Electron ionization time of flight mass spectrometry historical review and current applications

ELECTRON IONIZATION TIME-OF-FLIGHT MASS
SPECTROMETRY: HISTORICAL REVIEW AND
CURRENT APPLICATIONS

Nasrin Mirsaleh-Kohan,

Wesley D. Robertson,

1,2

and Robert N. Compton

1,3

Department of Physics, The University of Tennessee, Knoxville, TN 37996

Department of Physics, Emory University, Atlanta, GA 30322

Department of Chemistry, The University of Tennessee, Knoxville,

TN 37996

Received 27 September 2007; accepted 12 December 2007

Published online 4 March 2008 in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/mas.20162

This review presents an overview of electron ionization time-of-
flight mass spectroscopy (EITOFMS), beginning with its early
development to the employment of modern high-resolution
electron ionization sources. The EITOFMS is demonstrated to
be ideally suited for analytical and basic chemical physics
studies. Studies of the formation of positive ions by electron
ionization time-of-flight mass spectroscopy have been respon-
sible for many of the known ionization potentials of molecules
and radicals, as well as accepted bond dissociation energies for
ions and neutral molecules. The application of TOFMS has
been particularly important in the area of negative ion physics
and chemistry. A wide variety of negative ion properties have
been discovered and studied by using these methods including:
autodetachment lifetimes, metastable dissociation, Rydberg
electron transfer reactions and field detachment, SF

Scavenger

method for detecting temporary negative ion states, and many
others. # 2008 Wiley Periodicals, Inc., Mass Spec Rev 27:
237–285, 2008
Keywords: TOF mass spectrometry; positive and negative
ions; negative ion lifetimes; metastable negative ions; SF
(sub)6-Scavenger technique; electrostatic energy analyzer;
trochoidal electron guns; Rydberg electron transfer; retarding
potential difference technique; Wiley-McLaren space -focusing

I. INTRODUCTION

In recent years, time-of-flight mass spectroscopy (TOFMS) has
represented one of the fastest growing areas of mass spectrometry
(see, e.g., Cotter, 1994; Price & Milnes, 1990). Introduced
commercially in the early 1960s, TOFMS has witnessed a
resurgence in interest in the past 15 years due primarily to new
methods to produce pulsed sources of ionization such as laser
desorption ionization and its cousin matrix-assisted laser desorp-
tion/ionization (MALDI; Karas et al., 1987). It is noteworthy that
the 2002 Nobel Prize in Chemistry was awarded to John Fenn
(electrospray) and Koichi Tanaka (MALDI) for ‘‘their development
of soft ionization methods for mass spectrometric analysis of
biological macromolecules.’’ TOFMS has played an important role

in the development of both of these techniques. Fenn and Tanaka
shared one-half the Nobel Prize with Kurt Wu¨thrich. Wu¨thrich was
recognized for his development of nuclear magnetic resonance
spectroscopy for determining the three-dimensional structure of
biological macromolecules in solution.

The development of methods for the introduction of

continuous sources of ionization into the TOFMS has seen
increasing importance (Guilhaus, Selby, & Milynski, 2000).
These instruments are generally referred to as orthogonal
acceleration TOFMS (oa-TOFMS). This method is also finding
applications in such areas as electrospray, MALDI, and plasma
ionization mass spectroscopy, among others. Multiphoton
ionization (MPI) and resonantly enhanced multiphoton ioniza-
tion (REMPI) mass spectroscopy have also contributed to the
exceptional growth in TOFMS (Cooper et al., 1980; Gobeli,
Yang, & El-Sayad, 1985). The development of the pulsed laser,
nanosecond and shorter pulsed lasers (picosecond and femto-
second), has allowed for efficient and often gentle (non-
dissociation of the precursor ion) methods of ionization. The
introduction of the reflectron time-of-flight mass spectrometer
(Mamyrin et al., 1973; Mamyrin, 1974) which is used to
compensate for energy spread from the initial ion velocities as
well as the development of delayed pulse extraction (DPE) for
laser desorption ionization, has resulted in mass resolving power
approaching m/Dm

10,000 [Dm ¼ full-width half-maximum

(FWHM)]. As a result of this unique ability to gently ionize small
samples of ‘‘soft’’ materials (i.e., thermally sensitive, low
volatility, etc.) with high efficiencies, the applications of TOFMS
in the biological sciences are rapidly expanding.

Electron ionization has been widely used in mass spectrom-

etry, and is commonly used in negative ion formation studies
along with electrospray and chemi-ionization methods. In fact,
a significant fraction of the available atomic and molecular
structural information, such as ionization potentials, electron
affinities, dissociation energies, etc. has been determined with
electron ionization. Electron ionization has become somewhat
less attractive as a tool in this area because the development of
the laser due primarily to the lack of resolution inherent in
conventional electron beam sources. The application of new
methods to produce high-resolution electron beams has the
promise of rejuvenating this field of research. In this brief review,
we will summarize some of the unique capabilities that TOFMS
can offer the field of electron ionization. Page restriction will not

Mass Spectrometry Reviews, 2008,

27, 237– 285

# 2008 by Wiley Periodicals, Inc.

————

Contract grant sponsor: National Science Foundation.
*Correspondence to: Robert N. Compton, Departments of Physics and
Chemistry, The University of Tennessee, Knoxville, TN 37996.
E-mail: rcompton@utk.edu

allow a complete review of this field, but we will rely heavily
upon the contributions of our group in this area over the past
40 years. One of the authors (RNC) has been continually involved
in many forms of TOFMS since 1962 to the present, and has been
especially interested in the applications of pulsed lasers, electron
beams, and neutral beams to TOFMS. This article will draw
heavily upon our previous work, and we apologize in advance for
the partial neglect of the many other worthy contributors to this
area of science and technology.

In this article, we will discuss many of the advantages

that electron ionization time-of-flight mass spectrometry
(EITOFMS) has to offer the basic and applied sciences.
Historically, the first time-of-flight mass spectrometers employed
electron ionization as the primary source of ionization. In this
mode of operation, an electron beam is pulsed into the ion source
between two plates (repeller plate and extraction grid). After
passing through the interaction region, the electron beam is
collected on a Faraday cup. There are no electric fields in the
interaction region during the electron transit through the ion
source, which produces ideal conditions to study electron
ionization under well-defined conditions. Many TOFMS instru-
ments employ small magnetic fields to assist in the collimation
of the electrons, especially at low energy; however often it is
necessary to minimize the magnetic field to sub-milli gauss for
some applications; for example, electrostatic energy analyzers
for electron beams. There are no radio frequency fields or ion
draw-out fields present as are required in other types of mass
spectrometers. As we will see later, near-zero field conditions in
the source is especially important for the application of electron-
attachment mass spectroscopy. The formation of negative ions by
electron ionization is a resonance process (i.e., the cross-section
peaks in a narrow range of electron energies) that requires
well-defined electron beams (resolution, energy, etc.). These
conditions can be ideally accommodated by TOFMS.

The pulsed nature of the ion source in TOFMS also allows

for ion–molecule interactions to occur before extraction of the
precursor and secondary ions into the flight tube. By recording
the time evolution of the primary and secondary ions, ion–
molecule reaction rate constants can be determined (Stockdale,
Compton, & Reinhardt, 1969 and others cited therein). Often, the
linear flight tube may contain electrodes (retarding grids or an
Einzel lens), which permit further analysis of the ions while in
flight. For example, using these electrodes, metastable decay or
collisional dissociation can be easily detected by a second (or
third) flight tube that exists at different potentials along the
original flight tube. This allows for mass analysis of the decay
products of the primary ions, and constitutes a type of tandem
mass spectrometer; that is, MS/MS. Often, the increased time
width inherent in the time-of-flight of the decaying ions
(or neutrals) provides information on the energy shared by the
fragment ion and neutral upon dissociation. As we will see,
EITOFMS is especially well-suited for studies of the autodetach-
ment and dissociation of metastable negative ions formed by
unimolecular electron attachment.

II. BRIEF HISTORY OF TOFMS

In 1946, Stephens (1946) of the University of Pennsylvania,
speaking at a meeting of the American Physical Society at
the Massachusetts Institutes of Technology; proposed the possi-
ble construction of ‘‘a pulsed mass spectrometer, using time
dispersion’’ in a Friday afternoon session presided over by W.P.
Allis, one of the fathers of the field of gaseous electronics (J.J.
Thompson being the grandfather). Stephens promised that a mass
spectrometer of this type was under construction. Two years later,
Cameron and Eggers (1948), working at the Oak Ridge Y-12
plant (Clinton Engineering Works), reported the first TOFMS, an

FIGURE 1.

Faithful reproduction of the ‘‘Velocitron’’ time-of-flight mass spectrometer originally

presented and constructed by Cameron and Eggers (1948).

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

238

Mass Spectrometry Reviews DOI 10.1002/mas

instrument that did not involve a magnetic field, and fulfilled
the proposal by Stephens. The authors dubbed this instrument a
‘‘Velocitron.’’ The ions were accelerated to 300 eV and traveled
down a 3-m flight tube to the detector (Fig. 1). In Figure 1, we
have taken the liberty of faithfully re-drawing the original
schematic appearing in the publication of Cameron and
Eggers. For this set-up the mass calculation reduces to m

ðV=ð5:25 10

ÞÞT

, where V is the accelerating voltage and

the numerical constant contains the flight path distance (317 cm).
The correction factor is necessary to give the expression in the
specified units. The singly charged masses in their spectra were
calculated from the single relation m

¼ 2 eVðL

Þ. The mass

resolution was poor, but the principle had been demonstrated.
Singly and multiply charged ions of mercury were resolved, but
not their isotopes.

A working non-magnetic TOFMS by Wolff and Stephens

(1953) finally appeared in 1953. The method employed by these
authors was interesting in that the acceleration region was
designed such that the pulsed voltage used to accelerate the ions
was turned off before the ions reached the full acceleration
voltage. Under these conditions, all ions acquire the same
momentum instead of the same energy as in the conventional
TOFMS arrangement. Their experimental geometry is shown in
Figure 2. Notice that a ten-stage copper-beryllium dynode
electron multiplier was used in this early instrument. The time-
of-flight for ions of charge Q is equal to (Lm/ET

Q), where L is the

path length (100 cm), E is the acceleration field V/d {V is the pulse
voltage (300 V) and d is the length of the acceleration field},
and T

is the pulse length (5–50 msec). The theoretical resolution

is given by m/Dm

¼ t/Dt which, for their conditions, is m=Dm ¼

5:75

ﬃﬃﬃﬃ

. The 5.75 constant is a conversion factor, which allowed

them to use atomic mass units and centimeters for their tube
length. We will note later that the resolution of a TOFMS in which

the ions achieved full acceleration in the field is m/Dm

¼ t/2Dt.

In that 1953 article, reference was also made to a TOFMS
constructed at the Esso Laboratories of the Standard Oil
Company by W. Priestly, Jr. and E.C. Rearick (in consultation
with W.E. Stephens), and a TOF spectrum for hydrocarbons and
air recorded by this instrument was presented.

Apparently, the first report of a magnetic time-of-flight

mass spectrometer is credited to Bleakney and Hipple in 1938
(Bleakney & Hipple, 1938). As early as 1948, Goudsmit (1948)
proposed a magnetic time-of-flight mass spectrometer in which
ions are accelerated at right angles to a magnetic field. The pulsed
ions make a 3608 circular orbit to a detector as shown below in
Figure 3.

A small initial velocity in the direction of the magnetic field

produces a helical motion for the ions as shown on the right hand
side of Figure 3. Goudsmit pointed out that the angular velocity of
the helical motion is independent of the velocity of the ion or its
initial direction of motion. The time for a complete revolution is
given by T

¼ 670 m/B, where T is in msec if the magnetic field,

B, is in Gauss and m is in atomic units. The dimensions of the
apparatus are determined by the radius R of the helical path,
where R

¼ 145

ﬃﬃﬃﬃﬃﬃﬃ

=B and R is in cm if V is in volts. Another

attractive feature of this design is the fact that the helical path can
allow for many revolutions before detection. Hays, Richardson
and Goudsmit (1951) reported in 1951 on the construction and
properties of the magnetic TOFMS. In more conventional mass
spectrometers, the accurate determination of the mass decreases
with increasing mass; however, in this 1948 method, the mass can
be determined with the same precision for all masses. Figure 4
shows time-of-flight mass spectra for the case of Rubidium and
Xenon. It should be noted that the discussions by these authors do
not mention the fact that the ion path is not a simple circle of
length 2pR, but rather a helical path. The length, L, along a helix

FIGURE 2.

Schematic diagram of the arrangement for linear mass dispersion. Ions do not escape the ion

acceleration region before the acceleration pulse ceases. Reproduced from the work of Wolff and Stephens
(1953) with permission from American Institute of Physics, Copyright 1953.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

239

of projected radius R and pitch, a, (distance between arcs) is given
by one of the Frenet–Serret formulas: L

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

þ a

. If

R the increase in path length, DL, due to travel along one

cycle of the helix can be written as DL

¼ ða

=4pR

Þ. The value of

a/R in Figure 3 is made especially large for illustration. For most
applications a/R would be on the order of 0.01 making
ða

=4pR

Þ 10

a. Thus, the helical motion will change the

total TOF only slightly from that of circular orbits and should be
considered in the instrument design and operation.

The mass accuracy of all of the peaks was found to be within

one or two milli-mass units. As one can see from Figure 4, the
resolution exhibited by this instrument is excellent (we estimate
m/Dm

¼ 10,000 from this figure), and is probably limited by

the electronics employed as the detector at that time. According
to the above equation, the mass is directly proportional to
the time-of-flight; however, a careful study afforded by the high
resolution and accurate mass determination showed that small
perturbing electric fields were present in the flight path that
caused slight deviations from linearity. The stray fields were
caused by deposits on the walls. The surface deposits were
also found to vary with time. The authors suggested a correction
for these potentials as T

¼ amð1 þ emÞ. The correction co-

efficient, e, is less than 10

. From this analysis, they estimated

the perturbing surface field to be less than 2 mV/cm. Surface
potentials were believed to vary by

20 mV. The use of aerodag

(graphite) on the surfaces or the construction of the electrodes
with molybdenum would be expected to help correct this
deficit in a modern instrument. Nevertheless, the Goudsmit
magnetic TOFMS is a remarkable instrument worthy of further
consideration.

In 1951, Smith (1951) described another version of the

‘‘magnetic period mass spectrometer,’’ which is shown schemati-
cally in Figure 5.

He referred to this instrument as the ‘‘synchrometer’’

because of its resemblance to a synchrotron. He also mentioned
that Goudsmit had named his instrument the ‘‘chronotron.’’
Smith’s spectrometer employed a three-part electrostatic lens to
form and deflect pulses of ions so that their orbit missed the ion
source from which they originated. The first and last elements of
the lens, seen in Figure 5, were grounded, and a square wave
voltage pulse was applied to the center element. The pulse
was timed so that the ions were decelerated slightly with each
revolution and consequently each revolution had a smaller radius.
The applied pulse sequence prevents the ions from striking the

FIGURE 3.

Illustration of the Goudsmit magnetic Time-of-Flight mass spectrometer. Drawn based upon

information from the work of Goudsmit (1948).

FIGURE 4.

Top, Time-of-Flight mass spectrum of Rubidium: bottom, Xenon isotopes. Modified from the

work of Hays, Richardson, and Goudsmit (1951).

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

240

Mass Spectrometry Reviews DOI 10.1002/mas

ion source or walls of the vacuum chamber, and allows for a
greater number of revolutions and enhanced resolution. The
detector is located near the center of the apparatus, and does not
intercept any ions until they have been significantly decelerated
and thus their orbit made smaller. Smith reported a mass
resolution for this instrument, m/Dm of 8,500, which is
comparable to that of the Goudsmit spectrometer.

III. PRINCIPLES OF TOFMS

A. Wiley– McLaren Space-Focusing TOFMS

The development of the modern commercial TOFMS began with
the seminal article by Wiley and McLaren published in 1955
(Wiley & McLaren, 1955). However, similar TOFMS proposals
also appeared in the Russian literature slightly earlier (Ionov,
Mamyrin, & Fiks, 1953). Although this new design did not
improve upon the resolution of the magnetic TOFMS, it
represented a great advance in the case of the non-magnetic
TOFMS. In this modification, an electron beam of finite spatial
width was directed between the ion draw-out plates of the
TOFMS, as shown in Figure 6. Ions are created over the full
extent of the electron beam. Thus ions created at different
positions within the spatial width of the electron beam will fall
through different voltages and acquire a range of ion velocities,
resulting in a spread in the time-of-arrival distribution, which will
degrade the resolution of the instrument. The Wiley–McLaren

focusing conditions attempted to partially correct for this spatial
dispersion.

The Wiley–McLaren double-field source consists of four

elements (reading left to right in Fig. 6); the backing plate, ion
draw-out grid, a grid defining the entrance to the field free drift
space, D, and the detector. Electric fields are established in the
region, and are labeled as a result of the application of a positive
(negative) pulse voltage, V

, on the backing plate for the analysis

of positive (negative) ions, followed by a negative (positive)
voltage applied to the flight tube. The first grid is generally at
ground potential. Alternately, the backing plate can be held at
ground potential, and the pulsed-voltage is applied to the first
grid. In some applications in which positive and negative ions
are both studied, the same voltage pulse can be switched to either
the backing plate or first grid to obviate the need for two
expensive pulse generators. Assuming that the ions are produced
with no initial kinetic energy, the major spread in the time-of-
flight distribution for ions of the same mass is due to the different
energies gained by ions created at different positions within the
electron beam. Those ions created closer to the backing plate will
acquire a higher terminal velocity in the flight tube region and
will eventually overtake and pass those ions created closer to the
first grid. There is, of course, a distance at which the two ions
arrive at the same time (i.e., the point at which the faster ions
overtake the slower ions). The applied voltages, which allow for
the fast- and slow-ions to arrive at the same time, is referred to as
the Wiley–McLaren space-focusing condition. In the following,
we will repeat the Wiley–McLaren derivation of the expression
for the total time-of-flight, T. The equations presented by Wiley

FIGURE 5.

Schematic cross-section of ‘‘synchrometer’’ introduced by Lincoln Smith in 1951. Reproduced

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

241

and McLaren contained a curious factor of 1.02. Although this
factor is not discussed by the authors, it is a conversion factor, in
which the time is calculated in microseconds if one uses atomic
mass units, centimeters for distances, electric fields in volts per
centimeter, and ion energy in electron volts. When an ion with
initial energy U

is moving through the source, its energy will

increase to a value, U given by:

¼ U

þ qsE

þ qdE

ð1Þ

The total time-of-flight is the sum of the time spent in the

source, T

, the acceleration region, T

, and the drift space, T

(see

Fig. 6):

ðU

; s

Þ ¼ T

þ T

ð2Þ

where

¼ 1:02

ð2mÞ

1=2

½ðU

þ qsE

1=2

ðU

1=2

ð3Þ

¼ 1:02

ð2mÞ

1=2

½U

1=2

ðU

þ qsE

1=2

ð4Þ

and

¼ 1:02

ð2mÞ

1=2

ð5Þ

To simplify the equation for T, T can be driven when U

¼ 0

and s

¼ s

. If one define

¼ qs

þ qdE

ð6Þ

and

þ dE

ð7Þ

with assumption in Equations (6) and (7), the Equation (2)
becomes

ð0; s

Þ ¼ 1:02

1=2

1=2
0

þ 1

þ D

ð8Þ

The position D

for which the two ions overlap is determined

by setting dT/ds

¼ 0 or

¼ 2s

3=2
0

þ k

1=2
0

ð9Þ

Therefore, if s

, d, and D are fixed values, then the ratio of

voltages E

is uniquely determined by Equation (9). A plot of

versus the voltage ratio E

for the case of s

¼ 1/2 cm and

¼ 1 cm is shown in Figure 7.

In the practical application of Equation (9) one would

choose the dimensions for s

, s, d, and D and adjust the pusher-

plate pulse voltage and/or the flight-tube voltage, E

, to obtain

maximum resolution by monitoring the minimum widths of the
mass peaks.

Typical resolving power of the Wiley–McLaren TOFMS

for a 2-m instrument is approximately m/Dm

300 to 500. One

remaining contribution to lowering the mass resolution is the
initial energy (velocity) spread of the ions. We will consider later
corrections to the energy spread, but first we will discuss further
corrections to the dual-source TOFMS design of Wiley–
McLaren.

B. Higher-Order Corrections to Space Focusing

A number of authors have proposed higher-order corrections to
the Wiley–McLaren dual-source space-focusing condition (see
for example, Eland, 1993; Even & Dick, 2000a,b; Seccombe &

FIGURE 6.

Schematic drawing of the Wiley–McLaren double-field TOFMS. Modified from the work of

Wiley and McLaren (1955).

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

242

Mass Spectrometry Reviews DOI 10.1002/mas

Reddish, 2001, and others cited therein). All of these articles
make reference to other methods for improvement upon the basic
Wiley–McLaren TOFMS. The introduction of additional ion-
grid elements in the ion-source region can lead to higher-order
space focusing and considerably improved resolution. The
ultimate goal is to make ions of the same mass created anywhere
in the ion-source region arrive simultaneously at the detector.
We will briefly discuss higher-order space focusing below.

Using the illustration of Even and Dick (2000a), we show the

first- and second-order focusing geometries in Figure 8.

A third grid is added (with voltage V

) as shown in Figure 8b

to produce second-order focusing. As an example, these authors
calculate the total transit time for an ion, that passes through all

three regions that originated at position x in region 1 as

ðxÞ ¼

ðxÞ

ðxÞ þ

ðxÞ

ð10Þ

The authors then calculate the parameter, which serves

to minimize the sum of the square of the deviations of the
arrival times from that calculated for an ion that started at the
midpoint of the ionization volume (i.e., center of the electron
beam in our case); that is,

Min

ðx

Þ t

ð11Þ

The equally spaced steps {x

} cover the whole ionization

volume x

¼ 0 to x ¼ L

. Figure 9 is an example of the arrival-time

differences in nanoseconds as a function of initial position of the
ions in the ionization region for the case of L

¼ 2.5 cm,

¼ 2.5 cm, L

¼ 70 cm, L

¼ 1 m, V

¼ 1,000 V. The optimum

value V

for this case is calculated to be 810 V. They have applied

the same analysis for the case of second-order space focusing.
Figure 10 below shows a similar optimization for the case
of second-order space focusing. The optimum voltages for this
case are V

¼ 1,000, V

¼ 868, and V

¼ 737 V. One notices the

much smaller time spread over a considerably larger distance for
the case of second-order focusing. As a result, second-order
focusing is much more forgiving of an extended ion source. A
similar theoretical treatment of the general nth-order space-
focusing condition, using n fields, has also been presented by
Seccombe and Reddish (2001).

An EITOFMS in which the flight tube has been replaced by

several tandem cylindrical lenses was introduced by Srivastava,
Iga, and Rao (1995). The authors attribute their exceptional
resolution to the ability of the segmented lenses to confine the ion
path close to the axis of the flight path. Although the segmented

FIGURE 7.

Shows a plot of D

versus the voltage ratio (E

) for the

case of s

¼ 1/2 cm and d ¼ 1 cm.

FIGURE 8.

: The ionization region extends from x

¼ 0 to x ¼ L

, where x

¼ 0 is at the second extraction

plate with voltage V

. The acceleration regions have imposed voltages V

and V

, length L

and L

, and flight

tube length is L

. Reproduced from the work of Even and Dick (2000a) with permission from American

Institute of Physics, Copyright 2000. b: The second-order focusing agreement for improved resolution.
Reproduced from the work of Even and Dick (2000a) with permission from American Institute of Physics,
Copyright 2000.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

243

FIGURE 9.

Difference in time-of-arrival of ions in the ionization region with ions produced in the center of

the ionization region, t(x)

t (L

/2), plot for first-order space focusing as a function of initial ion position in

ionization region. The position x

¼ 0 is the second extraction plate with voltage V

, distance is in mm and

changed from an error in the original publication as per the authors. Reproduced from the work of Even and
Dick (2000a) with permission from American Institute of Physics, Copyright 2000.

FIGURE 10.

Difference in time-of-arrival of ions at detector with ions produced in the center of the

ionization region, t(x)

t (L

/2), for second-order space focusing as a function of initial ion position in

ionization region. The position x

¼ 0 is the second extraction plate with voltage V

, distance is in cm and

changed from an error in the original publication. Reproduced from the work of Even and Dick (2000a) with
permission from American Institute of Physics, Copyright 2000.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

244

Mass Spectrometry Reviews DOI 10.1002/mas

flight paths can produce severe ‘‘end effects,’’ the methods
to describe second- and third-order space-focusing conditions
could be applied to this TOFMS (see Seccombe & Reddish,
2001).

C. Time-Lag Velocity Focusing

Wiley and McLaren also demonstrated that the resolution of the
double-field ion draw-out system was less affected by the initial
kinetic energies of the ions than that characteristic of the earlier
designs. Higher resolution is also expected from the application
of the higher-order corrections in multiple source electrode
geometries. The main contribution to the lack of resolution of the
linear TOFMS is due to the initial random velocity distribution
of the ions. To partially correct for the initial velocity, they
introduced a technique called time-lag energy focusing in which
the negative potential well created by the electron beam is used to
‘‘trap’’ most of the ions until the electron beam is shut off. A time-
lag is introduced before the ion pusher-plate voltage is applied
to inject the ions down the flight tube toward the detector.
The time-lag energy focusing effect can be best explained with
the aide of Figure 11.

First, we consider ions created at one position, s, in the

source. Ion 1 is sitting still, and ions 2 and 3 are moving with a
velocity toward (

þ) and away () from the flight path,

respectively. During this delay time, t, ions will move to new
positions (s

þ t and s t) because of their initial velocity.

Without a time-lag, the final TOF of all three ions will be
different. After the lag period, the ions with

þ will be

accelerated to a lower energy than those traveling with

because they will fall through a slightly smaller voltage. The
change in flight time due to the time-lag is approximately (dT/
ds)t. The ions will move to a new position on the flight time
curve as shown. From the simple illustration, it is possible to
choose a value of the time-lag that will allow all three ions to
arrive simultaneously. The difference in time between T(,s) and

T(o,s) is given by (m/E

)v. Thus, the delay time, t

, for optimum

focusing is derived from

¼ 0

ð12Þ

ðdT=dsÞ

ð13Þ

Note that the time-lag is proportional to

ﬃﬃﬃﬃ

because

dT=ds

ﬃﬃﬃﬃ

, and one can only obtain corrections over a narrow

range of masses. Although the time-lag is independent of the
initial velocity v, time-lag focusing is only possible if dT/ds is
negative. However, under these conditions the space-focusing
condition dT/ds

¼ 0 is invalid. Therefore, the choice of a proper t

and dT/ds represents a compromise between the two. RNC has
used time-lag focusing for many years with generally good
results by empirically varying the time-lag and V

to obtain

optimum resolution.

Time-lag focusing when using negative ions presents

obvious problems. One of the authors (RNC) has found that
application of the time-lag condition with negative ions will
produce broadening to the extent that two separate peaks will
appear in the TOF spectrum. The two peaks occur as a result of
the space charge of the electron beam repelling the negative ions.
The separation in time increases with electron beam intensity
and time-delay. In this case, the ions receive a repulsive impulse
from the electron beam pulse so that the ion spatial and TOF
separation continues long after the electron beam has exited the
ion source region; that is, the ions receive an impulse from
the initial electron beam. Thus, time-lag focusing cannot be
employed for negative ions unless positron ionization is
considered. However, this time-spread can be used to estimate
the charge density of the electron beam. A simple relationship
can be derived that gives the TOF spread due to the time-lag in
terms of the electronic charge density in the beam. Measurement
of the time-spread of the negative signal with delay time
represents a straightforward method to determine the electron-
beam charge distribution. This may prove useful for some
applications.

Time-lag focusing has made very important contributions to

the fields of laser desorption ionization (LDI) and MALDI from
surfaces. In this method, a pulsed laser is used to eject ions from a
surface into a TOFMS. Often the expression ‘‘the laser causes the
ions to ‘fly’ from the surface’’ is used to describe this process. The
resolution is greatly improved by delaying the time between
when the laser hits the surface and when the ion draw-out field is
applied. Many groups have adopted the name time-lag focusing
or DPE for this application. In this case, all of the ions leave from
the same point (the surface); however, ions of different velocities
will travel different distances into the vacuum. Those with higher
velocities will travel farther into the vacuum than the slower ones
and will therefore acquire a correspondingly smaller energy from
the draw-out pulse. Proper choice of the delay time will cause
ions of different energies (and same mass-to-charge ratio) to
arrive at the detector at the same time. Time-lag focusing can be
easily envisioned by considering only those ions which are

FIGURE 11.

Curves of flight time versus initial position used in the

discussion of time-lag focusing. Modified from the work of Wiley and
McLaren (1955).

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

245

moving with

þv and those that originated from the surface v ¼ 0.

Those moving with

v do not escape the surface. In this case,

space- and energy-focusing are possible. In Figure 12, we show
an example of TOF spectrum for laser desorption ionization of
C

ions from a surface with and without DPE. A mass resolving

power of 1,000 is routine and, with special care, one of 10,000 is
possible. Time-lag focusing (or DPE) has a huge impact in the
area of laser desorption and MALDI TOFMS. This method often
obviates the use of the reflectron TOFMS. This brief description
of time-lag focusing, or DPE, does not do justice to the enormous
impact this method is having on TOFMS, especially in the
biomaterial area. Further details of DPE can be found in articles
by Brown and Lennon (1995) and Vestal, Juhasz, and Martin
(1995).

D. Orthogonal Acceleration (oa)/or
Off-Axis Ion TOFMS

In recent years, use has increased of so-called orthogonal
acceleration (oa) or off-axis injection of ions into the draw-out
region of the TOFMS, referred to as oa-TOFMS. Orthogonal
acceleration (oa) TOFMS represents a method for gating ions
from a beam of ions into the source region of a TOF mass
spectrometer. Numerous groups have employed this method
beginning with those in the 1950s including the Bendix
Corporation. A seminal article by Guilhaus in 1994 (Guilhaus,
1994) chronicles the development of this field. In this review
Guilhaus points to the important contribution by Soviet scientists
(Chernushevich, Dodonov, & Dodonova, 1987; Dodonov et al.,
1989; Dodonov, Chernushevich, & Laiko, 1991) and others from
the Bendix Corporation (O’Halloran et al., 1964).

In the oa method, ions with initial velocities perpendicular to

the flight path will exhibit a smaller spread in their time-of-flight
distributions. If the perpendicular velocity is and the final
longitudinal ion velocity down the flight tube is V, then the final
ion velocity can be approximated as V

/2V because /V is

small. Introducing the ions into the ion source perpendicular to
the ion flight path produces a small spread in TOF in comparison
to the case where the initial ion velocity, v, is in the same direction
and in the opposite direction of the flight tube direction. In this
case, the TOF spread corresponds to 2v and is much larger (4V/v)
in comparison to

/2V. A further review by the Guilhaus group

provides a complete discussion of the modern adaptations of

oa-TOFMS (Guilhaus et al., 2000), and will not be discussed
further in this review. This technique finds many applications
in electrospray and related atmospheric pressure ionization,
MALDI, EI and gas-chromatograph/mass spectroscopy, and
elemental analysis.

In our own work, UF

ions formed from surface ionization

of UF

on a gently heated uranium wire were injected between the

backing plate and the first ion grid of a TOFMS ion source
(Compton, Reinhardt, & Garrett, 1976). UF

ions were also

produced by surface chemical reactions between a bare uranium
wire and molecular fluorine. Operating at 10K Hz repetition, this
simple arrangement produced exceptional mass resolution. This
simple geometry is shown in Figure 13. There are many other
examples of early experiments of orthogonal acceleration time-
of-flight mass spectroscopy (oa-TOFMS). However, this area of
mass spectroscopy became a major component of TOFMS,
beginning with the work of Guilhaus and others.

Another method employed to reduce the ion velocity

component along the ion flight tube direction is through the use
of a nozzle-jet expansion and a jet skimmer employed for the
source of molecules. Under these conditions, the initial velocity
of the molecules prior to ionization is mainly perpendicular to the
flight path. Using this method, it is necessary to use x–y deflector
voltages in the flight tube to correct for the greater transverse ion
velocity inherent in the nozzle-jet. Details of this method will be
discussed in the section describing electron attachment to
molecular clusters.

E. Reflectron TOFMS

As discussed above, since space focusing has been achieved, the
major source of spread in the time-of-flight and corresponding
degradation of mass resolution results from the distribution of

FIGURE 12.

Illustration of delayed pulse extraction of laser desorbed

ions. Lower trace represents the TOF for ions desorbed in a constant

draw-out field and the upper trace is the TOF for ions following delayed
pulse extraction (DPE).

FIGURE 13.

Diagram of the experimental apparatus to study the

negative ions produced from the reaction of uranium and fluorine. A fast
(0–40 eV) cesium beam could be directed through the interaction region
to produce UF

, UF

, and F

by collisional ionization with UF

for

mass calibration. Reproduced from the work of Compton et al. (1976)
with permission from American Institute of Physics, Copyright 1977.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

246

Mass Spectrometry Reviews DOI 10.1002/mas

initial ion velocities (kinetic energy). Velocities along or against
the direction of the flight tube represent a major source of spread
in TOF; velocities orthogonal to the flight path contribute to a
lesser extent. As explained above, the oa-TOFMS serves to
partially reduce the on-axis velocities by injecting the ions
perpendicular to the direction of the TOF. Another attractive
method to reduce the spread in TOF is to first pass the ions
through an energy or velocity filter prior to the TOF analysis.
Energy section of the initial ion beam is particularly important for
high-energy ion sources such as spark sources or laser ablation
methods. Using a spherical-sector electrostatic-energy analyzer
coupled to a TOFMS is particularly useful to understand the
physical processes that occur during laser ablation/desorption
(Shea, Compton, & Hettich, 1990). For example, Shea and
Compton (1993) employed this technique to detect the ejection of
silver ions at 2.5 eV produced from surface plasmon decay
produced by laser ablation of a roughened silver surface. While
adding another dimension to the understanding of the physics of
the ablation processes, these methods usually result in decreased
sensitivity due to the loss of ions in the process of energy or
velocity analysis.

In 1966, Mamyrin presented a doctoral dissertation to

describe a mass-reflectron TOFMS (Mamyrin, Doctoral Dis-

sertation, Physico-technical Institute, Academy of Sciences,
USSR, Leningrad, 1996). A reflectron TOFMS is a magnet-
free time-of-flight mass spectrometer, which is designed to
achieve second-order time focusing with respect to variation of
ion energies and angle of divergence of their exit from the ion
source. The principle of operation and the results of a working
model were published in 1973 (Mamyrin et al., 1973). The
principle can be described with reference to Figure 14.

Mamyrin chose to define the resolution of the TOFMS as the

full width at one-half of the peak maximum, R

50%

. For a linear

TOFMS and taking the ‘‘thickness’’ of the ion packet at the
detector as DL, the total flight path as L, the spread in initial ion
kinetic energy as DE, and the final ion energy in the flight tube as
E, the resolution becomes

50%

ﬃﬃﬃﬃﬃﬃﬃ

ð14Þ

Thus the resolution of a conventional, linear TOFMS cannot

be improved by a simple increase in the path length, L. In the
Mamyrin design, the spread in TOF as the ions traverse through
the first field free drift space is compensated by a region in
which the ions are ‘‘reflected’’ by a uniform electrostatic field.

FIGURE 14.

Reproduction of the mass-reflectron TOFMS introduced by Mamyrin et al. in 1973 with

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

247

Conventional space focusing corrections results in ions of the
same mass arriving at the focal point, the detector for linear TOF,
with a minimal spatial spread. These ions will have a large range
of kinetic energies, originating in part by the particle’s velocity
distributions following injection into the ionization region. The
large range of velocities of the ions limits the extent to which
space focusing of the ions can occur. An ion reflecting mirror
system, a ‘‘reflectron,’’ is used to correct for velocity distributions
of the ions of the same mass thus increasing the mass resolution.
The reflectron is used to refocus the ions from the focal point
obtained from space focusing, which is adjusted to be before
the reflectron, onto the ion detector. The ions then arrive at the
detector with a minimal velocity distribution, which allows for a
narrower spatial distribution at the detector as well. The time the
ions spend in the reflectron field is proportional to

ﬃﬃﬃﬃ

and thus

the time of transit in the drift region is proportional to 1=

ﬃﬃﬃﬃ

Therefore, using an appropriate choice of the ion energy and
the reflectron field one can compensate for the spread in ion
velocities and the TOF to the detector due to the initial velocity
distribution of the particles. Second-order focusing is can be
accomplished by applying a retarding field at the entrance to the
reflectron. Under these conditions, the resolution of an instru-
ment can be increased in proportion to its length. Mamyrin et al.
(1973) reported mass spectra for electron ionization of rhenium
bromide of R

50%

> 3,000, which is approximately an order of

magnitude greater than that of a typical linear TOFMS. Mamyrin
has reviewed the growing field of laser assisted reflectron time-
of-flight mass spectroscopy, which contains references up to
1994 (Mamyrin, 1994). In this review, it was stated that mass
resolution in the range of 300–30,000 has been attained using
different ion sources: electron ionization, fast atoms, laser
radiation, etc. The review by Mamyrin (1994) contains a very
complete account of the mass-reflectron and will not be discussed
further here.

IV. EXPERIMETAL METHODS

Research on electron ionization or electron attachment in
TOFMS heavily relies upon the production and precise control
of electron beams with narrow energy spread. The center-of-mass
energy (E

com

) between an electron of mass m and molecule of

mass M colliding with a relative velocity V is given by

com

þ M

ð15Þ

In most cases, the electron velocity is much larger than the
velocity of the atom or molecule; however, for low-energy
attachment studies (e.g., electron energies <10 meV) with
seeded nozzle-jet expanded molecules (light molecules with
energies >1 eV), the velocity of the heavy particle might have to
be taken into account as well. Of course, the total energy available
for a reaction will also include the internal energy (i.e.,
ro-vibrational energy) of the target molecules. The use of a
nozzle-jet expansion can also reduce the internal ro-vibrational
energy content of the molecules.

Because the electron mass is 1/1840 that of a proton mass, it

is clear from Equation (15) that the center-of-mass energy is

equal to the laboratory energy of the colliding electron to less
than one part in 1841, or less than 0.1% for all atoms and
molecules. Thus, for most purposes it is possible to equate the
center-of-mass energy to the laboratory energy of the electron.
Likewise, the resolution in collision energy is that of the
resolution of the laboratory energy of the electron. Methods for
the production of energy-resolved electron beams have been
discussed in many textbooks and the development of more
sophisticated methods is an ongoing technical endeavor (Hasted,
1972; McDaniel, 1989). This article will review only those
methods that have been employed in TOFMS. Unlike many other
methods of mass spectroscopy, which rely upon continuous ion
extraction, TOFMS is ideal for high-resolution electron ioniza-
tion studies because a pulsed beam of electrons is introduced
into a field-free ionization region. The ions are pulsed out of
the ionization region following a controlled delay time after the
electron beam has exited the ionization region; thus, the ions are
formed under field-free conditions.

Calibration of the electron energy scale for positive and

negative ion electron ionization at low energies (<20 eV) can be
carried out by observing the positive ion onset for many rare gas
atoms or negative ion resonances, respectively. Many dissocia-
tive (e.g., Cl

/HCl) and non-dissociative (e.g., SF

) resonances

have been well-characterized. Calibration of the electron energy
scale for positive ionization at higher energies can also be
accommodated by the two well-studied triply excited negative
ion resonances in helium in the region of 55–60 eV. Electron
ionization of He, leading to He

ions in this region, exhibit an

interference contribution due to excitation of the He

[2s

2p(

P)] and He

[2s

, 2p

(

D)] negative ions. Interference of

these short-lived negative ions with the background continuum
results in two features that exhibit a Fano line shape in the
ionization continuum. Figure 15 shows these resonances from the
research of Grissom (1970). (See also Grissom et al., 1969).
Many succeeding studies have been devoted to the character-
ization of the position and shape of these resonances with general
agreement (a recent article summarizes these results (Fiegele
et al., 2001)).

A. Retarding Potential Difference (RPD) Method

Retarding potential difference involves the manipulation of
electrons thermionically emitted from a filament in such a way
that a beam of ‘‘quasi-monoenergetic’’ electrons is simulated
(see Fox, Hickam, & Grove, 1951; Fox, Hickam, & Kjeldaas,
1953; Fox et al., 1955). The energy distribution of electrons
from a filament possesses a wide energy distribution (

0.5 eV

FWHM) due to the one-dimensional Maxwell Boltzmann energy
distribution. This energy spread will also be increased (to

1 eV

FWHM) as a result of the voltage drop across the filament if an
indirectly heated cathode is not employed. An indirectly heated
cathode is preferred because there is no voltage gradient along the
electron emitter. The RPD method utilizes the differential
retardation of this electron beam, and is best illustrated by
examination of Figure 16. A ribbon-like beam of electrons is
created by thermionic emission from a filament. Next the beam
passes through a series of acceleration and deceleration slits, and
is finally accelerated into the ionization region. The final energy

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

248

Mass Spectrometry Reviews DOI 10.1002/mas

of an electron (in eV) entering the grounded ion draw-out region
is equal to the electronic charge times the voltage (relative to
ground) at the point from which the electron was emitted plus its
‘‘thermal’’ energy. A retardation voltage is applied on electrode 3
in Figure 16 such that only about half of the electrons are
transmitted and half are retarded back toward the filament. The
retardation voltage produces an energy distribution, which is
‘‘chopped-off’’ on the low energy side. Note that even this simple
retardation can reduce the initial energy spread of the electron
beam by approximately one-half. Increasing or decreasing this
retardation voltage will produce a second distribution with a
different cut-off. Assuming a sharp edge on the two distributions,
a subtraction of the ionization produced by both of these electron
beams will simulate a signal representative of the electrons
contained within the ‘‘sliver’’ of electrons sliced from the
distribution. It is also desirable to apply a square-wave or sine-
wave voltage of magnitude DV to electrode 3 and to use a lock-in
amplifier to detect the ac component of the signal due to this
‘‘sliver’’ of electrons (see e.g., Stockdale et al., 1969; Frey
et al., 1973; Huebner, Frey, & Compton, 1973, and others cited
therein). Typical energy resolution of this method with DVof 0.05
and 0.1 eV is between 0.05 and 0.2 eV. The major attraction of this
method is the ease of use and relatively large electron currents
available, especially at low energy. Figure 17 is an example of the
negative-ion yields versus the electron energy for a mixture of
sulfur hexaflouride and fluoranthene. A RPD sine-wave ‘‘ripple’’
of 0.1 eV was used, and the apparent resolution is

0.15 eV as

evidenced by the FWHM of the SF

signal. As will be discussed

later, the cross-section for electron attachment to SF

in the

TOFMS exhibits a ‘‘delta function’’-like feature at

zero energy.

ions created by electrons with energy above

0.02 eV

decay before they can be accelerated out of the source. Thus, the
SF

signal gives a faithful representation of the electron-energy

distribution. Of course, this distribution is ‘‘reversed’’ in energy;
that is, the high-energy side of the distribution is at the low-
energy side of the true energy distribution. It is also possible to
record the transmitted electron current, which will exhibit a rapid
drop to zero as the electron-energy approaches zero. The
derivative of this retardation curve will also provide a reasonable
estimate of the electron energy distribution, which is found to
be identical to that derived from the SF

signal. Note that the

fluoranthene negative-ion cross-section begins at

0.1 eV and

peaks at

0.2 eV (Frey et al., 1973). The SF

ion signal onset at

0.0 eVand peaks at 0.4 eV. A small peak is also seen at 0.0 eV
due to electron attachment to vibrationally exited SF

. Heating

the SF

dramatically increases this ‘‘zero energy’’ peak (see Chen

& Chantry, 1979 and articles therein).

Figure 18 shows the electron ionization thresholds for the

precursor ion and precursor minus H atom ion signal for the
acenaphthene molecule, using the RPD method. The Kr

ion

threshold is used to calibrate the electron energy scale to within
0.05 eV. These data highlight another major advantage of the
RPD technique, namely that the electron beam is approximately
constant in intensity from over 100 eV down to less than 0.1 eV.

Some of the potential technical problems that might occur

with the RPD method have been discussed (Grobe, 1963;
Anderson & Eggleton, 1967; Anderson, Eggleton, & Kessing,
1967; Massey, Burhop, & Gilbody, 1969a). Although care must
be exercised in its use, the RPD method can provide adequate
electron-energy resolution and accurate energy thresholds to

FIGURE 15.

Negative ion resonances in the He ionization cross-section. Reproduced from J. T. Grissom’s

PhD thesis, 1970 with permission from the author.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

249

within 0.1 eV or better. Other methods to be discussed below
exhibit higher energy resolution and are believed to be more
reliable; however, the RPD method remains a reliable alternative
to the more difficult-to-implement methods described below.
One of the present authors (RNC) has published over 30 reports
on the application of the RPD method using a slightly modified
Bendix Model 14-206 TOFMS.

B. Trochoidal Electron Monochromator

The trochoidal electron monochromator (TM) involves the
motion of electrons in crossed electric and magnetic fields.
Electrons, which are collimated from a filament, are directed into
a region of crossed E

and B

fields

ð E

? B

Þ to produce a spatial

dispersion of the electrons according to their energy. The
magnetic field is along the final direction of the electron beam.
An aperture is used to select electrons from this distribution to
produce a monoenergetic electron beam with intensities on the
order of nanoamperes and energy resolution from 0.1 to 0.02 eV.

The first TM was employed to energy analyze electrons

from a plasma (Barr & Perkins, 1966). However, Stamatovic and
Schulz (1968) were the first to design a TM for specific use as a
monoenergetic electron source. The original articles of Stama-
tovic and Schulz reports a resolution of 0.02 eV down to collision
energies of almost 0 eV. Many researchers have employed the TM
in electron atom/molecule collision studies, but typical reso-
lution is on the order of 50–75 meV, which is sufficient for most
cases of interest. Ma¨rk and co-workers have considered the TM in
some detail, and find that the energy resolution may vary with the
electron energy (Grill et al., 2001) and have designed a TM with
energy resolution of

45 meV. The energy resolution was also

independent of the electron energy.

When performing electron ionization research, there arise

many situations in which it is necessary to use a more precisely
defined beam of electrons that interact with the atom or molecule
in question. Typically, electrostatic energy analyzers are used for
these purposes and are available commercially. But, as Ma¨rk has
pointed out, electrostatic energy analyzers have difficulty with

FIGURE 16.

Typical ion-source configuration that employs the RPD technique. Often, the difference

retarding voltage is a sine wave with

0.1 eV amplitude. Reproduced from R. N. Compton’s PhD thesis,

1966 with permission from the author.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

250

Mass Spectrometry Reviews DOI 10.1002/mas

very low-energy electrons. These low-energy electrons are very
sensitive to any stray electric field, and have a difficult time
making it through the analyzer. The typical minimal operating
energy for these analyzers is

4 or 5 eV. This low-energy region

happens to be very interesting, especially near zero eV, and
is particularly important for electron-attachment studies. The
trochoidal electron monochromator, introduced by Barr and
Perkins (1966), uses a magnetic field to keep electrons from
straying on their original path in combination with an electro-
static deflection region to disperse the beam in energy so that a
narrow energy can be selected. This magnetic field also allows
one to operate with electron energies very close to zero energy.

The trochoidal electron monochromator uses are twofold. It

can be used as a monochromatic energy electron source as we will
discuss in this review, but also as an electron energy analyzer. The
first was constructed by Barr and Perkins (1966) to analyze
electrons emitted from plasma. Since the work of Barr and
Perkins, many versions and improvements to the trochoidal
electron monochromator have been made (see Grill et al., 2001),
but the basic principle remains the same. Helmholtz coils are
used to introduce a magnetic field (50–100 gauss) along the
direction of the electron beam. Many standard cathode materials
(tungsten, thoriated irridium, lanthanum hexaborate, etc.) are
used to produce a large electron beam current (micro amps) with
an energy distribution of 0.5–1 eV. Two to three electrostatic lens
with positive voltages applied to them are then used to focus
the beam across the E

region. In the dispersion region,

there are typically two plates with a small deflection voltage
applied to them so that the electric field is at a right angle to the
magnetic field. An offset exit hole (S

) on the first plate past

the interaction region is followed by another set of focusing lens.
The exit hole picks off a small section of the dispersed electrons
to yield a very narrowly defined energy for the electrons that

FIGURE 17.

Metastable precursor negative-ion current of fluoranthene

as a function of electron energy in comparison with SF

* and SF

. The

RPD technique was employed. Reproduced from the work of Frey et al.
(1973) with permission from Elsevier Limited, Copyright 1973.

FIGURE 18.

Appearance potential for positive ions in acenaphthene. The retarding potential difference

(RPD) method was used. The appearance potential of Kr

from Krypton was used to establish the energy

scale. Reproduced from the work of Frey et al. (1973) with permission from Elsevier Limited, Copyright
1973.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

251

continue through the analyzer. The next set of plates focus the
electrons. The last plate is held at ground potential to define the
electrons energy before they enter the interaction region. A
typical experimental setup is shown in Figure 19.

The resolution for a typical TM is usually in the 30–75 meV

range. Of course, the process of producing such a narrowly
defined energy range limits the amount of current through the
apparatus, typically in the nano-ampere range. The graph below,
Figure 20, from Stamatovic and Schulz (1968), shows the
resolution as a function of electron beam current for their
apparatus. Such low current beam from a TM adds to the
difficultly of collecting ion signal with the time-of-flight system,
which typically uses micro-channel plates to collect a real-time
voltage pulse. An ion-counting system could be used, especially
for very small signals, to build up a mass and cross-section
energy-dependence over time.

Although TMs have been widely used with Quadrupole

Mass Spectrometers, to our knowledge, only two groups have
employed a TM with a TOFMS. The group of Deinzer and
Voinov et al. at Oregon State University has developed a resonant
electron capture TOF MS with a trochoidal electron mono-
chromator (Voinov et al., 2003, 2004). A schematic of this
instrument is shown in Figure 21. Note that the ions are injected
orthogonal to the flight path and a high repetition rate for
ion analysis is employed. The later version also used a GC
introduction system (Voinov et al., 2004). This instrument
provides an important new tool for the analysis of environmental
and biological samples (Vasil’ev et al., 2006, 2007). Later,
Robertson et al. (2005) introduced a more conventional TOFMS
using a trochoidal electron source (see Fig. 22) in which a second
electron beam is collinear with, but traveling in opposite
direction to, the TM electron beam. This apparatus will allow
the study of sequential electron-collision processes between
electrons and positive or negative ions preformed by the first
electron beam.

C. Electrostatic Energy Analyzers

The electrostatic energy analyzer (ESA) essentially employs an
electrostatic field at right angles to the trajectory of electrons
emitted from a filament to disperse electrons that travel with
different energies. A slit or an aperture at the exit of the

electrostatic field selects a narrow energy band from the electron-
energy distribution. The electrostatic field can result from either a
1278—cylindrical sector analyzer or an 1808—hemispherical
analyzer. As the name implies, the 1278—cylindrical analyzer
employs two partial cylindrical plates, and the hemispherical
analyzer employs two hemisphere electrodes, to disperse the
electron beam. The 1278 and 1808 refers to the angle through
which the electron beam must pass in the electric fields for
optimum focusing at the exit slit. These analyzers have been
described in detail by Simpson and Kuyatt (1963), which remain
today as the best reference for their performance and operation
(see also Hasted, 1972; McDaniel, 1989). It is crucial to reduce
external magnetic fields (Earth’s field, etc.) in these studies to
values less than milligauss. This reduction is done with mu-metal
shielding or Helmholtz fields or preferably both. It is also
imperative to remove any material in the spectrometer and its
environments that may become magnetized such as screws, etc.
Screwdrivers that have become magnetized can play havoc with
permeable material in the spectrometer.

The 1278 cylindrical electrostatic energy analyzer was

developed in the late 1950s by Marmet and Kerwin (1960), and
provided energy resolution of

60 meV. The 1278 cylindrical

analyzer consists of a filament followed by a series of slits to
collimate the electron beam before entering the cylindrical
capacitor. This instrument was the first to provide an energy
resolution

100 meV with electron-beam currents within

0.1–1 nanoampere.

The spherical sector electron analyzer has been widely used

in atomic and molecular physics (Hasted, 1972; McDaniel,
1989). An example of the hemispherical sector electron analyzer
applied to a TOFMS is shown in Figure 23 (Huang, 1995; Huang,
Carman, & Compton, 1995). Figure 23 shows the spherical sector
positioned so that the electron beam passes through the source
region of the TOFMS. The molecules in the interaction region
might consist of a diffuse gas, molecular beam, or a nozzle-
jet expansion. By way of illustration (Huang, 1995), we will
briefly discuss electron ionization and attachment to a thermal
(T

4008C) beam of C

molecules introduced at

in Figure 23.

is anomalous in its low cross-section for positive ionization

in the threshold region and its exceptionally large electron-
attachment cross-section from

0 eV to over 10 eV. Figure 24

shows the weak onset of ionization at the known ionization
potential of C

at 7.6 eV. The electron energy scale was

calibrated with the onset for electron ionization of Ar at 15.76 eV
(see Fig. 25). One will note the ‘‘break’’ in the ionization cross-
section at

16.6 eV due to the opening of the ionization channel

due to excited Ar

. Both gases are present during the measure-

ments, and the onsets can be determined simultaneously. The
relative cross-section for the attachment of low-energy electrons
to C

is shown in Figure 26. The low-energy resonance cross-

section for the attachment of slow electrons (<0.02 eV) to sulfur
hexaflouride is employed to establish the absolute electron
energy scale. Using known calibration gases, one can determine
the electron energy scale to a precision comparable to the electron
energy resolution (

0.2 to 0.05 eV).

There are continuing efforts to improve the performance of

hemispherical electrostatic energy analyzers. A recent article has
appeared that describes theoretical calculations designed to
improve the inherent resolution of hemispherical analyzers

FIGURE 19.

Shows principle of operation of the trochoidal electron

monochromator (TM). Reproduced from the work of Stamatovic and
Schulz (1970) with permission from American Institute of Physics,
Copyright 1970.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

252

Mass Spectrometry Reviews DOI 10.1002/mas

(Sagara et al., 2000). The article by Sagara et al. also summarizes
previous experimental and theoretical attempts in this area.

The resolution of the hemispherical analyzer is greatly

limited by space-charge conditions of the electron beam. The
maximum current that can pass through a field-free space is
proportional to the 1.5 power of the electron energy. Combining
this fundamental fact of space charge (also a consequence of
Louisville’s theorem) with the equations of motion through the
analyzer, it can be shown (Simpson & Kuyatt, 1963; Simpson,
1964) that the maximum current is inversely proportional to the
square root of the electron energy. A curve of electron beam
current through a spherical analyzer versus the electron beam
resolution at full-width half-maximum (FWHM) shows that the
maximum current available for an electron energy resolution of
0.027 is 1 nanoampere (see Fig. 27). As in the cases of the
RPD and the TM methods, both of which employ magnetic fields,
the transmission of electron current through a hemispherical
analyzer versus electron energy is not constant (especially at low
energy). An identical relation can also be found on page 230 of
the book by Hasted (Hasted, 1972) where both hemispherical and
cylindrical analyzers are treated. As one example, the maximum
current available for an electron-energy resolution of

0.027 is

1 nanoampere. Also, the transmission of the electron current
versus electron energy is not constant (especially at low energy)
as in the case of the RPD or the TM methods, which employ
magnetic fields.

D. Photoionization Electron Source

Many groups have employed photoionization of atoms to
produce photoelectrons of very high resolution and well defined
energy. These sources also suffer from the space-charge
limitation, although the positive ion formed mitigates the space

charge to some extent. The initial energy distribution from these
sources can be very narrow and the mean energy of the narrow
electron beam distribution can be precisely controlled by
changing the photon frequency. This method represents the most
desirable method to date for performing high-resolution electron
ionization experiments. The Hartmut Hotop group at Kaiserlau-
tern and the David Field group at Aarhus have made great
improvements in electron-energy resolution, using this technique
in the study of low-energy electron molecule/atom collision
physics (see, for example, Hoffmann et al., 2002; Boemmels
et al., 2005). The Smith group at the University of Arizona has
also successfully employed this method to study electron
attachment to SF

in a TOFMS (Garrec, Steinhurst, & Smith,

2001). This group employed a Nd:YAG-pumped dye laser to
ionize sodium atoms to produce low-energy electrons. A mixture
of DCM and rhodamine 101 (lambdachrome) laser dyes were
used to produce 45 mJ/pulse radiation, which was frequency
doubled and frequency mixed with the 1,064 nm fundamental
from the Nd:YAG laser. Typically, the mixing-after-doubling
output power was

1 mJ/pulse at 240 nm. The ionization scheme

þ Na ! Na

þ e(e) produces free electrons e of energy e.

The electron energy is determined by the difference h(v–v

where hv and hv

are the energies of the laser light and the

threshold ionization energy for sodium (5.139 eV), respectively.

V. FUNDAMENTALS OF ELECTRON IONIZATION

Books and review articles on the interaction of electrons with
atoms and molecules abound (Massey, Burhop, & Gilbody,
1969a,b; Hasted, 1972; Smirnov, 1982; Compton & Bardsley,
1984; McDaniel, 1989). In this review, we focus only on those
topics pertinent to EITOFMS. Electron ionization and resonance

FIGURE 20.

Electron beam resolution as function of beam current. Reproduced from the work of

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

253

electron attachment will be of primary concern. However, we will
broaden our treatment to include the interaction of highly excited
Rydberg atoms. Under certain conditions, Rydberg atoms in a
high state of excitation exhibit the properties of slow free
electrons; that is, the electron excited into an orbit distant from
the ion core moves slowly and can be treated as a slow-free
electron.

A. Positive Ionization

As discussed earlier, positive ion electron ionization was
employed in some of the very first applications of time-of-flight
mass spectrometry (see, e.g., Jatzenstein & Friedland, 1955). The
immediate advantages of EITOF to the mass spectroscopy

community were: (1) well-defined ionization region with precise
control of the electron energy, especially at low energy, (2) very
large mass range, from H

to m/q

50,000, (3) ease of

conversion from positive to negative ions, (4) good transmission
(

100%) and very little discrimination as a function of mass, (5)

ability to vary the residence time in the TOF ion source to
investigate ion molecule reactions, etc., and (6) ions created on
the filament of the EI source can be identified using the delayed
draw out.

The sensitivity of detecting gas-phase molecules by electron

ionization is determined by the ionization cross-section as a
function of electron energy. The energy dependence of electron
ionization cross-sections rises approximately linearly from zero
at the ionization potential (IP) of the atom or molecule and
exhibits a maximum at an electron energy that is approximately

FIGURE 21.

Schematic diagram of the TEM/TOF mass spectrometer, including TOF electronics.

Reproduced from the publication of Voinov et al. (Voinov et al., 2003 2003) with permission from American
Chemical Society, Copyright 2003.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

254

Mass Spectrometry Reviews DOI 10.1002/mas

two to three times the IP. At higher energy, the cross-section falls
approximately as (1/E) ln E. The ion source region of a typical
TOFMS is ideally suited to provide an estimate to the ionization
cross-section provided that the pressure in the ionization region
can be determined. Alternately, if the ionization cross-section (s)
and the electron path length (‘) are known, then the gas density in
the ionization region can be determined by measuring the
electron beam (I

) and ion beam (I

) current. The ion current can

be collected on the electrodes (ion pusher plate). Provided that

single electron–molecule conditions exist, the gas density (n) can
be calculated from n

¼ (I

)(1/s‘). We have employed a

TOFMS in this manner to demonstrate how to measure cross-
sections in undergraduate Physical Chemistry and Graduate
Physics laboratories for the past 6 years. The measurement of
electron ionization cross-sections together with a theoretical
understanding of the physics involved has played a major role in
atomic, molecular, and optical physics over the past 50 years
(Massey, Burhop, & Gilbody, 1969a,b).

FIGURE 22.

Schematic of time-of-flight with trochoidal electron gun and standard electron gun.

Reproduced from the work of Robertson et al. (2005) with permission from American Institute of Physics,
Copyright 2005.

FIGURE 23.

Schematic diagram of an electron ionization time-of-flight mass spectrometer. The

monochromatic electron beam is produced by a 1608 spherical-sector electrostatic energy analyzer with
a three element Einzel lens (L) used to focus the low-energy electron beam into the C

beam shown by

The electron current is collected by the Faraday collector (C) and recorded with the aid of an electrometer.
Not shown are the defectors in the flight tube (T) used to compensate for the C

beam velocity. Reproduced

from the work of Huang, Carman, and Compton (1995) with permission from American Chemical Society,
Copyright 1995.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

255

FIGURE 24.

Cross-section of electron ionization of C

. Reproduced from J. Huang’s PhD thesis, 1995

with permission from the author.

FIGURE 25.

Illustration of threshold electron ionization of Ar using the ESA. Reproduced from J. Huang’s

PhD thesis, 1995 with permission from the author.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

256

Mass Spectrometry Reviews DOI 10.1002/mas

One of the first theoretical calculations of total ionization

cross-section was presented by Khare and co-workers (Jain &
Khare, 1976b) who used two different collision theories (soft and
hard collision). Although many more approaches have been
developed since then, three models are widely used to model
ionization cross-sections. For example, Deutsch, Ma¨rk, and their

co-workers (Deutsch & Ma¨rk, 1987; Margreiter et al., 1990;
Margreiter et al., 1994) have developed a semiempirical method
(DM) to calculate electron ionization cross-sections. The binary-
encounter-Bethe (BEB; Kim and Rudd, 1994; Hwang, Kim, &
Rudd, 1996) model was proposed by Kim and Rudd by using an
analytical formula for the cross-section. Shortly afterward,
Vallance et al. introduced an electrostatic model (EM; Vallance,
Harland, & Maclagan, 1996; Vallance, Maclagan, & Harland,
1997).

A book chapter by Harland and Vallance (1998) described a

number of experimental methods that have been employed to
determine cross-sections as well as theoretical methods to
calculate the electron ionization cross-section in more detail.
According to Harland and Vallance, although the EM method
provides a better calculation of the maximum absolute ionization
cross-section, the DM and BEB methods provide a more accurate
description of the cross-section as a function of the electron
energy (see Fig. 28). From their detailed comparisons, they arrive
at the conclusion that, to provide a better cross-section model,
it would be necessary to consider all three methods to obtain a
comprehensive description of the experimental data. In their
review, they argue that the BEB method is the best model for
small molecules and it fits better with the experimental data at the
ionization threshold whereas the DM method represents a better
fit for the heavier molecules.

In 2000, Deutsch et al. (2000) also summarized theoretical

approaches to model the ionization cross-sections in a separate
review article. In that review, they compared experimental
cross sections with theoretical models for 31 molecules and free
radicals. Here, we will only discuss two of their comparisons.

FIGURE 26.

Relative cross-section for electron attachment in the range

from 0 to 12 eV to a thermal (T

¼ 673 K) beam of C

. The SF

energy

scale calibration is also shown for reference. An expanded view of the
low-energy attachment is shown above. Reproduced from the work of
Huang, Carman, and Compton (1995) with permission from American
Chemical Society, Copyright 1995.

FIGURE 27.

Electron beam current through a spherical electrostatic

energy analyzer as a function of the electron energy resolution defined by
the full-width at half-max. Constructed based upon information in the
articles by Simpson and Kuyatt (1963) and Hasted et al. (1972).

FIGURE 28.

A comparison between measured maximum ionization

cross-sections and the theoretical calculations, using the BEB (

þ), DM

(O), and EM (*) methods. The heavy line represents a perfect
correspondence between experiment and theory. Reproduced from the
work of Harland and Vallance (1997) with permission from Elsevier
Limited, Copyright 1997.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

257

In Figure 29, four experimental cross-sections of C

are

compared with DM (Deutsch et al., 2000), BEB (Hwang, Kim, &
Rudd, 1996), and the model by Jain and Khare (1976b). All four
experimental data sets are in very good agreement. In terms
of theoretical models, the BEB calculation agrees with the
experimental data for the entire range of electron energy. The DM
calculation as well as Jain and Khare’s model also show quite
good agreement with the experimental data, but they over-
estimate the maximum cross section by

10–15% (Deutsch

et al., 2000).

In Figure 30, experimental ionization cross section data

points of CO

for four different experiments (Rapp & Englander-

Golden, 1965; Shyn & Sharp, 1979; Orient & Srivastava, 1987;
Straub et al., 1996) are compared with the theoretical models
(Jain & Khare, 1976a; Hwang, Kim, & Rudd, 1996; Saksena,
Kushwaha, & Khare, 1997a,b). As can be seen, all four
calculations give a reasonable estimate of the experimental data;
however, the DM curve drops off faster than the other cross-
section models at higher energies.

For the past few decades, the theoretical models for electron

ionization cross-section have explained well the experimental
cross-sections; however, all of these models rely on the
experimental data to validate the theory.

Correlation between maximum electron ionization cross-

section and atomic or molecular polarizability attracted the
interest of Franklin and co-workers in 1957 (Lampe, Franklin,
& Field, 1957). Recently, Nishimura and Tawara (1994)
found a direct relation among the electron ionization cross-
section, number of electrons, and molecular dipole polar-
izability for some simple hydrocarbon molecules. Further
investigation shows a strong correlation between the maximum
electron ionization cross-section and the quantity (a/E

)

1/2

where a is the polarizability volume and E

is the ionization

potential.

According to Harland and Vallance (1997), this correlation

holds quite well for most atoms and molecules in general. For an
atom, its position in the periodic table defines the proportionality
constant (Fig. 31), whereas for a molecule, s

max

correlates

linearly with (a/E

)

1/2

(Fig. 32) The theoretical determination of

this correlation has been discussed in the review by Harland and
Vallance (1998). Thus, from these studies it is clear that one can
estimate the amount of the precursor ion signal that one might
expect from electron ionization of a given atom or molecule. For
molecules; however, one caveat should be stated. Namely, some
molecules primarily dissociate into fragment ions and neutrals
upon electron ionization. For example, electron ionization of the

FIGURE 29.

A comparison of the electron ionization cross-section for C

as a function of electron

energy. The experimental data points are from Duric, Cadez, and Kurepa (1991) [filled circles], (Schram
et al., 1966) [filled triangles], (Chatham et al., 1984) [filled squares], and (Grill et al., 1993)] [filled
diamonds]. The thick solid line represents DM calculation (Deutsch et al., 2000), the thin solid line denotes
the BEB calculation (Hwang, Kim, & Rudd, 1996) and the dotted line represents the calculation of Jain and
Khare (1976). Reproduced from the work of Deutsch et al. (2000) with permission from Elsevier Limited,
Copyright 2000.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

258

Mass Spectrometry Reviews DOI 10.1002/mas

molecule produces mainly fragment ions (UF

, with n

¼ 5

to 0), which peaks at

150 eV (Compton, 1977). Using the

polarizability of 12.6 A

˚´

(CRC Handbook) and the ionization

threshold as 15 eV, one obtains a cross-section of

14 A

˚´

, which

agrees well with the measured value of 17 A

˚´

(Compton, 1977). A

TOFMS was used to determine positive and negative ionization
cross-sections for UF

in the manner described above. Interest-

ingly, that study also showed that, despite the electron affinity of
a UF

molecule being over 5.1 eV, UF

does not attach slow

electrons. This example illustrates an important distinction
between positive and negative ionization of molecules by
electrons. Whereas simple descriptions can provide relatively
accurate positive ionization cross-sections, the prediction of
electron attachment cross-sections is far from even a preliminary
understanding.

B. Negative Ionization

1. Low-Energy Electron-Attachment Studies That
Employ a Time-of-Flight Mass Spectrometer

A time-of-flight mass spectrometer has proven to be ideally
suited for the study of low-energy electron attachment to
molecules. Electron-attachment cross-sections can be orders of
magnitude larger than the electron ionization cross-sections
discussed above. The attachment cross-section for many
molecules approaches the theoretical geometrical maximum.
The geometrical maximum for attachment of an electron with
angular momentum ‘ and de Broglie wavelength

l to a point

object is

ð2‘ þ 1Þ

and the total cross-section is s

max

ð2‘ þ 1Þ

. In the case that the

‘‘

size,

’’

R, of the molecule

is on the order of the de Broglie wavelength, this maximum can be
extended to

ð2‘ þ 1Þpð

þ RÞ

(Blatt & Weisskopf, 1952). In

addition, the presence of resonance features in the attachment
cross-section makes for much more interesting science and
potential applications. There are many books and review articles
devoted to the physics of electron attachment to molecules
(Schulz, 1973; Massey, 1976; Smirnov, 1982), and this review
will concentrate only on the utility of TOF-MS in this area of
chemical physics and analytical chemistry. The attachment of an
electron to a molecule occurs when the electron is temporarily
‘‘held-up’’ in its transit by a molecule. In order for the electron to
be ‘‘held-up’’ by the molecule, one of three types of conditions
must occur (Schulz, 1973):

1. Nuclear excited Feshbach Resonance: This attachment

process is similar to the Bohr model for the capture of a
neutron by a nucleus in which the incident neutron shares
its incident energy with the other nucleons of the nucleus
to produce a long-lived ‘‘compound nucleus.’’ By analogy,
the incoming electron enters an unoccupied valence
molecular orbital and is temporarily attached to the
molecule as a result of vibrational excitation imparted to
the molecular anion. Unimolecular electron attachment
represents a breakdown of the Born–Oppenheimer
approximation, in which the incident electron energy, E

and the electron-affinity energy, EA, are shared among the
available vibrational modes of the molecular anion.
Because this process is unimolecular, the electron will

FIGURE 30.

A comparison of electron ionization cross-section of CO

as a function of electron energy. The

experimental data points are from (Straub et al., 1996) [filled circles], (Rapp et al., 1965) [open circles],
(Shyn et al., 1979) [stars] and (Orient et al., 1987) [open triangles]. The thick solid line represents DM
calculation (Deutsch et al., 2000), the thin solid line denotes the BEB calculation (Hwang, Kim, & Rudd,
1996), the dotted line represents the calculation of Jain and Khare (1976a) and the dashed line denotes the
calculation of Saksena, Kushwaha, and Khare (1997a,b). Reproduced from the work of Deutsch et al. (2000)
with permission from Elsevier Limited, Copyright 2000.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

259

undergo autodetachment with a mean lifetime, t. This
attachment/autodetachment process was first treated
with the principal of detailed balance (Compton et al.,
1966a), which relates the autodetachment lifetime to the
attachment cross-section for electrons of velocity , s(),
and the density of states of the negative ion, r

, through

the relationship

ð16Þ

where

h is the Planck constant divided by 2p. The lifetime

is determined primarily by the density of states of the anion,
which can be conveniently calculated from (Whitten &
Rabinovitch, 1963).

ðE þ aE

ðNÞPðhu

ð17Þ

where E is the total internal energy (electron affinity plus
electron kinetic energy, EA

þ E

), E

is the total zero-point

energy for the anion, a is an empirical factor, N is the total
number of degrees of vibrational freedom (3 times the

number of atoms minus 6 for a non-linear molecule), G(N)
is the gamma function, and P(hu

) represents the product of

all the vibrational energies, hu

. It is easily seen by

inspection of Equation (16) that long lifetimes can exist
when the anion has a large density of states; that is, those
molecules that posses a large number of degrees of freedom
and which exhibit a large electron affinity. This process will
occur for essentially any molecule with a positive electron
affinity; however, such ions might not exhibit an autode-
tachment lifetime long enough to be observed in a typical
mass spectrometer (

1 m sec).

2. Electronically excited Feshbach Resonance: In this

process, the incident electron excites one or more of the
electrons in the molecule, and subsequently falls into an
available empty orbital of the excited state. Once again,
the molecular anion will have a finite lifetime described by
a greatly modified form of Equation (16). In this case, the
incident electron energy is shared between the electronic
and vibrational densities of state of the negative ion.

3. Shape Resonance: Electrons can be temporarily attached

to a molecule (or atom) by what is termed a ‘‘shape
resonance,’’ in which the electron –molecule system is in a

FIGURE 31.

Correlation of atomic maximum electron ionization cross-sections with the quantity (a/E

)

1/2

reproduced from the work of Harland and Vallance (1997) with permission from Elsevier Limited,
Copyright 1997.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

260

Mass Spectrometry Reviews DOI 10.1002/mas

quasi-stationary state as a result of the combined repulsive
centripetal potential ‘(‘

þ 1)h

/2 mr

and attractive polar-

ization—a/r

potential, where a is the molecular polar-

izability. These ions are typically short-lived; however, in
the special case of electron attachment to a negative ion,
it has been shown that a doubly charged negative ion
(e.g., C

) can exist in a long-lived unbound state for

many seconds as a result of the long-range coulomb
repulsion and short-range polarizability attraction (Comp-
ton et al., 1997). The potential ‘‘hill’’ created by this long-
range repulsion and short-range attraction has been
referred to as a ‘‘coulomb barrier.’’ The term ‘‘coulomb
barrier’’ first arose to describe the decay of a metastable
nucleus to emit an alpha particle (doubly charged helium
ion). In a multiply charged negative ion and a metastable
nucleus, the long-range repulsion is due to the coulomb
repulsion. For the metastable nucleus, the coulomb barrier
results from the short-range nuclear force, whereas for the
multiply charged negative ion the coulomb barrier results
primarily from the polarizability attraction. Although
electron attachment to a negative ion has only been

verified in two cases, it is expected to be one of the
primary mechanisms for the formation of multiply
charged anions. This subject is important to the field of
electrospray ionization, where multiply charged anions are
used for mass analysis. Two recent reviews of the
relatively new field of multiply charged negative ions
have appeared (Scheller, Compton, & Cederbaum, 1995;
Boldyrev, Gutowski, & Simons, 1996).

Electrons can also induce dissociative electron attachment

for molecules in which any of the above resonances is the initial
step in the reaction. For example, an electron might be initially
attached into an electronically excited Feshbach resonance and
rapidly decay by dissociation into a negative ion fragment and a
neutral. Dissociative electron attachment represents a severe
breakdown of the Born–Oppenheimer approximation, and
the lifetime for the dissociation of the resonance generally
occurs during the time required for one vibration time (i.e.,
10

13 –15

sec). However, when the incident electron energy and

the electron affinity energy are shared among the various
vibrational degrees of freedom, metastable dissociation on

FIGURE 32.

Correlation of molecular maximum electron ionization cross-sections with the quantity

(a/E

)

1/2

reproduced from the work of Harland and Vallance (1997) with permission from Elsevier Limited,

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

261

the time-scale of microseconds can also occur as will be
illustrated below.

We should also include the formation of ion-pairs by

electron ionization to the list of possible mechanisms that can
produce negative ions. The incoming electron excites an
electronically excited ion-pair state of the molecule that leads
to dissociation into fragment ion-pairs. Ion-pair formation
does not involve a resonance, and the cross-section exhibits a
somewhat linear increase with an onset close to the energy that
corresponds to the energetic threshold (ionization potential of the
positive species plus the bond dissociation energy for the neutral
fragments minus the electron affinity of the negative fragment).
The incoming electron excites an electronically excited ion-pair
state of the molecule that leads to rapid dissociation into fragment
ion-pairs.

Finally, we include one other electron-attachment mecha-

nism, which to our knowledge has only recently been pointed out.
It has recently been suggested that an electron can attach to
molecules to result in molecular bond breaking (e.g., ring
opening), but not complete dissociation, leaving the anion
‘‘partially dissociated’’ (Robertson et al., 2005). In this
incomplete dissociative electron attachment (IDEA) processes,
the anion can be left in a configuration that is not susceptible to
autodetachment until the broken bond is reformed. This process
may be especially important in the case of large biomolecules.
For example, Martin et al. (2004) have also presented evidence
for a ‘‘partial’’ dissociation of negative ion-shape resonances.
Specifically, they reported single-strand breaks in the DNA super
molecule induced by low-energy (0–4 eV) electrons. This IDEA
should be tested for other ring-opening electron attachment
processes.

The study of electron attachment requires the use of energy-

resolved electrons under well-controlled conditions at low
energy (0 to

10 eV). A typical TOF-MS apparatus is shown

in Figure 33, where a pulsed (

10 nsec) beam of electrons enters

a field-free interaction region. Ions produced by electron
attachment are pulsed out of the interaction region and directed
down the flight tube for mass analysis. It is possible to introduce
electrons with energies near zero (<0.1 eV) if proper precautions
are taken. The most important condition to be met is to keep
the interaction region ‘‘field-free.’’ Schulz first pointed out the
importance of stray electrostatic potentials in the interaction
region (Schulz, 1960a,b). There are many sources of stray
electrostatic fields in the ion-source region. For example, the
pulsed voltage applied to the draw-out electrode (backing plate or
extraction grid) can introduce a slight positive potential (voltage
over-shoot) onto this electrode. If this slight voltage is negative,
then the result is to shift the electron energy scale (as monitored
by the acceleration voltage) to higher energy. ‘‘Zero-energy’’
electrons can still be introduced into the chamber by ‘‘climbing’’
this negative bias. On the other hand, if the potential of
the interaction is positive, then zero-energy electrons will be
accelerated into the chamber to make it impossible to observe the
true nature of a very low-energy resonance. Higher-energy
electrons that enter the chamber will simply be increased in
energy by this small amount. However, when the electrons enter
the region with energies less than this potential, the electron
energies will be increased and never reach ‘‘zero-energy.’’ The
voltage on the flight tube can also be a source of positive potential
creeping into the interaction region. If the grids that define the
ion draw-out potentials are not of sufficient density to prevent
field penetration, then a positive potential will exist in the
interaction region and again the electrons will never reach zero
energy. Molecules that coat the surfaces of the collision chamber
can also introduce an unwanted surface potential. The
importance of maintaining the potential of the interaction at
ground potential cannot be overstated. Many errors in low-
energy electron attachment studies have occurred and continue
to occur as a result of this technical problem. It is also imperative
to coat the interaction region with aerodag, carbon soot,

FIGURE 33.

Schematic diagram of a typical time-of-flight mass spectrometer. Reproduced from the work

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

262

Mass Spectrometry Reviews DOI 10.1002/mas

platinum black, or other to reduce irregular surface potentials
and especially to reduce scattered electrons. Heating the ion
source to remove insulating surface molecules is also highly
desirable.

Although we have emphasized the importance of making the

interaction region at ground potential, there can be special
applications where the introduction of a negative potential can be
beneficial. If a slight negative potential is introduced into the
interaction region, then electrons that enter this region can be
turned around and spend some period of time near zero energy as
the electrons ‘‘turn around.’’ Early researchers in this field
noticed that under these conditions very large negative ion signals
could be observed for molecules, which attach with a cross-
section that approaches ‘‘infinity’’ as the electron velocity
approaches zero. Anyone employing a battery bias on the
collision electrodes has noticed an enormous increase in the ion
current as a result of the ‘‘turn around’’ of slow electrons that
increase the interaction time and cross-section. This simple
but powerful technique has found analytical applications
(Boumsellek, et al., 1992a,b).

The approximate electron-beam energy can be determined

by measuring the voltage difference between the center-point of
the filament and the electron/molecule interaction region (usually
ground). The actual or absolute energy scale is ‘‘calibrated’’ by
observing an energy threshold for a known reaction. Also, the
electron-energy scale can be calibrated by observing the SF

ion

signal at ‘‘zero-energy.’’ Hickam and Fox (1956) were the first to
utilize this method, and it has been used by most experimentalists
in the field of low-energy electron attachment. This sharp signal
peaking at

0 eV, has also been equated to the derivative of the

electron-retarding potential curve (i.e., the appearance curve of

the electron current upon going through ‘‘zero energy’’). The
peak in the SF

current is often taken to be

0.05 eV. Of course,

if the condition of having no positive voltages in the interaction
region is not met, then the use of the position of the SF

0.05

eV is void. There are many well-known electron-attachment
resonances at energies from 0.2 to

10 eV that can be used to

calibrate the electron-energy scale. The most accurate electron-
energy scales are determined by using atomic electronic energy
levels, using what is called the SF

Scavenger Technique. The

Scavenger Technique was first introduced by Curran in

1963 (Curran, 1963), and later by Jacobs and Henglein in 1964
(Jacobs and Henglein, 1964) and widely used by Compton et al.
(1968) to study electronically excited states of atoms and
molecules as well as to detect temporary negative ion states of
molecules. The technique involves the introduction of SF

together with another gas into the electron–molecule interaction
region. Electrons that undergo a completely inelastic collision
with the second gas molecules produce

zero energy electrons

that are captured or ‘‘scavenged’’ by SF

molecules. Thus,

scanning the electron energy and recording the SF

ion signal

produces a spectrum that can be called a ‘‘threshold electron
impact excitation’’ spectrum. This method is especially useful to
study the excitation and existence of triplet states as well as
temporary negative ion resonances. This technique is illustrated
by the threshold impact excitation spectrum for benzene shown in
Figure 34. The large peak at

1.5 eV is the ‘ ¼ 3 (f-wave)

negative ion shape resonance of the benzene negative ion. The
lowest triplet state of benzene is also clearly visible. Another
example of the SF

Scavenger Technique is shown in Figure 35

for hydrogen chloride. Various electronic excited states are seen
including a molecular Rydberg series, leading up to the

FIGURE 34.

Electron-impact threshold excitation spectrum of C

. Produced from the work of Compton

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

263

ionization potential. Notice that the Cl

ion from HCl is also

recorded, which illustrates another powerful aspect of the SF

Scavenger Technique. In many cases, temporary negative ion
states leading to slow electron and dissociation can both be
studied. For example, the C

Cl molecule shows short-lived

autodetaching states and Cl

that occur from a common

resonance. This and other examples are shown in Figure 36
which records the SF

Scavenger signal along with the negative

ion yields for a series of benzene derivatives. One will note that a
single resonance decays to slow electrons as well as dissociates
into a negative ion. The SF

Scavenger Technique can also be

used as an excellent means to calibrate the electron energy scale.
For electron ionization studies, by recording the SF

scavenger

spectrum for rare gases such as helium, argon, or krypton in
mixtures with the gas under study, the energy scale can be
calibrated by using the well-known spectroscopic values for
the energy levels of the rare gas. This method was used in many of
the studies of dissociative electron attachment over the years. The
SF

Scavenger Technique is very similar to the Trapped-

Electron Method first introduced by Schulz (1958, 1959,
1960b) in which the ‘‘well-depth’’ is

0.02 eV. Also, the SF

Scavenger Technique is a fore-runner to the zero kinetic energy
electron spectroscopy (ZEKE) commonly used today in laser
spectroscopy, the difference being that one uses electrons
instead of light to produce the ZEKE spectra (Mu¨ller-Dethlefs,
Sander, & Schlag, 1984). It should be added that atoms or
molecules that have been excited to high Rydberg levels can also

charge exchange to SF

to produce SF

, making the

‘‘

Zero-

energy

’’

that results from autodetachment of SF

* ions

production method uncertain by

0.05 eV. However, if one

monitors the neutral SF

molecules produced by slow electron

attachment (see later in the review), then one can avoid those
SF

ions produced by high Rydberg species. Thus, the SF

Scavenger Technique, employed under these conditions can
be considered as a true zero kinetic energy electron (ZEKE)
detection method.

The spherical sector electrostatic energy analyzer has also

been used in one TOFMS study to observe the electron
attachment cross-section for the attachment of low-energy
electrons to the C

molecule in a molecular beam (Huang,

Carman, & Compton, 1995). The apparatus is shown in Figure 23
in which the earth’s magnetic field is reduced to <0.001 Gauss by
individually adjusting the current through three mutually
orthogonal Helmholz coils. The spherical sector and cylindrical
electron energy analyzers have been used in many experiments
with a quadrupole mass analyzer (QMS). However, one of the
authors (RNC) has found the use of these analyzers difficult to
employ due to the presence of radio frequency voltages that
appear on the inner and outer spheres of the analyzer. On the
positive side, the high duty cycle offered by the QMS and the ease
of single ion counting has definite advantages over the TOF. On
the negative side, the QMS does require ion extraction, which can
adversely affect the electron energy as discussed above. Effects
of the ion draw out of field can be alleviated by employing pulsed

FIGURE 35.

Electron-impact threshold excitation spectrum of HCl. Produced from the work of Compton

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

264

Mass Spectrometry Reviews DOI 10.1002/mas

ion extraction while turning off the electron beam, as is similarly
done in the case of EITOF.

2. SF

Lifetime

Autodetachment lifetimes from SF

ions formed by the

attachment of low-energy electrons to the SF

molecule; that is,

þ SF

$ SF

ð18Þ

was first observed by Edelson, Griffiths, and McAfee (1962) in
a time-of-flight mass spectrometer. The method consisted of
attaching ‘‘low-energy’’ electrons to gaseous SF

and measuring

the ratio of SF

neutrals compared to SF

ions that arrive at the

detector after a predetermined flight time down the flight-tube of
a TOF mass spectrometer. The kinetic energy of the SF

particles that arrive at the detector produced pulse heights as large
as those of the SF

. Four years later, Compton et al. (1966a)

determined this ratio as a function of the TOF by varying the ion

FIGURE 36.

Electron-energy loss resonances in benzene and benzene derivatives and comparison with the

dissociative-attachment resonances for several halogenated benzenes. The dotted curve represents a
subtraction of the primary SF

current from the total SF

current with SF

and o-dichlorobenzene in the

mass spectrometer. Reproduced from the work of Compton et al. (1966b) with permission from Elsevier
Limited, Copyright 1966.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

265

velocity down the flight-tube to measure the lifetime of
exponential decay. A typical time-of-arrival distribution of the
SF

and SF

particles that arrive at the detector is shown in

Figure 37.

Five years after this study, Harland and Thynne (1971a)

again measured the autodetachment lifetime with a TOFMS for
three different ion energies. Later, Delmore and Applehans
(1986) also used the TOF method to determine the autodetach-
ment lifetime by measuring the SF

neutrals that result from

autodetachment of SF

ions as a function of the distance along

the flight path. These authors also found that the ‘‘effective
lifetime increased from 15 to 20 msec as the temperature of the
SF

was reduced from

475 to 375 K. Two years later,

Applehans and Delmore (1988) reported a ‘‘refinement’’ of their
previous measurement to include two other lifetime components
in the beam. Thirteen years later, Garrec, Steinhurst, and Smith
(2001) measured the autodetachment lifetime of SF

* as a

function of the incident electron energy for SF

cooled in a free-

jet expansion, again with the TOF technique. They report no
change in the autodetachment lifetime in the range from

to100 meV. Thus, autodetachment studies of SF

* with the

TOFMS technique over the past 40 years have reported lifetimes
that vary from

10 to 60 msec as summarized in Table 1.

In the mid-seventies, Odom, Smith, and Futrell (1975) used

an ion cyclotron resonance (ICR) technique to measure the

‘‘decay’’ of SF

* as a function of the ‘‘residence time’’ in the

ICR. These authors found that the ‘‘apparent’’ autodetachment
lifetime of SF

* varied as a function of the observation time

(non-single exponential decay) from 50 msec to 10 msec. These
authors argue that those ions that exist for as long as 10 msec
might undergo radiative stabilization. Likewise, Foster and
Beauchamp (1975) rationalized their observation of stable SF

in an ICR mass spectrometer to radiative stabilization. Henis and
Mabie (1970) also used an ICR to measure a lifetime of
300 msec for SF

* formed by slow electron attachment. In

support of the increase of lifetime for a packet of ions due to
radiative decay it should be noted that emission of one IR
quantum of radiation of SF

* ions made by electron attachment

to a ground state SF

will be stable.

Compton et al. (1966) and Klots (1967) introduced a theory

of electron autodetachment based upon detailed balancing and
quasi-equilibrium theory to relate the electron attachment rate
and autodetachment rate. In its simplest embodiment, the
autodetachment lifetime, t, for attachment of electrons with
velocity v for an attachment cross-section s is given by

ðE

ð19Þ

where r(E*)

is the ratio of densities of states of the negative

ion that have a total energy E* to that of the density of states for

FIGURE 37.

Separation of neutral and negative ion currents of SF

. The neutral current results from

autodetachment of SF

* formed by low-energy electron attachment. Reproduced from R. N. Compton’s

PhD thesis, 1966 with permission from the author.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

266

Mass Spectrometry Reviews DOI 10.1002/mas

the initial electron molecule system, e

þ SF

. E* is the excess

vibrational energy of the negative ion (electron affinity plus
incident electron energy plus any initial vibrational energy in the
neutral before attachment). Effects of rotational motion are not
addressed in this treatment and all vibrations are considered to be
effective.

The density of states for the SF

* ion, r(E*)

, is calculated

by a modified version of the Marcus–Rice quasi-equilibrium
theory (QET) expression proposed by Whitten and Rabinovitch
(1963).

In this approximation, the sum of states for a molecule that

has s degrees of freedom, and excess energy of E is given by

ðEÞ ¼

ðE þ aE

ð20Þ

where a is an empirical factor between 0 and 1, and E

is the

vibrational zero point energy. The factor a is a function of E

: E/

and the dispersion in the molecular vibrational frequencies.

Differentiating the density of state Equation (20) with respect to
E results in the density of states:

ðEÞ ¼

ð1 þ da=dE

ÞðE þ aE

ðs 1Þ!P

ð21Þ

If one assumes that the initial molecules are in their ground

ro-vibrational states, then the density of states for e

þ SF

is just

that of a free particle and Equation (19) becomes

ðEÞ

ð22Þ

This simple relationship was used to calculate the electron

affinity of SF

, using the measured attachment rate and assuming

the vibrational frequencies for the ion to be the same as those of
the neutral. Under these assumptions, Equation (22) gives a value
of

1 eV for the electron affinity of SF

. The electron affinity for

is reported to be 1.05

0.1 eV (Grimsrud, Chowdhury, &

Kebarle, 1985). However, reliable calculations for the vibrational
frequencies for SF

ions are now available, and it is clear this

initial agreement is fortuitous.

Using the calculated vibrational frequencies for SF

with

symmetry gives lifetimes of

5 msec (see Fig. 38), in

contradiction to the lifetimes measured by the TOF methods. The
major remaining assumption is that the initial molecules are in
the ground ro-vibrational state. The lifetime calculation, using
Equation (22), is very sensitive to the electron affinity of SF

and

vibrational frequencies of the negative ion. To calculate the
lifetime, we used an electron affinity

¼ 1.05 eV (Grimsrud,

TABLE 1. Summary of reported SF

* Lifetime Measurements Using the TOFMS

Technique

FIGURE 38.

The autodetachment lifetime for SF

as a function of electron energy from Quasi-

equilibrium calculation.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

267

Chowdhury, & Kebarle, 1985). The vibrational frequencies used
are given by Gutsev and Bartlett (1998), using the MP2 method.
Absolute cross-sections for associative electron attachment
to SF

used here are those reported by Hotop and co-workers

(Braun et al., 2005). A more complete treatment of the lifetime
calculations, regarding different temperatures and various
negative ion vibrational frequencies, can be found in a recent
article by Cannon et al. (2007).

Until recently, the autodetachment lifetime for SF

appeared to be reasonably well-understood. However, recent
studies by Dunning and co-workers (Suess, Parthasarathy, &
Dunning, 2002; Liu, Suess, & Dunning, 2005), using a Penning
ion-trap, report lifetimes for SF

* ions formed by high Rydberg

electron transfer (RET) to be

1 msec to 10 msec. Although

one might attribute these long lifetimes to stabilization by the
Rydberg ion core, these authors also report autodetachment
lifetimes from free-electron attachment to be

10 msec. Thus,

there seems to be a great discrepancy in the reported autodetach-
ment lifetimes for SF

*. Some of these differences can be

ascribed to ions that contain different internal energies; however
the lifetimes greater than 10 msec cannot be reconciled unless
one evokes radiative stabilization or other effects. In attempting
to resolve these differences between the ion-trap and TOF
measurements, it should be kept in mind that ion trap experiments
contain magnetic fields of

0.3 Tesla, whereas the TOF

experiments are performed in zero or very low (

300 G)

magnetic field. Realizing that the autodetachment process occurs
by ejecting a very slow electron out of a highly vibrationally
excited anion (all of the internal energy goes into the out-going
electron) the magnetic field can manifest itself by shifting energy
levels, affecting the autodetachment process or inhibiting the
escape of electrons from the anion. It has been argued by Cannon
et al. (2007) that this is not the case.

Obviously, more studies are clearly needed in a number of

areas. Further scrutiny needs to be directed toward the TOF
experiments. For example, what processes could favor the neutral
signal over the ion signal as they travel down the flight-tube?
Also better control of the internal energy of the neutral before
attachment as well as higher electron energy resolution would
be desirable. A key experiment would be the measurement of
the decay of the negative ion in an ion trap with magnetic field-
free environment. Theoretical studies of the possible effects
of a magnetic field on autodetachment from SF

would be of

interest.

We will illustrate the utility of TOFMS through the detailed

study of low-energy electron attachment to the tetracyanoquino-
dimethane (TCNQ) molecule (Compton et al., 1977). Many of
the interesting properties of autodetachment and metastable
decay are inherent in this molecule. TCNQ represents the
‘‘negative half’’ of the tetrathiofulvalene tetracyanoquinodi-
methane (TTF-TCNQ) charge transfer salt, which has received
considerable attention in the solid-state community. A typical
TOFMS is shown in Figure 33 where a cylindrical lens positioned
down the flight tube can be biased positive or negative to separate
in time any neutral species that results from autodetachment of
the precursor or fragment ions. This lens can also separate any
ions created by metastable decay from its precursor ion. Also, if
metastable dissociation as well as dissociation of an ion occurs as
it travels down the flight tube, then the decay products, neutrals,

and fragment ions can be separated in time-of-flight and the
decay products can be identified. Of course, the neutral product
from the decay is unaffected by the potential barrier and the peak
due to the neutral will remain at the time of arrival of the ion peak
without the potential barrier. In essence, the second ‘‘Einzel
lens’’ acts as a second flight tube for mass analysis. Employment
of the Einzel-lens as an ion retardation (or acceleration) can be
considered a type of MS/MS. The Bendix Model 14-206 TOFMS
was equipped with two such lenses to allow for the study of the
further (secondary) decay of ions that had undergone primary
decay. Although it was never published, one of us (RNC) used
this to study the sequential metastable decay (i.e., the further
decay of a metastable decay product) of many tetrathiofulvalene
and tetraselenofulvalene positive ions, and others.

3. Metastable Ions: Dissociation and Autodetachment

Dissociation of metastable positive or negative ions can also be
effectively observed and studied with time-of-flight mass
spectroscopy. Precursor ions that decay into product ions and
neutral fragments have been examined by application of a second
flight tube that operated at a different potential than the initial
flight tube voltage in a manner identical to that used for the
autodetachment lifetimes discussed above. In this case, the
second flight region has a larger (for acceleration) or smaller (for
retardation) potential on the second flight tube as compared to the
original flight tube. The product ions will be separated in time
from the precursor ions, and appear at a shorter (or longer) time-
of-flight, respectively, than the undissociated precursor ions. The
neutral fragments will appear at the same time-of-flight as the
ions observed when no voltage is applied to the second flight tube.
Metastable ion dissociation studies from TOFMS were first
proposed by Wacks (1959) in a PhD Thesis in 1959, and later
employed by Hunt et al. (1964a,b). McLafferty, Gohlke, and
Golesworthy (1964) derived a convenient expression that related
the time of arrival of each fragment as a function of the voltages
on the first (V) and second (V’) flight tubes:

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

ðm

Þð1 V

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

ðm

Þð1 V

ð1

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

ð23Þ

where t denotes the TOF and m denotes the masses of the
subscript fragment (f), neutral (n) or precursor (p) ions. An
excellent example of autodetachment and metastable dissocia-
tion of a molecular anion is seen in the electron attachment to the
TCNQ molecule. TCNQ is known to attach free electrons from
0 to 3 eV to form precursor anions with lifetimes that vary
from

2 msec at 0 eV to 200 msec at 3 eV (Compton &

Cooper, 1977). Above

3 eV, the TCNQ anion undergoes

metastable dissociation. Figure 39 shows the three peaks (neutral,
fragment, and precursor) from the decomposition of the
metastable tetracyanoquinodimethan anion formed by the
unimolecular attachment of

3–4 eV electrons.

Using Equation (23), the product anion was identified as due

to (TCNQ–HCN)

. The neutral signal is due to HCN and TCNQ

because autodetachment and metastable dissociation both occur
at this electron energy.

The utility of studying electron attachment with time-of-

flight mass spectroscopy can best be illustrated by considering

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

268

Mass Spectrometry Reviews DOI 10.1002/mas

the case of electron attachment to a series of cyclic anhydrides
(Cooper & Compton, 1973). Maleic anhydride (250 msec),
phthalic anhydride (313 msec), and pyromellitic anhydride
(8 msec) all attach thermal electrons to form long-lived negative
ions with autodetachment lifetimes that decrease with electron
energy above

0 eV. The lifetime given here is that measured

at the peak in the ‘‘cross-section,’’ that is,

0 eV. The most

abundant fragment negative ions observed at higher electron
energy are the metastable RCO

* anions that result from the

loss of CO, where R represents C

, etc. Short-lived compound

negative ion states were also observed for succinnic anhydride
(1.3 eV), cyclobutane-dicarboxylic anhydride (1.3 eV), glutaric
anhydride (0.6 eV), and succinimide (1.1 eV) by the SF

Scavenger Technique described earlier. The energy cited in
parenthesis gives the peak observed for the compound state and
can be interpreted as the (negative) vertical electron affinity. In
addition, metastable CO

* anions are observed for all of the

molecules. We will use maleic and succinic anhydride to
illustrate the TOFMS results. Figure 40 shows the negative ion
yields as a function of electron energy for maleic anhydride. The
energy scale was calibrated with the attachment peak for SF

0 eV as well as the SF

Scavenger peak for Kr at 10.0 eV as

discussed earlier. Unimolecular electron attachment to maleic
anhydride produces a long-lived anion whose lifetime varies
from

300 msec at 0 eV to 100 msec at 0.4 eV. Metastable

* anions were observed to peak at 3.5 eV. The C

ion that peaked at 2.3 eV is seen to decay into C

þ CO

þ e.

Similar results are seen for succinic anhydride (C

Figure 41 shows the time-of-arrival distributions for the
metastable ions, C

*, and their decay products as well as

the autodetachment of CO

. The broad distribution for the

þ CO

neutrals shows that there is kinetic energy release in

the decay as well. This energy was determined for all of the
decays studied. Autodetachment lifetimes for CO

* have been

reported earlier (Cooper et al., 1972; Cooper et al., 1973;
Compton, Reinhardt, & Cooper, 1975) under other conditions,
and the lifetimes are reported to vary from

90 msec (Compton,

Reinhardt, & Cooper, 1975) to

50 msec (Cooper & Compton,

1972; Cooper & Compton, 1973). The differences probably
represent the different internal energy states of the CO

* anions

formed. The long lifetimes have been attributed to poor overlap
between the bent and extended CO

* anion with the linear CO

neutral molecule.

The EITOF has provided most of the known autodetachment

lifetimes for negative ions. Tables 2 and 3 provide a summary of
these measurements for approximately 70 molecules.

4. Proposed Experiment to Measure Lifetimes of
Negative Ions with Short Lifetimes (

< 10

sec)

In the previous section, we considered the measurement of the
lifetime of long-lived (t > msec) negative ions, which decay as
they travel down the flight path of a TOFMS. We now consider
the possibilities of determining shorter lifetime negative ions;
that is, lifetimes of autodetaching negative ions, which decay
in the acceleration field. Ions that autodetach in the first
(acceleration) region of the TOFMS will result in a fast-neutral
species, which could have sufficient energy to produce signal
counts at the detector. The final energy of the neutral will depend

FIGURE 39.

Time-of-arrival distribution of TCNQ

* together with the peaks that correspond to a neutral

resulted from autodetachment plus metastable decomposition products, and the ion products that resulted
from metastable decomposition TCNQ

into HCN

þ C

. Reproduced from the work of Compton

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

269

upon how long the negative ion exists in the field. Assuming that
such neutrals could be detected (energy 0 kilovolt), the lifetime
could be determined by measuring the energy gained in the
acceleration field. Longer-lived ions will produce faster neutral
species; that is, the energy gained in the field would be directly
related to the time spent in the field. In our considerations below,
we will neglect the time of acceleration of the ions because
their time can be made small in comparison to the total
flight time. We need only consider the acceleration region, as
shown in Figure 42.

The electron would arrive immediately at the detector

followed by the fast neutral M

. As will be shown below,

the time-of-arrival distribution for the neutrals can be related to
the auto-detachment lifetime. This method could be used to
determine lifetimes in the range from

to 10

sec. The

major equipment requirement aside from the conventional TOF
mass spectrum in this experiment is a pulse generator to provide a
rise time much less than the ion lifetime. Sub-nanosecond voltage
pulse generators are presently available for this purpose.
Metastable negative ions would be produced by a pulsed electron
source followed by a fast-pulsed acceleration of the ions. A
microsecond pulse of electrons would produce a statistical
distribution of sub-microsecond lifetime ions. Those ions that
survive the time delay just before the applied voltage pulse would
be accelerated up to an energy determined by their lifetime in the
electric field.

Consider a single ion of mass m accelerated by an electric

field V/d to a velocity v that travels a distance L to a detector. The
lifetime of the negative ion (t) is related to the flight time (T) of the
neutral m

through the relation:

mLd

ð24Þ

FIGURE 40.

Negative ion production cross-sections versus electron energy for maleic anhydride. The

energy scale was calibrated with the krypton energy-loss peak (SF

Scavenger Technique) at 10.0 eV. An

effective electron-energy distribution of 0.2 eV was obtained with the RPD method. Reproduced from the
work of Cooper and Compton (1973) with permission from American Institute of Physics, Copyright 1973.

FIGURE 41.

Time-of-arrival distributions for N

and N

for CO

* and

*. The top curves show autodetachment of electron from for

*, and the bottom show dissociation of metastable C

* into

þ CO

þ e. Reproduced from the work of Cooper and Compton

(1973) with permission from American Institute of Physics, Copyright
1973.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

270

Mass Spectrometry Reviews DOI 10.1002/mas

Let there be N

ions present in the source at t

¼ 0 so that an

exponential decrease in concentration would result in N

(t)

ions

present at time t given by:

ðtÞ

¼ N

t=t

ð25Þ

where t is the mean lifetime.

Because there is a distribution of negative ion lifetimes,

there will be a corresponding distribution of flight times.
Substituting Equation (24) into Equation (25) gives the number
of ions that arrive at the detector as a function of the flight time:

ðtÞ

¼ N

ðmLd=eVTtÞ

ð26Þ

TABLE 2. Lifetime of Complex Negative Ions Produced Via Low-Energy Electron Attachment (N is the Number of
Vibrational Degrees of Freedom)

formed in bimolecular collisions with cesium or potassium atoms.

produced by dissociative electron attachment to 1,2-cyclobutane dicarboxylic anhydride.

produced by dissociative electron attachment in succinic and maleic anhydrides.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

271

TABLE 3. Energy Dependence of the Autodetachment Lifetime of Complex Negative Ions Produced Via Low-Energy Electron
Attachment (N is the Number of Vibrational Degrees of Freedom)

Ions produced by dissociative electron attachment from succinic anhydride.

Ions produced by electron attachment to maleic anhydride.

(Continued )

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

272

Mass Spectrometry Reviews DOI 10.1002/mas

Experimentally, one would measure the number of ions that

arrive at time T within an interval DT. The number of ions with
lifetime t to t

þ Dt is found from Equation (25)

ðtÞ

t=t

ð27Þ

because T

¼ ðmdL=eVTÞ we have

mdL

eVT

ð28Þ

Therefore, the number of ions that arrive at the detector and

have a flight time T within an interval DT is

ðtÞ

mdL

eVT

ðmLd=eVTtÞ

ð29Þ

The functional form of this distribution, DN

(t)

, is shown in

Figures 43 and 44 for a variety of experimental conditions and
assumed autodetachment lifetimes. An estimate of the mean
lifetime can be calculated from the peak of this distribution, T

max

mLd

2 eV

max

ð30Þ

Inspection of Figures 43 and 44 verifies this estimate of

autodetachment lifetimes. This method could determine life-
times in the nanosecond range. Measurements for lifetimes less
than this value would probably prove difficult because of the
small number of ions present at onset of the acceleration voltage.
There is no doubt that many small molecules (five or six atoms)
such as CH

form negative ions with lifetimes where this

method could be more easily employed. Of course, this method
could be applied in general to any metastable negative ion,
including ions with microsecond or millisecond lifetimes in
which the decay occurs in the acceleration field and the neutral

can be detected by the electron multiplier or where re-ionization
and detection is possible. The proposed method of observing
metastable decay of anions in an electric field is analogous to
studies of metastable dissociation (the so-called m* peaks) of
ions that decay in the electric or magnetic fields of sector mass
magnetic spectrometers.

5. Short-Lived Negative Ions in Mass Spectrometry

Following the above treatment of temporary negative ions, it is
important to mention mechanism of negative ion fragmentation
under high-pressure conditions. Molecules, which attach slow-
electron to form short-lived negative ion states, can often be
stabilized by collision with a third body. For example, in
atmospheric pressure non-thermal discharge in air, negative ions
are created in which three-body attachment plays a key role.
There are many other examples. In 1935, Bloch and Bradbury
(1935) presented a three-body attachment process for molecular
oxygen, which has become known as the Bloch–Bradbury
mechanism. In this scheme, slow-electrons attach to molecular
oxygen with a rate coefficient of k

producing a short-lived

vibrationally excited (O

)* ion with a lifetime t, that is,

þ O

ðX

4Þ

The anions are formed with n

4 as indicated. These ions,

once formed can be stabilized by a third-body with a rate constant
k

, or detached with a rate k

, that is,

ðX

4Þ þ M !

ðX

< 4

Þ þ M

ðX

¼ 4Þ þ M !

ðX

Þ þ M

þ e

TABLE 3. (Continued )

Ions produced by dissociative electron attachment from maleic anhydride.

Ions produced by dissociative electron attachment from glutaric anhydride.

Ions produced by dissociative electron attachment from 1,2-cyclobutane dicarboxylic anhydrides.

Ions produced by electron attachment to phthalic anhydride.

Ions produced by electron attachment to pyromellitic anhydride.

Ions produced by electron attachment to maleimide.

Ions produced by dissociative electron attachment from tetrafluorosuccinic anhydride.

Ions produced by dissociative electron attachment from hexafluoroglutaric anhydride.

Ions produced by electron attachment to fluoranil.

Ions produced by dissociative electron attachment from fluoranil.

Ions produced by electron attachment to chloranil.

Ions produced by electron attachment to bromanil.

FIGURE 42.

Illustration of proposed method to determine negative ion lifetimes for short-lived negative

ions. The metastable negative ion m

* is assumed to decay at the point of the arrow

! in the acceleration

field.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

273

where M* is internally or translationally excited. Once the O

contains less than 4 quanta of vibration energy, the ions become
stable. Under steady state conditions, the overall rate coefficient,
k (M), for forming O

can be written as:

ðMÞ ¼

ðk

þ k

Þn

þ ð1=tÞ

=sec

where n

is the concentration of the third body M. Under

conditions in which (k

þ k

)

1/t then k ¼ k

Some typical rate coefficients are k (O

)

¼ 2.5 10

sec and k (Ar)

¼ 3 10

/sec. The Bloch–Bradbury

mechanism was revisited by Herzenberg in 1969 (Herzenberg,
1969) in light of more recent experiments and collision theory. In
Herzenberg’s treatment if one supposes that all collisions are
stabilizing and an upper bound of the order of 10

/sec is

obtained. This is approximately that obtained for k(O

Collisional stabilization at high-pressure occurs under many

condition of atmospheric mass spectroscopy. Short-lived neg-
ative ions play an equally important role in experiments in which
high-energy electrons are injected into a gas at moderate to high
pressure. Most molecules capture electrons at low energy and
therefore it is important to somehow reduce the high-energy
electrons to thermal energy. Collisions with atoms or molecules

FIGURE 43.

Distribution of the number of negative ions v.s. time-of-flight. (m

¼ 20 amu, d ¼ 0.5 cm,

¼ 200 cm, eV ¼ 1,000 eV, t ¼ 10

sec, and DT

¼ 10

sec.)

FIGURE 44.

Plot of calculated fast-neutral molecule signal (Neutral

that resulted from autodetachment of metastable negative ions with mean
lifetime t) for ions with lifetime t

¼ 1 msec (filled diamond), t ¼ 10 nsec

(filled square), and t

¼ 1 nsec (filled triangle) that arrived with

interval times of DT

¼ 0.2 msec, DT ¼ 20 msec, and DT ¼ 200 msec,

respectively. (m

¼ 20 amu, d ¼ 0.5 cm, L ¼ 20 cm, eV ¼ 2,000 eV.)

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

274

Mass Spectrometry Reviews DOI 10.1002/mas

can be degraded in energy via electronic, vibrational, and
rotational energy loss collisions. In addition, collisional ioniza-
tion results primarily in a slow and a fast electron. The energy loss
collisions are most effective in which a temporary negative ion is
formed as discussed in reference to the SF

Scavenger Technique

(see Low-Energy Electron-Attachment Studies That Employ
a Time-of-Flight Mass Spectrometer Section). Thus in experi-
ments where high-energy electrons are initially present collisions
in which a temporary negative ion can be formed (e.g., N

, CO

benzene, etc.), thermal electrons can be produced with large
cross-sections. In a sense, one is performing an SF

Scavenger

experiment in which the molecules of interest replace SF

In the next section, we consider a method, which allows for

stabilization of short-lived anions as result of a near-by positive
ion, namely Rydberg electron attachment. In this case collisions
of atoms in highly excited Rydberg states with O

can result in

stable O

via stabilization with the Rydberg core. This would

correspond to a type of Bloch–Bradbury attachment in which the
ion core in the ‘‘third-body.’’

C. Rydberg Electron Transfer (RET) TOFMS

As discussed previously, the cross-section for the attachment of
slow free electrons to molecules often exhibit a 1/v velocity (or
alternately a 1=

ﬃﬃﬃﬃ

energy) dependence at low energy. Such

reactions can also be studied by collisions between a highly
excited Rydberg atom and an electron-accepting molecule.
Fermi (1934) was the first to point out that the reaction between
an atom in a high Rydberg state and a ground-state atom or a
molecule target can be approximated by that of the interaction of
a slow, quasi-free electron with the target, together with the
polarization induced upon the target by the Rydberg ion core.
Amaldi and Segre (1934) had previously reported shifts of the
high Rydberg energy levels (n

¼ 20–30) of alkali atoms at

densities such that about ten thousand rare gas atoms were
contained within the Rydberg orbit. The absorption lines were
shifted to higher or lower energies (depending upon the rare gas)
and showed little broadening. The observed shifts implied that
collisional ionization was small, and that the electron was
essentially elastically scattered by many atoms during one
classical orbit. Fermi correlated the magnitude and direction of
the shift to the known scattering lengths for slow free electrons in
the rare gases. This observation, along with the inclusion of the
effects of polarization of the Rydberg core on the nearby atoms,
led Fermi to propose that the Rydberg-neutral reaction can be
thought of as elastic scattering of a quasi-free electron with an
atom that has been polarized to some extent by the ion core. The
success and simplicity of this model prompted others (Flannery,
1970, 1973; Matsuzawa, 1971,1972a,b, 1974; Priest, 1972) to
apply the Fermi ‘‘free-electron model’’ to other Rydberg atom
reactions, beginning in the 1970s; a practice that continues today.
For more details on collisions between Rydberg atoms and
molecules, the reader is referred to the review articles by
Matsuzawa (1983) and Fabrikant (1996).

Since Fermi’s original contribution, there has been a

considerable body of theoretical (Matsuzawa, 1971,1972a,b,
1974) and experimental (Stockdale et al., 1974; West et al., 1976;
Foltz et al., 1977) evidence to suggest that the rate constant for the

transfer of a loosely bound Rydberg electron to an electron-
attaching molecule should be equal to the electron attachment
rate for free electrons at the equivalent energy. Perhaps the best
experimental evidence for this equivalence comes from the
unpublished Rydberg electron-transfer reaction rate data of W.A.
Chupka (private communication) and the electron attachment
data of Davis, Compton, and Nelson (1973) and Jarvis, Kennedy,
and Mayhew (2001) for the series of molecules SF

, SeF

, and

TeF

(Fig. 45). The electron attachment rate and the Rydberg

electron-transfer rate for this series show a linear decrease over
five orders of magnitude (see also Compton, 1980). Interestingly,
the electron affinity for this series decreases dramatically over
this same series. In Figure 45, we show these data along with data
for a number of other molecules. The straight-line fit in the figure
equates the Rydberg rate to the attachment rate, which we will
call the Fermi–Matsuzawa model. Not only does this model
accurately account for the rates of exchange and attachment, the
ions observed in the two cases are identical (e.g., SF

/SF

;

SeF

/SeF

; TeF

/TeF

; Cl

/CCl

, etc.).

Hotop and Niehaus (1967) were the first to study ion-pair

formation in Rydberg collisions with molecules. They reported a
cross-section of

for the reaction A**

þ SF

. Under these conditions, we know that the lifetime of

the negative ion is very (>1 msec) long-lived, whereas the SF

formed under slow ‘‘free’’ electron attachment has an auto-
detachment lifetime of

10 to 50 msec (see above). Thus,

interactions with the ion core can lead to vibrational relaxation
of the anion resulting in stable anion formation. Measurements
of the internal temperature of the anion following the charge
exchange collision have not been performed. The core certainly
plays a role in these collisions-at least for atoms excited to low
lying excited states. The strong Coulombic attraction between the
nascent positive and negative ions (especially at lower n) acts to
prevent the ions from separating, resulting in a rapid decrease in
the rate constants for ion production as n decreases (Zollars et al.,
1986; Zheng, Smith, & Dunning, 1988; Carman, Klots, &
Compton, 1989; Harth, Ruf, & Hotop, 1989). Using faster SF

molecules in a nozzle-jet expansion aids in the ion-ion separation
and partially restores the ion signal at lower principal quantum
numbers. In these experiments, the ion-pair production rate for
collision energies of 0.15 eV was noticeably larger than for 0.05
eV, illustrating the higher degree of ‘‘dissociation’’ for the greater
kinetic energy. The anion interaction with the positive-ion core
can also act to stabilize (vibrationally relax) short-lived anions
formed by Rydberg electron transfer to small molecules (e.g., O

, NO

, CH

, etc.). There are many instances for which

the lifetime of the anion formed under unimolecular electron
attachment is too short to be observed in a conventional mass
spectrometer (<microsecond); however, the anion formed under
Rydberg electron-transfer conditions is very long-lived or stable.
In order to further understand the effects of stabilization, let us
first discuss the physics of electron attachment to molecules.

As discussed earlier in Short-Lived Negative Ions in Mass

Spectrometry Section, three-body collisions can effectively
stabilize temporary negative ions formed by unimolecular
electron attachment by the removal of vibrational energy. A
vivid example of stabilization of a temporary negative ion formed
by high-Rydberg electron transfer comes from studies of HI

formation in reactions of alkali Rydberg atoms with hydrogen

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

275

iodide (Carman, Klots, & Compton, 1993). At high principal
quantum numbers, ‘‘dissociative attachment’’ reactions that lead
to I

dominate the reaction cross-section in qualitative agreement

with the free-electron attachment data of Alajajian and Chutjian
(1988) and the Fermi–Matsuzawa model (see also Klar et al.,
1994). For n < 13, non-dissociative reactions that lead to HI

are

prominent. Calculations of the potential energy surface for this
anion (Chapman, Balasubramanian, & Lin, 1988; Horacek
et al., 1997) show a broad, flat, and shallow minimum at large
internuclear separation (

5 a.u.). It is obvious that the presence

of the ion core near the ‘‘dissociating’’ HI

anion acts to stabilize

the anion into the very shallow potential well. The fact that ions
are observed demonstrates that at least one vibrational level
is supported by this potential; that is, HI

is stable. In this

connection, Tuinman and Compton (1996) also searched for
hydrogen halide negative ions produced by dissociative electron
attachment from a host of alkyl halides. Only HI

was observed

from 2-iodomethane.

Rydberg electron transfer time-of-flight mass spectroscopy

has played a major role in the study of dipole- and quadrupole-
bound negative ions. It is sometimes convenient to represent the
interaction of an electron with an atom or molecule in terms of its
static and dynamic electrostatic energies. If we consider only the
dipole and quadrupole terms, then the potential of interaction

between an electron and a molecule can be conveniently
expressed as

ðrÞ ¼

ðr; RÞ

ð31Þ

where m is the dipole moment and Q is the electric quadrupole
moment of the molecule. P

(r, R) and P

(r, R) are the first-

and second-order Legendre polynomials, respectively. The last
term represents the polarizability attraction and a represents the
polarizability of the molecule. It is now well-known (see reviews
by Desfrancois, Abdoul-Carmine, & Schermann, 1996 and
Compton & Hammer, 2001) that any molecule with a dipole
moment that exceeds

2.5 Debye will bind an excess electron

into a stable negative-ion state with electron-binding energies
between 0 and 20 meV (Hammer et al., 2003) for molecules with
dipole moments between 2.5 and 4.5 Debye, and as high as
50 meV for a molecule with a dipole moment of

5.4 Debye

(Hammer et al., 2004). Almost all of these studies employed high
Rydberg states excited by a laser to specific n and l states to
produce dipole-bound negative ions in the source region of a TOF
mass spectrometer. The apparatus employed by Hammer et al.
(Hammer et al., 2003, 2004) is shown below. Dipole-bound
anions formed by charge transfer from a Rydberg atom exhibit a

FIGURE 45.

Correlation of the free-electron attachment rate with the Rydberg charge-transfer rate for a

series of molecules. Plotted using the data from the work of Davis et al. (1973), and Jarvis, Kennedy, and
Mayhew (2001).

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

276

Mass Spectrometry Reviews DOI 10.1002/mas

marked dependence on the effective quantum number, n*, of the
electron in the Rydberg atom. The diffuse Rydberg electron
gently ‘‘moves’’ over to the diffuse orbital of the dipole-bound
anions, loosely stated as a resonance charge transfer. Desfrancois
et al. (see again the review by Desfrancois, Abdoul-Carmine, &
Schermann, 1996) provide a convenient semi-empirical expres-
sion that relates the maximum ion yield versus n* to the electron
affinity, EA as EA

¼ 23/n*

2.8

. The value for n* is taken for the n*

that corresponds to the maximum ion signal or n*

max

. In addition,

the electron affinity (EA) can be deduced from field detachment
of the weakly bound electron. The Desfrancois group has
employed a retardation (or an acceleration) grid down the flight
tube bounded by two grids at the flight tube potential. The
Hammer/Compton group employed field detachment in the ion-
acceleration region of the ion source (first two grids of the
apparatus shown in Fig. 46).

In this case, the electric field was very precisely and

accurately determined by using the field detachment of highly
excited alkali atom Rydberg states. Both of these methods,
measurement of n*

max

and field detachment, have provided

values for electron affinities with high precision. These precise
values have been calibrated by the more accurate, but less
precise, photodetachment values for a few high electron-affinity
dipole-bound anions in collaboration with the Bowen group
(see Hammer et al., 2004). A good example of these two methods
can be seen in a study of the isotope effects on the dipole-bound
anions of acetone and deuterated acetone (Hammer et al., 2005).
Figure 47 shows the n* dependence and field-detachment
thresholds for the case of acetone and deuterated acetone. The
data in Figure 47 show that the electron affinity of acetone and
deuterated acetone are

2.5 meV and that the difference in EA is

55 10 meV, the deuterated acetone EA being smaller. The
lower affinity was attributed theoretically to a slight reduction
(0.5%) of the average dipole moment upon deuteration.

D. Electron Collisions with Clusters that Lead to
Positive and Negative Ions

The investigations of gas-phase molecular clusters began with
studies of ionic clustering in high-pressure mass spectrometer ion
sources. In these studies, precursor ions undergo three-body
collisions with molecules to produce molecular cluster ions. In
recent decades, the development of the supersonic nozzle
expansion has allowed researchers to directly prepare atomic
and molecular clusters of ever increasing sizes. The development
of supersonic nozzle-jet expansion methods has led to the
production of molecules and van der Waals clusters with low
ro-vibrational states of internal energies. These low-temperature
expansion techniques also allow the preparation of clusters in
reasonably well-defined (lowest energy) conformations. Atoms
and molecules have also been ‘‘injected’’ into expanding nozzle-
jets to great benefit. For example, it can be said that the ‘‘nano-
material revolution’’ presently underway in science began with
the Smalley–Curl–Kroto studies of laser ablation of graphite
into a rare-gas nozzle-jet to produce the fullerene molecules; that
is, C

, C

, etc. (Kroto et al., 1985). In these experiments,

laser ablation of graphite in a nozzle-jet expansion produced
carbon clusters of ever-increasing sizes. Proper adjustment of the
experimental conditions produced magic-number clusters at C

, and other larger carbon cage isolated-pentagon fullerenes.

Textbooks that describe fullerenes abound (e.g., Billups &
Ciufolini, 1993; Kroto, Fischer, & Cox, 1993; Kadish & Ruoff,
1994; Dresselhaus, Dresselhaus, & Eklund, 1995). There are also
books (e.g., Gonzalez-Moraga, 1993) and review articles on
molecular clusters per se. For example, Castleman and Bowen
(1996) have provided a comprehensive review of the field of
clusters and intermediate structures of matter. That 1996 review
contains references to over 690 articles in this field. Time-of-
flight mass spectroscopy has played a major role in the analysis of

FIGURE 46.

Rydberg charge-exchange apparatus employed in the work of Hammer et al. (2003, 2004)

used to produce dipole-bound anions from nozzle-jet expanded polar molecules. Reproduced from N. I.
Hammer’s PhD thesis, 2003 with permission from the author.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

277

cluster masses in a large majority of these studies. In keeping with
the body of this review, we will discuss one of the authors (RNC)
endeavors in the area of electron collisions with molecular
clusters formed via nozzle-jet expansions. In a series of four
articles, Klots and Compton described studies of electron
collisions with clusters of CO

(Klots & Compton, 1977,

1978a), N

O (Klots & Compton, 1978a), H

O (Klots &

Compton, 1978b), and CH

I (Klots & Compton, 1980). In these

studies, a nozzle-jet source was attached to a modified

commercial time-of-flight mass spectrometer (Bendix Model
14-206). Figure 48 shows the nozzle-jet attachment with the
differential pumping stage and skimmer entrance to the TOF
source region. The nozzle-skimmer distance could be continu-
ously varied to produce optimum results. The diameter of the
nozzle was 25 mm. The neutral jet-expanded molecules and
molecular clusters travel a few centimeters into the source region
of the TOFMS where it crosses an electron beam. The nozzle-jet
is injected between the pusher plate and first grid of the TOFMS.

FIGURE 47.

Upper two figures show the acetone and deuterated acetone dipole-bound anion signal as a

function of rubidium ns and nd Rydberg states. The lower figure shows the ion signal as a function of the ion
draw-out electric field. Notice that both ions are field detached at

700 V/cm. Reproduced from the work of

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

278

Mass Spectrometry Reviews DOI 10.1002/mas

Positive or negative ions are pushed down the flight tube and
detected with an electron multiplier. The initial molecular beam
travels at right angle to the flight path, and, therefore, the velocity
component along the flight path is small, whereas the transverse
velocity is increased over the normal thermal distribution. Thus,
it is necessary to adjust the steering voltage on the ion deflector
field order to correct for this transverse velocity. The small
velocity component along the flight path allows for greater mass
resolution in this arrangement. The nozzle-jet introduction into

the ion source could be considered as a variation of the off-axis
(oa TOF) ion injection method described earlier; the difference is
that the neutrals are injected off-axis. One example of these
studies (Klots & Compton, 1978a) is shown in Figure 49 in which
electron attachment to carbon dioxide clusters are shown as a
function of electron energy. The RPD technique was employed to
obtain electron-energy resolution of

0.15 eV. At low energy, the

dimer anion of CO

is observed to peak at

3 eV produced from

electron attachment to CO

clusters. Unlike CO

anions, the

FIGURE 48.

Time-of-flight mass spectrometer ion source, modified to incorporate gas nozzle and pumping

system. The TOF path is into the page. Reproduced from the work of Klots and Compton (1978a) with
permission from American Institute of Physics, Copyright 1978.

FIGURE 49.

Primary negative ions from carbon dioxide as a function of electron energy. Higher clusters

ions of the type (CO

)

and CO

(CO

)

are discussed in the text. Reproduced from the work of Klots and

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

279

dimer was observed to be stable (i.e., no autodetachment) on
the time scale of up to 2 msec. The dimer (CO

)

anion as well as

other CO

cluster anions of the type (CO

)

with n

¼ 3 to 6

served to illustrate the evaporative electron attachment process;
that is,

þ ðCO

! ðCO

þ CO

The onset of O

from ultra-cold (nozzle-jet expanded) CO

molecules occurs at the thermodynamic threshold of 3.98 eV.
Previous studies for CO

at room temperature shows a lower

onset (

0.3 eV) for O

/CO

due to ‘‘hot bands’’ (Schultz &

Spence, 1969). Many other dissociative attachment studies at
room temperature have been plagued by ‘‘hot band’’ effects.
These studies clearly demonstrate the utility of studying electron
attachment to nozzle-jet expanded molecular beams.

Ions of the type O

(CO

)

with n

¼ 1–6 are also observed.

One such ion, O

(CO

) is shown in Figure 49. These ions derive

from a process that can be described as ion–molecule half-
reactions; that is,

þ ðCO

! O

ðCO

þ CO

ðCO

! CO

ðCO

followed by evaporation; that is,

ðCO

! CO

ðCO

þ CO

This brief review illustrates the great utility of studying

electron collision with nozzle-jet expanded molecules by using
electron collisional ionization time-of-flight mass spectroscopy.

VI. SUMMARY

Studies of electron collisions with atoms and molecules have
occupied a central role in modern science beginning with the
discovery of the electron by J.J. Thompson and the experiments
of Franck and Hertz, which ushered in the quantum era in physics
and chemistry. In the 1970s Massey and his colleagues Burhop
and Gilbody completed a four volume series of books sum-
marizing the field of ‘‘Electronic and Ionic Impact Phenomena’’
(Massey, Burhop, & Gilbody, 1969a, Vol. I; Massey, Burhop, &
Gilbody, 1969b, Vol. II; Massey, Burhop, & Gilbody, 1971, Vol.
III; Massey, Burhop, & Gilbody, 1974, Vol. IV) at that time.
Electron ionization was a growing and vibrant field prior to the
1980s, especially so in the United States. With the advent of the
tunable laser and other emerging technologies, investigations
of electron ionization have waned somewhat in the US in
recent years. It is our observation that a greater concentration on
electron ionization research is now seen in European laborato-
ries. This review has attempted to emphasize the advantages
of time-of-flight-mass spectrometry for the study of electron –
molecule collisions. The field-free ionization region allows for
high precision in the electron energy and as well as high electron
energy resolution. Application of electric fields down the ion
flight tube allows for the study of autodetachment and metastable
decomposition as well as field detachment of weakly bound
negative ions (e.g., multi-pole bound anions). It is worth
emphasizing that the SF

Scavenger Technique discussed in

Low-Energy Electron-Attachment Studies That Employ a Time-
of-Flight Mass Spectrometer Section, when monitoring the neutral
SF

molecules resulting from autodetachment SF

* ions can be

considered a true zero kinetic energy (ZEKE) detection scheme.
Using this method along with nozzle-jet expanded molecules
would allow for a new era of electron–molecule collision studies.

Finally, a historical review of some of the early types of

TOFMS reveals a number of machines that should be re-
considered in modern times. In particular, the exceptional mass
resolution exhibited by some of the magnetic TOF mass
spectrometers presented could provide a new type of ‘‘ion-
storage’’ mass spectrometer. We encourage further consideration
of these simple, yet elegant, designs.

ACKNOWLEDGMENTS

This work was supported by the National Science Foundation.
We especially appreciate the many suggestions/corrections of the
editor and five referees.

REFERENCES

Alajajian SH, Chutjian A. 1988. Measurement of electron-attachment line

shapes, cross sections, and rate constants in HI and DI at ultralow
electron energies. Phys Rev A 37:3680–3684.

Amaldi E, Segre E. 1934. Effect of pressure on the high terms of the alkalies.

Nuovo Cimento 11:145–156.

Anderson N, Eggleton PP. 1967. A theoretical analysis of the retarding

potential difference technique. Int Electron 22:497–511.

Anderson N, Eggleton PP, Keesing RGW. 1967. Magnetic effects on the

transmission of electrons through an aperture and their application to the
retarding-potential-difference technique. Rev Sci Instrum 38:924–927.

Applehans AD, Delmore JE. 1988. Refinement of the autoneutralization

lifetimes of short lived states of SF

. J Chem Phys 88:5561–5570.

Barr WE, Perkins WA. 1966. E X B energy analyzer for electrons. Rev Sci

Instrum 37:1354–1359.

Begun GM, Compton RN. 1973. Electron impact ionization studies of

ferrocene, cobaltocene, nickelocene, and magnesocene. J Chem Phys
58:2271–2280.

Billups WE, Ciufolini MA. 1993. Buckminsterfullerenes. Weinheim, Germany:

Wiley-VCH.

Blatt J, Weisskopf VF. 1952. Theoretical nuclear physics.

Bleakney W, Hipple JA, Jr. 1938. A new mass spectrometer with improved

focusing properties. Phys Rev 53:521–529.

Bloch F, Bradbury NE. 1935. On the mechanism of unimolecular electron

capture. Phys Rev 48:689–695.

Boemmels J, Franz K, Hoffmann TH, Gopalan A, Zatsarinny O, Bartschat K,

Ruf M-W, Hotop H. 2005. Low-lying resonances in electron-neon
scattering: Measurements at 4-meV resolution and comparison with
theory. Phys Rev A 71:012704/1–012704/12.

Boldyrev AI, Gutowski M, Simons J. 1996. Small multiply charged anions as

building blocks in chemistry. Acc Chem Res 29:497–502.

Boumsellek S, Chutjian A. 1992a. Increased response of the reversal electron

attachment detector and modelling of ion space-charge effects. Anal
Chem 64:2096–2100.

Boumsellek S, Alajajian SH, Chutjian A. 1992b. Negative-ion formation in

the explosives RDX, PETN, and TNT by using the reversal electron
attachment detection technique. J Am Soc Mass Spectrom 3:243–
247.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

280

Mass Spectrometry Reviews DOI 10.1002/mas

Braun M, Barsotti S, Marienfeld S, Leber E, Weber JM, Ruf M-W, Hotop H.

2005. High resolution study of anion formation in low-energy electron
attachment of SF6 molecules in a seeded supersonic beam. Eur Phys J D
35:177–191.

Brown RS, Lennon JJ. 1995. Mass resolution improvement by incorporation

of pulsed ion extraction in a matrix-assisted laser desorption/ionization
linear time-of-flight mass spectrometer. Anal Chem 67:1998–2003.

Cameron AE, Eggers DF, Jr. 1948. An ion‘‘velocitron’’. Rev Sci Instrum

19:605–607.

Cannon M, Liu Y, Suess L, Dunning FB, Steill J, Compton RN. 2007.

Temperature dependence of negative ion lifetimes. J Chem Phys
127:064314 (10 pages).

Carman HS, Jr., Klots CE, Compton RN. 1989. Observation of l-dependent

rate constants for ion production in Rydberg electron transfer reactions.
J Chem Phys 90:2580–2584.

Carman HS, Jr., Klots CE, Compton RN. 1993. Rydberg electron transfer to

hydrogen iodide: Dissociative and nondissociative electron capture. J
Chem Phys 99:1734–1743.

Castleman AW, Bowen K. 1996. Clusters: Structure, energetics, and

dynamics of intermediate states of matter. J Phys Chem 100:12911–
12944.

Chaney EL, Christophorou LG, Collins PM, Carter JG. 1970. Electron

attachment in the field of the ground and excited electronic states of the
azulene molecule. J Chem Phys 52:4413–4417.

Chapman DA, Balasubramanian K, Lin SH. 1988. Theoretical study of the

negative ions of HBr and HI. Phys Rev A 38:6098–6106.

Chatham H, Hils D, Robertson R, Gallagher AC. 1984. Total and partial

electron collisional ionization cross sections for CH

, C

, SiH

, and

. J Chem Phys 81:1770–1777.

Chen CL, Chantry PJ. 1979. Photon-enhanced dissociative electron attach-

ment in SF

and its isotopic selectivity. J Chem Phys 71:3897–3907.

Chernushevich IV, Dodonov AF, Dodonova TF. 1987. Inventors; USSR Chem

Phys Energy Problems Inst., Assignee. Soviet patent No. SU 1681340
A1, Mass spectrometric analysis according to transit time of continuous
ion beam-involves feeding ions into modulator in direction perpendic-
ular to axis of drift tube.

Christophorou LG, Carter JG, Christodoulides AA. 1969. Long-lived parent

negative ions p-benzoquinone formed by electron capture in the field of
the ground and excited states. Chem Phys Lett 3:237–240.

Collins PM, Christophorou LG, Chaney EL, Carter JG. 1970. Energy

dependence of the electron attachment cross section and the transient
negative ion lifetime for p-benzoquinone and 1, 4-naphthoquinone.
Chem Phys Lett 4:646–650.

Compton RN. 1966. Electron capture cross section and negative ion lifetimes.

PhD Thesis. The University of Tennessee.

Compton RN, Christophorou LG, Hurst GS, Reinhardt PW. 1996a.

Nondissociative electron capture in complex molecules and negative-
ion lifetimes. J Chem Phys 45:4634–4639.

Compton RN, Christophorou LG, Huebner . 1996b. Temporary negative ion

resonances in benzene and benzene derivatives. Phys Lett 23:656–658.

Compton RN, Heubner RH, Reinhardt PW, Christophorou LG. 1968.

Threshold electron impact excitation of atoms and molecules: Detection
of triplet and temporary negative ion states. J Chem Phys 48:901–909.

Compton RN, Huebner RH. 1969. Temporary attachment of electrons to

azulene-h

and azulene-d

. J Chem Phys 51:3132–3133.

Compton RN, Reinhardt PW, Cooper CD. 1975. Collisional ionization of Na,

K, and Cs by CO

, COS, and C S

: Molecular electron affinities. J Chem

Phys 63:3821–3827.

Compton RN, Reinhardt PW, Garrett WR. 1976. Uranium hexaflouride

negative ions from surface reactions involving uranium and flourine. J
Chem Phys 66:4712–4713.

Compton RN. 1977. On the formation of positive and negative ions in gaseous

. J Chem Phys 66:4478–4485.

Compton RN, Cooper CD. 1977. Negative ion properties of tetracyanoqui-

nodimethan: Electron affinity and compound states. J Chem Phys
66:4325–4329.

Compton RN. 1980. Electron attachment to molecules In: Oda N,

Takakayanagi K, editors. Electronic and ionic collisions. North
Holland. Publishing Company, Amsterdam.

Compton RN, Bardsley JN. 1984. Dissociation of molecules by slow

electrons. In: Shimamura Isao, Takayanagi Kazuo, editors. Electron-
molecule collisions. Plenum Press.

Compton RN, Tuinman AA, Klots CE, Pederson MR, Patton DC. 1997.

Electron attachment to a negative ion: e

þ C

$ C

. Phys Rev Lett

78:4367–4370.

Comptom RN, Hammer NI. 2001. Multipole-bound molecular anions. In:

Adams N, Babcock L, editors. Advances in gas phase ion chemistry,
volume 4. Amsterdam: Elsevier Science. p. 257–305.

Cooper CD, Williamson AD, Miller JC, Compton RN. 1980. Resonantly

enhanced multiphoton ionization of pyrrole, n-methyl Pyrrole and
furan. J Chem Phys 73:1527–1537.

Cooper CD, Compton RN. 1973. Electron attachment to cyclic anhydrides

and related compounds. J Chem Phys 59:3550–3565.

Cooper CD, Compton RN. 1972. Metastable anions of CO

. Chem Phys Lett

14:29–32.

Cooper CD, Naff WT, Compton RN. 1975. Negative ion properties of

p-benzoquinone: Electron affinity and compound states. J Chem Phys
63:2752–2757.

Cooper CD, Frey WF, Compton RN. 1978. Negative ion properties of fluoranil,

and bromanil: Electron affinities. J Chem Phys 69:2367–2374.

Cooper CD, Compton RN. 1974. Electron attachment and cesium collisional

ionization studies of tetrafluorosuccinic and hexafluoroglutaric anhy-
drides: Molecular electron affinities. J Chem Phys 60:2424–2429.

Cotter RJ. 1994. Time-of-flight mass spectrometry. Washington: American

Chemical Society. 233 p.

Curran RK. 1963. Formation of SF

by inelastically scattered electrons. J

Chem Phys 38:780–781.

Davis FJ, Compton RN, Nelson DR. 1973. Thermal energy electron

attachment rate constants for some polyatomic molecules. J Chem
Phys 59:2324–2329.

Delmore JE, Applehans AD. 1986. Measured variation in the autoneutraliza-

tion lifetime of SF

. J Chem Phys 84:6238–6246.

Desfrancois C, Abdoul-Carmine H, Schermann JP. 1996. Ground-state

dipole-bound anions. Int J Mod Phys B 10:1339–1395.

Deutsch H, Ma¨rk TD. 1987. Cluster ions: Production, detection and stability.

Int J Mass Spectrom Ion Processes 79:1–59.

Deutsch H, Becker K, Matt S, Ma¨rk TD. 2000. Theoritical determination of

absolute electron-impact ionization cross-sections of molecules. Int J
Mass Spectrom 197:37–69.

Dodonov AF, Chernushevich IV, Dodonova TF, Raznikov VV, Talroze VL.

1989. Inventors USSR Chem Phys Energy Problems Inst. Assignee.
Soviet patent No. WO, 91/03071 (PCT/SU89/00228), Method and
device for continuous-wave ion beam time-of-flight mass spectrometric
analysis.

Dodonov AF, Chernushevich IV, Laiko VV. 1991. Extended abstract from the

proceedings of the 12th international mass spectrometry conference,
Amsterdam, The Netherlands.

Dresselhaus MS, Dresselhaus G, Eklund PC. 1995. Science of fullerenes and

carbon nanotubes. Academic Press, San Diego, CA.

Duric N, Cadez I, Kurepa M. 1991. Electron impact total ionization cross-

section for methane, ethane and propane. Int J Mass Spectrom Ion
Processes 108:R1–R10.

Edelson D, Griffiths JE, McAfee KB. 1962. Autodetachment of electrons in

sulfur hexafluoride. J Chem Phys 37:917–918.

Eland JHD. 1993. Second-order space focusing in two-field time-of-flight

mass spectrometers. Meas Sci Technol 4:1522–1524.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

281

Elhamidi O, Pommier J, Abouaf R. 1997. Low-energy electron attachment to

fullerenes C

and C

in the gas phase. J Phys B 30:4633–4642.

Elhamidi O, Pommier J, Abouaf R. 2001. Low-energy electron impact on C

and C

: Excitation, metastable anion formation, and lifetime. Int J

Mass Spectrom 205:17–25.

Even U, Dick B. 2000a. Computer optimization for high-resolution time-of-

flight mass spectrometer. Rev Sci Instrum 71:4415–4420.

Even U, Dick B. 2000b. Optimization of a one-dimensional Time-of Flight-

Mass Spectrometer. Rev Sci Instrum 71:4421–4430.

Fabrikant II. 1996. Electron scattering by neutral tergets at milli- and

submilli-electron-volt energies. At Mol Phys 32:267–279.

Fermi E. 1934. The displacement by pressure of the high terms of spectral

series. Nuovo Cimento 11:157–166.

Fiegele T, Mason N, Foltin V, Lukac P, Stamatovic A, Scheier P, Mark TD.

2001. The energies of the triply excited n

¼ 2 intrashell He

resonances

2p and 2s 2p

revisited. Int J Mass spectrom 209:23–29.

Flannery MR. 1970. Semiquantal theory of heavy-particle excitation,

deexcitation, and ionization by neutral atoms. I. Slow and intermediate
energy collisions. Ann Phys 61:465–487.

Flannery MR. 1973. The semiquantal theory of heavy particle inelastic

collisions with neutral atoms and molecules. II. Thermal, intermediate
and high-energy collisions. Ann Phys 79:480–517.

Foltz GW, Latimer GJ, Hilterbrant GF, Kellert FG, Smith KA, West WP,

Dunning FB, Stebbings RF. 1977. Ionization of xenon atoms in high
Rydberg states by collision with molecules. J Chem Phys 67:1352–1359.

Foster MS, Beauchamp JL. 1975. Electron attachment to sulphur hexa-

fluoride: Formation of stable SF

at low pressure. Chem Phys Lett

31:482–486.

Fox RE, Hickam WM, Grove DJ. 1951. Ionization potentials and probabilities

using a mass spectrometer. Phys Rev 84:859–860.

Fox RE, Hickam WM, Kjeldaas T. 1953. Ionization probability curves for

krypton and xenon near threshold. Rev Phys 89:555–558.

Fox RE, Hickam WM, Kjeldaas T, Grove DJ. 1955. Ionization in a mass

spectrometer by monoenergetic electrons. Rev Sci Instrum 26:1101–1107.

Frey WF, Compton RN, Naff WT, Schweiler HC. 1973. Electron impact

studies of some cyclic hydrocarbons. Int J Mass Spectrom Ion Phys
12:19–32.

Garrec JL, Steinhurst DA, Smith MA. 2001. Measurement of the autodetach-

ment lifetime of SF

as a function of electron energy in a free jet

expansion. J Chem Phys 114:8831–8835.

Gobeli DA, Yang JJ, El-Sayad MA. 1985. Laser multiphoton ionization-

dissociation mass spectrometry. Chem Rev 85:529–554.

Gonzalez-Moraga G. 1993. Cluster chemistry. Springer-Verlag, New York.

Goudsmit SA. 1948. A time-of-flight mass spectrometer. Phys Rev 74:622–623.

Grill V, Walder G, Scheier P, Kurdel M, Ma¨rk TD. 1993. Absolute partial and

total electron impact ionization cross sections for C

from threshold

up to 950 eV. Int J Mass Spectrom Ion Processes 129:31–42.

Grill V, Drexel H, Sailer W, Lezius M, Ma¨rk TD. 2001. Operating principle of

an electron monochromator in an axial magnetic field. J Mass Spectrom
36:151–158.

Grissom JT, Compton RN, Garrett WR. 1969. Electron impact excitation and

ionization of helium near 60 ev. Phys Lett A 30:117–118.

Grissom JT. 1970. Electron impact excitation and ionization studies of the

rare gases. PhD Thesis, The University of Tennessee.

Grimsrud EP, Chowdhury S, Kebarle P. 1985. Electron affinity of SF

and

perfluoromethylcyclohexane. The unusual kinetics of electron transfer
reactions A

þ B ¼ A þ B

, where A

¼ SF

or perfluorinated cyclo-

alkanes. J Chem Phys 83:1059–1068.

Grobe V. 1963. Formation of negative ions in SbCl

. J Chem Phys 39:972–978.

Guilhaus M. 1994. Spontaneous and deflected drift-trajectories in orthogonal

acceleration time-of-flight mass spectrometry. Am Soc Mass Spectrom
5:588–595.

Guilhaus M, Selby D, Milynski V. 2000. Orthogonal acceleration time-of-

flight mass spectrometry. Mass Spectrom 19:65–107.

Gutsev GL, Bartlett RJ. 1998. Adiabatic electron affinities of PF5 and S F6: A

coupled-cluster study. Mol Phys 94:121–125.

Hammer NI. 2003. Dipole-bound anions. PhD Thesis. The University of

Tennessee.

Hammer NI, Diri K, Jordan KD, Desfrancois C, Compton RN. 2003. Dipole-

bound anions of carbonyl, nitrile, and sulfoxide containing molecules. J
Chem Phys 119:3650–3660.

Hammer NI, Hinde RJ, Compton RN, Diri K, Jordan KD, Radisic D, Stokes ST,

Bowen KH. 2004. Dipole-bound anions of highly polar molecules:
Ethylene carbonate and vinylene carbonate. J Chem Phys 120:685–690.

Hammer NI, Compton RN, Adamowicz L, Stepanian SG. 2005. Isotope effects

in dipole-bound anions of acetone. Phys Rev Lett 94:153004 (4 pages).

Harland PW, Thynne JCJ. 1971. Autodetachment lifetime, attachment cross-

section and negative ions formed by sulfur hexafluoride and sulfur
tetrafluoride. J Phys Chem 75:3517–3523.

Harland PW, Thynne JCJ. 1971. Decomposition of the metastable SF

ion. J

Inorg Nucl Chem Lett 7:29–31.

Harland PW, Thynne JCJ. 1972. The effect of substituents upon the

autodetachment lifetimes of the perfluorobenzene and nitrobenzene
Anions. J Chem Soc Chem Commun 8:476–477.

Harland PW, Thynne JCJ. 1972/1973. Ionisation of perfluorocyclobutane by

electron impact. Int J Mass Spectrom Ion Phys 10:11–23.

Harland PW, Thynne JCJ. 1973. Negative-ion formation by perfluoro-n-

butane as the result of low energy electron impact. Int J Mass Spectrom
Ion Phys 11:445–454.

Harland PW, Vallance C. 1998. Positive ion-electron impact ionization cross-

sections. Adv Gas-Phase Ion Chem 3:319–358.

Harland PW, Vallance C. 1997. Ionization cross-section and ionization

efficiency curves from polarizability volumes and ionization potentials.
Int J Mass Spectrom Ion Processes 171:173–181.

Harth K, Ruf M-W, Hotop H. 1989. Electron transfer from laser excited

Rydberg atoms to molecules. Absolute rate constants at low and
intermediate principal quantum numbers. Z Phys D 14:149–165.

Hasted JB. 1972. Physics of atomic collisions Second edition. Butterworth

and Co. Ltd., London.

Hays EE, Richards PI, Goudsmit SA. 1951. Mass measurements with a

magnetic time-of-flight mass spectrometer. Phys Rev 84:824–829.

Henis JMS, Mabie CA. 1970. Determination of autoionization lifetimes by

ion cyclotron reonances linewidths. J Chem Phys 53:2999–3013.

Herzenberg A. 1969. Attachment of slow electrons to oxygen molecules. J

Chem Phys 51:4942–4950.

Hickam WM, Fox RE. 1956. Electron attachment in sulfur hexafluoride using

monoenergetic electrons. J Chem Phys 25:642–647.

Hoffmann SV, Lunt SL, Jones NC, Field D, Ziesel J-P. 2002. An undulator-based

spherical grating monochromator beamline for low energy electron-
molecule scattering experiments. Rev Sci Instrum 73:4157–4163.

Horacek J, Domcke W, Nakamura H. 1997. Electron attachments and

vibrational excitation in hydrogen iodide: Calculation based on the
nonlocal resonance model. Z Phys D 42:181–185.

Hotop H, Niehaus A. 1967. Collisional ionization of long-lived highly excited

atoms. J Chem Phys 47:2506–2507.

Huang J. 1995. Low energy photon and low energy electron interactions with

. PhD Thesis, University of Pennsylvania.

Huang J, Carman HS, Jr., Compton RN. 1995. Low-energy electron

attachment to C

. J Phys Chem 99:1719–1733.

Huebner RH, Frey WF, Compton RN. 1973. Threshold electron excitation of

azulene. Chem Phys Lett 23:587–591.

Hunt WW, Jr., Huffman RE, McGhee KE. 1964a. Observation and

identification of ion dissociation processes occurring in the drift tube
of a time-of-flight mass spectrometer. Rev Sci Instrum 35:82–87.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

282

Mass Spectrometry Reviews DOI 10.1002/mas

Hunt WW, Jr., Huffman RE, Saari J, Wassel G, Betts JF, Paufve EH, Wyess W,

Fluegge RA. 1964b. Time-of-flight mass spectrometer adapted for
studying charge transfer, ion dissociation, and photoionization. Rev Sci
Instrum 35:88–95.

Hwang W, Kim YK, Rudd ME. 1996. New model for electron-impact

ionization cross sections of molecules. J Chem Phys 104:2956–2966.

Ionov NI, Mamyrin BA, Fiks VB. 1953. A resonance magnetic mass

spectrometer with a high resolving power. Zh Tekh Fiz 23:2104–2106.

Jacobs G, Henglein A. 1964. A method for detection of slow electrons from

excitation and ionization of molecules by electron collisions. Z
Naturforch 19a:906–910.

Jaffke T, Illenberger E, Lezius M, Matejcik S, Smith D, Ma¨rk TD. 1994.

Formation of C

and C

by free electron capture. Activation energy

and effect of the internal energy on lifetime. Chem Phys Lett 226:213–
218.

Jain DK, Khare SP. 1976a. Ionizing collisions of electrons with CO

, CO,

O, CH

and NH

. J Phys B 9:1429–1438.

Jain DK, Khare SP. 1976b. Ionizing collisions of electrons with ethylene and

ethane. Indian J Pure Appl Phys 14:201–202.

Jarvis GK, Kennedy RA, Mayhew CA. 2001. Investigations of low energy

electron attachment to ground state group 6B hexafluorides (SF

, SeF

and TeF

) using an electron-swarm mass spectrometric technique. Int J

Mass Specrom 205:253–270.

Kadish KM, Ruoff RS. 1994. Fullerenes: Recent advances in the chemistry

and physics of fullerenes and related material. The Electrochemical
Society, Inc., New Jersey.

Karas M, Bachmann D, Bahr U, Hillenkamp F. 1987. Matrix-assisted

ultraviolet laser desorpton of non-volatile compounds. Int J Mass
Spectrom Ion Processes 78:53–68.

Katzenstein HS, Friedland SS. 1955. New time-of-flight mass spectrometer.

Rev Sci Instrum 26:1150–1157.

Kim YK, Rudd ME. 1994. Binary-encounter-dipole model for electron-

impact ionization. Phys Rev A 50:3954–3967.

Klar D, Mirbach B, Korsch HJ, Ruf MW, Hotop H. 1994. Comparison of rate

coefficients for Rydberg electron and free electron attachment. Z Phys
Chem D 31:235–244.

Klots CE. 1967. Statistical aspects of autoionization lifetimes. J Chem Phys

46:1197–1199.

Klots CE, Compton RN. 1977. Electron attachment to carbon dioxide clusters

in a supersonic beam. J Chem Phys 67:1779–1780.

Klots CE, Compton RN. 1978a. Electron attachment to van der Waals

polymers of carbon dioxide and nitrous oxide. J Chem Phys 69:1636–
1643.

Klots CE, Compton RN. 1978b. Electron attachment to van der Waals

polymers of water. J Chem Phys 69:1644–1647.

Klots CE, Compton RN. 1980. Self-scavenging of electrons in van der Waals

molecules of methyl iodide. Chem Phys Lett 73:589–591.

Kroto HW, Heath JR, O’Brien SC, Curl RF, Smalley RE. 1985. C

Buckminsterfullerene. Nature 318:162–163.

Kroto HW, Fischer JE, Cox DE. 1993. The fullerenes. Pergamon Press,

Oxford, UK.

Lampe FW, Franklin JL, Field FH. 1957. Cross sections for ionization by

electrons. J Am Chem Soc 79:6129–6132.

Liu Y, Suess L, Dunning FB. 2005. Rydberg electron transfer to S F

: Product

ion lifetimes. J Chem Phys 122:214313 (6 pages).

Mamyrin BA, Karataev VI, Shmikk DV, Zagulin VA. 1973. The mass–

reflectron, a new nonmagnetic time-of-flight mass spectrometer with
high resolution. Sov Phys JETP Lett 37:45–48.

Mamyrin BA. 1994. Laser assisted reflectron time-of-flight mass spectrom-

etry. Int J Mass Spectrom Ion Processes 131:1–19.

Marmet P, Kerwin L. 1960. An improved electronstatic energy selector. Can J

Phys 38:787–796.

Margreiter D, Deutsch H, Schmidt M, Ma¨rk TD. 1990. Electron impact

ionization cross sections of molecules Part II. Theoretical determination
of total (counting) ionization cross sections of molecules: A new
approach. Int J Mass Spectrom Ion Processes 100:157–176.

Margreiter D, Deutsch H, Schmidt M, Ma¨rk TD. 1994. A semiclassical

approach to the calculation of electron impact ionization cross-sections
of atoms: From hydrogen to uranium. Int J Mass Spectrom Ion
Processes 139:127–139.

Martin F, Burrow PD, Cai Z, Coutier P, Hunting D, Sanche L. 2004. DNA

strand breaks induced by 0–4 eV electrons: The role of shape
resonances. Phys Rev Lett 93:068101–068103.

Massey HSW, Burhop EHS, Gilbody HB. 1969. Electronic and ionic impact

phenomena Vol. I. Oxford University Press, UK.

Massey HSW, Burhop EHS, Gilbody HB. 1969. Electronic and ionic impact

phenomena Vol. II. Oxford University Press, UK.

Massey HSW, Burhop EHS, Gilbody HB. 1971. Electronic and ionic impact

phenomena Vol. III. Oxford University Press, UK.

Massey HSW, Burhop EHS, Gilbody HB. 1974. Electronic and ionic impact

phenomena Vol. IV. Oxford University Press, UK.

Massey HSW. 1976. Negative ions. Cambridge University Press, UK.

Matsuzawa M. 1971. Ionization of long-lived highly excited atoms by

collision with molecules. J Chem Phys 55:2685–2689.

Matsuzawa M. 1972a. Electron transfer from highly excited atoms to

molecules. J Phys Soc Japan 32:1088–1094.

Matsuzawa M. 1972b. Reaction of highly excited atoms with molecules

A**

þ BC ! A

þ B þ C

. J Phys Soc Japan 33:1108–1119.

Matsuzawa M. 1974. Ionization of long-lived highly excited atoms by

collision with molecules. J Electron Spectrosc Relat Phenom 4:1–12.

Matsuzawa M. 1983. Theoretical studies of collisions of Rydberg atoms with

molecules. In: Stebbings RF, Dunning FB, editors. Rydberg states of
atoms and molecules. Cambridge University Press, UK.

McDaniel EW. 1989. Atomic collisions: Electrons and proton projectiles.

Wiley-Interscience, New York.

McLafferty FW, Gohlke RS, Golesworthy RC. 1964. 12th Annual Conference

on Mass Spectrometry and Allied Topics, ASTM Committee E-14.
Montreal.

Mu¨ller-Dethlefs K, Sander M, Schlag EW. 1984. Two-colour photoionization

resonance spectroscopy of NO: Complete separation of rotational levels
of NO

at the ionization threshold. Chem Phys Lett 112:291–294.

Naff WT, Cooper CD, Compton RN. 1968. Transient negative-ion states in

alicyclic and aromatic fluorocarbon molecules. J Chem Phys 49:2784–
2788.

Naff WT, Compton RN, Cooper CD. 1971. Attachment of electrons to

substituted benzenes. J Chem Phys 54:212–222.

Naff WT, Compton RN, Cooper CD. 1972. Electron attachment and

excitation process in selected carbonyl compounds. J Chem Phys
57:1303–1307.

Nishimura H, Tawara H. 1994. Total electron impact ionization cross sections

for simple hydrocarbon molecules. J Phys B 27:2063–2074.

Odom RW, Smith DL, Futrell JH. 1975. A study of electron attachment to SF

and autodetachment and stabilization of SF

. J Phys B 8:1349–1366.

O’Halloran GJ, Fluegge RA, Betts JF, Everett WL. 1964. Technical

documentary report, No. ASD-TDR-62-644, Part I & II, prepared
under contract Nos. AF33(616)-8374 and AF33(657)-11018 by the
Bendix Corporation, Research Laboratories Division, Southfield,
Michigan.

Orient OJ, Srivastava SK. 1987. Electron impact ionization of water, carbon

monoxide, carbon dioxide, and methane. J Phys B 20:3923–3936.

Priest TW. 1972. Interpretation of collisional ionization rates in flames. J

Chem Soc Farad Trans I 68:661–669.

Price D, Milnes GJ. 1990. The renaissance of time-of-flight mass

spectrometry. Int J Mass Spectrm Ion Processes 99:1–39.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

283

Rapp D, Englander-Golden P. 1965. Total cross sections for ionization and

attachment in gases by electron impact. I. Positive ionization. J Chem
Phys 43:1464–1479.

Robertson WD, Hammer NI, Bartmess JE, Compton RN, Diri K, Jordan KD.

2005. Negative ions of ethylene sulfite. J Chem Phys 122:204319 (6
pages).

Saksena V, Kushwaha MS, Khare SP. 1997a. Ionization cross section of

molecules due to electron impact. Physica B 233:201–212.

Saksena V, Kushwaha MS, Khare SP. 1997b. Electron impact ionization of

molecules at high energies. Int J Mass Spectrom Ion Processes 171:L1–
L5.

Sagara T, Boesten L, Nishida S, Okada K. 2000. Resolution improvements for

hemispherical energy analyzers. Rev Sci Instrum 71:4201–4207.

Scheller MK, Compton RN, Cederbaum LS. 1995. Gas-phase multiply

charged anions. Science 270:1160–1166.

Schram BL, van der Wiel MJ, de Heer FJ, Moustafa HR. 1966. Absolute gross

ionization cross sections for electrons (0.6–12 keV) in hydrocarbons. J
Chem Phys 44:49–54.

Schultz GJ, Spence D. 1969. Temperature dependence of the onset for

negative-ion formation inCO

. Phys Rev Lett 22:47–50.

Schulz GJ. 1958. Measurement of atomic and molecular by a trapped-electron

method. Phys Rev 112:150–154.

Schulz GJ. 1959. Measurement of excitation of N

, CO and He by electron

impact. Phys Rev 116:1141–1147.

Schulz GJ. 1960a. Excitation and negative ions in H

O. J Chem Phys

33:1661–1665.

Schulz GJ. 1960b. Use of SF6

for calibration of the electron energy scale. J

Appl Phys 31:1134–1135.

Schulz GJ. 1973. Resonances in electron impact on diatomic molecules. Rev

Mod Phys 45:423–486.

Seccombe DP, Reddish TJ. 2001. Theoretical study of space focusing in

linear time-of-flight mass spectrometers. Rev Sci Instrum 72:1330–
1338.

Shea MJ, Compton RN, Hettich RL. 1990. Laser ablation studies of palladium

electrolytically loaded with hydrogen and deuterium. Phys Rev A
42:3579–3586.

Shea MJ, Compton RN. 1993. Surface-plasmon ejection of Ag

ions from

laser irradiation of a roughened silver surface. Phys Rev B 47:9967–
9970.

Shyn TW, Sharp WE. 1979. Doubly differential cross section of secondary

electrons ejected from gases by electron impact: 50–400 eV on CO

Phys Rev A 20:2332–2339.

Simpson JA. 1964. High-resolution spectrometer for low-energy electrons.

Rev Sci Instrum 35:1698–1704.

Simpson JA, Kuyatt C. 1963. Design of low voltage electron guns. Rev Sci

Instrum 34:265–268.

Smith LG. 1951. A new magnetic period mass spectrometer. Rev Sci Instrum

22:115–116.

Smirnov BM. 1982. Negative ions. McGraw-Hill, New York.

Srivastava SK, Iga I, Rav MVVS. 1995. Segmented time-of-flight mass

spectrometer. Meas Sci Technol 6:1379–1382.

Stamatovic A, Schulz GJ. 1968. Trochoidal electron monochromator. Rev Sci

Instrum 39:1752–1753.

Stamatovic A, Schulz GJ. 1970. Characteristics of the trochoidal electron

monochromator. Rev Sci Instrum 41:423–427.

Stephens WE. 1946. A pulsed mass spectrometer with time dispersion. Bull

Am Phys Soc 21(2):22.

Stockdale JA, Compton RN, Reinhardt PW. 1969. Studies of negative-ion-

molecule reactions in the energy region from 0 to 3 eV. Phys Rev
184:81–93.

Stockdale JA, Davis FJ, Compton RN, Klots CE. 1974. Production of negative

ions from CH

X molecules (CH

, CH

CN, CH

I, CH

Br) by

electron impact and by collisions with atoms in excited Rydberg states. J
Chem Phys 60:4279–4285.

Straub HC, Lindsay BG, Smith KA, Stebbings RF. 1996. Absolute partial

cross sections for electron-impact ionization of CO

from threshold to

1000 eV. J Chem Phys 105:4015–4022.

Suess L, Parthasarathy R, Dunning FB. 2002. Nondissociative low-energy

electron attachment to SF

, C

, and c-C

: Negative ion

lifetimes. J Chem Phys 117:11222–11227.

Tuinman AA, Compton RN. 1996. A Search for long-lived hydrogen halide

anions. Rapid Commun Mass Spectrom 10:389–392.

Vallance C, Harland PW, Maclagan RGAR. 1996. Quantum mechanical

calculation

maximum

electron

impact

single

ionization

cross sections for the inert gases and small molecules. J Phys Chem
100:15021–15026.

Vallance C, Maclagan RGAR, Harland PW. 1997. Ionization surfaces for

small molecules. J Phys Chem A 101:3505–3508.

Vasil’ev YV, Absalimov RR, Nasibullaev SK, Lobach AS, Drewello T. 2000.

Formation and characterization of long-lived negative molecular ions of
C

. J Chem Phys A 105:661–665.

Vasil’ev YV, Muftakhov MV, Tuimedov GM, Khatymov RV, Abzalimov RR,

Mazunov VA, Drewello T. 2001. Specific formation of (M-H)

ions from

OH-group-containing molecules. Int J Mass Spectrom 205:119–135.

Vestal ML, Juhasz P, Martin SA. 1995. Delayed extraction matrix-assisted

laser desorption time-of-flight mass spectroscopy. Rapid Commun
Mass Spectrom 9:1044–1050.

Voinov VG, Vasil’ev YV, Moore´ J, Barofsky DF, Deinzer ML, Gonin M, Egan

TF, Fu¨hrer K. 2003. A resonant electron capture time-of flight MS with
trochoidal electron monochromator. Anal Chem 75:3001–3009.

Voinov VG, Vasil’ev YV, Ji H, Figard B, Moore´ J, Egan TF, Barofsky DF,

Deinzer ML. 2004. A gas chromatograph/resonant electron capture-
TOF mass spectrometer for four dimensions of negative ion analytical
information. Anal Chem 76:2951–2957.

Vasil’ev YV, Figard BJ, Voinov VG, Barofsky DF, Deinzer ML. 2006.

Resonant electron capture by some amino acids and their methyl esters.
J Am Chem Soc 128:5506–5515.

Vasil’ev YV, Figard BJ, Voinov VG, Barofsky DF, Deinzer ML. 2007.

Resonant electron capture by some amino acid esters. Int J Mass
Spectrom 268:106–121.

Wacks M. 1959. Ph.D. Thesis, University of Utah.

West WP, Foltz GW, Dunning FB, Latimer CT, Stebbings RF. 1976. Absolute

measurements of collisional ionization of xenon atoms in well-defined
high Rydberg states. Phys Rev Lett 36:854–855.

Whitten GZ, Rabinovitch BS. 1963. Accurate and facile approximation for

vibrational energy-level sums. J Chem Phys 38:2466–2473.

Wiley WC, McLaren IH. 1955. Time-of-flight mass spectrometer with

improved resolution. Rev Sci Instrum 26:1150–1157.

Wolff MM, Stephens WE. 1953. A pulsed mass spectrometer with time

dispersion. Rev Sci Instrum 24:616–617.

Zheng Z, Smith KA, Dunning FB. 1988. Postattachment interactions in

K(nd)–SF

collisions at intermediate n. J Chem Phys 89:6295–6297.

Zollars BG, Walter CW, Lu F, Johnson CB, Smith KA, Dunning FB. 1986.

Ionization in K(nd)–SF

and K(nd)–CCl

collisions at intermediate n. J

Chem Phys 84:5589–5593.

MIRSALEH-KOHAN, ROBERTSON, AND COMPTON

284

Mass Spectrometry Reviews DOI 10.1002/mas

Nasrin Mirsaleh Kohan

was born in Tehran, Iran where she received

her Bachelor of Science degree in Physics at the University of Tehran.
She completed her Masters degree in computational Physics at the
Bowling Green State University and is currently completing a Ph.D.
degree in Chemical Physics at the University of Tennessee. The focus
of her research is in the area of gas phase multiply charged negative
ions. Her research involves various mass spectrometry techniques such
as Time-of-Flight and Electrospray ionization mass spectrometry.

Wesley D. Robertson

was born in Castalian Springs, Tennessee. He

completed his Bachelor of Science degree (2001) and his Masters
degree (2003) in physics at the University of Tennessee. He has worked
extensively with many variations of time-of-flight mass spectrometers
in chemical physics research as well as built an electron impact time-
of-flight mass spectrometer with a high-resolution trochoidal electron
monochromator. He is currently a physics Ph.D. candidate at Emory
University in Atlanta, Georgia.

Robert N. Compton

was born November 28, 1938 in Metropolis, IL.

He received degrees in Physics from Berea College (BA), the
University of Florida (MS) and the University of Tennessee (PhD).
He was a Senior Corporate Fellow at the Oak Ridge National
Laboratory from 1965 to 1995 and has been a Professor of both Physics
and Chemistry at the University of Tennessee from 1995 to the present
date. He has been a Visiting Professor at the University of Aarhus,
University of Paris, and the FOM Institute in Amsterdam. In 2001,

he was an Erskine Fellow at the University of Christchurch, New Zealand. In 1979, he and
J.A.D. Stockdale co-founded a small scientific instruments company (Comstock, Inc.). His
research interests include negative ions, laser spectroscopy, and molecular chirality.

ELECTRON IONIZATION TOFMS

Mass Spectrometry Reviews DOI 10.1002/mas

285

Wyszukiwarka

Podobne podstrony:
Szewczyk, Rafał i inni Rapid method for Mycobacterium tuberculosis identification using electrospra
English Summary of The Invisible Rainbow A History of Electricity and Life 3
wright the historiography of the mass worker
The Promise for AGB Stars Physics and Chemistry of the Inner Circumstellar Envelope, and the Mass L
In vivo MR spectroscopy in diagnosis and research of
UTC?te and time of solstices and equinoxes
The time of my life Bill Medley
Weis & Hickman Dragonlance Legends Time of Twins
Dirty Dancing Time Of My Life
time of my life
Book 1 The time of the?rk heat
Book 3 The time of the cold sun
Dirty?ncing The time of my life
Megamorphs 2 In The Time of Dinosaurs
A time of prosperous change
The History of the USA 7 American Expansion (units 9,, and)
by the time of world war ii NZRMQ22WVNW73JEDUAEMIMBG2CLONPKSGVQGVJA

więcej podobnych podstron