CA 02227806 1998-O1-23
B&P File No. 571-453
BERESKIN & PARK CANADA
Title: SPECTROMETER PROVIDED WITH PULSED ION SOURCE AND
TRANSMISSION DEVICE TO DAMP ION MOTION AND METHOD OF
USE
Inventors: Andrew N. Krutchinsky, Alexandre V. Loboda, Victor L. Spicer,
Erich W. Ens, Kenneth G. Standing
CA 02227806 2005-05-05
-1-
Title: SPECTROMETER PROVIDED WITH PULSED ION SOURCE AND
TRANSMISSION DEVICE TO DAMP ION MOTION AND METHOD OF USE
FIELD OF THE INVENTION
This invention relates to mass spectrometers and ion sources therefor.
More particularly, this invention is concerned with pulsed ion sources and the
provision of a transmission device which gives a pulse ion source many of the
characteristics of a continuous source, such that it extends and improves the
application of Time of Flight Mass Spectrometry (TOFMS) and that it
additionally can be used with a wide variety of other spectrometers, in
addition
to an orthogonal injection time-of-flight mass spectrometer.
BACKGROUND OF THE INVENTION
Ion sources for mass spectrometry may be either continuous, such as
ESI (electrospray ionization) sources or SIMS (secondary ion mass
spectrometry) sources, or pulsed, such as MALDI (matrix-assisted laser
desorption/ionization sources). Continuous sources have normally been used
to inject ions into most types of mass spe~ctrameter, such as sector
instruments, quadrupoles, ion traps and ion cyclotron resonance
spectrometers. Recently it has also become possible to inject ions from
continuous sources into time-of flight (TOF) mass spectrometers through the
use of "orthogonal injection", whereby the continuous beam is injected
orthogonally to the main TOF axis and is converted to the pulsed beam
required in the TOF technique. This is most efficiently carried out with the
addition of a collisional damping interface between the source and the
spectrometer.
On the other hand, pulsed sources, MALDI sources for example, have
usually been coupled directly to TOF mass spectrometers, to
CA 02227806 1998-O1-23
-2-
take advantage of the discrete or pulse nature of the source. TOF mass
spectrometers have several advantages over conventional quadrupole or
ion tarp mass spectrometers. One advantage is that TOF mass spectrometers
can analyze a wider mass-to-charge range than do quadrupole and ion trap
mass spectrometers. Another advantage is that TOF mass spectrometers can
record all ions simultaneously without scanning, with higher sensitivity
than quadrupole and ion trap mass spectrometers. In a quadrupole or other
scanning mass spectrometer, only one mass can be transmitted at a time,
leading to a duty cycle which may typically be 0.1%, which is low (leading to
low sensitivity). A TOF mass spectrometer therefore has a large inherent
advantage in sensitivity.
However, TOF mass spectrometers encounter problems
with many widely used sources which produce ions with a range of energies
and directions. The problems are particularly acute when ions produced by
the popular MALDI (matrix-assisted laser desorption/ionization) technique
are u~~ed. In this method, photon pulses from a laser strike a target and
desorb ions whose masses are measured in the mass spectrometer. The
target material is composed of a low concentration of analyte molecules,
which, usually exhibit only moderate photon absorption per molecule,
embedded in a solid or liquid matrix consisting of small, highly-absorbing
species. The sudden influx of energy is absorbed by the matrix molecules,
causing them to vaporize and to produce a small supersonic jet of matrix
molecules and ions in which the analyte molecules are entrained. During
this Ejection process, some of the energy absorbed by the matrix is
transferred to the analyte molecules. The analyte molecules are thereby
ionized, but without excessive fragmentation, at least in the ideal case.
Because a pulsed laser is normally used, the ions also
appear as pulses, facilitating their convenient measurement in a
time-of-flight spectrometer. However, the ions acquire a considerable
amount of energy in the supersonic jet, with velocities of the order of 700
m/s, and they also may lose energy through collisions with the matrix
molecules during acceleration, particularly in high accelerating fields. These
CA 02227806 1998-O1-23
-3-
and similar effects lead to considerable peak broadening and consequent loss
of resolution in a simple linear time-of-flight instrument, where the ions
are extracted from the target nearly parallel to the spectrometer axis. A
partial solution to the problem is provided by a reflecting spectrometer,
which. partially corrects for the velocity dispersion, but a more effective
technique is the use of delayed extraction, either by itself or in combination
with ~~ reflector. In delayed extraction, the ions are allowed to drift for a
short period before the accelerating voltage is applied. This technique
partially decouples the ion production process from the measurement,
making the measurement less sensitive to the detailed pattern of ion
desorption and acceleration in any particular case. Even so, successful
operation requires careful control of the laser fluence (i.e. the amount of
power supplied per unit area) and usually some hunting on the target for a
favorable spot. Moreover, the extraction conditions required for optimum
performance have some mass dependence; this complicates the calibration
procedure and means that the complete range of masses cannot be observed
with optimum resolution at any given setting. Also, the technique has had
limited success in improving the resolution for ions of masses greater than
about 20,000 Da and due to the pulsed nature of MALDI ion sources it is
difficult to obtain high performance in MS/MS instruments.
Although coupling to a TOF instrument is used as an
example above, similar problems arise in coupling MALDI and other pulsed
sources to other types of mass spectrometer, such as quadrupole (or other
multipole), ion trap, magnetic sector and FTICRMS (Fourier Transform Ion
Cyclotron Resonance Mass Spectrometer). Further, it is also desirable to be
able to couple MALDI or other pulsed sources to tandem mass
spectrometers, e.g. a triple quadrupole or a QqTOF, which allows MS/MS of
MALI~I ions to be formed. Standard MALDI instruments cannot be
configured to carry out high performance MS/MS. The dispersion in
energ~~ and angle of ions produced by a MALDI source, or similar source,
acceni:uate the difficulty of ion injection. Also, because the residence times
of ions in most other types of mass spectrometer are considerably longer
CA 02227806 1998-O1-23
-4-
than i:n TOF instruments, the large space charge in the pulse can introduce
additional problems. These instruments are all designed to operate with
continuous sources, so conversion of the pulsed source to a quasi-
continuous one solves most of the problems.
BRIEF SUMMARY OF THE PRESENT INVENTION
Accordingly, it is desirable to provide an apparatus and
method enabling a pulse source, such as a MALDI source, to be coupled to a
variety of spectrometer instruments, in a manner which more completely
decouples the spectrometer from the source and provides a more
continuous ion beam with smaller angular and velocity spreads.
More particularly, it is desirable to provide an improved
TOF mass spectrometer with a pulsed ion source, in which the energy
spread in the ion beam is reduced, in which the source is more completely
decoupled from the spectrometer than in existing instruments, in which
problems resulting from ion fragmentation are reduced, enabling new types
of measurement, and in which the results obtained from the mass
spectrometer and its ease of operation are consequently improved.
It is also desirable to provide a TOF mass spectrometer
with both continuous and pulsed sources, for example both ESI and MALDI
sourcE~s, so either source can be selected.
In accordance with the present invention, there is
provided a mass spectrometer system comprising:
a pulsed ion source, for providing pulses of analyte ions;
a mass spectrometer;
an ion path extending between the ion source and the
mass ;>pectrometer; and
an ion transmission device located in said ion path and
having a damping gas therein for reducing the spatial and energy spread of
ions travelling from said ion source to said mass spectrometer.
The invention has particular applicability to time of flight
mass spectrometers. As these require a pulsed beam, conventional teaching
CA 02227806 1998-O1-23
-5-
is that a pulsed source should be coupled maintaining the pulsed
chara<aeristics. However, the present inventors have now realised that
there are advantages to, in effect converting a pulsed beam into a
continuous, or at least quasi-continuous, beam, and than back into a pulsed
beam.The advantages are: improvement in beam quality through
collisional damping; decoupling of the ion production from the mass
measurement; ability to measure the beam current by dingle-ion counting
because it is converted from a few large pulses to many small pulses, for
examf>le from about 1 Hz. to about 4 kHz., or a factor of 4,000; compatibility
with a continuous source , such as ESI, offering the possibility of running
both sources on one instrument.
The invention also has applicability to mass spectrometers
that work with or require a continuous beam. Then, the advantage is that a
pulsed source can indeed be used with such spectrometers.
Preferably, the ion source provides the analyte for
ionization by radiation, and wherein there is provided a source of
electromagnetic radiation, more preferably a pulsed laser, directed at the ion
sourcE~, for generating radiation pulses to cause desorption and ionization of
analyte molecules.
Advantageously, the ion source comprises a target material
composed of a matrix and analyte molecules in the matrix, the matrix
comprising a species adapted to absorb radiation from the radiation source,
to promote desorption and ionization of the analyte molecules.
Preferably, the transmission device comprises a multipole
rod set.
there can be two or more multipole rod sets and means for supplying
different RF and DC voltages to the rod sets.
The mass spectrometer system can include a continuous
ion source, and means for selecting one of the pulsed ion source and the
continuous ion source, and this then provides the characteristics of two
separate instruments in one instrument. The two ion sources can comprise
a MA1~DI source and an ESI source.
CA 02227806 1998-O1-23
-6-
Another aspect of the present invention provides a
method of generating ions and delivering ions to spectrometer, the method
comprising:
(1) providing an ion source;
(2) causing the ion source to produce pulses of ions;
(3) providing an ion transmission device along an ion path
extending from the ion source, and providing the ion transmission device
with a~ damping gas to reduce the spatial and energy spread of ions from the
ion source; and
(4) passing the ions from the ion transmission device into
a spectrometer, for mass analysis.
Preferably, step (3) comprises providing an RF rod set
within the transmission device. Further, a DC field can be provided
between the ion source and the spectrometer to promote movement of ions
towards the spectrometer.
The method can include providing two or more rod sets in
the ion transmission device, and operating at least one rod set with a DC
offset to enable selection of ions with a desired mass-to-charge ratio. A
potential difference can be provided between two adjacent rod sets
significant to accelerate ions into the downstream rod set, to cause
collisionally induced dissociation in to the downstream rod set.
When a pulsed laser is used, for each laser pulse, a
plurality of pulses of ions are delivered into the time-of-flight mass
spectrometer.
BRIEF DESCRIPTION OF THE DRAWING FIGURES
For a better understanding of the present invention and to
show more clearly how it may be carried into effect, reference will now be
made, by way of example, to the accompanying drawings, which show
preferred embodiments of the present invention and in which:
Figure 1 shows a block diagram of a mass spectrometer
system;
CA 02227806 1998-O1-23
-7-
Figure 2 is a schematic diagram showing a MALDI-TOF
mass spectrometer with orthogonal injection of the MALDI ions into the
spectrometer through a collisional damping interface (quadrupole ion
guide) according to the present invention;
Figure 3 shows a mass spectrum of a mixture of several
peptides and proteins leucine-enkephalin-Arg (Le-R), substance P (Sub P),
melittin (ME), CD4 fragment 25-58 (CD4), and insulin (INS) ) produced in
the spectrometer of Figure 2;
Figure 4 shows plots of transit times through the interface
for different ions;
Figure 5 shows a mass spectrum of substance P;
Figure 6 shows a mass spectrum of a tryptic digestion of
citrate synthase;
Figure 7A shows a schematic of part of spectrometer of
Figure 2, showing the collisional interface and indicating applied voltages;
Figures 7B, 7C and 7D show different operating regimes of
the m;~ss spectrometer of Figure 2;
Figures 8A, 8B, 8C, and 8D are mass spectra obtained from
substance P recorded in the different operation regimes, according to Figures
7B, 7C, and 7D;
Figure 9 shows the behaviour of the ion current from a
single target spot as a function of time; and
Figure 10 shows schematically combined ESI and MALDI
sources for a mass spectrometer.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
The first embodiment shown in Figure 1 is a block
diagrarn of a general mass spectrometer system. Here 1 represents any sort
of puh,ed ion source (for instance MALDI), 2 is a collisional focusing
chambE~r or region filled with a buffer gas and with a multipole 3 driven at
some RF voltage. This is followed by an optional manipulation stage 4 and
then a :mass analyzer 5. The collisional ion guide 3, in accordance with the
CA 02227806 1998-O1-23
-g_
present invention, spreads the pulsed ion beam in time, and improves its
beam quality (i.e. space and velocity distributions) by damping the initial
velocity and focusing the ions toward the central axis. The beam is then
quasi-c~~ntinuous and may enter an optional manipulation stage 4, where
ions ca.n be subjected to any sort of further manipulation. Finally the
resultalit ions are analyzed in the mass analyzer 5.
A simple example of further manipulation in stage 4 is
dissoci<~tion of the ions by collisions in a gas cell, so that the resulting
daughter ions can be examined in the mass analyzer. This may be adequate
to determine the molecular structure of a pure analyte. If the analyte is a
complex mixture, stage 4 needs to be more complicated. In a triple
quadrupole or a QqTOF instrument (as disclosed in A. Shevchenko et al,
Rapid (~ommun. Mass Spectrom. 11, 1015, (1997)), stage 4 would include a
quadrupole mass filter for selection of a parent ion of interest and a
quadrupole collision cell for decomposition of that ion by collision-induced
dissociation (CID). Both parent and daughter ions are then analyzed in
section 5, which is a quadrupole mass filter in the triple quadrupole, or a
TOF spectrometer with orthogonal injection in the QqTOF instrument. In
both cases stages 1 and 2 would consist of a pulsed source and a collisional
damping ion guide.
It will be appreciated that the collisional focusing
chamber 2 is shown with a multipole rod set 3, which could be any suitable
rod set, e.g. a quadrupole, hexapole or octopole. The particular rod set
selected will depend upon the function to be provided.
Figure 2 shows a preferred embodiment of a MALDI-TOF
mass spectrometer 10 according to the present invention. The spectrometer
10 includes a conventional MALDI target probe 11, a shaft seal chamber 12,
pumped in known manner, and a target installed in the target-holding
electrode 13. A mixture of the sample to be investigated and a suitable
matrix are applied to the sample probe following the usual procedure for
preparing MALDI targets. A pulsed laser 14 is focused on the target surface
15 by lens 16, and is run at a repetition rate of some tens of pulses per
CA 02227806 1998-O1-23
-9-
second,, more specifically at a rate of 13 Hz. An inlet 18 is provided for
nitrogen or other neutral gas. Each laser shot produces a plume of neutral
and charged molecules. Ions of the sample analyte are produced and
entrained in the plume which expands into vacuum chamber 30, which
contains two quadrupole rod sets 31 and 32. Chamber 30 is pumped by a
pump (not shown) connected to port 34 to about 70 mTorr but the pressure
can be varied over a substantial range by adjusting the flow of gas through a
controllLed leak valve 18. Lower pressures could be used, and an important
characteristic is the product of pressure and rod length. Thus, a total length
x pressure value of 22.5 mTorr-cm could be used, as in U.S. patent 4,963,736.
The ga;s in chamber 30 (typically nitrogen or argon or other suitable gas,
preferable an inert gas) will be referred to as a damping gas or cooling gas
or
buffer g;as.
In the embodiment tested, the quadrupole rod sets 31 and
32 were made of rods 4.45 cm in length and 11 mm in diameter, and were
separated by 3 mm, i.e. the spacing between rods on adjacent corners of the
rod set. The quadrupoles 31 and 32 are driven by a power supply which
providE~s operating sine wave frequencies from 50 to 2 MHz, and output
voltages from 0 to 1000 volts peak-to-peak. Typical frequencies are 200 kHz
to 1 Mlaz, and typical voltage amplitudes are 100 to 1000 V peak-to-peak.
Both guadrupoles are driven by the same power supply through a
transformer with two secondary coils. Different amplitudes may be applied
to the quadrupoles by using a different number of turns in the two
secondary coils. D.C. Bias or offset potentials are applied to the rods of
quadrupoles 31 and 32 and to the various other components by a
multiple-output power supply. The RF quadrupoles 31 and 32, with the
damping gas between their rods can be run in an RF-only mode, in which
case they serve to reduce the axial energy, the radial energy, and the energy
spreads., of the ions which pass through it, as will be described. This
process
substantially spreads the plume of ions out along the ion path, changing the
initial beam, pulsed at about 13 Hz, into a quasi-continuous beam as
described in more detail below. The first quadrupole 31, can also be run in a
CA 02227806 1998-O1-23
-10-
mass-filtering mode by the application of a suitable DC voltage. The second
quadrupole 32 can then be used as a collision cell (and an RF-guide) in
collision-induced dissociation experiments (see below).
From chamber 30, the ions pass along an ion path 27 and
through a focusing electrode 19 and then pass through orifice 38, into a
vacuum chamber 40 pumped by a pump (not shown) connected to a port 42.
There, ithe ions are focused by grids 44 through a slot 46 into an ion storage
region ~I8 of a TOF spectrometer generally indicated at 50.
In known manner, ions are extracted from the storage
region 48 and are accelerated through a conventional accelerating column
51 whir_h accelerates the ions to an energy of approximately 4000 electron
volts per charge (4 keV). The ions travel in a direction generally orthogonal
to the ion path 27 between the ion storage region, through a pair of
deflection plates 52. The deflection plates 52 can serve to adjust the ion
trajectories, so that the ions are then directed toward a conventional
electrostatic ion mirror 54, which reflects the ions to a detector 56 at which
the ions are detected. The ions are detected using single-ion counting and
recorded with a time-to-digital converter (TDC). The accelerating column
51, plates 52, mirror 54 and detector 56 are contained in a main TOF
chamber 58 pumped to about 2 x 10-~ Torr by a pump (not shown) connected
to a port 60.
The use of orthogonal-injection of MALDI ions from
source 13 into the TOF spectrometer 50 has some potential advantages over
the usual axial injection geometry. It serves to decouple the ion production
process from the mass measurement to a greater extent than is possible in
the usual delayed-extraction MALDI. This means that there is greater
freedom to vary the target conditions without affecting the mass spectrum,
and they plume of ions has more time to expand and cool before the electric
field is applied to inject them into the spectrometer. Some improvement in
performance might also be expected because the largest spread in velocities
is along the ejection axis, i.e. the ion path 27, normal to the target, which
in
this cage is orthogonal to the TOF axis. However, orthogonal injection of
CA 02227806 1998-O1-23
-11-
MALD1 ions into the TOF 50 without collisional cooling has several
problems which appear to make the geometry impractical, namely:
(1) The radial energy distribution, while much smaller
than th.e axial energy is still sufficient to cause substantial spreading and
expansion of the beam as it leaves the quadrupole rod set 32 and travels
toward the TOF axis. The spatial spread of the beam along the TOF axis
limits the resolution. The effect can be reduced with collimation but only at
a significant sacrifice in sensitivity; a collimating slit must be placed
sufficiently far from the TOF axis to avoid distorting the extraction field,
and so the target must be placed far enough from the collimation slit to
produce a reasonably parallel beam;
(2) The axial velocity of the ions, i.e. velocity along the
ion path 27, in the plume is largely independent of mass which means the
energy is mass dependent. Since the axial energy determines the direction of
the trajjectory after acceleration into the TOF spectrometer, instrumental
acceptance (or acceptance by the TOF spectrometer) is mass dependent; i.e.
there is mass discrimination. The same effect is observed when ESI ions are
injected without collisional cooling as explained in detail in the prior
application mentioned above; and
(3) The width of the axial energy distribution is
comparable in magnitude with the axial energy itself, so the beam spreads
out along its axis by an amount comparable to the separation between the
target and the TOF axis. The size of the aperture which admits ions from the
storage region into the spectrometer must clearly be much smaller than this
to maintain a uniform extraction field, particularly if a slit is placed
between
the target and the TOF axis. This further reduces the sensitivity.
In delayed extraction MALDI in the usual axial geometry,
i.e. not the orthogonal configuration shown, acceptance is nearly complete,
and while the largest velocity spread is along the TOF axis, the well-defined
target-ylane perpendicular to the TOF axis allows a combination of time-lag
focusing (delayed extraction with optimized values of delay and applied
voltage) and electrostatic focusing (optimized value of the reflector voltage)
CA 02227806 1998-O1-23
-12-
in an ion mirror to produce resolution well above 10,000 in some cases.
Experiments carried out by the present inventors suggest
that competitive resolution could not be obtained with an acceptable signal
using orthogonal injection, unless collisional cooling according to the
present invention is employed. Moreover, some disadvantages of
delayed-extraction MALDI -- the dependence of optimum extraction
conditions on mass, and the more complex calibration required -- are still
present in orthogonal injection MALDI without cooling although to a lesser
extent i:han with axial injection.
The introduction of an RF quadrupole or other multipole
with collisional cooling of the ions between the MALDI target and the
orthogonal injection geometry avoids the problems described above while
offering additional advantages. These are detailed below with reference to
the remaining figures.
By reducing the radial energies of the ions, an
approximately parallel beam can be produced, greatly reducing the losses
that re~~ult from collimation before the ions enter the storage region. This
allows the use of a larger entrance aperture to the TOF spectrometer 50,
further reducing losses. By reducing the axial energies of the ions, and then
reaccel,erating them to a uniform energy, the mass discrimination
mentioned above is not present.
The uniform energy distributions of the ions after cooling
remove any mass dependence on the optimum extraction conditions and
allow tlhe simple quadratic relation between TOF and mass to be used for
calibration with two calibrant peaks. Figure 3 shows a spectrum of an
equimolar mixture of several peptides and proteins from mass 726 to 5734
Da in an a-cyano-4-hydroxy cinnamic acid matrix. The spectrum was
acquired in a single run and shows uniform mass resolution (M/~M~HM)
of about 5000 throughout the mass range. Using a simple external
calibration with substance P and melittin, the mass determination for each
of the molecular ions is accurate within about 30 ppm. Here, the peaks for
the various substances are identified as: peak 60 for Leucine-enkephalin;
CA 02227806 1998-O1-23
-13-
peak 61 for substance P; peak 62 for Melittin; peak 63 for CD fragment 25-28;
and peak 64 for insulin. All peaks are identified both on the overall
spectrum and as an enlarged partial spectrum. The resolution
demon;~trated in Figure 3 is rather close to the resolution obtainable with
the same instrument using an ESI source. In the present embodiment, the
entrance orifice was made slightly larger than normally used in ESI,
approximately 1 mm diameter as compared to a normal diameter of around
1 /3 mrn, to make adjustments easier in the preliminary experiments. This
does not appear to have been necessary so it is reasonable to expect
improved resolution is expected if a smaller orifice is used. Resolution up
to 10,000 has been obtained with ESI ions in the same instrument.
The decreasing relative intensity of the molecular ions
with mass is to some extent a reflection of the decreasing detection
efficiency with increasing mass. Detection efficiency depends strongly on
velocity, which decreases with mass for singly-charged ions at a given
energy. In this embodiment the energy of singly-charged ions is only about 5
keV (compared to 30 keV in typical MALDI experiments), so the detection
efficiency limits the practical range of application to less than about 6000
Da.
The relative intensities of the molecular ion peaks in Fig. 3 is consistent
with that observed from the same sample when analyzed in a conventional
MALDI experiment using 5 kV acceleration. The detection efficiency in the
present embodiment can be increased by increasing the voltage which
accelerates the ions into the spectrometer, or by increasing the voltage on
the detE~ctor.
As mentioned above, the collisional cooling spreads the
ions out along the ion beam axis changing the initial beam pulsed at 13 Hz
into a duasi-continuous beam. This is illustrated in Figure 4 which shows
the count rate as a function of time after the laser pulse; i.e. the
distribution
of transit times through the spectrometer. The width of the time
distribution is on the order of 20 ms which represents an increase in the
time spread by a factor of at least 10~ as each laser pulse is about 2 ns
long.
Dispersion along the axis is a disadvantage in orthogonal-injection MALDI
CA 02227806 1998-O1-23
-14-
without cooling, but with the present invention, since optimum extraction
conditions do not depend on the time delay after the laser shot, multiple
injection pulses into the TOF storage region 48 can be used for each laser
shot. In the present embodiment, 256 injection pulses into the TOF storage
region ~I8 were used for every laser shot. The losses are then determined by
the duty cycle of the instrument which in this case is about 20%. The duty
cycle is the percentage of the time that ions can be injected from the storage
region into the TOF spectrometer; here, it effectively means the fraction of
the tirrte, the TOF storage region 48 is available to accept ions. A
quasi-continuous beam is in fact an advantage in this mode of operation.
Approximately 104 to 106 ions are ejected from the target probe with every
laser shot at a repetition rate of 13 Hz, but as a result of spreading along
the
beam axis or ion path 27 (and some losses) approximately 2 to 5 ions are
injected into the instrument with every injection pulse less than one ion
on average of a particular species. This allows single-ion counting to be used
with a TDC (Time to Digital Converter), which makes the combination of
high timing resolution (0.5 ns) and high repetition rate (essential for
maximum duty cycle) technically much simpler than using a transient
recorder which is necessary in conventional MALDI experiments. In
addition, the use of single-ion counting eliminates problems with detector
shadowing from intense matrix peaks, and problems with peak saturation
which require attention in conventional MALDI because of the strong
dependence of the signal on laser fluence and the shot-to-shot variation.
Finally, single-ion counting places much more modest demands on the
detector and amplifier time resolution because the electronic reduction and
digitization of the pulse is quite insensitive to the detector pulse shape.
In Figure 4, four graphs are shown of the count rate
against time, for leucine-enkephalin shown at 70, substance P shown at 72,
Melittin shown at 74 and insulin shown at 76. Additionally, for each of
these substances, graphs or spectra 71, 73, 75 and 77, are inserted showing
normal TOF spectra, similar to Figure 3.
Assuming 104 ions of a single molecular ion species are
CA 02227806 1998-O1-23
-15-
produced with each laser shot, the transmission efficiency of the
RF-quadrupole is in the range of 10 %. Taking account of the duty cycle,
about ;? % of the ions produced at the target are detected in the mass
spectrometer. This represents significant losses compared to the
conventional axial MALDI experiment in which transmission is probably
50% or more. However, from the point of view of data rate, the losses can be
compensated to a large extent by the higher repetition rate and higher
fluence of the laser. In these experiments, the repetition rate was 13 Hz, but
can easily be increased to 20 Hz with the current laser, or in principle up to
at least 100 Hz before the counting system becomes saturated. In contrast,
the usual MALDI experiment is run at about 1 or 2 Hz. The laser fluence in
a conventional MALDI experiment must be kept close to threshold to
achieve the best performance, the threshold being the energy necessary to
cause vaporization of the sample. In the present invention, the laser
fluence can be increased to the fluence at which the ion production process
saturates. As the quadrupole serves to smooth out the ion burst produced
by the laser, a short intense burst of ions can be accepted. From the point of
view of absolute sensitivity, it seems that the independence of the spectrum
on laser conditions (see below) allows more efficient usage of the sample
deposited on the target. Using fluence several times higher than threshold
produces ions until the matrix is completely removed from the target probe.
Figure 5 shows that the practical sensitivity achieved with substance P is in
the sane range as that obtainable with conventional MALDI. Five
femtom.oles of substance P were applied to the target using 4HCCA as the
matrix. The left hand side of the spectrum is indicated at 80, and the right
hand side is shown enlarged by a factor of 44 as indicated at 81. A portion of
this spectrum is shown enlarged at 82 showing the molecular ion (MH+) .
Figure 6 shows the spectrum 85 obtained from a tryptic
digest of citrate synthase again showing the uniform mass resolution over
the mass range; the inset 8b shows the spectrum obtained from 20 (moles
applied to the target.
These results indicate that the performance of the
CA 02227806 1998-O1-23
-16-
inventi~~n for peptides is comparable to conventional MALDI experiments
but wii:h the advantage of a mass-independent calibration, and a simple
calibration procedure. However, the most important advantages result from
the nearly complete decoupling of the ion production from the mass
measurement. In a conventional MALDI experiment, the location of the
laser s~~ot on the target and the laser fluence and location must be carefully
selected for optimum performance, and these conditions are typically
different for different matrices and even for different target preparation
methodLs. The situation was improved with the introduction of delayed
extraction but even so, many commercial instruments have implemented
software to adjust laser fluence, detector gain, and laser position, and to
reject shots in which saturation occurs. None of these techniques are
necessary with the present invention.The performance obtained shows no
dependence on target or laser conditions. The laser is simply set to
maximum fluence (several times the usual threshold) and left while the
target i;> moved to a fresh position occasionally. This means that alternative
targets can easily be tried (including insulating targets), and alternative
lasers with different wavelengths or pulse widths can be used.
The decoupling of the ion production from the mass
measurement also provides an opportunity to perform various
manipulations of the ions after ejection but before mass measurement. One
example is parent ion selection and subsequent fragmentation (MS/MS).
This is most suitably done with an additional quadrupole mass filter as
described below, but even in the present embodiment of Figure 2, some
selectivity and fragmentation is possible.
Figures 7A, 7B and 7C show three different modes of
operation of the instrument shown in Figure 2. The reference numerals of
Figure 2 are provided along the z axis to indicate correspondence between
potential level and the different elements of the apparatus. Voltages for the
quadru~pole sections 31, 32 are indicated respectively at Ul (t) and U2 (t).
Figure 7A shows the simple collisional ion guide mode
that was used in obtaining the results shown in Figures 4-6. Here the same
CA 02227806 1998-O1-23
-17-
amplitudes of RF voltage and no DC offset voltages are applied to different
sections of the quadrupole. Potential differences in the longitudinal
direction are kept small to minimize fragmentation due to CID.
Figure 7B shows a mass filtering mode, which is
analogous to the same filtering mode implemented in conventional
quadrupole mass filters. Here a DC offset voltage V is added to the first
section of the quadrupole to select an ion of interest, while the second
section again acts as an ordinary ion guide since there is no CID because of
the smell potential difference between the sections. The amplitude of the
voltage applied in the second quadrupole section 32 is only one third of the
voltage applied in the first section 31.
Figure 7C is an MS-MS mode which differs from the
mode of Figure 7B by a higher potential difference between the quadrupole
sections> 31, 32, so ions are accelerated in that region and enter the second
section with high kinetic energy, the additional energy being indicated as 0
collision energy. In that case the second section acts as an collision cell
and
parent ions are decomposed there by collisions with the buffer gas (CID).
Again, the amplitude of the RF voltage in the second section is only one
third oi= the amplitude of the RF voltage in the first section, which allows
daughter ions much lighter than the parent ions to have stable trajectories
and to he transmitted through the second quadrupole.
Figure 8 shows examples of the spectra obtained in the
different modes illustrated in Figure 7, and in particular gives an example
of possible beam manipulation. All the spectra were acquired using the
same initial sample.
Figure 8A is a mass spectrum where ions were cooled in a
collisional focusing ion guide (the mode of Figure 7A).
Figure 8B is an example where ions of interest were
selectedL in the first quadrupole 31 and cooled in the second quadrupole 32
section (the mode of Figure 7B). Once ions of interest have been selected,
they can be used for fragmentation in CID to obtain detailed information on
composition and structure.
CA 02227806 1998-O1-23
-18-
Figure 8C presents an MS/MS spectrum of substance P
obtained in this way. Molecular ions if substance P are selected in the first
quadrupole section and fragmented by collisions in the second quadrupole
section (according to the mode of Figure 7C). The potential difference, 0
collision energy, between the first and second quadrupoles was 100V. The
intensities of the fragment ions were small in comparison with intensity of
the primary ion so the region inside dotted lines is expanded by a factor of
56. Figure 8B shows the spectrum obtained in the same mode but where the
potenti~~l difference between the quadrupoles 31, 32 was 150 V. In this case,
more fragment ions are observed and the parent ion peak is substantially
reduced.
Figure 9 shows how long a signal from the same spot on a
MALDI: target can last. In this experiment, a given spot was irradiated by a
series of shots from the laser, running at 13 Hz. The laser intensity was two
or three times the "threshold" intensity. On average the sample lasted for
about one minute. The shape of the curve suggests that the laser shots dig
deeper and deeper into the sample until it is exhausted. At that point the
laser irradiates the metallic substate, so no signal is observed.
In the past it has not been possible to use both continuous
sources, such as electrospray ionization (ESI), and pulsed sources, such as
MALDI, in the same instrument, which would have significant advantages.
To the inventor's knowledge, the only successful ESI-TOF instruments to
date have been the orthogonal injection spectrometers (by the present
inventors, Dodonov, and now the commercial PerSeptive machines), so it
appear; that orthogonal injection is necessary for ESI-TOF, with or without
collisional damping, although the former improves the situation, as
detailed in the earlier application. Up to now, the only attempts to put
MALDI_ on an orthogonal injection instrument have been made by the
present inventors (first) and Guilhaus, both without collisional damping
and neither giving very promising results. The present invention enables
two such sources to be available in one instrument. Here, the MALDI probe
11 in Figure 2 can be replaced by an ESI source to enable measurement of ESI
CA 02227806 1998-O1-23
-19-
spectra in the instrument. The instrument would then be essentially the
same as the one illustrated in the earlier patent application noted above.
This change could of course be carried out by actually taking off one source
and replacing it by the other, but a number of more convenient
arrangements can be provided.
For instance Figure 10 shows a further embodiment
where i:he electrospray ion source 94 is attached to the input of a
collisional
damping interface 92, including a quadrupole, or other multipole, rod set
93. A MALDI ion source 94 is introduced on a probe 95 that enters from the
side, arid can be displaced in and out; for this purpose, a shaft end 96 is
slidingly and sealingly fitted into the housing of the collisional interface
92.
The M~~LDI ion source 94 is similar to the one shown in Figure 2 except in
this case the sample is deposited onto a flat surface machined on the side of
the probe shaft 95, instead of onto the end of a cylindrical probe. The
sample is irradiated by a laser with corresponding optics, generally
indicatE~d at 97, and ions are transmitted to a spectrometer indicated at 98.
When the ESI source is operating, shaft 96 is pulled out far enough to clear
the path of the ESI ions. When the MALDI ion source 94 is operating the
shaft 96 is inserted back so the MALDI target 94 is in the central position.
Presently, MALDI and ESI techniques are often
considered to be complementary methods for biochemical analysis, so many
biochennical or pharmaceutical laboratories have two instruments in use.
Obviously there are significant benefits of combining both ion sources in
one instrument, as in the embodiments above. In particular, the cost of a
combined instrument is expected to be little more than half the cost of two
separai:e instruments. In addition. similar procedures for ion
manipulation, detection and mass calibration could be used, since the ion
production is largely decoupled from the ion measurement. This would
simplify the analysis and processing of the separate spectra and their
comparison.
The ability the use both MALDI and ESI sources on a
single instrument is not restricted to the spectrometer shown in Figure 1,
CA 02227806 1998-O1-23
-20-
but is applicable to any mass spectrometer with a collisional damping
interface. In particular it is applicable to the QqTOF instrument discussed
above.
While specific embodiments of the invention have been
described, it will be appreciated that a number of variations are possible
within the scope of the present invention. Thus, the apparatus could
include a single multipole rod set as shown in Figure 1, or two rod sets as
shown in Figure 2. While quadrupole rod sets are preferred, other rod sets,
such as hexapole and octopole are possible, and the rod set can be selected
based on the known characteristics of the different rod sets. Additionally, it
is possible that three or more rod sets could be provided. Further, while
Figure :? shows the two rod sets, 31 and 32 provided in a common chamber,
the rod sets could, in known manner, be provided in separate chambers
operating at different pressures, to enable different operations to be
preformed. Thus, to perform conventional mass selection, there could be
one chamber operating at a very low pressure so that there is little or no
collisional activity between the ions and the damping gas. Further, the
pressure of the gas could be varied, between different chambers, to meet the
requirements for collisional damping, where a relatively large number of
collisions are desired as opposed to collision induced fragmentation, where
excessi~Te collisions are not desirable.
