samples.81,82 In both cases, ions of
higher molecular weight are predominately metastable. Neutral particles in the keV range have also
been used, but the results are similar to those with
ions of comparable energy.83 Since it is generally
understood that ions approaching a surface with keV
energies are probably neutralized by long-range
electron transfer as they approach the surface, it is
not surprising that the charge on the particle is
unimportant.

F. Low-Energy Impact with Liquid Surfaces

Another successful approach to improving the
utility of low-energy impact ionization for nonvolatile
molecules was developed by Barber and co-workers.84
This technique differs from SIMS in two respects.
First, the primary beam is a keV-energy neutral
beam produced by charge exchange neutralization of
an ion beam, and second, the sample is introduced
as a solution or suspension in a relatively nonvolatile
liquid such as glycerol. It is now clear that the
presence or absence of charge on the incident particle
has little effect on the desorption process, but the use
of the neutral beam is more convenient with magnetic
instruments where the source is at high potential.
Replacing the solid surface with a liquid matrix has
pronounced effects.

Chemical Reviews, 2001, Vol. 101, No. 2 371

The most striking effect of the liquid matrix is

removal of the requirement for working at low
current densities to avoid damage to molecules on
the surface. Current densities several orders of
magnitude larger than those suggested by Benninghoven78 have been used routinely, and in many cases,
production of protonated molecular ions from nonvolatile samples has persisted at stable intensities
for 30 min or more.85 The rationale for this effect is
that solution presents a mobile, constantly renewed
surface to the bombarding beam. This provides
continuous replenishment of undamaged sample
molecules to the surface to be ionized. Work by Field86
showed that substantial radiation damage to the
glycerol does occur; his results indicate that approximately 100 molecules of radiation-damaged
products are produced per incident 5-keV argon atom.
One disadvantage of the liquid matrix is that a
substantial background is produced by ionization of
the glycerol and its radiation-induced fragments.
Also, substantial clustering of glycerol with sample
ions is observed together with mixed cluster ions
when impure samples or mixtures are analyzed. This
technique, known as fast atom bombardment (FAB)
or liquid SIMS, depending of the charge on the
incident beam, has been successfully applied to a
wide variety of difficult samples.87 Part of this success
can be ascribed to its compatibility with highperformance magnetic instruments; these were the
instruments most widely used at the time this
technique was developed, and FAB or LSIMS sources
were commercially available from all of the major
manufacturers almost immediately after the initial
publications.

G. Flow FAB
Continuous-flow FAB was developed to correct
some of the difficulties with the standard FAB
technique while retaining the essential advantages.88
This technique uses a sample introduction probe that
provides a continuous flow of liquid into the mass
spectrometer ion source and onto the sample stage
where atom bombardment occurs. This reduces the
requirement for a viscous carrier and allows the use
of more volatile solvents such as water, methanol,
and acetonitrile. The liquid flow rate is generally in
the range of 1-20 L/min, and samples may be
analyzed by flow injection, or following on-line separation by capillary LC,89 or capillary electrophoresis.90
In addition to the obvious advantages of convenience
and speed of sample introduction, the flow technique
also provides lower limits of detection91 and decreased
ion suppression effects92 compared to conventional
FAB.

H. Laser IonizationMALDI
Laser ionization of organic solids was investigated
by Mumma and Vastola93 in the late 1960s and early
1970s, but most of the early work on laser ionization
was focused on inorganic samples. The early work
has been reviewed by Conzemius and Capellen.,94
During the next decade, several groups conducted
research on improved methods for applying lasers to

372 Chemical Reviews, 2001, Vol. 101, No. 2

Figure 10. MALDI mass spectrum of -D-galactosidase

in the parent molecule range. Sum of 100 laser shots (266
nm) with nicotinic acid matrix. This is the first mass
spectrum reported showing intact molecular ions in the gas
phase with MW greater than 100 000. (Reprinted with
permission from ref 98. Copyright 1989.)

ionization of organic samples,95,96 but until the breakthrough by Karas and Hillenkamp in 1988, these
techniques were not widely applied.97,98 By embedding large bioorganic molecules in a suitable matrix
which strongly absorbs the radiation from a UV laser,
they showed that singly charged molecular ions with
masses greater than 100 000 could be efficiently
desorbed and measured in a time-of-flight mass
spectrometer. An example of one of the spectra from
the original publication is shown in Figure 10.98 At
the time, this example of -D-galactosidase, molecular
mass 116 900, was by far the largest intact molecular
ion that had been observed. This matrix-assisted
laser desorption/ionization (MALDI) technique has
rapidly evolved to become one of the predominant
methods for analyzing large, nonvolatile molecules.
Since the ions are produced in short pulses using a
pulsed laser, the technique is particularly well suited
for time-of-flight analyzers, but it is also used quite
successfully with FT-ICR and with quadrupole ion
traps.
A schematic diagram of a MALDI source is shown
in Figure 11.103 Samples in solution are mixed with
a matrix solution containing a large molar excess (ca.
104-105) of a UV-absorbing matrix material. This
liquid is deposited in a small droplet, a few L or less
in volume, on a sample probe or plate and allowed
to dry. The early work employed a single sample
probe, similar to solids probes used in other areas,
but most commercial systems now use a sample plate
holding a number of samples and a mechanism for
sequentially positioning the samples in front of the
laser. A laser with pulse duration of a few nanoseconds is preferred. Nitrogen lasers emitting at 337 nm

Figure 11. Schematic drawing of a MALDI ion source.

(Reprinted with permission from ref 103. Copyright 1997.)

Vestal

are the most widely used, but frequency tripled and

quadrupled Nd:YAG lasers emitting at 355 and 266
nm are also used. The angle of irradiance to the
sample surface varies between 15 and 70 and is not
critical. The laser is focused to a spot size between
30 and 500 m, and the required irradiance is
typically between 106 and 107 W/cm2 depending on
the spot size and matrix employed.101 Both neutral
and ion desorption show a strong dependence on
applied laser irradiance. For a given sample, there
is a distinct threshold value for producing detectable
ion intensities and above that value the ion yield
increases rapidly with increasing irradiance. Typically, the ion current is proportional to the fifth or
sixth power of irradiance above threshold.101 Best
results are obtained with laser irradiances not more
than 20-50% above threshold; thus, a carefully
controlled attenuator is required to adjust the laser
intensity.
Some of the details of the mechanisms involved in
MALDI are still somewhat controversial and are the
subject of research by several investigators.99-102
Nevertheless, the major features are established and
have been reviewed recently by Karas and Bahr.103
The matrix serves several functions. First, it must
provide strong absorption of the incident laser light.
This allows transfer of energy from the laser to the
solid sample in a controllable way. Typical operating
conditions are laser irradiance in the 106-107 W/cm2
range, laser pulse length of a few nanoseconds, and
molar absorption coefficients for the matrix of a few
thousand L mol-1 cm-1. This corresponds typically
to the order of 1 photon absorbed per matrix molecule
in the outermost layer. This results in exponentially
decaying in-depth excitation with typical penetration
depths of ca. 100 nm. This energy is sufficient to
induce the ablation of a small volume, setting free
intact matrix and analyte neutrals and ions. Another
important function of the matrix is to isolate individual sample molecules and prevent aggregation.
The analyte is typically diluted in a large excess of
matrix (ca. 10 000-fold or more on a molar basis), and
it appears that molecules are isolated as a solid
solution in the matrix. Incorporation of analyte
molecules into matrix crystals formed upon evaporation of the solvent is desirable and for some matrixes
may be essential. The matrix also appears to play a
role in the ionization process, but this is not fully
understood.
The understanding of the MALDI process is not yet
sufficient to allow selection of a matrix on theoretical
considerations alone. The required absorption at the
UV-laser wavelength is not a very restrictive criterion; it is necessary but not sufficient for a good
matrix. A large number of materials have been
evaluated, and only a few have been found satisfactory. Compatibility of the solvents required for analyte and matrix, absence of chemical reactivity with
analytes, and stability in a vacuum are some of the
properties required of a matrix. A number of the
compounds that have been tried empirically can work
as matrixes, but very few are presently regarded as
high-quality matrixes for UV-MALDI. These include
sinapic acid and some related cinnamic acid deriva-

Methods of Ion Generation

Chemical Reviews, 2001, Vol. 101, No. 2 373

tives,104,105 2,5-dihydroxybenzoic acid (DHB)106 often

with an admixture of 10% 2-hydroxy-5-methoxybenzoic acid(super DHB),107 R-cyano-4-hydroxycinnamic
acid (4-HCCA),108 3-hydroxypicolinic acid (HPA),109
and di- and triacetophenones.110 The first three are
generally used for proteins and peptides and the
latter two for DNA oligomers. A number of other
matrixes satisfying the co-solubility requirement
have been employed with industrial polymers.111

I. IR MALDI
The use of infrared (IR) lasers instead of the UV
lasers traditionally employed in MALDI has been
studied by several research groups.112-114 It has been
demonstrated that approximately the same performance can be obtained with either laser source with
respect to mass resolution, mass accuracy, and
analytically useful sensitivity.114 The major differences observed are that IR generally gives less
fragmentation of large molecules, but much more
sample is consumed per laser shot. The ionization
efficiency is generally several orders of magnitude
lower with IR than with UV, but the sensitivity is
generally limited by chemical noise from matrix and
other impurities, and a much smaller fraction of the
sample is desorbed in typical UV measurements. The
IR laser also penetrates more deeply into the matrix
and allows sample to be successfully analyzed from
gels and membranes. On the other hand, UV MALDI
detects sample present at or near the surface and
does not detect molecules embedded more deeply. The
major limitation on the use of IR lasers in MALDI is
that the available lasers are rather expensive and
somewhat more difficult to use routinely.

J. Delayed Extraction
Electrospray was more rapidly accepted than
MALDI for many routine applications primarily
because it was readily adapted to the quadrupole and
magnetic mass analyzers predominately in use. The
pulsed ion beam produced by MALDI was particularly compatible with time-of-flight (TOF) techniques,
but the resolution and mass accuracy obtained with
MALDI-TOF on larger molecules was not as good as
that obtained with electrospray on either quadrupole
or magnetic analyzers. The use of reflecting TOF
instruments such as that described originally by
Mamyrin115 improved the performance for peptides
and other small molecules, but the resolution and
accuracy for oligonucleotides and proteins was similar for both linear and reflecting analyzers and
clearly inferior to that obtained by electrospray. Early
studies showed that the ions produced by MALDI
exhibit a rather broad velocity distribution and that
the initial velocity of desorbed ions is nearly independent of mass.116,117 Also, when desorption occurs
in a strong electric field, energy is lost presumably
by collisions with the neutral plume.
In pioneering work, Wiley and McLaren118 described a two-field, pulsed ion source for time-of-flight
analyzers that provided first-order correction for the
initial space distribution of ions. They also described
a technique which they named time-lag energy

Figure 12. Molecular ion region of the MALDI-TOF mass

spectrum of two small proteins, bovine ubiquitin and
cytorchrome c, obtained using a state-of-the-art reflecting
analyzer with a delayed extraction source. Measured
resolving powers (fwhm) are indicated on the spectra.

focusing to correct for initial velocity distributions.

The ions are produced in a field-free region, and the
accelerating field is turned on by application of a fast
pulse at a predetermined delay time. The correct
choice of delay time and accelerating field allows the
initially slow ions to receive enough additional energy
to catch the initially fast ions at the ion detector. In
MALDI, the ions are produced from a nearly flat
surface; thus, space focusing is not required, but
velocity focusing greatly improves performance. The
improvement in performance resulting from delayed
extraction has been independently demonstrated in
a number of laboratories119-123 and is now used for
almost all analytical applications of MALDI. An
example of the resolving power presently possible by
MALDI-TOF is shown in Figure 12. Mass measurement accuracy with internal calibration is typically
better than 10 ppm for such spectra. Higher resolving
power can be obtained at low m/z by FT-ICR and
double-focusing magnetic mass spectrometry, but for
singly charged ions produced by MALDI, the results
of Figure 12 represent the current state of the art
for any available mass spectrometric technique.124

K. New Mass Analyzers

It is difficult to discuss methods of ionization
without occasionally straying across the boundary
with mass analyzers. The evolution of mass analyzer
designs has strongly interacted with that of ionization methods. Also, the potential applications have
sometimes driven the development of the instrumentation, but more often, the applications have been
determined by the capabilities of commercially available systems. MALDI and electrospray have made
mass spectrometry applicable to very large molecules
which were not previously accessible. This has driven
the development of improved mass analyzers with
larger mass range and improved resolving power,
mass accuracy, and sensitivity. Since these new
ionization techniques often produce molecular ions
with little or no fragmentation, development of
tandem mass spectrometers compatible with electrospray and MALDI has been essential for obtaining
structural information. Most of these developments
have been in the areas of TOF,125,126 quadrupole ion
traps,127 and Fourier transform ion cyclotron resonance (FT-ICR).128 Dramatic improvements in these

374 Chemical Reviews, 2001, Vol. 101, No. 2

techniques have occurred during the past few years,

particularly for biological applications, and coupled
with the new ionization techniques these have revolutionized the practice of analytical mass spectrometry for nonvolatile molecules.

V. Present Status and Future Prospects

The methods of ionization presently available for
mass spectrometry are briefly summarized above. It
is now possible to generate either atomic ions or
molecular ions from almost any material and to
efficiently introduce these ions into the vacuum of
the mass spectrometer for analysis. Except for large,
nonpolar organic polymers that are difficult to ionize,
there appears to be no limit on the size and complexity of molecules that can be ionized and analyzed.
During the past decade no new ionization methods
have been introduced, but enormous advances have
been made in the sensitivity and utility of existing
methods. These developments have occurred in parallel with dramatic improvements in mass analyzers
with extended mass range, higher resolution, improved mass accuracy, and vastly improved sensitivity and speed. Applications for these techniques have
grown exponentially, particularly in the life sciences.
It appears that mass spectrometer technology may
be approaching maturity, but the applications are
still in their infancy.

VI. References
(1) Thomson, J. J. Rays of Positive Electricity; Longmans, Green
and Co.: London, 1913.
(2) DeLaeter, J. R. Mass Spectrom. Rev. 1994, 13, 3-22.
(3) Ionov, N. I. Surface Ionization and Its Applications. In Progress
in Surface Science; Pergamon: New York, 1972; Vol. I, pp 237354.
(4) Inghram, M. G.; Chupka, W. A. Rev. Sci. Instrum. 1953, 24. 518.
(5) Cornides, I. Advances in Mass Spectrometry; Todd, J. F. J., Ed.;
Wiley: 1985; p 133.
(6) Deron, S. Advances in Mass Spectrometry; Longevialle, Heydon,
Ed.; London, 1988; Vol. 11B, p 1746.
(7) Dempster, A. J. Rev. Sci. Instrum. 1936, 7, 46.
(8) Chakravarty, B.; Venkatasubramanian, V. S.; Duckworth, H. E.
Advances in Mass Spectrometry; Elliot, R. M., Ed.; Pergamon:
New York, 1963; Vol. 2, pp 128-134.
(9) Brown, R.; Craig, R. D.; Elliott, R. M. Advances in Mass
Spectrometry; Elliot, R. M., Ed.; Pergamon: New York, 1963;
Vol. 2, pp 141-156.
(10) Hieftje, G. M.; Norman, L. A. Int. J. Mass Spectrom. Ion
Processes 1992, 118, 519-573.
(11) Harrison, W. W.; Hess, K. R.; Marcus, R. K.; King, F. L. Anal.
Chem. 1986, 58, 341A.
(12) Mason, R.; Milton, D. Int. J. Mass Spectrom. Ion Processes 1989,
91, 209-225.
(13) Duckworth, D. D.; Marcus, R. K. Anal. Chem. 1989, 61, 1879.
(14) Chapman, B. Glow Discharge Processes; Wiley: New York, 1980.
(15) Harrison, W. W.; Barshick, C. M.; Klinger, P. H.; Ratliff, P. H.;
Mei, Y. Anal. Chem. 1990, 62, 943A.
(16) Fassel, V. A. Science 1978, 202, 183-191.
(17) Pertel, R. Int. J. Mass Spectrom. Ion Processes 1975, 16, 3952.
(18) Gray, A. L. Analyst 1975, 100, 289-299.
(19) Houk, R. S.; Fassel, V. A.; Flesch, G. D.; Svec, H. J.; Gray, A. L.;
Taylor, C. E. Anal. Chem. 1980, 52, 2283-2289.
(20) Houk, R. S. Anal. Chem. 1986, 58, 97A.
(21) Gray, A. L.; Williams, J. G. J. Anal. At. Spectrom. 1987, 2, 595.
(22) Montaser, A.; Chan, S. Anal. Chem. 1987, 59, 1240.
(23) Kopenaal, D. W.; Quinton, L. F. J. Anal. At. Spectrom. 1988, 3,
667.
(24) Creed, J. T.; Davidson, T. M.; Shen, W.-L.; Caruso, J. A. J. Anal.
At. Spectrom. 1990, 5, 109.
(25) Dempster, A. J. Phys. Rev. 1921, 18, 415.
(26) Nier, A. O. Rev. Sci. Instrum. 1947, 18, 398.
(27) Rosenstock, H. M. Int. J. Mass Spectrom. Ion Phys. 1976, 20,
139.

Vestal
(28) Rosenstock, H. M.; Wallenstein, M. B.; Warhaftig, A. L.; Eyring,
H. Proc. Natl. Acad. Sci. U.S.A. 1952, 38, 667.
(29) Robinson, P. J.; Holbrook, K. A. Unimolecular Reactions;
Wiley: New York, 1972.
(30) Vestal, M. L. Ionic Fragmentation Processes. Fundamental
Processes in Radiation Chemistry; Ausloos, P., Ed.; Interscience: New York, 1968; Chapter 2.
(31) McLafferty, F. W. Interpretation of Mass Spectra; University
Science Books: Mill Valley, CA, 1980.
(32) Ausloos, P.; Clifton, C. L.; Lias, S. G.; Mikaya, A. I.; Stein, S.
E.; Tchekhovskoi, D. V.; Sparkman, O. D.; Zaikin, V.; Zhu, D. J.
Am. Soc. Mass Spectrom. 1999, 10, 287-299.
(33) Stein, S. E. J. Am. Soc. Mass Spectrom. 1999, 10, 770-781.
(34) Knapp, D. R. Chemical Derivatization for Mass Spectrometry.
In Methods in Enzymology; McCloskey, J. A., Ed.; Academic
Press: New York, 1990; Vol. 193, pp 314-329.
(35) Bluehler, R. J.; Flanigan, E.; Greene, L. J.; Friedman, L. J. Am.
Chem. Soc. 1974, 96, 3990.
(36) Cotter, R. J. Anal. Chem. 1980, 52, 1589A-1606A.
(37) Munson, M. S. B.; Field, F. H. J. Am. Chem. Soc. 1966, 88, 26212630.
(38) Fales, H. M.; Milne, G. W. A.; Vestal, M. L. J. Am. Chem. Soc.
1969, 91, 3682.
(39) Talroze, V. L.; Skurat, V. E.; Karpov, G. V. Russ. J. Phys. Chem.
1969, 43, 214.
(40) Willoughby, R. C.; Browner, R. F. Anal. Chem. 1984, 56, 2626.
(41) Baldwin, M. A.; McLafferty, F. W. Org. Mass Spectrom. 1973,
7, 1353.
(42) Vestal, M. L. Science 1984, 221, 275.
(43) Horning, E. C.; Carroll, D. I.; Dzidic, I.; Haegele, K. D.; Horning,
M. G.; Stilwell, R. N. J. Chromatogr. Sci. 1974, 12, 1974.
(44) Covey, T. R.; Lee, E. D.; Bruins, A. P.; Henion, J. D. Anal. Chem.
1986, 58, 1451A.
(45) Hurzeler, H.; Inghram, M. G.; Morrison, J. D. J. Chem. Phys.
1958, 28, 76.
(46) Boesl, U.; Grotemeyer, J.; Mueller-Dethlefs, K.; Neussser, H. J.;
Selze, H. L.; Schlag, E. W. Int. J. Mass Spectrom. Ion Processes
1992, 191-220.
(47) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin,
R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17, 1-861.
(48) Inghram, M. G.; Gomer, R. J. J. Chem. Phys. 1954, 22, 1279.
(49) Beckey, H. D. Principles of Field Ionization and Field Desorption
Mass Spectrometry; Pergamon Press: Oxford, 1977.
(50) Torgerson, D. F.; Skowronski, R. P.; Macfarlane, R. D. Biochem.
Biophys. Res. Commun. 1974, 60, 616-621.
(51) Macfarlane, R. D.; Torgerson, D. F. Science 1976, 191, 920.
(52) Macfarlane, R. D.; McNeal, C. J.; Hunt, J. E. Adv. Mass
Spectrom. 1980, 8A, 349.
(53) McNeal, C. J.; Macfarlane, R. D. J. Am. Chem. Soc. 1981, 103,
1609.
(54) Loeb, L. Static Electrification; Springer-Verlag: Berlin, 1958.
(55) Chapman, S. Phys. Rev. 1937, 52, 184. Chapman, S. Phys. Rev.
1938, 54, 520. Chapman, S. Phys. Rev. 1938, 54, 528.
(56) Dole, M.; Mack, L. L.; Hines, R. L.; Mobley, R. C.; Ferguson, L.
D.; Alice, M. B. J. Chem. Phys. 1968, 49, 2240. Mack, L. L.;
Kralick, P.; Rheude, A.; Dole, M. J. Chem. Phys. 1970, 52, 4977.
(57) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1976, 64, 2287.
(58) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1979, 71, 4451.
(59) Blakley, C. R.; Carmody, J. J.; Vestal, M. L. J. Am. Chem. Soc.
1980, 102, 5931.
(60) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4451.
Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4471.
(61) Mann, M.; Meng, C. K.; Fenn, J. B. Proceeding of 36th Annual
Conference on Mass Spectrometry and Allied Topics; San Francisco, 1988; pp 1207-08.
(62) Whitehouse, C. M.; Dreyer, R. N.; Yamashita, M.; Fenn, J. B.
Anal. Chem. 1985, 57, 675.
(63) Whitehouse, C. M.; Yamashita, M.; Meng, C. K.; Fenn, J. B.
Proceeding of the 14th International Symposium on Rarified Gas
Dynamics, July 16-20, 1984; Oguchi, H., Ed.; University of
Tokyo Press: Tokyo, 1984; p 857.
(64) Lord Rayleigh, J. W. Philos. Mag. 1882, 14, 184.
(65) Schmelzeisen-Redeker, G.; Buftering, L.; Roellgen, F. W. Int. J.
Mass Spectrom. Ion Phys. 1989, 90, 139.
(66) Chowdhury, S. K.; Katta, V.; Chait, B. T. Rapid Commun. Mass
Spectrom. 1990, 4, 81.
(67) Allen, M. H.; Vestal, M. L. J. Am. Soc. Mass Spectrom. 1992, 3,
18.
(68) Bertrand, M.; Vayer, P.; Perez, J. C.; McDowell, M.; Preece, S.
Proc. 46th ASMS Conf. Mass Spectrom. 1998, p 13.
(69) Shaffer, S. A.; Tang, K.; Anderson, G. A.; Prior, D. C.; Udseth,
H. R.; Smith, R. D. Rapid Commun. Mass Spectrom. 1997, 11,
1813.
(70) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1-8.
(71) Beckey, H. D. Int. J. Mass Spectrom. Ion Phys. 1969, 2, 500.
(72) Giessmann, U.; Roellgen, F. W. Int. J. Mass Spectrom. Ion Phys.
1981, 38, 267.

Methods of Ion Generation

(73) Sundqvist, B.; Hedin, A.; Hakansson, P.; Kamensky, I.; Kjellberg,
J.; Salehpour, M.; Save, G.; Widdiyasekera, S. Int. J. Mass
Spectrom. Ion Phys. 1983, 53, 167-183.
(74) Sundqvist, B.; Macfarlane, R. D. Mass Spectrom. Rev. 1985, 4,
421-460.
(75) Benninghoven, A.; Jaspers, D.; Sichtermann, W. Appl. Phys.
1976, 11, 35-39.
(76) Liebl, H. Anal. Chem. 1974, 46, 22A.
(77) Benninghoven, A. Surf. Sci. 1973, 35, 427; 1975, 53, 596.
(78) Benninghoven, A.; Sichtermann, W. Anal. Chem. 1978, 50,
1180-1184.
(79) Grade, H.; Winograd, N.; Cooks, R. G. J. Am. Chem. Soc. 1977,
99, 7725.
(80) Ens, W.; Standing, K. G.; Westmore, J. B.; Ogilvie, K. K.; Nemer,
M. J. Anal. Chem. 1982, 54, 960-966.
(81) Ens, W.; Standing, K. G.; Chait, B. T.; Field, F. H. Anal. Chem.
1981, 53, 1241.
(82) Kamensky, I.; Hakansson, P.; Sundqvist, B.; McNeal, C. J.;
Macfarlane, R. D. Nucl. Instrum. Meth. 1982, 198, 65-68.
(83) Devienne, F. M.; Roustan, J.-C. Org. Mass Spectrom. 1983, 17,
171.
(84) Barber, M.; Bordoli, R. S.; Sedgwick, R. D.; Tyler, A. N.; Whalley,
E. T. Biomed. Mass Spectrom. 1981, 8, 337; 8, 492.
(85) Barber, M.; Bordoli, R. S.; Elliott, G. J.; Sedgwick, R. D.; Tyler,
A. N. Anal. Chem. 1982, 54, 645A-657A.
(86) Field, F. H. J. Phys. Chem. 1982, 26, 5115.
(87) Rinehart, K. L., Jr. Science 1983, 218, 254-260.
(88) Caprioli, R. M.; Fan, T.; Cottrell, J. S. Anal. Chem. 1986, 58,
2949.
(89) Caprioli, R. M.; DaGue, B.; Fan, T.; Moore, W. T. Biomed.
Biophys. Res. Commun. 1987, 146, 291.
(90) Caprioli, R. M.; Moore, W. T.; Wilson, K. B.; Moring, J. J.
Chromatogr. 1989, 480, 247.
(91) Caprioli, R. M.; Fan, T. Biomed. Biophys. Res. Commun. 1986,
141, 1058.
(92) Caprioli, R. M.; Fan, T. Rapid Commun. Mass Spectrom. 1987,
1, 1058.
(93) Mumma, R. O.; Vastola, F. J. Org. Mass Spectrom. 1972, 6,
1373-1376.
(94) Conzemius, R. J.; Capellen, J. M. Int. J. Mass Spectrom. Ion
Phys. 1980, 34, 197.
(95) Hillenkamp. F. Laser Induced Ion Formation from Organic
Solids. In Proceeding of 2nd International Workshop on Ion
Formation from Organic Solids; Benninghoven, A., Ed.; SpringerVerlag: NewYork, 1983.
(96) Hardin, E. D.; Vestal, M. L. Anal. Chem. 1981, 53, 1492.
(97) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2288.
(98) Hillenkamp, F. Adv. Mass Spectrom. 1989, 11A, 354.
(99) Johnson, R. E. Models for Matrix-Assisted Laser Desorption and
Ionization: MALDI. In Large Ions: Their Vaporization, Detection
and Structural Analysis; Baer, T., Ng, C. Y., Powis, I., Eds.;
Wiley: New York, 1996.
(100) Vertes, A.; Gijbels, R.; Levine, R. D. Rapid Commun. Mass
Spectrom. 1990, 4, 228.
(101) Dreisewerd, K.; Schurenberg, M.; Karas, M.; Hillenkamp, F. Int.
J. Mass Spectrom. Ion Processes 1995, 141, 127
(102) Beavis, R. C.; Chen, X.; Carroll, J. A. J. Am. Soc. Mass Spectrom.
1998, 9, 885.

Chemical Reviews, 2001, Vol. 101, No. 2 375

(103) Karas, M.; Bahr, U. Selected Topics in Mass Spectrometry in the
Biomolecular Sciences; Kluwar Academic: London, 1997; p 33.
(104) Beavis, R. C.; Chait, B. T. Rapid Commun. Mass Spectrom. 1989,
3, 436.
(105) Beavis, R. C.; Chait, B. T. Rapid Commun. Mass Spectrom. 1989,
3, 432.
(106) Strupat, K.; Karas, M.; Hillenkamp, F. Int. J. Mass Spectrom.
Ion Processes 1991, 111, 89.
(107) Karas, M.; Ehring, H.; Nordhoff, E.; Stahl, B.; Strupat, K.;
Hillenkamp, F. Org. Mass Spectrom. 1993, 28, 1476.
(108) Beavis, R. C.; Chaudhary, T.; Chait, B. T. Org. Mass Spectrom.
1992, 27, 156.
(109) Wu, K. J.; Steding, A.; Becker, C. H. Rapid Commun. Mass
Spectrom. 1993, 7, 142.
(110) Pieles, U.; Zurcher, W.; Schar, M.; Moser, H. Nucl. Acid Res.
1993, 7, 142.
(111) Danis, P. O.; Karr, D. E. Org. Mass Spectrom. 1993, 28, 923.
(112) Overberg, A.; Karas, M.; Bahr, U.; Kaufman, R.; Hillenkamp,
F. Rapid Commun. Mass Spectrom. 1990, 4, 293.
(113) Overberg, A.; Karas, M.; Hillenkamp, F. Rapid Commun. Mass
Spectrom. 1991, 5, 128.
(114) Cramer, R.; Burlingame, A. L. Mass Spectrometry in Biology and
Medicine; Burlingame, A. L., Carr, S. A., Baldwin, M. A., Eds.;
Humana: Totowa, NJ, 2000; p 289.
(115) Mamyrin, B. A.; Karatajev, V. J.; Smikk, D. V.; Zagulin, V. A.
JETP 1973, 37, 45.
(116) Beavis, R. C.; Chait, B. T. Chem. Phys. Lett. 1991, 181, 479.
(117) Zhou, J.; Ens, W.; Standing, K.; Verentchikov, A. Rapid Commun. Mass Spectrom. 1992, 6, 671.
(118) Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1953, 26, 1150.
(119) Brown, R. S.; Lennon, J.; Christie, D. Desorption 94-Mass
Spectrometry of Large Organic Ions by Particle and Photon
Induced Desorption; Sunriver, OR, March 27-31, 1994; p 63.
(120) Brown, R. S.; Lennon, J. J. Anal. Chem. 1995, 67, 1998.
(121) Colby, S. M.; King, T. B.; Reilly, J. P. Rapid Commun. Mass
Spectrom. 1994, 8, 865.
(122) Whittal, R. Ml; Li, L. Anal. Chem. 1995, 67, 1950.
(123) Vestal, M. L.; Juhasz, P.; Martin, S. A. Rapid Commun. Mass
Spectrom. 1995, 9, 1044.
(124) Vestal, M. L.; Juhasz, P. Unpublished results.
(125) Krutchinsky, A. N.; Chernushevich, I. V.; Loboda, A. V.; Ens,
W.; Standing, K. G. Mass Spectrometry in Biology and Medicine;
Burlingame, A. L., Carr, S. A., Baldwin, M. A., Eds.; Humana:
Totowa, NJ, 2000; p 17.
(126) Vestal, M.; Juhasz, P.; Hines, W.; Martin, S. Mass Spectrometry
in Biology and Medicine; Burlingame, A. L., Carr, S. A., Baldwin,
M. A., Eds.; Humana: Totowa, NJ, 2000; p 1.
(127) McLuckey, S. A.; Van Berkel, G. J.; Goeringer, D. E.; Glish, G.
L. Anal. Chem. 1994, 66, 689A.
(128) Williams, E. R. Anal. Chem. News Features 1998, 3/1, 179A.
(129) Sundqvist, B.; Hedin, A.; Hakansson, P.; Kamensky, I.; Kjellberg,
J.; Salehpour, M.; Save, G.; Widdiyasekera, S.; Fohlman, J.;
Peterson, P. A.; Roepstorff, P. Biomed. Mass Spectrom. 1984,
11, 242.

CR990104W