A Beginner's Guide To ICP-MS: Tutorial

Download as pdf or txt
Download as pdf or txt
You are on page 1of 79

S P E C T R O S C O P Y

TUTORIAL

A Beginners Guide to ICP-MS


Part I

ROBERT THOMAS

mazingly, 18 years after the com- sive than ICP-OES and GFAA, why hasnt a novice reader. There is no question in

A mercialization of inductively cou-


pled plasma mass spectrometry
(ICP-MS), less than 4000 systems
have been installed worldwide. If
you compare this number with another
rapid multielement technique, inductively
coupled plasma optical emission spec-
it been more widely accepted by the ana-
lytical community? I firmly believe that
the major reason why ICP-MS has not
gained the popularity of the other trace
element techniques is that it is still con-
sidered a complicated research tech-
nique, requiring a very skilled person to
my mind that the technique needs to be
presented in a more user-friendly way to
make routine analytical laboratories more
comfortable with it. Unfortunately, the
publishers of the for Dummies series of
books have not yet found a mass (excuse
the pun) market for writing one on ICP-
trometry (ICP-OES), first commercial- operate it. Manufacturers of ICP-MS MS. So until that time, we will be present-
ized in 1974, the difference is quite signif- equipment are constantly striving to ing a number of short tutorials on the
icant. In 1992, 18 years after ICP-OES was make the systems easier to operate, the technique, as a follow-up to the poster
introduced, more than 9000 units had software easier to use, and the hardware that was included in the February 2001
been sold, and if you compare it with the easier to maintain, but even after 18 years issue of Spectroscopy.
same time period that ICP-MS has been it is still not perceived as a mature, rou- During the next few months, we will be
available, the difference is even more dra- tine tool like flame AA or ICP-OES. This discussing the following topics in greater
matic. From 1983 to the present day, might be partially true because of the rel- depth:
more than 17,000 ICP-OES systems have ative complexity of the instrumentation; principles of ion formation
been installed more than four times however, in my opinion, the dominant sample introduction
the number of ICP-MS systems. If the reason for this misconception is that plasma torch/radio frequency genera-
comparison is made with all atomic spec- there has not been good literature avail- tor
troscopy instrumentation (ICP-MS, ICP- able explaining the basic principles and interface region
OES, graphite furnace atomic absorption benefits of ICP-MS in a way that is com- ion focusing
[GFAA] and flame atomic absorption pelling and easy to understand for some- mass separation
[FAA]), the annual turnover for ICP-MS one with very little knowledge of the ion detection
is less than 7% of the total atomic spec- technique. Some excellent textbooks (1, sampling accessories
troscopy market 400 units compared 2) and numerous journal papers (35) applications.
to approximately 6000 atomic spec- are available that describe the fundamen- We hope that by the end of this series,
troscopy systems. Its even more surpris- tals, but they tend to be far too heavy for we will have demystified ICP-MS, made it
ing when you consider that ICP-MS of-
fers so much more than the other
techniques, including two of its most at-
tractive features the rapid multiele-
ment capabilities of ICP-OES, combined
with the superb detection limits of GFAA.

ICP-MS ROUTINE OR RESEARCH?


Clearly, one of the reasons is price an
ICP-MS system typically costs twice as
much as an ICP-OES system and three
times more than a GFAA system. But in a
competitive world, the street price of an
ICP-MS system is much closer to a top-of-
the-line ICP-OES system fitted with sam-
pling accessories or a GFAA system that
has all the bells and whistles on it. So if
Figure 1. Generation of positively charged ions in the plasma.
ICP-MS is not significantly more expen-
38 SPECTROSCOPY 16(4) APRIL 2001
w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

Figure 2. Simplified schematic of a chromium ground-state atom Figure 3. Conversion of a chromium ground-state atom (Cr0) to an
(Cr0). ion (Cr1).

a little more compelling to purchase, and GENERATION OF IONS IN THE PLASMA typically in liquid form, is pumped into
ultimately opened up its potential as a Well start this series off with a brief de- the sample introduction system, which is
routine tool to the vast majority of the scription of the fundamental principle made up of a spray chamber and nebu-
trace element community that has not yet used in ICP-MS the use of a high- lizer. It emerges as an aerosol and eventu-
realized the full benefits of its capabilities. temperature plasma discharge to gener- ally finds its way by way of a sample in-
ate positively charged ions. The sample, jector into the base of the plasma. As it
travels through the different heating
zones of the plasma torch it is dried, va-
porized, atomized, and ionized. During
this time, the sample is transformed from
a liquid aerosol to solid particles, then
into a gas. When it finally arrives at the
analytical zone of the plasma, at approxi-
mately 60007000 K, it exists as excited
atoms and ions, representing the elemen-
tal composition of the sample.
The excitation of the outer electron of
a ground-state atom, to produce
wavelength-specific photons of light, is
the fundamental basis of atomic emission.
However, there is also enough energy in
the plasma to remove an electron from its
orbital to generate an ion. It is the genera-
tion, transportation, and detection of sig-
nificant numbers of these positively
charged ions that give ICP-MS its charac-
teristic ultratrace detection capabilities.
It is also important to mention that,
although ICP-MS is predominantly used
for the detection of positive ions, negative
ions (such as halogens) are also pro-
duced in the plasma. However, because
the extraction and transportation of nega-
tive ions is different from that of positive
ions, most commercial instruments are
not designed to measure them. The
process of the generation of positively
charged ions in the plasma is shown con-
ceptually in greater detail in Figure 1.

40 SPECTROSCOPY 16(4) APRIL 2001


Circle 28 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

Table I. Breakdown of the atomic structure of copper isotopes.

63Cu 65Cu

Number of protons (p1) 29 29


Number of electrons (e2) 29 29
Number of neutrons (n) 34 36
Atomic mass (p1 1 n) 63 65
Atomic number (p1) 29 29
Natural abundance 69.17% 30.83%
Nominal atomic weight 63.55*

* Calculated using the formulae 0.6917n 1 0.3083n 1 p1 (referenced to the


atomic weight of carbon)

Figure 4. Mass spectra of the two copper isotopes 63Cu1 and


65Cu1.

Figure 5. Relative abundance of the naturally occurring isotopes of all the elements (6). Reproduced with the permission of PerkinElmer
Instruments (Norwalk, CT).

APRIL 2001 16(4) SPECTROSCOPY 41


............... SPECTROSCOPY TUTORIAL ................

ION FORMATION plasma, two different ions of copper,


Figures 2 and 3 show the actual process 63Cu1 and 65Cu1, are produced, which
of conversion of a neutral ground-state generate different mass spectra one at
atom to a positively charged ion. Figure 2 mass 63 and another at mass 65. This can
shows a very simplistic view of the be seen in Figure 4, which is an actual
chromium atom Cr0, consisting of a nu- ICP-MS spectral scan of a sample contain-
cleus with 24 protons (p1) and 28 neu- ing copper. It shows a peak for the 63Cu1
trons (n), surrounded by 24 orbiting elec- ion on the left, which is 69.17% abundant,
trons (e2) (It must be emphasized that and a peak for 65Cu1 at 30.83% abun-
this is not meant to be an accurate repre- dance, on the right. You can also see
sentation of the electrons shells and sub- small peaks for two Zn isotopes at mass
shells, but simply a conceptual explana- 64 (64Zn) and mass 66 (66Zn) (Zn has a to-
tion for the purpose of clarity). From this tal of five isotopes at masses 64, 66, 67,
we can say that the atomic number of 68, and 70). In fact, most elements have
chromium is 24 (number of protons), and at least two or three isotopes and many
its atomic mass is 52 (number of protons elements, including zinc and lead, have
1 neutrons). four or more isotopes. Figure 5 is a chart
If energy is then applied to the that shows the relative abundance of the
chromium ground-state atom in the form naturally occurring isotopes of all the
of heat from a plasma discharge, one of elements.
the orbiting electrons will be stripped off During the next few months, we will
the outer shell. This will result in only 23 systematically take you on a journey
electrons left orbiting the nucleus. Be- through the hardware of an ICP mass
cause the atom has lost a negative charge spectrometer, explaining how each major
(e2) but still has 24 protons (p1) in the component works, and finishing the se-
nucleus, it is converted into an ion with a ries with an overview of how the tech-
net positive charge. It still has an atomic nique is being used to solve real-world ap-
mass of 52 and an atomic number of 24, plication problems. Our goal is to present
but is now a positively charged ion and both the basic principles and benefits of
not a neutral ground-state atom. This the technique in a way that is clear, con-
process is shown in Figure 3. cise, and very easy to understand. We
hope that by the end of the series, you
NATURAL ISOTOPES and your managers will be in a better po-
This is a very basic look at the process, sition to realize the enormous benefits
because most elements occur in more that ICP-MS can bring to your laboratory.
than one form (isotope). In fact,
chromium has four naturally occurring REFERENCES
isotopes, which means that the (1) A. Montasser, Inductively Coupled Plasma
chromium atom exists in four different Mass Spectrometry (Wiley-VCH, Berlin,
forms, all with the same atomic number 1998).
of 24 (number of protons), but with differ- (2) F. Adams, R. Gijbels, and R. Van Grieken,
Inorganic Mass Spectrometry (John Wiley
ent atomic masses (numbers of neu-
and Sons, New York, 1988.).
trons). (3) R.S. Houk, V. A. Fassel, and H.J. Svec, Dy-
To make this a little easier to under- namic Mass Spectrom. 6, 234 (1981).
stand, lets take a closer look at an ele- (4) A.R. Date and A.L. Gray, Analyst 106,
ment like copper, which has only two dif- 1255 (1981).
ferent isotopes one with an atomic (5) D.J. Douglas and J.B. French, Anal.
mass of 63 (63Cu) and the other with an Chem. 53, 37 (1982).
atomic mass of 65 (65Cu). They both have (6) Isotopic Composition of the Elements: Pure
the same number of protons and elec- Applied Chemistry 63(7), 9911002
trons, but differ in the number of neu- (1991).
trons in the nucleus. The natural abun-
dances of 63Cu and 65Cu are 69.1% and Robert Thomas is the principal of his own
30.9%, respectively, which gives copper a freelance writing and scientific consulting
nominal atomic mass of 63.55 the company, Scientific Solutions, based in
value you see for copper in atomic weight Gaithersburg, MD. He can be contacted by
reference tables. Details of the atomic email at thomasrj@bellatlantic.net or via
structure of the two copper isotopes are his web site at www.scientificsolutions1.
shown in Table I. com.
When a sample containing naturally oc-
curring copper is introduced into the
Circle 29
42 SPECTROSCOPY 16(4) APRIL 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
S P E C T R O S C O P Y
TUTORIAL

A Beginners Guide to ICP-MS


Part II: The Sample-Introduction System

ROBERT THOMAS

Part II of Robert Thomas series on induc-


tively coupled plasma mass spectrometry
looks at one of the most critical areas of the
instrument the sample introduction sys-
tem. He discusses the fundamental princi-
ples of converting a liquid into a fine-
droplet aerosol suitable for ionization in
the plasma, and provides an overview of
the different types of commercially avail-
able nebulizers and spray chambers.

he majority of inductively coupled

T plasma mass spectrometry (ICP-


MS) applications involve the analysis
of liquid samples. Even though spec-
troscopists adapted the technique
over the years to handle solids, it was de-
veloped in the early 1980s primarily to an-
alyze solutions. There are many ways of
introducing a liquid into an ICP mass
spectrometer, but they all basically Figure 1. ICP-MS system diagram showing the location of the sample introduction area.
achieve the same result they generate
a fine aerosol of the sample so it can be generation using a nebulizer and droplet gas flow (~1 L/min) smashing the liquid
efficiently ionized in the plasma dis- selection by way of a spray chamber. into tiny droplets, which is very similar to
charge. The sample-introduction area has Sharp carried out a thorough investiga- the spray mechanism of a can of deodor-
been called the Achilles heel of ICP-MS tion of both processes (2). ant. Although pumping the sample is the
because it is considered the weakest most common approach to introducing it,
component of the instrument, with only AEROSOL GENERATION some pneumatic nebulizers, such as the
12% of the sample finding its way into As mentioned previously, the main func- concentric design, dont need a pump be-
the plasma (1). Although there has re- tion of the sample introduction system is cause they rely on the natural venturi ef-
cently been much improvement in this to generate a fine aerosol of the sample. It fect of the positive pressure of the nebu-
area, the fundamental design of an ICP- achieves this purpose with a nebulizer lizer gas to suck the sample through the
MS sample introduction system has not and a spray chamber. The sample is nor- tubing. Solution nebulization is conceptu-
dramatically changed since the technique mally pumped at ~1 mL/min via a peri- ally represented in Figure 3, which shows
was first introduced in 1983. staltic pump into the nebulizer. A peri- aerosol generation using a nebulizer with
Before discussing the mechanics of staltic pump is a small pump with lots of a crossflow design.
aerosol generation in greater detail, let us minirollers that rotate at the same speed.
look at the basic components of a sample The constant motion and pressure of the DROPLET SELECTION
introduction system. Figure 1 shows the rollers on the pump tubing feed the sam- Because the plasma discharge is ineffi-
proximity of the sample introduction area ple to the nebulizer. The benefit of a peri- cient at dissociating large droplets, the
relative to the rest of the ICP mass spec- staltic pump is that it ensures a constant spray chambers function is primarily to
trometer, while Figure 2 represents the flow of liquid, irrespective of differences allow only the small droplets to enter the
individual components. in viscosity between samples, standards, plasma. Its secondary purpose is to
The mechanism of introducing a liquid and blanks. After the sample enters the smooth out pulses that occur during the
sample into analytical plasma can be con- nebulizer, the liquid is broken up into a nebulization process, due mainly to the
sidered as two separate events aerosol fine aerosol by the pneumatic action of peristaltic pump. Several ways exist to en-
56 SPECTROSCOPY 16(5) MAY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

sure only the small droplets get through,


but the most common way is to use a
double-pass spray chamber where the
aerosol emerges from the nebulizer and
is directed into a central tube running the
whole length of the chamber. The
droplets travel the length of this tube,
where the large droplets (greater than
~10 m in diameter) fall out by gravity
and exit through the drain tube at the end
of the spray chamber. The fine droplets
(~510 m in diameter) then pass be- Figure 3. Conceptual representation of
tween the outer wall and the central tube, aerosol generation with an ICP-MS
where they eventually emerge from the nebulizer.
spray chamber and are transported into
the sample injector of the plasma torch Figure 2. Diagram of the ICP-MS sample
(3). Although many different designs are introduction area.
available, the spray chambers main func-
tion is to allow only the smallest droplets
into the plasma for dissociation, atomiza- (4). Therefore, general-purpose ICP-OES
tion, and finally ionization of the samples nebulizers that are designed to aspirate
elemental components. Figure 4 presents 12% dissolved solids, or high-solids neb-
a simplified schematic of this process. ulizers such as the Babbington, V-groove,
Let us now look at the different nebu- or cone-spray nebulizers, which are de-
lizer and spray chamber designs that are signed to handle as much as 20% dis-
most commonly used in ICP-MS. This ar- solved solids, are not ideal for use with
ticle cannot cover every type available be- ICP-MS. The most common of the pneu-
cause a huge market has developed over matic nebulizers used in commercial ICP
the past few years for application-specific mass spectrometers are the concentric Figure 4. Simplified representation of the
customized sample introduction compo- and crossflow designs. The concentric de- separation of large and fine droplets in the
nents. This market created an industry of sign is more suitable for clean samples, spray chamber.
small OEM (original equipment manufac- while the crossflow is generally more tol-
turers) companies that manufacture parts erant to samples containing higher levels
for instrument companies as well as sell- of solids or particulate matter. common with crossflow nebulizers,
ing directly to ICP-MS users. Concentric design. In the concentric neb- forced through the tube with a peristaltic
ulizer, the solution is introduced through pump. In either case, contact between the
NEBULIZERS a capillary tube to a low-pressure region high-speed gas and the liquid stream
By far the most common design used for created by a gas flowing rapidly past the causes the liquid to break up into an
ICP-MS is the pneumatic nebulizer, end of the capillary. The low pressure and aerosol. Crossflow nebulizers are gener-
which uses mechanical forces of a gas high-speed gas combine to break up the ally not as efficient as concentric nebuliz-
flow (normally argon at a pressure of solution into an aerosol, which forms at ers at creating the very small droplets
2030 psi) to generate the sample the open end of the nebulizer tip. Figure needed for ICP-MS analyses. However,
aerosol. The most popular designs of 5 illustrates the concentric design. the larger diameter liquid capillary and
pneumatic nebulizers include concentric, Concentric pneumatic nebulizers can longer distance between liquid and gas
microconcentric, microflow, and cross- provide excellent sensitivity and stability, injectors reduce clogging problems.
flow. They are usually made from glass, particularly with clean solutions. How- Many analysts feel that the small penalty
but other nebulizer materials, such as ever, the small orifices can be plagued by paid in analytical sensitivity and precision
various kinds of polymers, are becoming blockage problems, especially if large when compared with concentric nebuliz-
more popular, particularly for highly cor- numbers of heavy matrix samples are ers is compensated by the fact that the
rosive samples and specialized applica- aspirated. crossflow design is far more rugged for
tions. I want to emphasize at this point Crossflow design. For samples that con- routine use. Figure 6 shows a cross sec-
that nebulizers designed for use with ICP- tain a heavier matrix or small amounts of tion of a crossflow nebulizer.
optical emission spectroscopy (OES) are undissolved matter, the crossflow design Microflow design. A new breed of nebu-
not recommended for ICP-MS. This fact is probably the best option. With this de- lizers is being developed for ICP-MS
results from a limitation in total dissolved sign the argon gas is directed at right an- called microflow nebulizers, which are
solids (TDS) that can be put into the ICP- gles to the tip of a capillary tube, in con- designed to operate at much lower sam-
MS interface area. Because the orifice trast to the concentric design, where the ple flows. While conventional nebulizers
sizes of the sampler and skimmer cones gas flow is parallel to the capillary. The have a sample uptake rate of about
used in ICP-MS are so small (~0.61.2 solution is either drawn up through the 1 mL/min, microflow nebulizers typically
mm), the concentration of matrix compo- capillary tube via the pressure created by run at less than 0.1 mL/min. They are
nents must generally be kept below 0.2% the high-speed gas flow or, as is most based on the concentric principle, but
MAY 2001 16(5) SPECTROSCOPY 57
............................... SPECTROSCOPY TUTORIAL ..............................

Figure 5. Diagram of a typical concentric nebulizer. Figure 6. Schematic of a crossflow nebulizer.

constructed from polymer materials such common, with the cyclonic type gaining
as polytetrafluoroethylene (PTFE), per- in popularity. Another type of spray cham-
fluoroalkoxy (PFA), or polyvinylidene flu- ber based on the impact bead design
oride (PVDF). In fact, their excellent cor- (first developed for flame AA and then
rosion resistance means that they have adapted for ICP-OES) was tried on the
naturally low blank levels. This character- early ICP-MS systems with limited suc-
istic, together with their ability to handle cess, but is not generally used today. As
small sample volumes such as vapor- mentioned earlier, the function of the
Figure 7. A typical concentric microflow phase decomposition (VPD) applications, spray chamber is to reject the larger
nebulizer. Printed with permission from Ele- makes them an ideal choice for semicon- aerosol droplets and also to smooth out
mental Scientific (Omaha, NE). ductor labs that are carrying out ultra- pulses produced by the peristaltic pump.
trace element analysis (5). A typical mi- In addition, some ICP-MS spray cham-
croflow nebulizer made from PFA is bers are externally cooled (typically to
they usually operate at higher gas pres- shown in Figure 7. 25 C) for thermal stability of the sam-
sure to accommodate the lower sample ple and to minimize the amount of solvent
flow rates. The extremely low uptake rate SPRAY CHAMBERS going into the plasma. This can have a
makes them ideal for applications with Let us now turn our attention to spray number of beneficial effects, depending
limited sample volume or where the sam- chambers. Basically two designs are used on the application, but the main benefits
ple or analyte is prone to sample intro- in commercial ICP-MS instrumentation are reduction of oxide species and the
duction memory effects. These nebuliz- double pass and cyclonic spray cham- ability to aspirate volatile organic
ers and their components are typically bers. The double pass is by far the most solvents.

Figure 9. Schematic of a cyclonic spray chamber (shown with


concentric nebulizer).

Figure 8. Schematic of a Scott double-pass spray chamber (shown


with crossflow nebulizer). Printed with permission of PerkinElmer
Instruments (Norwalk, CT).
58 SPECTROSCOPY 16(5) MAY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

Double pass. By far the most common with the gas stream into the ICP-MS, tutorial at the end of this series.
design of double-pass spray chamber is while the larger droplets impinge on the
the Scott design, which selects the small walls and fall out through the drain. It is REFERENCES
droplets by directing the aerosol into a generally accepted that a cyclonic spray (1) R. A. Browner and A.W. Boorn, Anal.
central tube. The larger droplets emerge chamber has a higher sampling effi- Chem. 56, 786798A (1984).
from the tube and, by gravity, exit the ciency, which, for clean samples, trans- (2) B.L. Sharp, Analytical Atomic Spectrome-
spray chamber via a drain tube. The liq- lates into higher sensitivity and lower de- try 3, 613 (1980).
(3) L.C. Bates and J.W. Olesik, Journal of An-
uid in the drain tube is kept at positive tection limits. However, the droplet size
alytical Atomic Spectrometry 5(3), 239
pressure (usually by way of a loop), distribution appears to be different from a (1990).
which forces the small droplets back be- double-pass design, and for certain types (4) R.S. Houk, Anal. Chem. 56, 97A (1986).
tween the outer wall and the central tube, of samples, can give slightly inferior pre- (5) E. Debrah, S. A. Beres, T.J. Gluodennis,
where they emerge from the spray cham- cision. An excellent evaluation of the ca- R.J. Thomas, and E.R. Denoyer, Atomic
ber into the sample injector of the plasma pabilities of a cyclonic spray chamber was Spectroscopy, 197202 (September 1995).
torch. Scott double-pass spray chambers made by Beres and co-workers (6). Fig- (6) S. A. Beres, P. H. Bruckner, and E.R. De-
come in a variety of shapes, sizes, and ure 9 shows a cyclonic spray chamber noyer, Atomic Spectroscopy, 9699
materials, but are generally considered connected to a concentric nebulizer. (March/April 1994).
the most rugged design for routine use. Many other nonstandard sample intro-
Figure 8 shows a Scott spray chamber duction devices are available that are not Robert Thomas is the principal of his own
made of a polysulfide-type material, cou- described in this particular tutorial, such freelance writing and scientific marketing
pled to a crossflow nebulizer. as ultrasonic nebulization, membrane de- consulting company, Scientific Solutions,
Cyclonic spray chamber. The cyclonic solvation, flow injection, direct injection, based in Gaithersburg, MD. He specializes
spray chamber operates by centrifugal electrothermal vaporization, and laser ab- in trace element analysis and can be con-
force. Droplets are discriminated accord- lation. However, they are becoming more tacted by e-mail at thomasrj@bellatlantic.
ing to their size by means of a vortex pro- and more important, particularly as ICP- net or via his web site at www.
duced by the tangential flow of the sam- MS users are demanding higher perfor- scientificsolutions1.com.
ple aerosol and argon gas inside the mance and more flexibility. For that rea-
chamber. Smaller droplets are carried son, they will be addressed in a separate

News Spectrum continued from page 13


TRAINING COURSES
Thermo Nicolet (Madison, WI) is offering a
free Spring 2001 Spectroscopic Solutions
Seminar Series. The seminars will cover
basic FT-IR spectroscopy, microspec-
troscopy, dispersive Raman microscopy,
Raman spectroscopy, and specialized
sampling techniques. This years sched-
ule includes the following seminars:
May 22 at the Syracuse Sheraton,
Syracuse, NY; May 24 at the Wynd-
ham Westborough Hotel, Westbor-
ough, MA; May 24 at the Marriott
Oak Brook, Oak Brook, IL; June 5 at
the East Lansing Marriott, East Lans-
ing, MI; June 5 at the Embassy Suites,
Overland Park, KS; June 7 at the Sher-
aton Indianapolis, Indianapolis, IN;
June 12 at the Delta Meadowvale
Conference Center, Mississauga,
Ontario, Canada; June 12 at the Em-
bassy Suites, Brookfield, WI; July 10
at the Coast Terrace Inn, Edmonton,
Alberta, Canada; and July 26 at the Ala
Moana Hotel, Honolulu, HI.
For more information, contact Thermo
Nicolet, (800) 201-8132 fax: (608) 273-
5046, e-mail: nicinfo@thermonicolet.
com, web site: www.thermonicolet.
com.

Circle 51
60 SPECTROSCOPY 16(5) MAY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
S P E C T R O S C O P Y
TUTORIAL

A Beginners Guide to ICP-MS


Part III: The Plasma Source

ROBERT THOMAS

Part III of Robert Thomas series on induc-


tively coupled plasmamass spectroscopy MS
( ICP-MS) looks at the area where the ions interface
Ion
are generated the plasma discharge. He detector
gives a brief historical perspective of some
of the common analytical plasmas used ICP torch
Ion optics
over the years and discusses the compo- Mass separation Spray
nents that are used to create the ICP. He device chamber
finishes by explaining the fundamental
principles of formation of a plasma dis-
charge and how it is used to convert the Nebulizer
sample aerosol into a stream of positively Radio frequency
power supply
charged ions.

nductively coupled plasmas are by far

I
Turbomolecular Turbomolecular
the most common type of plasma pump pump
sources used in todays commercial Mechanical
pump
ICPoptical emission spectrometry
(OES) and ICP-MS instrumentation.
However, it wasnt always that way. In the
early days, when researchers were at- Figure 1. Schematic of an ICP-MS system showing the location of the plasma torch and radio
tempting to find the ideal plasma source frequency (RF) power supply.
to use for spectrometric studies, it was
unclear which approach would prove to were better understood, the technique ing aspiration of liquid samples. For these
be the most successful. In addition to became more accepted. In fact, for those reasons, they have had limited success as
ICPs, some of the other novel plasma who want a DCP excitation source cou- an emission source, because they are not
sources developed were direct current pled with an optical emission instrument considered robust enough for the analysis
plasmas (DCP) and microwave-induced today, an Echelle-based grating using a of real-world, solution-based samples.
plasmas (MIP). A DCP is formed when a solid-state detector is commercially However, they have gained acceptance as
gas (usually argon) is introduced into a available (2). an ion source for mass spectrometry (3)
high current flowing between two or Limitations in the DCP approach led to and also as emission-based detectors for
three electrodes. Ionization of the gas the development of electrodeless plasma, gas chromatography.
produces an inverted Y-shaped plasma. of which the MIP was the simplest form. Because of the limitations of the DCP
Unfortunately, early DCP instrumenta- In this system, microwave energy (typi- and MIP approaches, ICPs became the
tion was prone to interference effects and cally 100200 W) is supplied to the plasma dominant focus of research for both opti-
also had some usability and reliability gas from an excitation cavity around a cal emission and mass spectrometric
problems. For these reasons, the tech- glass or quartz tube. The plasma dis- studies. As early as 1964, Greenfield and
nique never became widely accepted by charge in the form of a ring is generated co-workers reported that an atmospheric-
the analytical community (1). However, inside the tube. Unfortunately, even pressure ICP coupled with OES could be
its one major benefit was that it could as- though the discharge achieves a very used for elemental analysis (4). Although
pirate high levels of dissolved or sus- high power density, the high excitation crude by todays standards, the system
pended solids, because there was no re- temperatures exist only along a central fil- showed the enormous possibilities of the
strictive sample injector for the solid ament. The bulk of the MIP never gets ICP as an excitation source and most defi-
material to block. This feature alone hotter than 20003000 K, which means it nitely opened the door in the early 1980s
made it attractive for some laboratories, is prone to very severe matrix effects. In to the even more exciting potential of us-
and once the initial limitations of DCPs addition, they are easily extinguished dur- ing the ICP to generate ions (5).
26 SPECTROSCOPY 16(6) JUNE 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

Plasma Auxiliary (a) (b) (c)


Interface Outer Quartz
gas gas
tube
Plasma Middle torch
tube

Nebulizer gas
Sample injector Electromagnetic
RF Load field High voltage
coil RF power Tangential flow
coil spark
of argon gas
Figure 2. Detailed view of a plasma torch Sample introduced
(d) (e)
through sample injector
and RF coil relative to the ICP-MS interface.

Figure 3. (right) Schematic of an ICP torch


and load coil showing how the inductively
coupled plasma is formed. (a) A tangential
flow of argon gas is passed between the
outer and middle tube of the quartz torch. Collision-induced Formation of inductively
(b) RF power is applied to the load coil, ionization of argon coupled plasma
producing an intense electromagnetic field.
(c) A high-voltage spark produces free elec-
trons. (d) Free electrons are accelerated by THE PLASMA TORCH nents that are used to generate the
the RF field, causing collisions and ioniza- Before we take a look at the fundamental source: a plasma torch, a radio frequency
tion of the argon gas. (e) The ICP is formed principles behind the creation of an in- (RF) coil, and RF power supply. Figure 1
at the open end of the quartz torch. The ductively coupled plasma used in ICP- shows their proximity to the rest of the
sample is introduced into the plasma via MS, let us take a look at the basic compo- instrument; Figure 2 is a more detailed
the sample injector. view of the plasma torch and RF coil rela-
tive to the MS interface.
The plasma torch consists of three con-
centric tubes, which are usually made
from quartz. In Figure 2, these are shown
as the outer tube, middle tube, and sam-
ple injector. The torch can either be one-
piece with all three tubes connected, or it
can be a demountable design in which
the tubes and the sample injector are sep-
arate. The gas (usually argon) used to
form the plasma (plasma gas) is passed
between the outer and middle tubes at a
flow rate of ;1217 L/min. A second gas
flow, the auxiliary gas, passes between
the middle tube and the sample injector
at ;1 L/min and is used to change the
position of the base of the plasma relative
to the tube and the injector. A third gas
flow, the nebulizer gas, also flowing at
;1 L/min carries the sample, in the form
of a fine-droplet aerosol, from the sample
introduction system (for details, see Part
II of this series: Spectroscopy 16[5],
5660 [2001]) and physically punches a
channel through the center of the plasma.
The sample injector is often made from
materials other than quartz, such as alu-
mina, platinum, and sapphire, if highly
corrosive materials need to be analyzed.
It is worth mentioning that although ar-
gon is the most suitable gas to use for all
three flows, there are analytical benefits
in using other gas mixtures, especially in
the nebulizer flow (6). The plasma torch
Circle 21
28 SPECTROSCOPY 16(6) JUNE 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

6500 K 7500 K 8000 K Preheating


6000 K zone

Droplet (Desolvation) Solid (Vaporization) Gas (Atomization) Atom (Ionization) Ion

M(H2O)1 X2 (MX)n MX M M1

Initial From sample injector To mass spectrometer


Normal radiation
analytical 10,000 K zone
zone

Figure 4. Different temperature zones in the plasma. Figure 5. Mechanism of conversion of a droplet to a positive ion in the ICP.

is mounted horizontally and positioned induced ionization of the argon continues cause it uses slightly lower power, this
centrally in the RF coil, approximately in a chain reaction, breaking down the might be considered advantageous
1020 mm from the interface. It must be gas into argon atoms, argon ions, and when it comes to long-term use of the
emphasized that the coil used in an ICP- electrons, forming what is known as an generator.
MS plasma is slightly different from the inductively coupled plasma discharge. The more important consideration is
one used in ICP-OES. In all plasmas, The ICP discharge is then sustained the coupling efficiency of the RF genera-
there is a potential difference of a few within the torch and load coil as RF en- tor to the coil. The majority of modern
hundred volts produced by capacitive ergy is continually transferred to it solid-state RF generators are on the order
coupling between the RF coil and the through the inductive coupling process. of 7075% efficient, meaning that 7075%
plasma. In an ICP mass spectrometer, The sample aerosol is then introduced of the delivered power actually makes it
this would result in a secondary dis- into the plasma through a third tube into the plasma. This wasnt always the
charge between the plasma and the inter- called the sample injector. This whole case, and some of the older vacuum
face cone, which could negatively affect process is conceptionally shown in tubedesigned generators were notori-
the performance of the instrument. To Figure 3. ously inefficient; some of them experi-
compensate for this, the coil must be enced more than a 50% power loss. An-
grounded to keep the interface region as THE FUNCTION OF THE RF GENERATOR other important criterion to consider is
close to zero potential as possible. I will Although the principles of an RF power the way the matching network compen-
discuss the full implications of this in supply have not changed since the work sates for changes in impedance (a mater-
greater detail in Part IV of this series. of Greenfield (4), the components have ials resistance to the flow of an electric
become significantly smaller. Some of the current) produced by the samples matrix
FORMATION OF AN ICP DISCHARGE early generators that used nitrogen or air components or differences in solvent
Let us now discuss the mechanism of for- required 510 kW of power to sustain the volatility. In older crystal-controlled gen-
mation of the plasma discharge. First, a plasma discharge and literally took up erators, this was usually done with servo-
tangential (spiral) flow of argon gas is di- half the room. Most of todays generators driven capacitors. They worked very well
rected between the outer and middle tube use solid-state electronic components, with most sample types, but because they
of a quartz torch. A load coil, usually cop- which means that vacuum power ampli- were mechanical devices, they struggled
per, surrounds the top end of the torch fier tubes are no longer required. This to compensate for very rapid impedance
and is connected to a radio frequency makes modern instruments significantly changes produced by some samples. As a
generator. When RF power (typically smaller and, because vacuum tubes were result, the plasma was easily extin-
7501500 W, depending on the sample) is notoriously unreliable and unstable, far guished, particularly during aspiration of
applied to the load coil, an alternating more suitable for routine operation. volatile organic solvents.
current oscillates within the coil at a rate As mentioned previously, two frequen- These problems were partially over-
corresponding to the frequency of the cies have typically been used for ICP RF come by the use of free-running RF gen-
generator. In most ICP generators this generators: 27 and 40 MHz. These fre- erators, in which the matching network
frequency is either 27 or 40 MHz. This quencies have been set aside specifically was based on electronic tuning of small
RF oscillation of the current in the coil for RF applications of this kind, so they changes in frequency brought about by
causes an intense electromagnetic field to will not interfere with other communica- the sample solvent or matrix components.
be created in the area at the top of the tion-based frequencies. The early RF gen- The major benefit of this approach was
torch. With argon gas flowing through erators used 27 MHz, while the more re- that compensation for impedance
the torch, a high-voltage spark is applied cent designs favor 40 MHz. There changes was virtually instantaneous be-
to the gas, which causes some electrons appears to be no significant analytical ad- cause there were no moving parts. This
to be stripped from their argon atoms. vantage of one type over the other. How- allowed for the successful analysis of
These electrons, which are caught up and ever, it is worth mentioning that the 40- many sample types that would probably
accelerated in the magnetic field, then MHz design typically runs at lower power have extinguished the plasma of a
collide with other argon atoms, stripping levels, which produces lower signal inten- crystal-controlled generator.
off still more electrons. This collision- sity and reduced background levels. Be-
JUNE 2001 16(6) SPECTROSCOPY 29
............................... SPECTROSCOPY TUTORIAL ..............................

IONIZATION OF THE SAMPLE As mentioned previously, the sample zone before it eventually becomes a posi-
To better understand what happens to the aerosol enters the injector via the spray tively charged ion in the analytical zone.
sample on its journey through the plasma chamber. When it exits the sample injec- To explain this in a very simplistic way,
source, it is important to understand the tor, it is moving at such a velocity that it lets assume that the element exists as a
different heating zones within the dis- physically punches a hole through the trace metal salt in solution. The first step
charge. Figure 4 shows a cross-sectional center of the plasma discharge. It then that takes place is desolvation of the
representation of the discharge along goes through a number of physical droplet. With the water molecules
with the approximate temperatures for changes, starting at the preheating zone stripped away, it then becomes a very
different regions of the plasma. and continuing through the radiation small solid particle. As the sample moves
further into the plasma, the solid particle
changes first into a gaseous form and
then into a ground-state atom. The final
process of conversion of an atom to an
ion is achieved mainly by collisions of en-
ergetic argon electrons (and to a lesser
extent by argon ions) with the ground-
state atom (7). The ion then emerges
from the plasma and is directed into the
interface of the mass spectrometer (for
details on the mechanisms of ion genera-
tion, please refer to Part I of this series:
Spectroscopy 16[4], 3842 [2001]). This
process of conversion of droplets into
ions is represented in Figure 5.
The next installment of this series will
focus on probably the most crucial area
of an ICP mass spectrometer the inter-
face region where the ions generated
in the atmospheric plasma have to be
sampled with consistency and electrical
integrity by the mass spectrometer,
which is under extremely high vacuum.

REFERENCES
(1) A.L. Gray, Analyst 100, 289299 (1975).
(2) G.N. Coleman, D.E. Miller, and R.W.
Stark, Am. Lab. 30(4), 33R (1998).
(3) D.J. Douglas and J.B. French, Anal.
Chem. 53, 37-41 (1981).
(4) S. Greenfield, I.L. Jones, and C.T. Berry,
Analyst 89, 713720 (1964).
(5) R.S. Houk, V. A. Fassel, and H.J. Svec,
Dyn. Mass Spectrom. 6, 234 (1981).
(6) J.W. Lam and J.W. McLaren, J. Anal.
Atom. Spectom. 5, 419424 (1990).
(7) T. Hasegawa and H. Haraguchi, ICPs in
Analytical Atomic Spectrometry, A.
Montaser and D.W. Golightly, Eds., 2d ed.
(VCH, New York, 1992).

Robert Thomas is the principal of his own


freelance writing and scientific marketing
consulting company, Scientific Solutions,
based in Gaithersburg, MD. He specializes
in trace-element analysis and can be con-
tacted by e-mail at thomasrj@bellatlantic.
net or via his web site at www.
scientificsolutions1.com.

Circle 22
30 SPECTROSCOPY 16(6) JUNE 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
S P E C T R O S C O P Y
TUTORIAL

A Beginners Guide to ICP-MS


Part IV: The Interface Region

ROBERT THOMAS

he interface region is probably the sampling process works, which will give of 0.81.2 mm. From there they travel a

T most critical area of the whole induc-


tively coupled plasma mass spec-
trometry (ICP-MS) system. It cer-
tainly gave the early pioneers of the
technique the most problems to over-
come. Although we take all the benefits
of ICP-MS for granted, the process of tak-
readers an insight into the many prob-
lems faced by the early researchers.

SAMPLING THE IONS


Figure 1 shows the proximity of the inter-
face region to the rest of the instrument.
The role of the interface is to transport
short distance to the skimmer cone,
which is generally sharper than the sam-
pler cone and has a much smaller orifice
(0.40.8 mm i.d.). Both cones are usually
made of nickel, but they can be made of
materials such as platinum that are far
more tolerant to corrosive liquids. To re-
ing a liquid sample, generating an aerosol the ions efficiently, consistently, and with duce the effects of the high-temperature
that is suitable for ionization in the electrical integrity from the plasma, plasma on the cones, the interface hous-
plasma, and then sampling a representa- which is at atmospheric pressure (760 ing is water-cooled and made from a ma-
tive number of analyte ions, transporting Torr), to the mass spectrometer analyzer terial that dissipates heat easily, such as
them through the interface, focusing region, which is at approximately 1026 copper or aluminum. The ions then
them via the ion optics into the mass Torr. One first achieves this by directing emerge from the skimmer cone, where
spectrometer, finally ending up with de- the ions into the interface region. The in- they are directed through the ion optics,
tection and conversion to an electronic terface consists of two metallic cones and finally are guided into the mass sepa-
signal, are not trivial tasks. Each part of with very small orifices, which are main- ration device. Figure 2 shows the inter-
the journey has its own unique problems tained at a vacuum of ;2 Torr with a me- face region in greater detail; Figure 3
to overcome but probably the most chal- chanical roughing pump. After the ions shows a close-up of the sampler and
lenging is the movement of the ions from are generated in the plasma, they pass skimmer cones.
the plasma to the mass spectrometer. through the first cone, known as the sam-
Lets begin by explaining how the ion- pler cone, which has an orifice diameter CAPACITIVE COUPLING
This process sounds fairly straight-
forward but proved very problematic dur-
ing the early development of ICP-MS be-
MS
interface cause of an undesired electrostatic
Ion (capacitive) coupling between the load
detector coil and the plasma discharge, producing
ICP torch a potential difference of 100200 V. Al-
Ion optics though this potential is a physical charac-
Mass separation teristic of all inductively coupled plasma
device Spray
chamber discharges, it is particularly serious in an
ICP mass spectrometer because the ca-
Nebulizer pacitive coupling creates an electrical dis-
Radio
charge between the plasma and the sam-
power supply pler cone. This discharge, commonly
called the pinch effect or secondary dis-
charge, shows itself as arcing in the re-
Turbomolecular Turbomolecular gion where the plasma is in contact with
pump pump the sampler cone (1). This process is
Mechanical
pump shown very simplistically in Figure 4.
If not taken care of, this arcing can
cause all kinds of problems, including an
increase in doubly charged interfering
Figure 1. Schematic of an inductively coupled plasma mass spectrometry (ICP-MS) system, species, a wide kinetic energy spread of
showing the proximity of the interface region. sampled ions, formation of ions gener-
26 SPECTROSCOPY 16(7) JULY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

Interface Secondary
housing discharge

Sample Plasma
Skimmer cone torch
cone

Ion Sampler
optics cone

~2 Torr
vacuum

RF coil
Skimmer
cone Figure 4. Interface area affected by sec-
ondary discharge.
Figure 2. Detailed view of the interface Figure 3. Close-up view of the sampler and
region. skimmer cones. (Courtesy PerkinElmer
Instruments, Norwalk, CT.)

ated from the sampler cone, and a de- not be overestimated with respect to its ef- tering the mass spectrometer (typically
creased orifice lifetime. These problems fect on the kinetic energy of the ions being 2040 eV), which makes ion focusing far
were reported by many of the early re- sampled. It is well documented that the en- more complicated (8).
searchers of the technique (2, 3). In fact, ergy spread of the ions entering the mass
because the arcing increased with sam- spectrometer must be as low as possible to BENEFITS OF A WELL-DESIGNED
pler cone orifice size, the source of the ensure that they can all be focused effi- INTERFACE
secondary discharge was originally ciently and with full electrical integrity by The benefits of a well-designed interface
thought to be the result of an electro-gas- the ion optics and the mass separation de- are not readily obvious if simple aqueous
dynamic effect, which produced an in- vice. When the ions emerge from the ar- samples are analyzed using only one set
crease in electron density at the orifice gon plasma, they will all have different ki- of operating conditions. However, it be-
(4). After many experiments it was even- netic energies based on their mass-to- comes more apparent when many differ-
tually realized that the secondary dis- change ratio. Their velocities should all be ent sample types are being analyzed, re-
charge was a result of electrostatic cou- similar because they are controlled by quiring different operating parameters. A
pling of the load coil to the plasma. The rapid expansion of the bulk plasma, which true test of the design of the interface oc-
problem was first eliminated by ground- will be neutral as long as it is maintained at curs when plasma conditions need to be
ing the induction coil at the center, which zero potential. As the ion beam passes changed, when the sample matrix
had the effect of reducing the radio fre- through the sampler cone into the skim- changes, or when a dry sample aerosol is
quency (RF) potential to a few volts. This mer cone, expansion will take place, but its being introduced into the ICP-MS. Ana-
effect can be seen in Figure 5, taken from composition and integrity will be main- lytical scenarios like these have the po-
one of the early papers, which shows the tained, assuming the plasma is neutral. tential to induce a secondary discharge,
reduction in plasma potential as the coil is This can be seen in Figure 6. change the kinetic energy of the ions en-
grounded at different positions (turns) Electrodynamic forces do not play a tering the mass spectrometer, and affect
along its length. role as the ions enter the sampler or the the tuning of the ion optics. It is therefore
Originally, the grounding was imple- skimmer because the distance over
mented by attaching a physical ground- which the ions exert an influence on each
ing strap from the center turn of the coil other (known as the Debye length) is 100 n
to the interface housing. In todays instru- small (typically 10231024 mm) com- n
mentation the grounding is achieved in a pared with the diameter of the orifice 80
Plasma potential

number of different ways, depending on (0.51.0 mm) (8), as Figure 7 shows.


the design of the interface. Some of the It is therefore clear that maintaining a 60
most popular designs include balancing neutral plasma is of paramount impor- n
n
the oscillator inside the circuitry of the tance to guarantee electrical integrity of 40
RF generator (5); positioning a grounded the ion beam as it passes through the in- n n
shield or plate between the coil and the terface region. If a secondary discharge 20
plasma torch (6); or using two interlaced is present, it changes the electrical char-
n
coils where the RF fields go in opposing acteristics of the plasma, which will affect 0
directions (7). They all work differently the kinetic energy of the ions differently, 0.0 0.5 1.0 1.5 2.0 2.5 3.0
but achieve a similar result of reducing or depending on their mass. If the plasma is Load coil turns
eliminating the secondary discharge. at zero potential, the ion energy spread is
Figure 5. Reduction in plasma potential as
in the order of 510 eV. However, if a sec-
ION KINETIC ENERGY ondary discharge is present, it results in the load coil is grounded at different posi-
The impact of a secondary discharge can- a much wider spread of ion energies en- tions (turns) along its length.

JULY 2001 16(7) SPECTROSCOPY 27


............................... SPECTROSCOPY TUTORIAL ..............................

Skimmer Sampler

Ion optics

Interface Debye Orifice


cone length diameter

34 Torr 760 Torr


Expansion of
ion beam Sample
Ion beam aerosol

Figure 7. Electrodynamic forces do not af-


Figure 6. The composition of the ion beam is maintained, assuming a neutral plasma.
fect the composition of the ion beam enter-
ing the sampler or the skimmer cone.
critical that the interface grounding such as 40Ar16O, 40Ar, and 38ArH, in the
mechanism can handle these types of determination of difficult elements like
real-world applications, of which typical 56Fe, 40Ca, and 39K. Such dramatic tem to reduce the solvent loading on the
examples include changes from normal operating condi- plasma. In addition, higher RF power
The use of cool-plasma conditions. It is tions (1000 W, 0.8 L/min) will affect the (13001500 W) and lower nebulizer gas
standard practice today to use cool- electrical characteristics of the plasma. flow (0.40.8 L/min) are required to dis-
plasma conditions (500700 W power and Running volatile organic solvents. Ana- sociate the organic components in the
1.01.3 L/min nebulizer gas flow) to lyzing oil or organic-based samples re- sample. A reduction in the amount of sol-
lower the plasma temperature and reduce quires a chilled spray chamber (typically vent entering the plasma combined with
argon-based polyatomic interferences 220 C) or a membrane desolvation sys- higher power and lower nebulizer gas
flow translate into a hotter plasma and a
change in its ionization mechanism.
Reducing oxides. The formation of ox-
ide species can be problematic in some
sample types. For example, in geochemi-
cal applications it is quite common to sac-
rifice sensitivity by lowering the nebulizer
gas flow and increasing the RF power to
reduce the formation of rare earth ox-
ides, which can interfere spectrally with
the determination of other analytes. Un-
fortunately these conditions have the po-
tential to induce a secondary discharge.
Running a dry plasma. Sampling
accessories such as membrane desolva-
tors, laser ablation systems, and elec-
trothermal vaporization devices are being
used more routinely to enhance the flexi-
bility of ICP-MS. The major difference be-
tween these sampling devices and a con-
ventional liquid sample introduction
system, is that they generate a dry sam-
ple aerosol, which requires totally differ-
ent operating conditions compared with a
conventional wet plasma. An aerosol
containing no solvent can have a dramatic
effect on the ionization conditions in the
plasma.
Even though most modern ICP-MS in-
terfaces have been designed to minimize
the effects of the secondary discharge, it

Tutorial continued on page 34

Circle 17
28 SPECTROSCOPY 16(7) JULY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
F O C U S O N
QUALITY

Validation of Spectrometry Software


Part II: Roles of the Validation Plan and User Requirements Specification

R.D. MCDOWALL

n this months column we will explore going operational, thus failing to gain Ill outline my philosophy and rationale

I why the user requirements specifica-


tion (URS) and the validation plan are
so important for the validation of spec-
trometry software, and well cover the
specification and system selection from a
software perspective.
In the first installment of this series, we
benefit from the investment in the instru-
ment or going live with no regulatory
coverage. It depends on your approach to
risk and if can you sleep at night.

THE WAY IT SHOULD BE


However, as we discussed in the previous
now and you, dear reader, can accept this
as is, modify it, or ignore it.
First, you should write the validation
plan as either the first or second docu-
ment in the life cycle; I advise writing it af-
ter the first or second draft of the URS to
incorporate any implementation or roll-out
looked at the system development life cy- article in this series, a proactive approach issues in the overall validation strategy.
cle (SDLC) and some validation concepts to validation is necessary and, if done The rationale for this approach is that the
(1). One concept was that validation is a right, will actually save you money by en- validation plan provides documented evi-
process that covers the entire system de- suring that you buy the right instrument dence of intent of the validation. The doc-
velopment life cycle: Once started, you for the job. So well start at the beginning ument will set out the overall strategy of
cant stop. Now we will look in more de- and look at the first stages of the life cycle: the validation and define the life cycle
tail at the first part of the SDLC. Defining and controlling the validation phases and the documented evidence that
throughout the whole life cycle (writ- will be produced in each phase. If you
THE WAY IT WAS ing the validation plan). leave writing the validation plan until later
In the past, the spectrometer and soft- Specifying what you want the system in the project, one or more phases of the
ware were purchased and then, just be- to do (writing a user requirements life cycle will have passed and you may
fore they were put into operational use, specification). need to write documents retrospectively.
someone thought about validation. Some Selecting the system using the require- Furthermore, youll be out of compliance
common questions may have been ments defined in the URS as the basis, with 21 CFR 11.10(k)(2), which requires
Have we validated the system? No. rather than the salesperson bought a time-sequenced audit trail of systems
Does it matter? Probably. me a good meal. documentation.
Will we get caught? Dont even think Defining and controlling the overall valida- Content of a validation plan. The purpose
about answering no to this question. tion. The validation plan is one name for of a validation plan is to provide docu-
Considering validation at such a late the document that controls the validation mentation of intent for the whole valida-
stage of the life cycle will mean a delay in effort for your spectrometer software. tion, including a definition of the life cycle
However, the name for this document used, documentation to be produced dur-
varies from laboratory to laboratory. It ing the each stage of the life cycle, and
may be called the validation plan, master roles and responsibilities of everyone in-
Bob McDowall validation plan, validation master plan, or volved in the project.
is a visiting senior quality plan. To provide a better perspective, the
lecturer in the Regardless of what you call this docu- content of a validation plan is listed in the
Department of ment in your organization, it should cover sidebar. It is based on the Institute of
Chemistry at the all the steps you are going to take to Electronic and Electrical Engineers
University of demonstrate the quality of the spectrome- (IEEE) standard for validation and verifi-
Surrey, principal try software in your laboratory. cation plans (2).
of McDowall Ideally the validation plan should be This document is important because it
written as early as possible in the life cy- defines what you will do in the validation,
Consulting
cle to define the overall steps that are re- and you will be judged against it when
(Bromley, UK), and Questions of Quality
quired as well as the documents to be pro- your operation is inspected. Therefore,
column editor of LCGC Europe,
duced during each phase of the life cycle. read and understand it well dont write
Spectroscopys sister magazine. Address There are different approaches to writing the plan and forget it, because what you
correspondence to him at 73 Murray validation plans, and the document can be plan does not always come to pass. Usu-
Avenue, Bromley, Kent BR1 3DJ, UK. written in several stages in the life cycle. ally deviations from the plan occur that
30 SPECTROSCOPY 16(7) JULY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
................................. FOCUS ON QUALITY .................................

youll need to record, such as ment. Ideally, the emphasis is


documents not written, new doc- Validation Plan Outline Format on the required functions and
uments required that have not not the method of implementa-
been specified, or parts of the Based on IEEE Std. 1012-1986 tion, as this may be the identifi-
life cycle omitted or modified. Purpose cation of a solution. The aim of
These changes will need to be What is the scope of the validation: What is the spectrom- a URS is to make a statement of
noted under the deviation proce- etry system to be validated? requirements rather than a
dure that you have in place in Reference documents statement of a potential solu-
the plan. Noting the changes Include references to any regulations, documents, guide- tion. This allows users to look
sounds like a pain, but once the lines, or internal policies and procedures that affect the objectively at software from dif-
principles are understood, it is validation of this spectrometer. ferent vendors and make an ob-
relatively simple to do. Definitions jective decision as to which sys-
Define key words and terms. tem is required.
DESIGN: THE URS Validation overview Nature of the URS. The URS
How much money have you Define the roles and responsibilities for the validation, and should address the following
wasted on purchasing spectrom- include both internal and external people involved. basic issues:
eters that were not fit for pur- Meaning of signatures: Why are you signing a document? Functionality: What is the sys-
pose, did not do the job you Cross-reference to the project plan for the schedule of tem or function supposed to do?
wanted, or used software that work. (This should be a separate document and can be External interfaces: How
was not up to snuff? From a busi- updated as the validation progresses or not, as the does the system interact with
ness perspective, a document case may be.) other systems and the users?
that says what you want the in- Life cycle validation Performance: What are the
strument and software to do will Define the system development life cycle that youll be us- speed, availability, and response
be beneficial, because youll ing for the validation. For each phase, state what the activ- time of the various functions of
have a better chance of selecting ities will be and what documented evidence you will be the system?
the right instrument and producing. Dont forget: some of this may be electronic, Attributes: What are the secu-
software. especially during the qualification phases. rity considerations of each
From a regulatory perspec- Validation reporting function?
tive, remember that the defini- Outline how the validation will be reported. Design constraints: Must the
tion of validation presented in Validation administration procedures system work on specific hard-
the first part of this series (1) in- State how change control for deviations, software bugs, ware or use an operating sys-
cluded that phrase predefined and so forth will be handled. The validation plan is a con- tem, and are these consistent
specifications. The document trolled document, so it must be paginated correctly (for ex- with your organizations
that provides the laboratory with ample, page X of Y), signed by the author, authorized by standards?
the predefined specifications for two others (technical and compliance/release reviews), Prioritization: All require-
the spectrometer and the soft- and distributed to specified individuals. ments are ranked for impor-
ware is the URS. Without this tance as either mandatory or
document or an equivalent, you desirable (respectively, you
cannot validate your spectrometer soft- veal omissions, misunderstandings, and must use the system, or it would sim-
ware, because you dont have a prede- inconsistencies in the specification. This ply be nice to have it).
fined specification and therefore there is means that they can be corrected easily The URS should form the basis of the
nothing to test against. This is particu- before you purchase the system. Finally, solution to be delivered by the selected
larly important when you consider which a well-written URS provides the input to vendor. If this does not happen, you can
electronic record and electronic signa- user acceptance test specifications and leave yourself open to a poor-quality prod-
ture functions are pertinent to define and qualification of the system. uct because either you dont know what
test for the way that you will use the General guidelines for a URS. A user re- you want the system to do or you cant ar-
instrument. quirements specification clearly and pre- ticulate this need to the vendor.
The URS provides the answer to the cisely defines what the customer (that is, Writing the specification. The following
question, What do you want the system to you) wants the system to do, and it guidelines should be followed during the
do? This makes the assumption that you should be understood by both the cus- production of the specification:
know what you want the system to do. tomer and the instrument vendor. The Each requirement statement should be
A well-written URS provides several URS is a living document and must be up- uniquely referenced and no longer
specific benefits. For one thing, it serves dated, via a change control procedure, than 250 words.
as a reference against which off-the-shelf throughout the computer system life cy- The URS should be consistent; there-
commercial products are selected and cle. After purchase, when you upgrade fore, requirement statements should
evaluated in detail and any enhancements the software, also update the URS to re- not be duplicated or contradicted.
are defined. Also, you are less likely to be flect the changes and new functions in The URS should express requirements
seduced by technology or buy a poor sys- the latest version. and not design solutions.
tem. Furthermore, the URS reduces the A URS defines the functions to be car- Each requirement should be testable
total system effort and costs, because ried out, the data on which the system (this allows the tests to be designed as
careful review of the document should re- will operate, and the operating environ- soon as the URS is finalized).
JULY 2001 16(7) SPECTROSCOPY 31
................................. FOCUS ON QUALITY .................................

uniquely numbered (not bad), short


Table I: Example Specification of Audit Trail Requirements for Spectrometer Software
(good), and prioritized (getting better).
However, 21 CFR 11 states that every
Requirement Spectrometer Data System Priority change must not overwrite the original
Number Feature Specification (M/D)* result and must include the name of the
5.1.01 The system software requires an audit trail to monitor user, along with date and time of the
the creation, modification and deletion of all electronic change. This is not mentioned in this
records generated and managed by the system. M specification (bummer!). So be careful,
5.1.02 The audit trail covers all acquisition, control, calibration, specify the system, and review it carefully
calculation, display, reporting, and export functions and or something essential may be missed.
includes all file handling options such as open, copy, edit,
rename, and delete. M SYSTEM SELECTION: PART ONE
5.1.03 The audit trail is able to support the system during normal Because your requirements for the over-
operation without an excessive system overhead or loss all system are contained in the URS, the
of performance. D document can be used as a basis to
5.1.04 The audit trail once invoked cannot be switched off. M design the tests to evaluate the various
5.1.05 Archival of electronic records will have an audit trail entry. M systems offered by vendors. Can the sys-
5.1.06 Selected portions of the audit trail must be made available tems offered meet your requirements, es-
either by printing or viewing. These partial audit trail reports pecially for the mandatory functions? Us-
must be made available in a portable electronic format for ing the URS requirements for system
use by regulatory agencies. D selection helps ensure that the system se-
5.1.07 The audit trails must be maintained for as long as the lected matches your business needs.
electronic records they correspond to exist. M Dont forget that the tests you use for
5.1.08 When a record is changed, all previous versions must be system selection should also include com-
readable or available for inspection. M mon problems that you know happen in
your laboratory. What happens when
*Mandatory or desirable samples are switched and you notice the
error only after the analysis? Can the sys-
Both customer and vendor must under- more time you spend in the specification tem handle the changes easily and with
stand the document; therefore, jargon and design phase getting your ideas and suitable audit trail entries?
should be avoided, and key words concepts right, the quicker the rest of the The system you select will be based on
should be defined in a specific section life cycle will go if you know what you the practical experience of using it in
in the document. want. You will get a spectrometer and as- your laboratory environment. However,
Requirements should be prioritized as sociated software that meet your require- before you sign on the dotted line, you
mandatory or desirable. ments more fully, and there will be less may want to make sure that the software
The URS should be modifiable, but chance later in the life cycle of finding out was developed in a quality manner
changes should be made under a for- that what looked good early on does not through a vendor audit.
mal control procedure. meet certain key requirements now.
A URS is correct if every requirement Contrast this to selecting a spectrome- VENDOR CERTIFICATES AND AUDITS
stated has only one interpretation and is ter with no user requirements. (This bit Many spectrometer vendors will be certi-
met by the system. Unfortunately, this is should be easy, because we have all done fied to ISO 9000 of some description and
rare. it.) will offer you a certificate that the system
Organizing requirements: Go with the work- Dont forget the instrument specs! In this conforms to its quality processes. This is
flow. A URS can be extensive unless you series well concentrate on the software fine, but please remember that no re-
plan well, so careful consideration should elements, but dont forget the instrument quirement for product quality exists in
be given to organizing requirements in itself. The software and the instrument any ISO 9000 schedule, and if you look at
the easiest manner to understand. The must be an integrated system. So, the in- the warranty of any software product,
best framework for writing a user re- strument specification also needs to be there is no guarantee that the software is
quirements specification for most spec- included in the overall URS. What operat- stated to be either fit for purpose or error
trometers is to follow the process or ing requirements do you need from the free. The certificates are fine, but if the
workflow that the data system will be spectrometer, such as mass range and system is critical to your operation, my
automating. Therefore, if you have resolution or wavelength? Get them down advice is to consider a vendor audit.
mapped the process, it makes an ideal in the URS. The vendor audit should take place
prompt for the URS because the require- A specific example. Table I shows an ex- once the product has been selected. The
ments can be defined against each activ- ample of what a URS could look like. It purpose is simply to see if the ISO 9000
ity in the process. defines the requirements for audit trail quality system is operated effectively. The
This idea of documenting what we functionality in the spectrometer software evaluation and audit process is a very im-
want in sufficient detail sounds great, but to meet Part 11 requirements. Looks im- portant part of the life cycle, because it
it means more work, doesnt it? Yes, that pressive, doesnt it? Look at the table and shows whether design, building, and test-
is true, but consider the benefits. The youll see that each requirement is ing stages (which are under the control

32 SPECTROSCOPY 16(7) JULY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m


.................. FOCUS ON QUALITY ....................

of the vendor) have been checked to en- SYSTEM SELECTION: PART TWO
sure compliance with the regulations. If the vendor audit, price quote, instru-
The audit should be planned and should ment, and software are all acceptable,
cover items such as the design and pro- youll be raising a capital expenditure
gramming phases, product testing and re- request (or whatever it is called in your
lease, documentation, and support; a re- organization) and then generating a pur-
port of the audit should be produced after chase order. The quote and the purchase
the visit. Two published articles have cov- order are a link in the validation chain;
ered vendor audits in more detail (3, 4). they provide a link into the next phase of
The minimum audit is a remote vendor the validation life cycle: qualification. The
audit using a checklist that the vendor purchase order is the first stage in defin-
completes and returns to you. This is ing the initial configuration of the system,
usually easy to complete, but the writer of as well discover in the next article in this
the checklist must ensure that the ques- series.
tions are written in a way that can be un-
derstood by the recipient, because lan- REFERENCES
guage and cultural issues could affect a (1) R.D. McDowall, Spectroscopy 16(2),
remote checklist. Moreover, there is little 3243 (2001).
way of checking the answers you receive. (2) IEEE Standard 1012-1986, Software Vali-
However, for smaller software systems dation and Verification Plans, Institute of
Electronic and Electrical Engineers, Pis-
and some spectrometers fall into this cat-
cataway, NJ, USA.
egory a remote audit is a cost-effective (3) R.D. McDowall, Sci. Data Mgmt. 2(2), 8
way of getting information on how a ven- (1998).
dor carries out its development process, (4) R.D. McDowall, Sci. Data Mgmt. 2(3), 8
so long as you know and understand its (1998).
limitations.

Tutorial continued from page 34


shouldnt be taken for granted that they REFERENCES
can all handle changes in operating con- (1) A.L. Gray and A.R. Date, Analyst 108,
ditions and matrix components with the 1033 (1983).
same amount of ease. The most notice- (2) R.S. Houk, V. A. Fassel, and H.J. Svec, Dy-
able problems that have been reported namic Mass Spectrosc. 6, 234 (1981).
(3) A.R. Date and A.L. Gray, Analyst 106,
include spectral peaks of the cone mater-
1255 (1981).
ial appearing in the blank (9); erosion or (4) A.L. Gray and A.R. Date, Dynamic Mass
discoloration of the sampling cones; Spectrosc. 6, 252 (1981).
widely different optimum plasma condi- (5) S.D. Tanner, J. Anal. At. Spectrom. 10,
tions for different masses (10); and in- 905 (1995).
creased frequency of tuning the ion op- (6) K. Sakata and K. Kawabata, Spectrochim.
tics (8). Of all these, probably the most Acta 49B, 1027 (1994).
inconvenient problem is regular opti- (7) S. Georgitus, M. Plantz, poster paper pre-
mization of the lens voltages, because sented at the Winter Conference on
slight changes in plasma conditions can Plasma Spectrochemistry, FP4, Fort
Lauderdale, FL (1996).
produce significant changes in ion ener-
(8) D.J. Douglas and J.B. French, Spec-
gies, which require regular retuning of trochim. Acta 41B, 3, 197 (1986).
the ion optics. Even though most instru- (9) D.J. Douglas, Can. J. Spectrosc. 34, 2
ments have computer-controlled ion op- (1089).
tics, it becomes another variable that (10) J.E. Fulford and D.J. Douglas, Appl. Spec-
must be optimized. This isnt a major trosc. 40, 7 (1986).
problem but might be considered an in-
convenience for a highsample through- Robert Thomas has more than 30 years ex-
put lab. There is no question that the perience in trace element analysis. He is the
plasma discharge, interface region, and principal of his own freelance writing and
ion optics all have to be designed in con- scientific consulting company, Scientific So-
cert to ensure that the instrument can lutions, based in Gaithersburg, MD. He can
handle a wide range of operating condi- be contacted by e-mail at thomasrj@
tions and sample types. The role of the bellatlantic.net or via his web site at
ion optics will be discussed in greater de- www.scientificsolutions1.com.
tail in the next installment of this series.
Circle 22, 23
34 SPECTROSCOPY 16(7) JULY 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
A Beginners Guide to ICP-MS
Part V: The Ion Focusing System

ROBERT THOMAS

Part V of the series on inductively coupled


plasmamass spectrometry (ICP-MS) takes
a detailed look at the ion focusing system
a crucial area of the ICP mass spec-
trometer where the ion beam is focused be-
fore it enters the mass analyzer. Sometimes
known as the ion optics, it comprises one or
more ion lens components, which electro-
statically steer the analyte ions from the in-
terface region into the mass separation de-
vice. The strength of a well-designed ion
focusing system is its ability to produce low
background levels, good detection limits,
and stable signals in real-world sample
matrices.

lthough the detection capability of

A ICP-MS is generally recognized as


being superior to any of the other
atomic spectroscopic techniques, it
is probably most susceptible to the
samples matrix components. The inher-
ent problem lies in the fact that ICP-MS is
relatively inefficient; out of every million
Figure 1. Position of ion optics relative to the plasma torch and interface region.

ROLE OF THE ION OPTICS


Figure 1 shows the position of the ion op-
in concert with each other. It is absolutely
critical that the composition and electrical
ions generated in the plasma, only one ac- tics relative to the plasma torch and inter- integrity of the ion beam is maintained as
tually reaches the detector. One of the face region; Figure 2 represents a more it enters the ion optics. For this reason it
main contributing factors to the low effi- detailed look at a typical ion focusing is essential that the plasma is at zero po-
ciency is the higher concentration of ma- system. tential to ensure that the magnitude and
trix elements compared with the analyte, The ion optics are positioned between spread of ion energies are as low as
which has the effect of defocusing the the skimmer cone and the mass separa- possible (2).
ions and altering the transmission charac- tion device, and consist of one or more A secondary but also very important
teristics of the ion beam. This is some- electrostatically controlled lens compo- role of the ion optic system is to stop par-
times referred to as a space charge ef- nents. They are not traditional optics that ticulates, neutral species, and photons
fect, and it can be particularly severe we associate with ICP emission or atomic from getting through to the mass ana-
when the matrix ions have a heavier mass absorption but are made up of a series of lyzer and the detector. These species
than the analyte ions (1). The role of the metallic plates, barrels, or cylinders that cause signal instability and contribute to
ion focusing system is therefore to trans- have a voltage placed on them. The func- background levels, which ultimately af-
port the maximum number of analyte tion of the ion optic system is to take ions fect the performance of the system. For
ions from the interface region to the from the hostile environment of the example, if photons or neutral species
mass separation device, while rejecting as plasma at atmospheric pressure via the reach the detector, they will elevate the
many of the matrix components and non- interface cones and steer them into the background noise and therefore degrade
analyte-based species as possible. Let us mass analyzer, which is under high vac- detection capability. In addition, if particu-
now discuss this process in greater detail. uum. As mentioned in Part IV of the se- lates from the matrix penetrate farther
ries, the plasma discharge, interface re- into the mass spectrometer region, they
gion, and ion optics have to be designed have the potential to deposit on lens com-
38 SPECTROSCOPY 16(9) SEPTEMBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
approach is to set the ion lens or mass an-
alyzer slightly off axis. The positively
charged ions are then steered by the lens
system into the mass analyzer, while the
photons and neutral and nonionic species
are ejected out of the ion beam (4).
It is also worth mentioning that some
lens systems incorporate an extraction
lens after the skimmer cone to electro-
statically pull the ions from the interface
region. This has the benefit of improving
the transmission and detection limits of
the low-mass elements (which tend to be
pushed out of the ion beam by the heav-
ier elements), resulting in a more uni-
form response across the full mass range
of 0300 amu. In an attempt to reduce
these space charge effects, some older
designs have used lens components to ac-
celerate the ions downstream. Unfortu-
nately this can have the effect of degrad-
Figure 2. A generic ion focusing system, showing position of ion optics relative to the inter- ing the resolving power and abundance
face cones and mass analyzer. sensitivity (ability to differentiate an ana-
lyte peak from the wing of an interfer-
ence) of the instrument because of the
much higher kinetic energy of the accel-
erated ions as they enter the mass
analyzer (5).

DYNAMICS OF ION FLOW


To fully understand the role of the ion op-
tics in ICP-MS, it is important to have an
appreciation of the dynamics of ion flow
from the plasma through the interface re-
gion into the mass spectrometer. When
the ions generated in the plasma emerge
from the skimmer cone, there is a rapid
expansion of the ion beam as the pres-
sure is reduced from 760 Torr (atmos-
pheric pressure) to approximately 1023 to
1024 Torr in the lens chamber with a
turbomolecular pump. The composition
of the ion beam immediately behind the
cone is the same as the composition in
front of the cone because the expansion
at this stage is controlled by normal gas
dynamics and not by electrodynamics.
One of the main reasons for this is that,
in the ion sampling process, the Debye
length (the distance over which ions ex-
Figure 3. Extreme pressure drop in the ion optic chamber produces diffusion of electrons, re-
ert influence on each other) is small com-
sulting in a positively charged ion beam.
pared with the orifice diameter of the
sampler or skimmer cone. Consequently
there is little electrical interaction be-
ponents and, in extreme cases, get into making it into the mass spectrometer. tween the ion beam and the cone and rel-
the mass analyzer. In the short term this The first method is to place a grounded atively little interaction between the indi-
will cause signal instability and, in the metal stop (disk) behind the skimmer vidual ions in the beam. In this way,
long term, increase the frequency of cone. This stop allows the ion beam to compositional integrity of the ion beam is
cleaning and routine maintenance. move around it but physically blocks the maintained throughout the interface re-
Basically two approaches will reduce particulates, photons, and neutral species gion (6). With this rapid drop in pressure
the chances of these undesirable species from traveling downstream (3). The other in the lens chamber, electrons diffuse out
40 SPECTROSCOPY 16(9) SEPTEMBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
in preference to ions with medium (mid-
mass elements) or low kinetic energy
(low-mass elements). This is shown in
Figure 4. The second stage of charge sep-
aration is therefore to electrostatically
steer the ions of interest back into the
center of the ion beam by placing volt-
ages on one or more ion lens compo-
nents. Remember, however, that this is
possible only if the interface is kept at
zero potential, which ensures a neutral
gas-dynamic flow through the interface
and maintains the compositional integrity
of the ion beam. It also guarantees that
the average ion energy and energy
spread of each ion entering the lens sys-
tems are at levels optimum for mass sepa-
ration. If the interface region is not
grounded correctly, stray capacitance will
generate a discharge between the plasma
and sampler cone and increase the ki-
Figure 4. The degree of ion repulsion will depend on kinetic energy of the ions: those with netic energy of the ion beam, making it
high kinetic energy (green with red +) will be transmitted in preference to ions with medium very difficult to optimize the ion lens
(yellow with red +) or low kinetic energy (blue with red +). system (7).

COMMERCIAL ION OPTIC DESIGNS


Over the years, there have been many dif-
ferent ion optic designs. Although they all
have their own characteristics, they per-
form the same basic function: to discrimi-
nate undesirable matrix- or solvent-based
ions so that only the analyte ions are
transmitted to the mass analyzer. The
most common ion optics design used to-
day consists of several lens components,
which all have a specific role to play in
the transmission of the analyte ions (8).
With these multicomponent lens sys-
tems, the voltage can be optimized on
every lens of the ion optics to achieve the
desired ion specificity. Over the years this
type of lens configuration has proven to
be very durable and has been shown to
produce very low background levels, par-
ticularly when combined with an off-axis
mass analyzer. However, because of the
Figure 5. Schematic of a multicomponent lens system with extraction lens and off-axis interactive nature of parameters that af-
quadrupole mass analyzer (courtesy of Agilent Technologies [Wilmington, DE]). fect the signal response, the more com-
plex the lens system the more variables
have to be optimized, particularly if many
of the ion beam. Because of the small size positive charge of the ion beam means different sample types are being ana-
of the electrons relative to the positively that there is now a natural tendency for lyzed. This isnt such a major problem be-
charged ions, the electrons diffuse far- the ions to repel each other. If nothing is cause the lens voltages are all computer-
ther from the beam than the ions, result- done to compensate for this, ions with a controlled, and methods can be stored for
ing in an ion beam with a net positive higher mass-to-charge ratio will dominate every new sample scenario. Figure 5 is a
charge. This is represented schematically the center of the ion beam and force the commercially available multicomponent
in Figure 3. lighter ions to the outside. The degree of lens system, with an extraction lens and
The generation of a positively charged loss will depend on the kinetic energy of off-axis quadrupole mass analyzer, show-
ion beam is the first stage in the charge the ions: those with high kinetic energy ing how the ion beam is deflected into the
separation process. Unfortunately, the net (high mass elements) will be transmitted mass analyzer, while the photons and

42 SPECTROSCOPY 16(9) SEPTEMBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m


next part of the series we will discuss the
heart of the ICP mass spectrometer: the
mass analyzer.

REFERENCES
(1) J. A. Olivares and R.S. Houk, Anal. Chem.
58, 20 (1986).
(2) D.J. Douglas and J.B. French, Spec-
trochim. Acta 41B(3), 197 (1986).
(3) E.R. Denoyer, D. Jacques, E. Debrah, and
S.D. Tanner, At. Spectrosc. 16(1), 1
(1995).
(4) D. Potter, American Lab (July 1994).
(5) P. Turner, paper presented at Second In-
ternational Conference on Plasma Source
Figure 7. A calibration of optimum lens volt- Mass Spec, Durham, UK, 1990.
Figure 6. Schematic of a single ion lens and ages is used to ramp-scan the ion lens in (6) S.D. Tanner, D.J. Douglas, and J.B.
grounded stop system (not to scale concert with the mass scan of the analyzer. French, Appl. Spectrosc. 48, 1373 (1994).
[courtesy of PerkinElmer Instruments The signals have been normalized for com- (7) R. Thomas, Spectroscopy 16(7), 2634
{Norwalk, CT}]). parison purposes. (2001).
(8) Y. Kishi, Agilent Technologies Application
Journal, August (1997).
neutral species travel in a straight line These all work slightly differently but (9) S.D. Tanner, L.M. Cousins, and D.J. Dou-
and strike a metal plate. share similar components. By using a glas, Appl. Spectrosc. 48, 1367 (1994).
Another, more novel approach is to use combination of slightly different cone (10) I.B. Brenner, M. Liezers, J. Godfrey, S.
just one cylindrical ion lens, combined geometry, higher vacuum at the inter- Nelms, and J. Cantle, Spectrochim. Acta
with a grounded stop positioned just in- face, one or more extraction lenses, and Part 53B(68), 1087 (1998).
side the skimmer cone as shown in slightly modified ion optic design, they of- (11) B.C. Gibson, presented at Surrey Interna-
Figure 6 (9). fer as much as 10 times the sensitivity of tional Conference on ICP-MS, London,
With this design, the voltage is dynam- a traditional interface (10). However, this UK, 1994.
ically ramped on-the-fly, in concert with increased sensitivity is usually combined (12) T. Howe, J. Shkolnik, and R.J. Thomas,
Spectroscopy 16(2), 54 (2001).
the mass scan of the analyzer (typically a with inferior stability and an increase in
quadrupole). The benefit of this approach background levels, particularly for sam-
is that the optimum lens voltage is placed ples with a heavy matrix. To get around Robert Thomas has more than 30 years ex-
on every mass in a multielement run to this degradation in performance one perience in trace element analysis. He is the
allow the maximum number of analyte must usually dilute the samples before principal of his own freelance writing and
ions through, while keeping the matrix analysis, which limits the systems applic- scientific consulting company, Scientific So-
ions to an absolute minimum. This is rep- ability for real-world samples (11). How- lutions, based in Gaithersburg, MD. He can
resented in Figure 7, which shows a lens ever, they have found a use in non-liquid- be contacted by e-mail at thomasrj@
voltage scan of six elements: lithium, based applications in which high bellatlantic.net or via his web site at
cobalt, yttrium, indium, lead, and ura- sensitivity is crucial, for example in the www.scientificsolutions1.com.
nium, at 7, 59, 89, 115, 208, and 238 amu, analysis of small spots on the surface of a
respectively. We can see that each ele- geological specimen using laser ablation
ment has its own optimum value, which is ICP-MS. For this application, the instru-
then used to calibrate the system, so the ment must offer high sensitivity because
lens can be ramp-scanned across the full a single laser pulse is used to ablate the
mass range. This type of approach is typi- sample and sweep a tiny amount of the
cally used in conjunction with a grounded dry sample aerosol into the ICP-MS (12).
stop to act as a physical barrier to reduce The role of the ion focusing system
particulates, neutral species, and photons cannot be overestimated. It affects the
from reaching the mass analyzer and de- background noise level of the instrument.
tector. Although this design produces It has a huge impact on both long- and
slightly higher background levels, it of- short-term signal stability, especially in
fers excellent long-term stability with real-world samples, and it also dictates
real-world samples. It works well for the number of ions that find their way to
many sample types but is most effective the mass analyzer. However, it must be
when low mass elements are being deter- emphasized that the ion optics are only as
mined in the presence of high-mass good as the ions that feed it, and for this
matrix elements. reason it must be designed in concert
It is also worth emphasizing that a with both the plasma source and the in-
number of ICP-MS systems offer what is terface region. There is no question that
known as a high-sensitivity interface. this area is crucial to the design of the
whole ICP mass spectrometer. In the
44 SPECTROSCOPY 16(9) SEPTEMBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
S P E C T R O S C O P Y
TUTORIAL

Beginners Guide to ICP-MS


Part VI The Mass Analyzer
ROBERT THOMAS

column. Lets first begin with the most


Ion MS common of the mass separation devices
detector interface
used in ICP-MS, the quadrupole mass
Ion optics filter.
ICP torch
Mass separation
device Spray QUADRUPOLE MASS FILTER
chamber TECHNOLOGY
Developed in the early 1980s,
Nebulizer quadrupole-based systems represent ap-
RF proximately 90% of all ICP mass spec-
power trometers used today. This design was
supply the first to be commercialized; as a result,
todays quadrupole ICP-MS technology is
Turbomolecular Turbomolecular considered a very mature, routine, high-
pump pump
Mechanical throughput, trace-element technique. A
pump quadrupole consists of four cylindrical or
hyperbolic metallic rods of the same
length and diameter. They are typically
Figure 1. Schematic of an ICP-MS system showing the location of the mass separation device.
made of stainless steel or molybdenum,
and sometimes have a ceramic coating
Part VI of the series on inductively coupled development of alternative mass separa- for corrosion resistance. Quadrupoles
plasmamass spectroscopy (ICP-MS) fun- tion devices that pushed the capabilities used in ICP-MS are typically 1520 cm in
damentals deals with the heart of the sys- of ICP-MS so it could be used for more length and about 1 cm in diameter and
tem the mass separation device. Some- challenging applications. Before we dis- operate at a frequency of 23 MHz.
times called the mass analyzer, it is the cuss these different mass spectrometers
region of the ICP mass spectrometer that in greater detail, lets take a look at the lo- BASIC PRINCIPLES OF OPERATION
separates the ions according to their mass- cation of the mass analyzer in relation to By placing a direct current (dc) field on
to-charge ratio (m/z). This selection the ion optics and the detector. Figure 1 one pair of rods and a radio frequency
process is achieved in a number of different shows this in greater detail. (rf) field on the opposite pair, ions of a se-
ways, depending on the mass separation As we can see, the mass analyzer is po- lected mass are allowed to pass through
device, but they all have one common goal: sitioned between the ion optics and the the rods to the detector, while the others
to separate the ions of interest from all the detector and is maintained at a vacuum of are ejected from the quadrupole. Figure 2
other nonanalyte, matrix, solvent, and approximately 1026 Torr with a second shows this in greater detail.
argon-based ions. turbomolecular pump. Assuming the ions In this simplified example, the analyte
are emerging from the ion optics at the ion (black) and four other ions (colored)
lthough inductively coupled optimum kinetic energy (1), they are have arrived at the entrance to the four

A plasmamass spectroscopy (ICP-


MS) was commercialized in 1983,
the first 10 years of its development
primarily used traditional quadru-
pole mass filter technology to separate
the ions of interest. These worked excep-
tionally well for most applications but
ready to be separated according to their
mass-to-charge ratio by the mass ana-
lyzer. There are basically four kinds of
commercially available mass analyzers:
quadrupole mass filters, double focusing
magnetic sector, time-of-flight, and
collisionreaction cell technology. They
rods of the quadrupole. When a particu-
lar rf-dc voltage is applied to the rods, the
positive or negative bias on the rods will
electrostatically steer the analyte ion of
interest down the middle of the four rods
to the end, where it will emerge and be
converted to an electrical pulse by the de-
proved to have limitations in determining all have their own strengths and weak- tector. The other ions of different mass-
difficult elements or dealing with more- nesses, which we will discuss in greater to-charge ratios will pass through the
complex sample matrices. This led to the detail in the next few installments of this spaces between the rods and be ejected
44 SPECTROSCOPY 16(10) OCTOBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............... SPECTROSCOPY TUTORIAL ...................

+
Ion
+ flow
Quadrupole rods
+ + +
+

+
+

Figure 2. Schematic showing principles of a quadrupole mass filter.

from the quadrupole. This scanning corresponding to the spectral peak of


process is then repeated for another ana- 63Cu. In a multielement run, repeated

lyte at a completely different mass-to- scans are made over the entire suite of
charge ratio until all the analytes in a mul- analyte masses, as opposed to just one
tielement analysis have been measured. mass represented in this example.
The process for the detection of one par- Quadrupole scan rates are typically on
ticular mass in a multielement run is rep- the order of 2500 atomic mass units
resented in Figure 3. It shows a 63Cu ion (amu) per second and can cover the en-
emerging from the quadrupole and being tire mass range of 0300 amu in about
converted to an electrical pulse by the de- 0.1 s. However, real-world analysis speeds
tector. As the rf-dc voltage of the quadru- are much slower than this, and in practice
pole corresponding to 63Cu is re- 25 elements can be determined in dupli-
peatedly scanned, the ions as electrical cate with good precision in 12 min.
pulses are stored and counted by a multi-
channel analyzer. This multichannel data- QUADRUPOLE PERFORMANCE CRITERIA
acquisition system typically has 20 chan- Two very important performance specifi-
nels per mass, and as the electrical pulses cations of a mass analyzer govern its abil-
are counted in each channel, a profile of ity to separate an analyte peak from a
the mass is built up over the 20 channels, spectral interference. The first is resolv-

Detector

Quadrupole

Copper ion m/z 63


Mass scan (amu)
63
63
Cu n
1

63
Cu scan Quadrupole
mass scan
Multichannel data controller
acquisition system

Ions (electrical
pulses)

1 2 3 4 5 6 7 ............ 20
Ch l
Figure 3. Profiles of different masses are built up using a multichannel data acquisition system.
Circle 33
OCTOBER 2001 16(10) SPECTROSCOPY 45
............................... SPECTROSCOPY TUTORIAL .............................

Low resolution (3.0 amu) 30,000


B Resolution 0.70 amu
Normal resolution (1.0 amu) Resolution 0.50 amu
High resolution Low

Pulse intensity
A resolution High resolution 20,000
(0.3 amu)
F(dc)

Intensity
Inadequate 10,000
resolution
Overlap
0
F(rf) 62 63 64 65 66
Resolution Mass
Figure 4. Simplified Mathieu stability dia-
gram of a quadrupole mass filter, showing Figure 5. Sensitivity comparison of a Figure 6. Sensitivity comparison of two cop-
separation of two different masses, A (light quadrupole operated at 3.0, 1.0, and 0.3 per isotopes, 63Cu and 65Cu, at resolution
blue plot) and B (yellow plot). amu resolution (measured at 10% of its settings of 0.70 and 0.50 amu.
peak height).

ing power (R), which in traditional mass signal contribution of the tail of an adja- RESOLUTION
spectrometry is represented by the fol- cent peak at one mass lower and one Let us now discuss this area in greater de-
lowing equation: R 5 m/Dm, where m is mass higher than the analyte peak (3). tail. The ability to separate different
the nominal mass at which the peak oc- Even though they are somewhat related masses with a quadrupole is determined
curs and Dm is the mass difference be- and both define the quality of a quadru- by a combination of factors including
tween two resolved peaks (2). However, pole, the abundance sensitivity is proba- shape, diameter, and length of the rods, fre-
for quadrupole technology, the term reso- bly the most critical. If a quadrupole has quency of quadrupole power supply, oper-
lution is more commonly used, and is good resolution but poor abundance sen- ating vacuum, applied rf-dc voltages, and
normally defined as the width of a peak at sitivity, it will often prohibit the measure- the motion and kinetic energy of the ions
10% of its height. The second specifica- ment of an ultratrace analyte peak next to entering and exiting the quadrupole. All
tion is abundance sensitivity, which is the a major interfering mass. these factors will have a direct impact on
the stability of the ions as they travel down
the middle of the rods and thus the
quadrupoles ability to separate ions of dif-
fering mass-to-charge ratios. This is repre-
sented in Figure 4, which shows a simpli-
fied version of the Mathieu mass stability
plot of two separate masses (A and B) en-
tering the quadrupole at the same time (4).
Any of the rf-dc conditions shown un-
der the light blue plot will allow only
mass A to pass through the quadrupole,
while any combination of rf-dc voltages
under the yellow plot will allow only mass
B to pass through the quadrupole. If the
slope of the rf-dc scan rate is steep, repre-
sented by the light blue line (high resolu-
tion), the spectral peaks will be narrow,
and masses A and B will be well sepa-
rated (equivalent to the distance between
the two blue arrows). However, if the
slope of the scan is shallow, represented
by the red line (low resolution), the spec-
tral peaks will be wide, and masses A and
B will not be so well separated (equiva-
lent to the distance between the two red
arrows). On the other hand, if the slope
of the scan is too shallow, represented by
the gray line (inadequate resolution), the
peaks will overlap each other (shown by
the green area of the plot) and the
masses will pass through the quadrupole
without being separated. In theory, the
resolution of a quadrupole mass filter can
be varied between 0.3 and 3.0 amu. How-
Circle 34
46 SPECTROSCOPY 16(10) OCTOBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

Quadrupole
Doubly charged 151Eu (75.5 amu)
Ion entering Ion exiting
Mass (M) (a) (b)
Intensity

M 1 M 1

Mass (amu)

Figure 7. Ions entering the quadrupole are


slowed down by the filtering process and
75 75
produce peaks with a pronounced tail or
shoulder at the low-mass end.
1 ppm of monoisotopic arsenic (75As)
ever, improved resolution is always ac-
Figure 8. A low abundance sensitivity specification is critical to minimize spectral interfer-
companied by a sacrifice in sensitivity, as
seen in Figure 5, which shows a compari- ences, as shown by (a) a spectral scan of 50 ppm of 151Eu++ at 75.5 amu and (b) an ex-
son of the same mass at a resolution of panded view, which shows how the tail of the 151Eu++ elevates the spectral background of
3.0, 1.0, and 0.3 amu. 1 ppb of As at mass 75.
We can see that the peak height at 3.0 until now the slight improvement has not the spectral peaks drop off more rapidly
amu is much larger than the peak height become a practical benefit for most rou- at the high mass end of the peak com-
at 0.3 amu but, as expected, it is also tine applications. pared with the low mass end. The overall
much wider. This would prohibit using a peak shape, particularly its low mass and
resolution of 3.0 amu with spectrally com- ABUNDANCE SENSITIVITY high mass tail, is determined by the abun-
plex samples. Conversely, the peak width We can see in Figure 6 that the tails of dance sensitivity of the quadrupole,
at 0.3 amu is very narrow, but the sensi-
tivity is low. For this reason, a compro-
mise between peak width and sensitivity
usually has to be reached, depending on
the application. This can clearly be seen
in Figure 6, which shows a spectral over-
lay of two copper isotopes 63Cu and
65Cu at resolution settings of 0.70 and

0.50 amu. In practice, the quadrupole is


normally operated at a resolution of
0.71.0 amu for most applications.
It is worth mentioning that most
quadrupoles are operated in the first sta-
bility region, where resolving power is
typically ;400. If the quadrupole is oper-
ated in the second or third stability re-
gions, resolving powers of 4000 (5) and
9000 (6), respectively, have been
achieved. However, improving resolution
using this approach has resulted in a sig-
nificant loss of signal. Although there are
ways of improving sensitivity, other prob-
lems have been encountered, and as a re-
sult, to date there are no commercial in-
struments available based on this design.
Some instruments can vary the peak
width on-the-fly, which means that the res-
olution can be changed between 3.0 and
0.3 amu for every analyte in a multiele-
ment run. For some challenging applica-
tions this can be beneficial, but in reality
they are rare. So, even though
quadrupoles can be operated at higher
resolution (in the first stability region),
Circle 35
OCTOBER 2001 16(10) SPECTROSCOPY 47
............................... SPECTROSCOPY TUTORIAL ...............................

which is affected by a combination of fac- These are the fundamental reasons cies. Theory tells us that hyperbolic rods
tors including design of the rods, fre- why the peak shape is not symmetrical should generate a better hyperbolic (el-
quency of the power supply, and operat- with a quadrupole and explains why there liptical) field than cylindrical rods, result-
ing vacuum (7). Even though they are all is always a pronounced shoulder at the ing in higher transmission of ions at
important, probably the biggest impacts low mass side of the peak compared to higher resolution. It also tells us that a
on abundance sensitivity are the motion the high mass side as represented in higher operating frequency means a
and kinetic energy of the ions as they en- Figure 7, which shows the theoretical higher rate of oscillation and therefore
ter and exit the quadrupole. If one looks peak shape of a nominal mass M. We can separation of the ions as they travel
at the Mathieu stability plot in Figure 3, it see that the shape of the peak at one down the quadrupole. Finally, it is very
can be seen that the stability boundaries mass lower (M 2 1) is slightly different well accepted that a higher vacuum pro-
of each mass are less defined (not so from the other side of the peak at one duces fewer collisions between gas mole-
sharp) on the low mass side than they are mass higher (M 1 1) than the mass M. cules and ions, resulting in a narrower
on the high mass side (4). As a result, the For this reason, the abundance sensitivity spread in kinetic energy of the ions and
characteristics of ion motion at the low specification for all quadrupoles is always therefore less of a tail at the low mass
mass boundary is different from the high worse on the low mass side than on the side of a peak. However, given all these
mass boundary and is therefore reflected high mass side and is typically 1 3 1026 specification differences, in practice the
in poorer abundance sensitivity at the low at M 2 1 and 1 3 1027 at M 1 1. In other performance of most modern quadrupole
mass side compared with the high mass words, an interfering peak of 1 million ICP-MS instrumentation is very similar.
side. In addition, the velocity (and there- counts per second (cps) at M 2 1 would So even though these differences will
fore the kinetic energy) of the ions enter- produce a background of 1 cps at M, mainly be transparent to users, there are
ing the quadrupole will affect the ion mo- while it would take an interference of 107 some subtle variations in each instru-
tion and, as a result, will have a direct cps at M 1 1 to produce a background of ments measurement protocol and the
impact on the abundance sensitivity. For 1 cps at M. softwares approach to peak quantitation.
that reason, factors that affect the kinetic This is a very important area that we will
energy of the ions, like high plasma po- BENEFITS OF GOOD ABUNDANCE discuss it in greater detail in a future col-
tential and the use of lens components to SENSITIVITY umn. The next part of the series will con-
accelerate the ion beam, will degrade the Figure 8 shows an example of the impor- tinue with describing the fundamental
instruments abundance sensitivity (8). tance of abundance sensitivity. Figure 8a principles of other types of mass analyz-
is a spectral scan of 50 ppm of the doubly ers used in ICP-MS.
charged europium ion 151Eu11 at 75.5
amu (a doubly charged ion is one with REFERENCES
two positive charges, as opposed to a nor- (1) R. Thomas, Spectroscopy 16(9), 3844
mal singly charged positive ion, and ex- (2001).
hibits a m/z peak at half its mass). We can (2) F. Adams, R. Gijbels, and R. Van Grieken,
see that the intensity of the peak is so Inorganic Mass Spectrometry (John Wiley
and Sons, New York, 1988).
great that its tail overlaps the adjacent
(3) A. Montasser, Ed. Inductively Coupled
mass at 75 amu, which is the only avail- Plasma Mass Spectrometry (Wiley-VCH,
able mass for the determination of ar- Berlin, 1998).
senic. This is highlighted in Figure 8b, (4) P.H. Dawson, Ed., Quadrupole Mass Spec-
which shows an expanded view of the tail trometry and its Applications (Elsevier,
of the 151Eu11, together with a scan of 1 Amsterdam, 1976; reissued by AIP Press,
ppb of As at mass 75. We can see very Woodbury, NY, 1995).
clearly that the 75As signal lies on the (5) Z. Du, T.N. Olney, and D.J. Douglas, J.
sloping tail of the 151Eu11 peak. Mea- Am. Soc. Mass Spectrom. 8, 12301236
surement on a sloping background like (1997).
(6) P.H. Dawson and Y. Binqi, Int. J. Mass
this would result in a significant degrada-
Spectrom., Ion Proc. 56, 25 (1984).
tion in the arsenic detection limit, particu- (7) D. Potter, Agilent Technologies Application
larly as the element is monoisotopic and Note, 228349 (January, 1996).
no alternative mass is available. This ex- (8) E.R. Denoyer, D. Jacques, E. Debrah, and
ample shows the importance of a low S.D. Tanner, At. Spectrosc. 16(1), 1
abundance sensitivity specification in (1995).
ICP-MS.
Robert Thomas has more than 30 years ex-
DIFFERENT QUARDUPOLE DESIGNS perience in trace element analysis. He is the
Many different designs of quadrupole are principal of his own freelance writing and
used in ICP-MS, all made from different scientific consulting company, Scientific So-
materials with various dimensions, lutions, based in Gaithersburg, MD. He can
shapes, and physical characteristics. In be contacted by e-mail at thomasrj@
addition, they are all maintained at bellatlantic.net or via his web site at
slightly different vacuum chamber pres- www.scientificsolutions1.com.
sures and operate at different frequen-
48 SPECTROSCOPY 16(10) OCTOBER 2001 Circle 36 w w w. s p e c t r o s c o p y o n l i n e . c o m
S P E C T R O S C O P Y
TUTORIAL

A Beginners Guide to ICP-MS


Part VII: Mass Separation Devices Double-Focusing Magnetic-Sector Technology
ROBERT THOMAS

Although quadrupole mass analyzers repre-


sent more than 90% of all inductively cou-
pled plasma mass spectrometry (ICP-MS)
systems installed worldwide, limitations in Exit slit
their resolving power has led to the develop- Detector
ment of high-resolution spectrometers based
on the double-focusing magnetic-sector de-
sign. Part VII of this series on ICP-MS
takes a detailed look at this very powerful
mass separation device, which has found
its niche in solving challenging application
problems that require excellent detection Electrostatic
analyzer Entrance
capability, exceptional resolving power, or
slit
very high precision. Plasma torch

s discussed in Part VI of this series Electromagnet

A (1), a quadrupole-based ICP-MS


system typically offers a resolution
of 0.71.0 amu. This is quite ade-
quate for most routine applications,
but has proved to be inadequate for many
elements that are prone to argon-,
solvent-, or sample-based spectral inter-
Acceleration
optics
Focusing
optics

ferences. These limitations in Figure 1. Schematic of a reverse Nier-Johnson double-focusing magnetic-sector mass
quadrupoles drove researchers in the di- spectrometer (Courtesy of Thermo Finnigan [San Jose, CA]).
rection of traditional high-resolution,
magnetic-sector technology to improve found to be unsuitable as a separation de- lyzers a traditional electromagnet and
quantitation by resolving the analyte vice for an ICP system because it required an electrostatic analyzer (ESA). In the
mass away from the spectral interference a few thousand volts of potential at the standard (sometimes called forward) de-
(2). These ICP-MS instruments, which plasma interface area to accelerate the sign, the ESA is positioned before the
were first commercialized in the late ions into the mass analyzer. For this rea- magnet, and in the reverse design it is po-
1980s, offered resolving power as high as son, basic changes had to be made to the sitioned after the magnet. A schematic of
10,000, compared with a quadrupole, ion acceleration mechanism to optimize it a reverse Nier-Johnson spectrometer is
which had a resolving power of approxi- as an ICP-MS separation device. This was shown in Figure 1.
mately 300. This dramatic improvement a significant challenge when magnetic-
in resolving power allowed difficult ele- sector systems were first developed in the PRINCIPLES OF OPERATION
ments like Fe, K, As, V, and Cr to be de- late 1980s. However, by the early 1990s, With this approach, ions are sampled
termined with relative ease, even in com- instrument designers solved this problem from the plasma in a conventional man-
plex sample matrices. by moving the high-voltage components ner and then accelerated in the ion optic
away from the plasma and interface and region to a few kilovolts before they enter
TRADITIONAL MAGNETIC-SECTOR closer to the mass spectrometer. Todays the mass analyzer. The magnetic field,
INSTRUMENTS instrumentation is based on two different which is dispersive with respect to ion en-
The magnetic-sector design was first used approaches, commonly referred to as ergy and mass, focuses all the ions with
in molecular spectroscopy for the struc- standard or reverse Nier-Johnson geome- diverging angles of motion from the en-
tural analysis of complex organic com- try. Both these designs, which use the trance slit. The ESA, which is only disper-
pounds. Unfortunately, it was initially same basic principles, consist of two ana- sive with respect to ion energy, then fo-
22 SPECTROSCOPY 16(11) NOVEMBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m
............................... SPECTROSCOPY TUTORIAL ..............................

cuses the ions onto the exit slit, where magnet. This was not such a major prob-
the detector is positioned. If the energy lem for qualitative analysis, but proved to B

Magnetic field
dispersion of the magnet and ESA are be impractical for routine trace element
equal in magnitude but opposite in direc- analysis. This concept is shown in greater
tion, they will focus both ion angles (first detail in Figure 2, which is a plot of four A
focusing) and ion energies (second or parameters magnetic field strength,
double focusing), when combined to- accelerating voltage, mass, and signal in-

Acceleration potential
gether. Changing the electrical field in tensity against time for four separate
the opposite direction during the cycle masses (M1M4). Scanning the magnet
time of the magnet (in terms of the mass from point A to point B (accelerating volt-
Measurement time
passing the exit slit) has the effect of age is fixed) results in a scan across the
Scanning and
freezing the mass for detection. Then as mass range, generating spectral peaks for settling (dead) time
soon as a certain magnetic field strength the four different masses. It can be seen
that this increased scanning and settling
overhead time (often referred to as dead

Mass
time) would result in valuable measure-
ment time being lost, particularly for high
Changing the electric sample throughput that required ultra-
trace detection levels.
M4

Intensity
field in the opposite Changing the electric field in the oppo- M1 M2 M3
site direction to the field strength of the
direction to the field magnet during the cycle time of the mag- Time
strength of the magnet net has the effect of stopping the mass
that passes through the analyzer. Then, Figure 2. A plot of magnetic field strength,
during the cycle time of as soon as the magnetic field strength is accelerating voltage (fixed), mass, and sig-
passed, the electrical field is set to its nal intensity over time for four separate
the magnet has the original value and the next mass masses (M1M4). Note that only the mag-
effect of stopping the stopped in the same manner. The accel- net is scanned, while the accelerating volt-
erating voltage, as well as its rate of age is fixed resulting in long scan times
mass that passes change, has to be varied depending on between the masses.
the mass, but the benefit of this method
through the analyzer. is that only the mass peaks of interest are
registered. This process is seen in Figure
Magnetic field

B
3, which shows the same four masses
scanned. The only difference this time is A
that as well as scanning the magnet from
is passed, the electric field is set to its point A to point B, the accelerating volt-
original value and the next mass is age is also changed, resulting in a step-
Acceleration potential

frozen. The voltage is varied on a per- wise jump from one mass to the next.
mass basis, allowing the operator to scan This means that the full mass range is
only the mass peaks of interest rather covered much faster than just by scan-
than the full mass range (3, 4). ning the magnet alone (because of the in-
Because traditional magnetic-sector creased speed involved in electrically Measurement time
technology was initially developed for the jumping from one mass to another) (5).
structural or qualitative identification of Once the magnet has been scanned to a
organic compounds, there wasnt a real particular point, an electric scan is used Jump time
Mass

necessity for rapid quantitation of spectral to cover an area of  1030% of the mass,
peaks required for trace element analysis. either to measure the analyte peak or
They functioned by scanning over a large monitor other masses of interest. Peak
mass range by varying the magnetic field quantitation is typically performed by tak-
Intensity

over time with a fixed acceleration volt- ing multiple data points over a preset M1 M4
M2 M3
age. During a small window in time, mass window and integrating over a fixed
which was dependent on the resolution period of time. Time
chosen, ions of a particular mass-to- It should be pointed out that although
charge are swept past the exit slit to pro- this approach represents enormous time Figure 3. A plot of magnetic field strength,
duce the characteristic flat top peaks. As savings over older, single-focusing accelerating voltage (changed), mass, and
the resolution of a magnetic-sector instru- magnetic-sector technology, it is still sig- signal intensity over time for the same four
ment is independent of mass, ion signals, nificantly slower than quadrupole-based masses (M1M4). This time, in addition to
particularly at low mass, are far apart. instruments. The inherent problem lies in the magnet being scanned, the accelerating
The result was that a large amount of the fact that a quadrupole can be elec- voltage is also changed, resulting in rapid
time was spent scanning and settling the tronically scanned much faster than a electric jumps between the masses.
NOVEMBER 2001 16(11) SPECTROSCOPY 23
............................... SPECTROSCOPY TUTORIAL ..............................

(a)
Wide slit Lower resolution

Intensity
Ion beam

Entrance slit Analyzer Exit slit Detector

(b) Higher resolution

Intensity
Ion beam
Narrow slit

Entrance slit Analyzer Exit slit Detector

Figure 4. Resolution obtained using (a) wide and (b) narrow exit slit widths as the magnetic field is scanned. The entrance slit widths are the
same in (a) and (b).

magnet. Typical speeds for a full mass cial magnetic-sector ICP-MS systems of- in Figure 4, which shows two slit width
scan (0250 amu) of a magnet are in the fer as much as 10,000 resolving power scenarios. Figure 4a shows an example of
order of 400500 ms, compared with 100 (5% peak height/10% valley definition), a wide exit slit producing relatively low
ms for a quadrupole. In addition, it takes which is high enough to resolve the ma- resolution and a characteristic flat-topped
much longer for magnets to slow down jority of spectral interferences. Its worth peak. Figure 4b shows the same size en-
and settle to take measurements typi- emphasizing that resolving power (R) is trance slit, but a narrower exit slit, pro-
cally 3050 ms compared to 12 ms for a represented by the equation: R  m/m, ducing higher resolution with a charac-
quadrupole. So, even though in practice, where m is the nominal mass at which teristic triangular peak. The lowest
the electric scan dramatically reduces the the peak occurs and m is the mass dif- practical resolution achievable with a dou-
overall analysis time, modern double- ference between two resolved peaks (6). ble-focusing magnetic-sector instrument,
focusing magnetic-sector ICP-MS sys- In a quadrupole, the resolution is se- using the widest entrance and exit slits, is
tems, especially when multiple resolution lected by changing the ratio of the rf/dc approximately 300400, whereas the
settings are used, are significantly slower voltages on the quadrupole rods. How- highest practical resolution, using the
than quadrupole instruments. This ever, because a double-focusing narrowest entrance and exit slits, is ap-
makes them less than ideal for routine, magnetic-sector instrument involves fo- proximately 10,000. Most commercial
high-throughput applications or for sam- cusing ion angles and ion energies, mass systems operate at fixed resolution set-
ples that require multielement determina- resolution is achieved by using two me- tings for example, low is typically
tions on rapid transient signals. chanical slits one at the entrance to 300400; medium is typically 30004000,
the mass spectrometer and another at the and high is typically 800010,000 (the
RESOLVING POWER exit, before the detector. Varying resolu- choice of settings will vary depending on
As mentioned previously, most commer- tion is achieved by scanning the magnetic the instrumentation).
field under differ- However, it should be emphasized that,
ent entrance- and similar to optical spectrometry, as the
Table I. Resolution required to resolve some common polyatomic exit-slit width con- resolution is increased, the transmission
interferences from a selected group of isotopes. ditions. Similar to decreases. So even though extremely
optical systems, high resolution is available, detection lim-
Isotope Matrix Interference Resolution Transmission low resolution is its will be compromised under these con-
39K H2O 38ArH 5570 6% achieved by using ditions. This can be seen in Figure 5,
40Ca H2O 40 Ar 199,800 0% wide slits, which shows a plot of resolution against
44Ca HNO3 14 14
N N O 16 970 80% whereas high res- ion transmission. Figure 5 shows that a
56Fe H2O 40Ar16O 2504 18% olution is achieved resolving power of 400 produces 100%
31P H2O 15N16O 1460 53% with narrow slits. transmission, but at a resolving power of
34S H2O 16 18
O O 1300 65% Varying the width 10,000, only 2% is achievable. This dra-
75As HCl 40 35
Ar Cl 7725 2% of both the en- matic loss in sensitivity could be an issue
51V HCl 35Cl16O 2572 18% trance and exit if low detection limits are required in
64Zn H2SO4 32S16O16O 1950 42% slits effectively spectrally complex samples that require
24Mg Organics 12 12
C C 1600 50% changes the oper- the highest possible resolution; however,
52Cr Organics 40 12
Ar C 2370 20% ating resolution. spectral demands of this nature are not
55Mn HNO 40Ar15N 2300 20% This can be seen very common. Table I shows the resolu-
3

24 SPECTROSCOPY 16(11) NOVEMBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m


............................... SPECTROSCOPY TUTORIAL ..............................

pecially for high-mass elements like ura- isotope ratio work (7). Although preci-
nium where high resolution is generally sion is usually degraded as resolution is
not required, are typically an order of increased (because the peak shape gets
100
magnitude better than those provided by worse), modern instrumentation with
Transmission (%)

80 a quadrupole-based instrument. high-speed electronics and low mass bias


60 Besides good detection capability, an- is still capable of precision values of
40 other of the recognized benefits of the 0.1% RSD in medium- or high-resolution
20 magnetic-sector approach is its ability to mode (8).
quantitate with excellent precision. Mea- The demand for ultrahigh-precision
0
400 2000 4000 6000 8000 10,000 surement of the characteristically flat- data, particularly in the field of geochem-
topped spectral peaks translates directly istry, has led to the development of in-
Resolution
into high-precision data. As a result, in struments dedicated to isotope ratio
the low-resolution mode, relative stan- analysis. These are based on the double-
Figure 5. Ion transmission with a magnetic- dard deviation (RSD) values of focusing magnetic-sector design, but in-
sector instrument decreases as the resolu- 0.010.05% are fairly common, which stead of using just one detector, these in-
tion increases. makes magnetic-sector instruments an struments use multiple detectors. Often
ideal tool for carrying out high-precision referred to as multicollector systems,

tion required to resolve fairly common


polyatomic interferences from a selected
group of elemental isotopes, together (a) 56
Fe 40
Ar 16O
with the achievable ion transmission.
Figure 6 is a comparison between a
quadrupole instrument and a magnetic-
sector instrument with one of the most
common polyatomic interferences
40Ar16O on 56Fe, which requires a resolu-
Intensity

tion of 2504 to separate the peaks. Figure


6a shows a spectral scan of 56Fe using a
quadrupole instrument. What it doesnt
show is the massive polyatomic interfer-
ence 40Ar16O (produced by oxygen ions
from the water combining with argon
ions from the plasma) completely over-
lapping the 56Fe. It shows very clearly
that these two masses are unresolvable 55 56 57
with a quadrupole. If that same spectral m/z
scan is performed on a magnetic-sector
instrument, the result is the scan shown
in Figure 6b. To see the spectral scan on
(b)
the same scale, it was necessary to exam-
ine a much smaller range. For this rea- 40
Ar 16O
son, a 0.100-amu window was taken, as in-
0.100 amu scan
dicated by the dotted lines.

OTHER BENEFITS
Besides high resolving power, another at-
Intensity

tractive feature of magnetic-sector instru-


ments is their very high sensitivity com-
bined with extremely low background
levels. High ion transmission in low-
resolution mode translates into sensitivity
specifications of typically 100200 million 56
Fe
counts per second (mcps) per ppm, while
background levels resulting from ex-
tremely low dark current noise are typi- 55.935 55.957
cally 0.10.2 cps. This compares with sen- m/z
sitivity of 1050 mcps and background
levels of 10 cps for a quadrupole instru- Figure 6. Comparison of resolution between (a) a quadrupole and (b) a magnetic-sector in-
ment. For this reason detection limits, es- strument for the polyatomic interference of 40Ar16O on 56Fe.

26 SPECTROSCOPY 16(11) NOVEMBER 2001 w w w. s p e c t r o s c o p y o n l i n e . c o m


.........................

they offer the capability of detecting and


measuring multiple ion signals at exactly
the same time. As a result of this simulta-
neous measurement approach, they are
recognized as producing the ultimate in
isotope ratio precision (9).
There is no question that double-
focusing magnetic-sector ICP-MS sys-
tems are no longer a novel analytical
technique. They have proved themselves
to be a valuable addition to the trace ele-
ment toolkit, particularly for challenging
applications that require good detection
capability, exceptional resolving power,
and very high precision. They do have
their limitations, however, and perhaps
should not be considered a competitor for
quadrupole instruments when it comes to
rapid, high-sample-throughput applica-
tions or when performing multielement
determinations on fast transient peaks,
using sampling accessories such as elec-
trothermal vaporization (10) or laser
ablation (11).

REFERENCES
(1) R. Thomas, Spectroscopy 16(10), 4448
(2001).
Circle 20
(2) N. Bradshaw, E.F.H. Hall, and N.E.
Sanderson, J. Anal. At. Spectrom. 4,
801803 (1989).
(3) R. Hutton, A. Walsh, D. Milton, and J.
Cantle, ChemSA 17, 213215 (1991).
(4) U. Geismann and U. Greb, Fresnius J.
Anal. Chem. 350, 186193 (1994).
(5) U. Geismann and U. Greb, Poster Presen-
tation at Second Regensburg Symposium
on Elemental Mass Spectrometry (1993).
(6) F. Adams, R. Gijbels, and R. Van Grieken,
Inorganic Mass Spectrometry (John Wiley
and Sons, New York, 1988).
(7) F. Vanhaecke, L. Moens, R. Dams, and R.
Taylor, Anal. Chem. 68, 567 (1996).
(8) M. Hamester, D. Wiederin, J. Willis, W.
Keri, and C.B. Douthitt, Fresnius J. Anal.
Chem. 364, 495497 (1999).
(9) J. Walder and P. A. Freeman, J. Anal. At.
Spectrom. 7, 571 (1992).
(10) S. Beres, E. Denoyer, R. Thomas, and P.
Bruckner, Spectroscopy 9(1), 2026
(1994).
(11) S. Shuttleworth and D. Kremser, J. Anal.
At. Spectrom. 13, 697699 (1998).

Robert Thomas has more than 30 years of


experience in trace element analysis. He is
the principal of his own freelance writing
and scientific consulting company, Scientific
Solutions, based in Gaithersburg, MD. He
can be contacted by e-mail at thomasrj@
bellatlantic.net or via his web site at
www.scientificsolutions1.com.

Circle 21
NOVEMBER 2001 16(11) SPECTROSCOPY 27
S P E C T R O S C O P Y
TUTORIAL

A Beginners Guide to ICP-MS


Part VIII Mass Analyzers: Time-of-Flight Technology
ROBERT THOMAS

ontinuing with the discussion on Therefore, if a population of ions all

C mass analyzers used in inductively


coupled plasmamass spectrometry
(ICP-MS), lets now turn our atten-
tion to the most recent mass separa-
tion device to be commercialized time-
of-flight (TOF) technology. Although the
first TOF mass spectrometer was first de-
having different masses are given the
same kinetic energy by an accelerating
voltage (U ), the velocities of the ions will
all be different, based on their masses.
This principle is then used to separate
ions of different mass-to-charge ratios
(m/z) in the time (t) domain, over a fixed
Accelerating
voltage
Flight
tube Detector

scribed in the literature in the late 1940s flight path distance (D) represented by
Flight path distance
(1), it has taken more than 50 years to equation 2
adapt it for use in an ICP-MS system. The
recent growth in TOF ICP-MS sales is in m/z = 2Ut2/D2 [2] Figure 1. Principles of ion detection using
response to the technologys unique abil- TOF technology, showing separation of
ity to sample all ions generated in the This is shown schematically in Figure three masses in the time domain.
plasma at exactly the same time, which is 1, with three ions of different mass-to-
advantageous in three major areas: charge ratios being accelerated into a
Multielement determinations of rapid flight tube and arriving at the detector at into the mass filter in the conventional
transient signals generated by sam- different times. It can be seen that, based way, packets (groups) of ions are electro-
pling accessories such as laser ablation on their velocities, the lightest ion arrives statically injected into the flight tube at
and electrothermal vaporization first, followed by the medium mass ion, exactly the same time. With the orthogo-
devices and finally the heaviest one. Using flight nal approach, an accelerating potential is
High-precision, ratioing techniques tubes of 1 m in length, even the heaviest applied at right angles to the continuous
such as internal standardization and ions typically take less than 50 s to ion beam from the plasma source. The
isotope ratio analysis reach the detector. This translates into ion beam is then chopped by using a
Rapid multielement measurements, approximately 20,000 mass spectra/s pulsed voltage supply coupled to the or-
especially where sample volume is approximately 23 orders of magnitude thogonal accelerator to provide repetitive
limited. faster than the sequential scanning mode voltage slices at a frequency of a few kilo-
TOFs simultaneous nature of sampling of a quadrupole system. hertz. The sliced packets of ions, which
ions offers distinct advantages over tradi- are typically long and thin in cross sec-
tional scanning (sequential) quadrupole DIFFERENT SAMPLING APPROACHES tion (in the vertical plane), are then al-
technology for ICP-MS applications Even though this process sounds fairly lowed to drift into the flight tube where
where large amounts of data need to be straightforward, sampling the ions in a the ions are temporally resolved accord-
captured in a short amount of time. Be- simultaneous manner from a continuous ing to their velocities. Figure 2 shows this
fore we go on to discuss this in greater source of ions being generated in the process schematically.
detail, lets go through the basic princi- plasma discharge is not a trivial task. Ba- With the axial approach, an accelerat-
ples of TOF analyzers. sically two sampling approaches are used ing potential is applied axially (in the
in commercial TOF mass analyzers. They same axis) to the incoming ion beam as it
BASIC PRINCIPLES OF TOF are the orthogonal design (2), where the enters the extraction region. Because the
All TOF-MS instruments are based on flight tube is positioned at right angles to ions are in the same plane as the detec-
the same fundamental principle that the the sampled ion beam, and the axial de- tor, the beam has to be modulated using
kinetic energy (Ek) of an ion is directly sign (3), where the flight tube is in the an electrode grid to repel the gated
proportional to its mass (m) and velocity same axis as the ion beam. In both de- packet of ions into the flight tube. This
(v), represented by equation 1 signs, all ions that contribute to the mass kind of modulation generates an ion
spectrum are sampled through the inter- packet that is long and thin in cross sec-
Ek  12mv2 [1] face cones, but instead of being focused tion (in the horizontal plane). The differ-

36 SPECTROSCOPY 17(1) JANUARY 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m


............................... SPECTROSCOPY TUTORIAL .............................

ent masses are then resolved in the time


domain in a similar manner to the orthog- Orthogonal
onal design. The layout of an on-axis Ion Flight tube Detector
accelerator
packets
TOF system is shown schematically in
Figure 3.
Figures 2 and 3 represent a rather sim-
plistic explanation of TOF principles of
operation. In practice, the many complex
ion focusing components in a TOF mass
analyzer ensure that a maximum number
of analyte ions reach the detector and
also that undesired photons, neutral
species, and interferences are ejected Ion beam
from the ion beam. Some of these compo-
nents are shown in Figure 4, which
shows a more detailed view of a typical
orthogonal system. This design shows
that an injector plate is used to inject
packets of ions at right angles from the
Plasma torch
ion beam emerging from the MS inter-
face. These packets of ions are then di-
rected toward a deflectionsteering plate
where pulsed voltages steer the ions (or
throw out unwanted species) in the direc- Figure 2. Schematic of an orthogonal acceleration TOF analyzer.
tion of a reflectron. The packets of ions
are then deflected back 180, where they
are detected by a channel electron multi-
plier or discrete dynode detector. The
reflectron is a type of ion mirror and func- Flight tube Detector
tions as an energy compensation device, Ion
Plasma torch packet
so that different ions of the same mass ar-
rive at the detector at the same time.
Even though the on-axis design might
use slightly different components, the
principles are very similar.

DIFFERENCES BETWEEN ORTHOGONAL Ion beam


AND ON-AXIS TOF TECHNOLOGY
Although there are real benefits of using Ion beam
TOF over quadrupole technology for modulation
some ICP-MS applications, each type of
TOF design also has subtle differences in
its capabilities. (However, it is not the in- Figure 3. Schematic of an on-axis acceleration TOF analyzer.
tent of this tutorial to make any personal
judgement about the benefits or disad- lower than the latest commercial quadru- Duty cycle. Duty cycle is usually defined
vantages of either design.) Lets take a pole instruments. as the fraction (percentage) of extracted
look at some of these differences in Background levels. The on-axis design ions that actually make it into the mass
greater detail (4, 5). tends to generate higher background lev- analyzer. Unfortunately, with a TOF ICP-
Sensitivity. The axial approach tends to els because neutral species and photons MS system that has to use pulsed ion
produce higher ion transmission because stand a greater chance of reaching the packets from a continuous source of ions
the steering components are in the same detector. This results in background lev- generated in the plasma, this process is
plane as the ion generation system els in the order of 2050 counts/s not very efficient. It should be empha-
(plasma) and the detector. This means approximately 510 times higher than the sized that even though the ions are sam-
that the direction and magnitude of great- orthogonal design. However, because the pled at the same time, detection is not si-
est energy dispersion is along the axis of ion beam in the axial design has a smaller multaneous because different masses
the flight tube. In addition, when ions are cross section, a smaller detector can be arrive at the detector at different times.
extracted orthogonally, the energy dis- used, which generally has better noise The difference between the sampling
persion can produce angular divergence characteristics. In comparison, most com- mechanisms of orthogonal and axial TOF
of the ion beam resulting in poor trans- mercial quadrupole instruments offer designs translates into subtle differences
mission efficiency. However, the sensitiv- background levels of 110 counts/s, in their duty cycles.
ity of either TOF design is still generally depending on the design.
JANUARY 2002 17(1) SPECTROSCOPY 37
............................... SPECTROSCOPY TUTORIAL ..............................

olution. However, the resolving power of


both commercial TOF ICP-MS systems is
Deflectionsteering plate typically in the order of 5002000 (4), de-
Injector pending on the mass region, which
plate
makes them inadequate to resolve many
of the problematic polyatomic species en-
countered in ICP-MS (6). In comparison,
commercial high-resolution systems
based on the double-focusing magnetic-
sector design offer resolving power as
high as 10,000, while commercial
Reflectron quadrupoles typically achieve 300400.
Mass bias. This is the degree to which
Detector
ion transport efficiency varies with mass.
Interface
All instruments show some degree of
mass bias, which is usually compensated
for by measuring the difference between
Figure 4. A more detailed view of a typical orthogonal TOF analyzer, showing some of the ion the theoretical and observed ratio of two
steering components. isotopes of the same element. In TOF, the
velocity (energy) of the initial ion beam
will affect the instruments mass bias
characteristics. In theory, mass bias
should be less with the axial design be-
2500 cause the extracted ion packets dont
have any velocity in a direction perpen-
dicular to the axis of the flight tube,
which could potentially impact their
2000 transport efficiency.

BENEFITS OF TOF TECHNOLOGY


Counts

1500 FOR ICP-MS


It should be emphasized that these per-
formance differences between the two
1000
designs are subtle and should not detract
from the overall benefits of the TOF ap-
Time (s)

500 3.25 proach for ICP-MS. As mentioned earlier,


1.75 a scanning device such as a quadrupole
0 0.25 can only detect one mass at a time, which
20 29 43 50 76 138 181 207 means that a compromise always exists
Mass (amu) between number of elements, detection
limits, precision, and the overall measure-
ment time. However, with the TOF ap-
Figure 5. A full mass scan of a transient signal generated by 10 L of a 5-ppb multielement proach, the ions are sampled at exactly
solution using an electrothermal vaporization sampling accessory coupled to a TOF ICP-MS the same moment in time, which means
system (courtesy of GBC Scientific Equipment [Arlington Heights, IL]). that multielement data can be collected
with no significant deterioration in qual-
ity. The ability of a TOF system to capture
With the orthogonal design, duty cycle cycles for both orthogonal and axial de- a full mass spectrum, significantly faster
is defined by the width of the extracted signs are in the order of 1520%. than a quadrupole, translates into three
ion packets, which are typically long and Resolution. The resolution of the major benefits.
thin in cross section, as shown in Figure orthogonal approach is slightly better
2. In comparison, the duty cycle of the ax- because of its two-stage extraction/ RAPID TRANSIENT PEAK ANALYSIS
ial design is defined by the length of the acceleration mechanism. Because a pulse Probably the most exciting potential for
extracted ion packets, which are typically of voltage pushes the ions from the ex- TOF ICP-MS is in the multielement analy-
wide and thin in cross section, as shown traction area into the acceleration region, sis of a rapid transient signal generated
in Figure 3. Duty cycle can be improved the major energy dispersion lies along by sampling accessories such as laser ab-
by changing the cross-sectional area of the axis of ion generation. For this rea- lation (7), electrothermal vaporization,
the ion packet but, depending on the de- son, the energy spread is relatively small and flow injection systems (4). Even
sign, is generally improved at the ex- in the direction of extraction compared to though a scanning quadrupole can be
pense of resolution. In practice, the duty the axial approach, resulting in better res- used for this type of analysis, it struggles

38 SPECTROSCOPY 17(1) JANUARY 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m


............................... SPECTROSCOPY TUTORIAL .............................

Correction
n reference to Part VII of my tutorial series, A Beginners My statement that typical scan speeds for a full mass
I Guide to ICP-MS, which was published in the November
2001 issue of Spectroscopy, I would like to make a number
scan were 400500 ms is reflective of older magnetic sec-
tor technology. This is not representative of the ELEMENT2,
of corrections. Even though the intent of the article was to which has a scan speed in the order of 150200 ms.
give a general overview of double-focusing magnetic-sector My statement that typical sensitivity was in the order of
mass analyzers for beginners, Thermo Finnigan contacted 100200 million cps/ppm is not reflective of the
Spectroscopy to inform the editors and me that the column ELEMENT2, which has a specification for 115In of 1 billion
contained errors, and that it did not reflect the current per- cps/ppm.
formance of their instrument, the ELEMENT2. For that rea- My conclusion should therefore be modified to say that
son, I wish to make the following amendments to the article. if transient peak analysis is a requirement, modern double-
My statement that double-focusing magnetic sector focusing magnetic sector technology such as the
ICP-MS instruments are significantly slower than ELEMENT2, with its improved scan speeds, should be con-
quadrupole technology does not hold true today. Recent sidered a viable option to quadrupole technology.
improvements in the scan rate of the ELEMENT2 translates I wish to apologize for any inconvenience caused by
into speeds approaching that of quadrupole-based these statements.
instruments. Robert Thomas

to produce high-quality, multielement less than 50 s. Figure 5 shows a full multielement solution in less than 10 s.
data when the transient peak lasts only a mass scan of a transient peak generated TOF technology is probably better suited
few seconds. The simultaneous nature of by an electrothermal vaporization sam- than any other design of ICP-MS for this
TOF instrumentation makes it ideally pling accessory coupled to a TOF ICP- type of application.
suited for this type of analysis, because MS system. The technique has generated
the entire mass range can be collected in a healthy signal for 10 L of a 5-ppb IMPROVED PRECISION
To better understand how TOF technol-
ogy can help improve precision in ICP-
MS, it is important to know the major
sources of instability. The most common
source of noise in ICP-MS is flicker noise
associated with the sample introduction
process (from peristaltic pump pulsa-
tions, nebulization mechanisms, and
plasma fluctuations) and shot noise de-
rived from photons, electrons, and ions
hitting the detector. Shot noise is based
on counting statistics and is directly pro-
portional to the square root of the signal.
It therefore follows that as the signal in-
tensity gets larger, the shot noise has less
of an impact on the precision (% RSD) of
the signal. At high ion counts the most
dominant source of imprecision in ICP-
MS is derived from flicker noise gener-
ated in the sample introduction area.
One of the most effective ways to re-
duce instability produced by flicker noise
is to use a technique called internal stan-
dardization, where the analyte signal is
compared and ratioed to the signal of an
internal standard element (usually of sim-
ilar mass and ionization characteristics)
that is spiked into the sample. Even
though a quadrupole-based system can
do an adequate job of compensating for
these signal fluctuations, it is ultimately
limited by its inability to measure the in-
ternal standard at exactly the same time
as the analyte isotope. So to compensate

Circle 33
40 SPECTROSCOPY 17(1) JANUARY 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m
.............................. SPECTROSCOPY TUTORIAL ...............................

for sample introduction and plasma- techniques, and also has the capability for (6) N. Bradshaw, E.F. Hall, and N.E. Sander-
based noise and achieve high precision, high speeds of analysis. However, even son, J. Anal. At. Spectrom. 4, 801803
(1989).
the analyte and internal standard isotopes though it has enormous potential, TOF
(7) P. Mahoney, G. Li, and G.M. Hieftje, J.
need to be sampled and measured simul- was only commercialized in 1998, so it is Am. Soc. Mass Spectrom. 11, 401406
taneously. For this reason, the design of a relatively immature compared with (1996).
TOF mass analyzer is perfect for true quadrupole ICP-MS technology, which is (8) K.G. Heumann, S.M. Gallus, G.
simultaneous internal standardization re- almost 20 years old. For that reason, Radlinger, and J. Vogl, J. Anal. At. Spec-
quired for high-precision work. It follows, there is currently only a small number of trom. 13, 1001 (1998).
therefore, that TOF is also well suited for TOF instruments carrying out high-
high-precision isotope ratio analysis throughput, routine applications.
where its simultaneous nature of mea- Robert Thomas has more than 30 years of
surement is capable of achieving preci- REFERENCES experience in trace element analysis. He is
sion values close to the theoretical limits (1) A.E. Cameron and D.F. Eggers, The Re- the principal of his own freelance writing
of counting statistics. And unlike a scan- view of Scientific Instruments 19(9), 605 and scientific consulting company, Scientific
ning quadrupole-based system, it can (1948). Solutions, based in Gaithersburg, MD. He
measure ratios for as many isotopes or (2) D.P. Myers, G. Li, P. Yang, and G.M. can be contacted by e-mail at thomasrj@
Hieftje, J. Am. Soc. Mass Spectrom. 5,
isotopic pairs as needed all with excel- bellatlantic.net or via his web site at
10081016 (1994).
lent precision (8). (3) D.P. Myers, paper presented at 12th Asilo-
www.scientificsolutions1.com.
mar Conference on Mass Spectrometry,
ANALYSIS TIME Pacific Grove, CA, Sept. 2024, (1996).
As with a scanning ICPoptical emission (4) R.E. Sturgeon, J.W.H. Lam, and A. Saint,
spectroscopy system, the speed of a J. Anal. At. Spectrom. 15, 607616,
quadrupole ICP mass spectrometer is lim- (2000).
ited by its scanning rate. To determine 10 (5) F. Vanhaecke, L. Moens, R. Dams, L.
elements in duplicate with good precision Allen, and S. Georgitis, Anal. Chem. 71,
and detection limits, an integration time of 3297 (1999).
3 s/mass is normally required. When
overhead scanning and settling times are
added for each mass and each replicate,
this translates to approximately 2
min/sample. With a TOF system, the
same analysis would take significantly
less time because all the data are captured
simultaneously. In fact, detection limit lev-
els in a TOF instrument are typically
achieved using a 1030 s integration time,
which translates into a 510-fold improve-
ment in measurement time over a quadru-
pole instrument. The added benefit of a
TOF instrument is that the speed of analy-
sis is not impacted by the number of ana-
lytes being determined. It wouldnt matter
if the suite of elements in the method was
10 or 70 the measurement time would
be approximately the same. However, one
point must be stressed: A large portion of
the overall analysis time is taken up with
flushing an old sample out and pumping a
new sample into the sample introduction
system. This can be as much as 2
min/sample for real-world matrices. So
when this time is taken into account, the
difference between the sample through-
put of a quadrupole system and a TOF
ICP-MS system is not so evident.
TOF ICP-MS, with its rapid, simultane-
ous mode of measurement, excels at mul-
tielement applications that generate fast
transient signals. It offers excellent preci-
sion, particularly for isotope-ratioing

Circle 34
JANUARY 2002 17(1) SPECTROSCOPY 41
TUTOR IAL

A Beginners Guide to ICP-MS


Part IX Mass Analyzers: Collision/Reaction Cell Technology
Robert Thomas

The detection capability of traditional quadrupole mass analyzers for the interface in the normal
some critical elements is severely compromised by the formation of manner, where they are ex-
polyatomic spectral interferences generated by either argon, solvent, or tracted under vacuum into a
sample-based ionic species. Although there are ways to minimize these collision/reaction cell that is po-
interferences including correction equations, cool plasma technology, sitioned before the analyzer
and matrix separation they cannot be completely eliminated. How- quadrupole. A collision/reaction
ever, a new approach called collision/reaction cell technology has re- gas such as hydrogen or helium
cently been developed that virtually stops the formation of many of is then bled into the cell, which
these harmful species before they enter the mass analyzer. Part IX of consists of a multipole (a
this series takes a detailed look at this innovative new technique and quadrupole, hexapole, or octa-
the exciting potential it has to offer. pole), usually operated in the
radio frequency (rf)-only mode.
The rf-only field does not sepa-

Robert
A small number of
elements are recog-
nized as having poor
detection limits by
inductively coupled plasma
mass spectrometry (ICP-MS).
The cold/cool plasma ap-
proach, which uses a lower tem-
perature to reduce the forma-
tion of the argon-based
interferences, is a very effective
way to get around some of these
rate the masses like a traditional
quadrupole, but instead has the
effect of focusing the ions,
which then collide and react
with molecules of the collision/
reaction gas. By a number of
These elements are predomi- problems (1); however, it is different ion-molecule collision
Thomas nantly ones that suffer from sometimes difficult to optimize, and reaction mechanisms, poly-
has more than
30 years of major spectral interferences it is only suitable for a few of the atomic interfering ions like 40Ar,
40
experience in generated by ions derived from interferences, it is susceptible to Ar16O, and 38ArH, will either be
trace element the plasma gas, matrix compo- more severe matrix effects, and converted to harmless noninter-
analysis. He is nents, or the solventacid used it can be time consuming to fering species, or the analyte will
the principal of
to get the sample into solution. change back and forth between be converted to another ion
his own
freelance writing Examples of these interferences normal- and cool-plasma con- which is not interfered with.
and consulting include: ditions. These limitations and This is exemplified by the reac-
company, 40Ar16O on the determination the desire to improve perform- tion [1], which shows the use of
Scientific of 56Fe ance led to the development of hydrogen gas to reduce the
Solutions, based
in Gaithersburg,
ArH on the determination
38
collision/reaction cells in the 38
ArH polyatomic interference in
MD. He can be of 39K late 1990s. Designed originally the determination of 39K. Hy-
contacted by 40Ar on the determination of for organic MS to generate drogen gas converts 38ArH to the
e-mail at 40
Ca daughter species to confirm harmless H3 ion and atomic
thomasrj@ 40Ar40Ar on the determination identification of the structure of argon, but does not react with
bellatlantic.net
of 80Se the parent molecule (2), they the potassium. The 39K analyte
or via his web
site at www. Ar35Cl on the determination
40
were used in ICP-MS to stop the ions, free of the interference,
scientificsolutions1. of 75As appearance of many argon- then emerge from the
com. Ar12C on the determination
40
based spectral interferences. collision/reaction cell, where
of 52Cr they are directed toward the
Cl16O on the determination
35
Basic Principles of Collision/ quadrupole analyzer for normal
of 51V. Reaction Cells mass separation.
With this approach, ions enter
42 Spectroscopy 17(2) February 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m
Tutorial

and rejected, while the analyte ions,


Gas inlet which have a higher energy than the cell
bias, are transmitted.
The inability to adequately control
the secondary reactions meant that low
reactivity gases like He, H2, and Xe were
ICP the only option. The result was that
ion-molecule collisional fragmentation
(and not reactions) was thought to be
Detector Analyzing Multipole Ion lens ICP-MS the dominant mechanism of interfer-
quadrupole reaction system interface
collision cell
ence reduction. So even though the ion
transmission characteristics of a hexa-
pole were considered very good (with
Figure 1. Layout of a typical collision/reaction cell instrument. respect to the range of energies and
masses transmitted), background levels
were still relatively high because the in-
phy or electrospray MS studies was a terference rejection process was not
38
ArH  H2  H3  Ar disadvantage in inorganic MS, where very efficient. For this reason, its detec-
K  H2  39K  H2 (no reaction) [1]
39 
secondary reactionproduct ions are tion capability particularly for some
something to be avoided. There were of the more difficult elements, like Fe,
The layout of a typical collision/ ways to minimize this problem, but K, and Ca offered little improvement
reaction cell instrument is shown in they were still limited by the type of over the cool plasma approach. Table I
Figure 1. collision gas that could be used. Unfor- shows some typical detection limits in
tunately, highly reactive gases such as ppb achievable with a hexapole-based
Different Collision/Reaction ammonia and methane, which are more collision cell ICP-MS system (4).
Approaches efficient at interference reduction Recent modifications to the hexapole
The previous example is a very simplis- could not be used because of the limita- design have significantly improved its
tic explanation of how a collision/ tions of a non-scan-
reaction cell works. In practice, com- ning hexapole (in rf-
plex secondary reactions and collisions only mode) to Table I. Typical detection limits (in ppb)
take place, which generate many unde- adequately control the achievable with a hexapole-based collision cell
sirable interfering species. If these secondary reactions. ICP-MS system (4).
species were not eliminated or rejected, The fundamental rea- Elemental Detection
they could potentially lead to additional son is that hexapoles Sensitivity Limit
spectral interferences. Basically two ap- do not provide ade- Element Isotope (cps/[g/mL]) (ppb)
proaches are used to reject the products quate mass discrimi- Be 9 6.9  107 0.0077
of these unwanted interactions: nation capabilities to Mg 24 1.3  108 0.028
Discrimination by kinetic energy suppress the un- Ca 40 2.8  108 0.07
Discrimination by mass. wanted secondary re- V 51 1.7  108 0.0009
The major differences between the actions, which necessi- Cr 52 2.4  108 0.0007
two approaches are in the types of mul- tates the need for Mn 55 3.4  108 0.0017
tipoles used and their basic mechanism kinetic energy dis- Fe 56 3.0  108 0.017
for rejection of the interferences. Lets crimination to distin- Co 59 2.7  108 0.0007
take a closer look at how they differ. guish the collision Ni 60 2.1  108 0.016
product ions from the Cu 63 1.9  108 0.003
Discrimination by Kinetic Energy analyte ions. This is Zn 68 1.1  108 0.008
The first commercial collision cells for typically achieved by Sr 88 4.9  108 0.0003
ICP-MS were based on hexapole tech- setting the collision Ag 107 3.5  108 0.0003
nology (3), which was originally de- cell bias slightly less Cd 114 2.4  108 0.0004
signed for the study of organic mole- positive than the mass Te 128 1.3  108 0.009
cules using tandem MS. The more filter bias. This means Ba 138 5.9  108 0.0002
collision-induced daughter species that that the collision- Tl 205 4.0  108 0.0002
were generated, the better the chance of product ions, which Pb 208 3.7  108 0.0007
identifying the structure of the parent have the same energy Bi 209 3.4  108 0.0005
molecule; however, this very desirable as the cell bias, are dis- U 238 2.3  108 0.0001
characteristic for liquid chromatogra- criminated against

February 2002 17(2) Spectroscopy 43


Tutorial

14
107 104
No matrix
12
0.2% Cl matrix 106 103
0.06% C matrix

Signal at mass 80 (cps)


10 105 102
1ppb Se (cps)

BEC (ppb)
Cr/ 53Cr

104 101
8
103 100
52

6 Blank (cps)
102 101
BEC (ppb)
4 101 102

100 103
2 0 0.7 1.1 1.5 1.9 2.3 2.7 3.1 3.5 3.9 4.3 4.7 5.0
0 1 2 3 4 5 6 Hydrogen flow (mL/min)
Collision gas flow rate (mL/min)

Figure 2. The use of a helium/ammonia mixture with a hexapole-based Figure 3. Background reduction of the argon dimer (40Ar2) with
collision cell for the successful determination of 52Cr/53Cr isotopic ratios hydrogen gas using an octapole reaction cell (Courtesy of Agilent
(courtesy of Thermo Elemental, Franklin, MA). Technologies, Wilmington, DE).

52
collision/reaction characteristics. In ad- Cr/53Cr (52Cr is 83.789% and 53Cr is order multipoles. Similar in design to
dition to offering good transmission 9.401% abundant). It can be seen that the hexapole, collisional fragmentation
characteristics and kinetic energy dis- the 52Cr/53Cr ratio is virtually the same and energy discrimination are the pre-
crimination, they now appear to offer in the chloride and carbon matrices as dominant mechanisms for interference
basic mass-dependent discrimination it is with no matrix present when the reduction, which means that lower re-
capabilities. This means that the kinetic optimum flow of collision/reaction gas activity gases like hydrogen and helium
energy discrimination barrier can be is used (5). are preferred. By careful design of the
adjusted with analytical mass, which of- Another way to discriminate by ki- interface and the entrance to the cell,
fers the capability of using small netic energy is to use an octapole in the the collision/reaction capabilities can be
amounts of highly reactive gases. Figure collision/reaction cell instead of a hexa- improved, by reducing the number of
2 shows an example of the reduction of pole. The benefit of using a higher sample/solvent/plasma-based ions en-
both 40Ar12C and 37Cl16O using helium order design is that its transmission tering the cell. This enables the collision
with a small amount of ammonia, in characteristics, particularly at the low gas to be more effective at reducing the
the isotopic ratio determination of mass end, are slightly higher than lower interferences. An example of this is the
use of H2 as the cell gas to reduce the
argon dimer (40Ar2) interference in the
ArO  NH3 O  Ar  NH3 determination of the major isotope of
selenium at mass 80 (80Se). Figure 3
Gas inlet shows an example of a dramatic reduc-
Quadrupole tion in the 40Ar2 background at mass 80
Ar mass filter using an ICP-MS fitted with an octa-
Reactive gas (NH3) pole reaction cell. By using the opti-
mum flow of H2, the spectral back-
40
Ar16O ground is reduced by about six orders
+ + of magnitude, from 10,000,000 cps to
+ + 10 cps, producing a background equiva-
+ + + + lent concentration of approximately
56
Fe + ++ + +
+ + + 1 ppt for 80Se (6).

+ + Discrimination by Mass
Analyzer
quadrupole
56
Fe The other way to reject the products of
+ the secondary reactions/collisions is to
Ionmolecule
reactions and collisions discriminate them by mass. Unfortu-

NH3 O nately, higher order multipoles cannot
be used for efficient mass discrimina-
tion because the stability boundaries
Figure 4. Elimination of the ArO interference with a dynamic reaction cell.
are diffuse, and sequential secondary

44 Spectroscopy 17(2) February 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m


Tutorial

pole electrical fields, unwanted reac-



Ar  NH3 
NH3  Ar tions between the gas and the sample
matrix or solvent (which could poten-
(Exothermic reaction)
tially lead to new interferences) are pre-
 vented. Therefore, every time an analyte
 Ar
and interfering ions enter the dynamic

reaction cell, the bandpass of the
quadrupole can be optimized for that
NH3 specific problem and then changed on-
the-fly for the next one. Figure 4 shows
a schematic of an analyte ion 56Fe and
an isobaric interference 40Ar16O entering
  Ca the dynamic reaction cell. The reaction
 gas NH3 reacts with the ArO to form
Ca  NH3 no reaction atomic oxygen and argon together with
(Endothermic reaction) a positive NH3 ion. The quadrupoles
electrical field is then set to allow the
transmission of the analyte ion 56Fe to
Figure 5. The reaction between NH3 and Ar is exothermic and fast, while there is no reaction the analyzer quadrupole, free of the
between NH3 and Ca in the dynamic reaction cell. problematic isobaric interference,
40
Ar16O. In addition, the NH3 is pre-

reactions cannot be easily intercepted. dynamic reaction cell is a pressurized


The way around this problem is to use a multipole positioned before the ana-
quadrupole (instead of a hexapole or lyzer quadrupole. However, the similar- Collision/reaction cells
octapole) inside the reaction/collision ity ends there. In dynamic reaction cell
cell, and use it as a selective bandpass technology, a quadrupole is used in- have given a new lease on
filter. The benefit of this approach is stead of a hexapole or octapole. A
that highly reactive gases can be used, highly reactive gas, such as ammonia or life to quadrupole mass
which tend to be more efficient at inter- methane, is bled into the cell, which is a
ference reduction. One such develop- catalyst for ion molecule chemistry to analyzers used in ICP-MS.
ment that uses this approach is called take place. By a number of different re-
dynamic reaction cell technology (7, 8). action mechanisms, the gaseous mole-
Similar in appearance to the hexapole cules react with the interfering ions to vented from reacting further to produce
and octapole collision/reaction cells, the convert them either into an innocuous a new interfering ion. The advantage of
species different from the this approach is that highly reactive
analyte mass or a harmless gases can be used, which increases the
Table II. Typical detection limits in ppt of neutral species. The analyte number of ionmolecule reactions tak-
an ICP-MS system fitted with a dynamic mass then emerges from the ing place and therefore more efficient
reaction cell (9). dynamic reaction cell free of removal of the interfering species. Of
its interference and steered course, this also potentially generates
Detection Detection
into the analyzer quadru- more side reactions between the gas
Analyte Limit (ppt) Analyte Limit (ppt)
pole for conventional mass and the sample matrix and solvent;
Li 0.08 Co 0.07
separation. The advantages however, by dynamically scanning the
Be 0.6 60
Ni 0.4
of using a quadrupole in bandpass of the quadrupole in the reac-
B 1.1 Zn 1
the reaction cell is that the tion cell, these reaction by-products are
Na 0.3 As 1.2
stability regions are much rejected before they can react to form
Mg 0.6 Se* 5
better defined than a hexa- new interfering ions.
Al 0.07 Sr 0.02
pole or an octapole, so it is The benefit of the dynamic reaction
K* 1 Rh 0.01
relatively straightforward to cell is that by careful selection of the re-
40
Ca* 1 In 0.01
operate the quadrupole in- action gas, the user takes advantage of
V* 0.3 Sb 0.06
side the reaction cell as a the different rates of reaction of the an-
Cr* 0.25 Cs 0.03
mass or bandpass filter, and alyte and the interfering species. This
Mn* 0.09 Pb 0.03
not just an ion-focusing process is exemplified by the elimina-
56
Fe* 0.15 U 0.01
guide. Therefore, by careful tion of 40Ar by NH3 gas in the determi-
* Indicates elements determined in dynamic reaction cell mode. optimization of the quadru- nation of 40Ca. The reaction between

46 Spectroscopy 17(2) February 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m


Tutorial

however, the ionization reaction cell. The elements with an as-


109
potential of Ca (6.1 eV) terisk were determined using ammonia
108 is significantly less than or methane as the reaction gas, while
107 that of NH3, so the reac- the other elements were determined in
tion, which is endother- the standard mode (no reaction gas).
Ion signal (cps)

6
10
100 ppt Ca mic, is not allowed to Collision/reaction cells have given a
105 proceed (8). Figure 5 new lease on life to quadrupole mass
104 shows this process in analyzers used in ICP-MS. They have
10 3 greater detail. enhanced its performance and flexibil-
This highly efficient ity, and most definitely opened up the
102
Water blank reaction mechanism technique to more-demanding applica-
101 translates into a dra- tions that were previously beyond its
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
matic reduction of the capabilities. However, it must be em-
NH3 flow (mL/min) spectral background at phasized that even though differences
mass 40, which is exist between commercially available
Figure 6. A reduction of eight orders of magnitude in the 40Ar
shown graphically in instruments, they all perform very well.
background signal is achievable with the dynamic reaction cell
Figure 6. At the opti- The intent of this tutorial is to present
resulting in 1 ppt detection limit for 40Ca.
mum NH3 flow, a re- the benefits of the technology to begin-
duction in the 40Ar ners and give an overview of the differ-
NH3 gas and the 40Ar interference, background signal of about eight orders ent approaches available. If it has cre-
which is predominantly a charge ex- of magnitude is achieved, resulting in a ated an interest, I strongly suggest that a
change, occurs because the ionization detection limit of 0.51.0 ppt for 40Ca. performance evaluation is made based
potential of NH3 (10.2 eV) is low com- Table II shows some typical detection on your own sample matrices.
pared with that of Ar (15.8 eV). There- limits in parts per trillion (ppt) of an
fore, the reaction is exothermic and fast; ICP-MS system fitted with a dynamic References
1. K. Sakata and K. Kawabata, Spec-
trochimica Acta 49B, 1027 (1994).
2. B.A. Thomson, D.J. Douglas, J.J. Corr,
J.W. Hager, and C.A. Joliffe, Anal. Chem.
67, 16961704 (1995).
3. P. Turner, T. Merren, J. Speakman, and
C. Haines, in Plasma Source Mass
Spectrometry: Developments and Ap-
plications, eds. G. Holland and S. Tan-
ner (Royal Society of Chemistry, Cam-
bridge, UK, 1996).
4. I. Feldmann, N. Jakubowski, C. Thomas,
and D. Stuewer, Fresenius J. Anal.
Chem. 365, 422428 (1999).
5. Collision Cell Technology with Energy
Discrimination, Thermo Elemental Ap-
plication Note (Thermo Elemental,
Franklin, MA, September 2001).
6. E. McCurdy and D. Potter, Agilent Tech-
nologies ICP-MS Journal 10, October
2001.
7. Covered by U.S. Patent No. 6140638.
8. S.D. Tanner and V.I. Baranov, At. Spectr.
20(2), 4552, (1999).
9. U. Voellkopf, K. Klemm, and M. Pfluger,
At. Spectr. 20(2), 5359, (1999). 

48 Spectroscopy 17(2) February 2002 Circle 28 w w w. s p e c t r o s c o p y o n l i n e . c o m


TUTOR IAL

A Beginners Guide to ICP-MS


Part X Detectors
Robert Thomas

Part X of this series on ICP-MS discusses the detection system an more secondary electrons form.
important area of the mass spectrometer that counts the number of ions The potential gradient inside
emerging from the mass analyzer. The detector converts the ions into the tube varies based on posi-
electrical pulses, which are then counted by its integrated measurement tion, so the secondary electrons
circuitry. The magnitude of the electrical pulses corresponds to the move farther down the tube. As
number of analyte ions present in the sample. Trace element quantita- these electrons strike new areas
tion in an unknown sample is then carried out by comparing the ion of the coating, more secondary
signal with known calibration or reference standards. electrons are emitted. This
process is repeated many times.

S
The result is a discrete pulse that
ince inductively coupled emission spectroscopy (ICP- contains many millions of elec-
plasmamass spectrom- OES); however, instead of using trons generated from an ion
etry (ICP-MS) was first individual dynodes to convert that first hits the cone of the de-
introduced in the early photons to electrons, the chan- tector (1). This process is shown
1980s, a number of different ion neltron is an open glass cone simplistically in Figure 1.
detection designs have been This pulse is
used, the most popular being then sensed
Robert electron multipliers for low ion- For some applications where and detected
Thomas count rates, and Faraday collec- by a very fast
has more than
30 years of
tors for high-count rates. Today, ultratrace detection limits are preamplifier.
the majority of ICP-MS systems The output
experience in
trace element
used for ultratrace analysis use not required, the ion beam from pulse from the
analysis. He is detectors that are based on the preamplifier
the principal of active film or discrete dynode the mass analyzer is directed then goes to a
his own electron multiplier. They are digital dis-
freelance writing
and consulting
very sophisticated pieces of into a simple metal electrode, criminator and
company,
equipment compared with ear- counting cir-
Scientific lier designs and are very effi- or Faraday cup. cuitry, which
Solutions, based cient at converting ion currents counts only
in Gaithersburg, into electrical signals. Before we pulses above a
MD. He can be describe these detectors in coated with a semiconductor- certain threshold value. This
contacted by
e-mail at
greater detail, it is worth looking type material that generates threshold level needs to be high
thomasrj@ at two of the earlier designs electrons from ions impinging enough to discriminate against
bellatlantic.net the channel electron multiplier on its surface. For the detection pulses caused by spurious emis-
or via his web (channeltron) and the Faraday of positive ions, the front of the sion inside the tube, stray pho-
site at www. cup to get a basic under- cone is biased at a negative po- tons from the plasma itself, or
scientificsolutions1.
com.
standing of how the ICP-MS ion tential and the far end, nearest photons generated from fast
detection process works. the collector, is kept at ground. moving ions striking the
Channel electron multiplier. The When the ion emerges from the quadrupole rods.
operating principles of the quadrupole mass analyzer, it is Sometimes the rate of ions
channel electron multiplier are attracted to the high negative hitting the detector is too high
similar to those of a photomul- potential of the cone. When the for the measurement circuitry
tiplier tube used in ICPoptical ion hits this surface, one or to handle in an efficient man-
34 Spectroscopy 17(4) April 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m
Tutorial

in the early 1990s to develop an ICP- way the measurement circuitry handles
Ions from MS system using a Faraday cup detector low and high ion-count rates. When
mass analyzer for environmental applications, but its ICP-MS was first commercialized, it
Preamplifier sensitivity was compromised and, as a could only handle as many as five or-
Secondary result, it was considered more suitable ders of dynamic range; however, when
electrons
for applications requiring ICP-OES de- attempts were made to extend the dy-
( )
tection capability. However, Faraday namic range, certain problems were en-
cup technology is still used in some countered. Before we discuss how mod-
3 kV
magnetic sector instruments, particu- ern detectors deal with this issue, lets
Figure 1. The path of an ion through a channel larly where high ion signals are first take a look at how it was addressed
electron multiplier. encountered in the determination of in earlier instrumentation.

ner. This situation is caused by ions ar- When ICP-MS was first commercialized, it could only
riving at the detector during the output
pulse of the preceding ion and not handle as many as five orders of dynamic range;
being detected by the counting system.
This dead time, as it is known, is a when attempts were made to extend the dynamic
fundamental limitation of the multi-
plier detector and is typically 3050 ns, range, certain problems were encountered.
depending on the detection system.
Compensation in the measurement cir-
cuitry has to be made for this dead time
in order to count the maximum num- high-precision isotope ratios using a Extending the Dynamic Range
ber of ions hitting the detector. multicollector detection system. Traditionally, ICP-MS using the pulse
Faraday cup. For some applications Discrete dynode electron multiplier. These counting measurement is capable of
where ultratrace detection limits are detectors, which are often called active about five orders of linear dynamic
not required, the ion beam from the film multipliers, work in a similar way range. This means that ICP-MS calibra-
mass analyzer is directed into a simple to the channeltron, but use discrete tion curves, generally speaking, are lin-
metal electrode, or Faraday cup (1). dynodes to carry out the electron mul- ear from ppt levels to as much as a few
With this approach, there is no control tiplication (2). Figure 2 illustrates the hundred parts-per-billion. However, a
over the applied voltage (gain), so a principles of operation of this device. number of ways exist to extend the dy-
Faraday cup can only be used for high The detector is usually positioned off- namic range of ICP-MS another three
ion currents. Their lower working range axis to minimize the background from to four orders of magnitude to work
is in the order of 104 counts/s, which stray radiation and neutral species com- from sub-part-per-trillion levels, to as
means that if a Faraday cup is to be ing from the ion source. When an ion much as 100 ppm. Following is a brief
used as the only detector, the sensitivity emerges from the quadrupole, it sweeps overview of some of the different ap-
of the ICP mass spectrometer will be through a curved path before it strikes proaches that have been used.
severely compromised. For this reason, the first dynode. On striking the first Filtering the ion beam. One of the first
Faraday cups are normally used in con- dynode, it liberates secondary electrons. approaches to extend the dynamic
junction with a channeltron or discrete The electron-optic design of the range in ICP-MS was to filter the ion
dynode detector to extend the dynamic dynode produces acceleration of these beam by putting a non-optimum volt-
range of the instrument. An additional secondary electrons to the next dynode, age on one of the ion lens components
problem with the Faraday cup is that, where they generate more electrons. or the quadrupole itself to limit the
because of the time constant used in the This process is repeated at each dynode, number of ions reaching the detector.
dc amplification process to measure the generating a pulse of electrons that is fi- This voltage offset, which was set on an
ion current, it is limited to relatively low nally captured by the multiplier collec- individual mass basis, acted as an en-
scan rates. This limitation makes it un- tor or anode. Because of the materials ergy filter to electronically screen the
suitable for the rapid scan rates re- used in the discrete dynode detector ion beam and reduce the subsequent
quired for traditional pulse counting and the difference in the way electrons ion signal to within a range covered by
used in ICP-MS and also limits its abil- are generated, it is typically more sensi- pulse-counting ion detection. The main
ity to handle fast transient peaks. tive than channeltron technology. disadvantage with this approach was
The Faraday cup never became pop- Although most discrete dynode de- that the operator had to have prior
ular with quadrupole ICP-MS systems tectors are very similar in the way they knowledge of the sample to know what
because it wasnt suitable for very low work, there are subtle differences in the voltage to the apply to the high concen-
ion-count rates. An attempt was made tration masses.
April 2002 17(4) Spectroscopy 35
Tutorial

Figure 2 (left). Schematic of a discrete dynode


Ion path
electron multiplier.

Generation trometer twice for each sample. A first


of electrons scan, in which the detector is operated

Noise
in the analog mode, provides signals for
elements present at high concentra-
tions. A second scan, in which the de-
tector voltage is switched to digital-
pulse counting mode, provides high
sensitivity detection for elements pres-
Quadrupole rods ent at low levels. A major advantage of
Individual dynodes
this technology is that users do not
need to know in advance whether to use
analog or digital detection because the
Scan Data system system automatically scans all elements
Incoming ion controller
in both modes. However, its disadvan-
tage is that two independent mass scans
are required to gather data across an ex-
To quadrupole tended signal range. This not only re-
sults in degraded measurement effi-
Midpoint ciency and slower analyses, but it also
dynode means that the system cannot be used
Analog signal Counter for fast transient signal analysis of un-
1
known samples because mode switch-
Detector ing is generally too slow.
MCA The second way of extending the dy-
namic range is similar to the first ap-
Pulse signal Counter proach, except that the first scan is used
2 as an investigative tool to examine the
sample spectrum before analysis (4).
This first prescan establishes the mass
Figure 3 (above). Dual-stage discrete dynode detector measurement circuitry. (Figures 3, 4, and 5
positions at which the analog and pulse
are courtesy of PerkinElmer Instruments, Shelton, CT.)
modes will be used for subsequently col-
lecting the spectral signal. The second
analytical scan is then used for data col-
Using two detectors. Another technique range. By using the detector in both the lection; the system switches the detector
that was implemented in some of the pulse-counting and analog modes, high back and forth rapidly between pulse
early quadrupole ICP-MS instrumenta- and low concentrations can be deter- and analog mode depending on the
tion was to use two different detectors, mined in the same sample. Three ap- concentration of each analytical mass.
such as a channel electron multiplier to proaches use this type of detection sys- The main disadvantage of these two
measure low current signals, and a tem; two of them involve carrying out approaches is that two separate scans
Faraday cup to measure high ion cur- two scans of the sample, while the third are required to measure high and low
rents. This process worked reasonably uses only one scan. levels. With conventional nebulization,
well, but struggled with some applica- Using two scans with one detector. The this isn't such a major problem except
tions because it required rapid switch- first approach uses an electron multi- that it can impact sample throughput.
ing between the two detectors. The plier operated in both digital and ana- However, it does become a concern
problem was that the ion beam had to log modes (3). Digital counting pro- when it comes to working with tran-
be physically deflected to select the op- vides the highest sensitivity, while sient peaks found in electrothermal va-
timum detector. Not only did this de- operation in the analog mode (achieved porization, flow injection, or laser sam-
grade the measurement duty cycle, but by reducing the high voltage applied to pling ICP-MS. Because these transient
detector switching and stabilization the detector) is used to reduce the sen- peaks often last only a few seconds, all
times of several seconds also precluded sitivity of the detector, thus extending the available time must be spent meas-
fast transient signal detection. the concentration range for which ion uring the masses of interest to get the
The more modern approach is to use signals can be measured. The system is best detection limits. When two scans
just one detector to extend the dynamic implemented by scanning the spec- have to be made, valuable time is
36 Spectroscopy 17(4) April 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m
Tutorial

wasted, which is not contributing to


10 9 quality of the analytical signal.
Using one scan with one detector. These
limitations of using two scans led to
the development of a third approach
using a dual-stage discrete dynode de-
106 tector (5). This technology uses meas-
urement circuitry that allows both

Analog (cps)
Pulse (cps)

high and low concentrations to be de-


termined in one scan. This is achieved
by measuring the ion signal as an ana-
104 log signal at the midpoint dynode.
When more than a threshold number
of ions are detected, the signal is
processed through the analog cir-
cuitry. When fewer than the threshold
0 number of ions are detected, the signal
0.1 ppt 100 ppm
cascades through the rest of the dyn-
Analyte concentration
odes and is measured as a pulse signal
in the conventional way. This process,
Figure 4. The pulse-counting mode covers rates as high as 106 counts/s, and the analog circuitry is which is shown in Figure 3, is com-
suitable from 104 to 109 counts/s with a dual-mode discrete dynode detector. pletely automatic and means that both
the analog and the pulse signals are col-
lected simultaneously in one scan (6).
The pulse-counting mode is typically
linear from zero to about 106 counts/s,
while the analog circuitry is suitable
from 104 to 109 counts/s. To normalize
both ranges, a cross calibration is per-
formed to cover concentration levels,
which could generate a pulse and an
analog signal. This is possible because
the analog and pulse outputs can be de-
fined in identical terms of incoming
pulse counts per second, based on
knowing the voltage at the first analog
stage, the output current, and a conver-
sion factor defined by the detection cir-
cuitry electronics. By performing a
cross calibration across the mass range,
a dual-mode detector of this type is ca-
pable of achieving approximately eight
to nine orders of dynamic range in one
simultaneous scan. Figure 4 shows the
pulse-counting calibration curve (yel-
low) is linear up to 106 cps, and the ana-
log calibration curve (blue) is linear
from 104 to 109 cps. Figure 5 shows that
after cross calibration, the two curves
are normalized, which means the detec-
tor is suitable for concentration levels
between 0.1 ppt and 100 ppm typi-
cally eight to nine orders of magnitude
for most elements.
There are subtle variations of this
type of detection system, but its major
benefit is that it requires only one scan
38 Spectroscopy 17(4) April 2002 Circle 20 w w w. s p e c t r o s c o p y o n l i n e . c o m
Tutorial

to determine both high and low con-


centrations. Therefore, it not only offers 109
the potential to improve sample
throughput, it also means that the max-
imum data can be collected on a tran-
sient signal that only lasts a few sec-

Cross-calibrated counts (cps)


onds. This process will be described in
greater detail in the next installment of
this series, in which I will discuss differ-
ent measurement protocols and peak
integration routines.

References
1. Inductively Coupled Plasma Mass
Spectrometry, Ed. A. Montasser (Wiley-
VCH, Berlin, 1998).
2. K. Hunter, Atomic Spectroscopy 15(1),
0
1720 (1994).
0.1 ppt 100 ppm
3. R.C. Hutton, A.N. Eaton, and R.M.
Analyte concentration
Gosland, Applied Spectroscopy 44(2),
238242 (1990).
4. Y. Kishi, Agilent Technologies Applica- Figure 5. Using cross calibration of the pulse and analog modes, quantitation from sub-part-per-
tion Journal (August 1997).
trillion to high parts-per-million levels is possible.
5. E.R. Denoyer, R.J. Thomas, and L.
Cousins, Spectroscopy 12(2), 5661,
(1997).
6. Covered by U.S. Patent No. 5,463,219. 

Circle 21 Circle 22
April 2002 17(4) Spectroscopy 39
Conference Preview

An Overview of New Products


at the Denver X-ray Conference

A
self-contained miniature x-ray tube

M4 FLUXER/CORPORATION SCIENTIFIQUE CLAISSE


system has been introduced by
This article provides
Amptek (Bedford, MA) for portable
a brief overview of x-ray fluorescence (XRF) analysis.
some of the latest The LASER-X system houses the x-ray tube,
spectroscopy-related power supply, and control electronics in a sin-
equipment and gle, compact enclosure. Features include a
supplies scheduled to grounded anode; variable current and volt-
be shown at the age; a 35-kV, 100-A power supply with a
2002 Denver X-ray solid silver target; and an end window. The
Conference at Antlers system is designed to replace radioisotopes in
Adams Mark Hotel, applications such as process control, OEM
Colorado Springs, use, research, and teaching.
Bede (Englewood, CO) has just released
Colorado, July 29
the second version of its Instrument Control fused, homogenous, pure, and do not absorb
August 2, 2002.
software, which offers an improved scripting moisture, according to the company.
engine, special applications screens, an im- DEL Electronics (Hicksville, NY) offers the
proved scan set-up screen, and Windows XRD-3000 for x-ray diffraction (XRD) appli-
2000 compatibility. The company is also re- cations and the XRF-3000 for fluorescence.
leasing new versions of its RADS and REFS The 3-kW x-ray generators are compact and
software, which incorporate the companys lightweight. They can be fully adjusted from
Mercury technology. Mercury uses genetic the front panel or by remote control for beam
algorithms and provides an automatic pa- voltage and emission current. The units con-
rameter-optimization procedure for x-ray tain an integrated filament supply and the con-
reflectivity and diffraction data. RADS 4.0 trol circuitry required for filamentary emission
calculates high-resolution x-ray rocking control. Users determine whether to use con-
curves for epitaxial crystal structures and stant beam current or constant beam power
substrates with Takagi-Taupin x-ray scatter- modes. The company says it can tailor the sys-
ing equations of dynamical diffractions. tems to meet specific user requirements.
REFS 4.0 employs the recursive formulation The Eagle III -Probe is the latest addition
of dynamical theory for simulating grazing to the Eagle series of tabletop micro-XRF ele-
incidence x-ray reflectivity for both specular mental analyzers from EDAX (Mahwah, NJ).
and diffuse and fluorescence curves. The system uses capillary optics to concen-
The M4 Fluxer from Corporation Scien- trate the x-ray beam size down to a 100-m
tifique Claisse (Sainte-Foy, Quebec, Canada) diameter on the sample. Look-down, high-
is designed for preparing glass disks for XRF intensity x-ray optics, CCD video imaging
analysis. The system employs fusion to pre- cameras, and motorized xyz stage allow non-
pare oxides, sulfides, alloys, metal powders, destructive, simultaneous NaU analysis of
and precious metals in the processing of ce- solids, liquids, and powders. The analyzer of-
ment, ores, slags, sediments, soils, rocks, ce- fers an optic protection sensor and optional
ramics, pigments, catalysts, ferroalloys, and 30-mm2 Si(Li) detector for spectral resolution
other inorganic materials. Burners produce of 80 mm2. The Eagle III XPL -Probe equips
stable flames at a maximum temperature of the Eagle III with an x-ray polycapillary lens
1630 C. Gas settings of 30% and higher pro- to generate an ultrahigh-intensity x-ray beam
duce a linear increase in temperature, making down to approximately a 40-m diameter on
the system easy to control. The fluxers are the sample. It can be equipped with variable
22 Spectroscopy 17(7) July 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m
Conference Preview

line overlaps. Data collection is peak- dard propane. Two independently ad-

EAGLE III -PROBE/EDAX


based, so detection limits and accuracy justable gas supplies for each burner give
for trace elements are comparable with consistent and reproducible conditions
those of conventional methods. of oxidation and fusion for each station.
Handley also offers reengineered x- Kratos Analytical (Chestnut Ridge,
ray tubes made by Australian X-ray NY) has developed the XRF-1800 wave-
Tubes and XRF sample-preparation length-dispersive sequential XRF spec-
equipment from HD Elektronik und trometer for qualitative and quantitative
spot size of 40 m to 150 m and auto- Elektrotechnik. Reengineered side- analysis. The instrument measures ele-
mated primary filter control. window tubes are available for a range ments from Be to U in 3 min and per-
Three new products are available of analyzers, including the Philips forms local area analysis for wavelength-
from distributor Handley Analytical PW1400 series, Siemens SRS1 and dispersive analysis. A high-sensitivity
Services (Houston, TX). UniQuant 5 is SRS200, and Toshiba-compatible side- aperture and an accurate sample-
the latest version of Omega Data Sys- window tubes to fit Rigaku instru- positioning mechanism allow selected
tems software package for XRF spec- ments. The tubes reportedly have sev- areas of the sample to be tested. The
trometers equipped with a goniometer. eral advantages over OEM tubes, such analysis area can be stepped down to
The software uses an extended form of as improved vacuum and gettering and 250 m, selected freely from within a
the fundamental parameters method to lower cost. They include a warranty 30-mm diameter area.
overcome the problem of spectral line comparable to that of new tubes. The company is also introducing the
overlap. Graphic supported, weighted The Vulcan2 and Vulcan4 automatic EDX-900 Benchtop EDXRF spectrome-
regression concentration is accompa- gas fusion machines are designed to pro- ter, which uses a new thermoelectric
nied by an indication of a practical con- vide highly flexible heating and cooling cooling system to eliminate the need for
fidence interval for true quantitative parameters, making them suitable for liquid nitrogen. The instrument main-
analysis. Theoretical standard devia- any mineral, ore, or ferroalloy sample tains the energy resolution of Si(Li) de-
tions are given based on counting errors preparation. The systems can reach tem- tectors, operates at 10 C, and handles
and propagation of such errors due to peratures as high as 1600 C using stan- count rates as high as 106.
The new XRD-7000 x-ray diffrac-
tometer is designed to accommodate
extra-large samples using a vertical
 goniometer. The goniometer gen-
erates scan rates as high as 1000 /min
and provides angle reproducibility of
0.0002 . The system supports analysis
of samples in various states.
Materials Data, Inc. (MDI, Liver-
more, CA) is now shipping Jade 6.1, the
latest version of its software for XRD
powder pattern processing. The soft-
wares new Xplorer feature finds and
organizes all relevant XRD files from
throughout an organizations network.
For XRD phase analysis, the software
now includes a full physics phase-
modeling option. The system supports
Microsoft XP as well as previous Mi-
crosoft platforms.
Also new from MDI is Riqas 4, a
stand-alone software tool for quantita-
tive analysis using Rietveld methods that
can be used to refine models against
patterns to determine exact composi-
tion and atomic coordinates. FilmScan
3, new film acquisition software for
film-based XRD analysis, is said to de-
liver the same level of automation to
XRD films as electronic diffraction sys-
tems at a much lower price.
24 Spectroscopy 17(7) July 2002 Circle 20
w w w. s p e c t r o s c o p y o n l i n e . c o m
Conference Preview

The new Bullet miniature standard XML language, it offers users


x-ray tube from Moxtek complete control over XRD measure-
(Orem, UT) is a compact ment data. The new XRDML data plat-
x-ray source designed for low- form stores ASCII-format XRD measure-
power applications such as ment data in XML-based files that
battery-operated, portable in- contain all measurement data as well as
struments. It features an inte- information required to reproduce the
grated, high-voltage power data, including instrument type and set-
supply that can provide 40 kV tings, and it allows XRD system users to
and 100 A of output with a share measurements with others. The lat-

MICRONX/THERMO NORAN
maximum input power of 8 est versions of XPert Data Collector,
W. The design allows very HighScore, and Epitaxy are available
close coupling of the source now; XPert Plus, Industry, and Stress
and sample, providing analyt- will be released later this year.
ical results comparable to MiniPal 2 is Philips Analyticals
higher-powered x-ray tubes. newest benchtop EDXRF spectrometer.
The end-window version is It performs nondestructive NaU analy-
available with a palladium or silver the intensity and peak-to-background sis in concentrations from 100% down
anode. Copper and tungsten anodes are ratios as the companys previous to parts-per-million levels. The thermo-
available in a side-window configura- Mo/B4C multilayer systems, it reports, electrically cooled system analyzes liq-
tion. The tube is available in 30- and and is approximately 20% improved uids, powders, and surface coatings
40-kV options. over La/B4C multilayers. using an integrated deconvolution func-
Osmic (Auburn Hills, MI) has intro- Philips Analytical (Almelo, The tion that separates closely spaced peaks
duced the OV160Y boron analyzer, a Netherlands) has developed a new data in the spectrum. A Microsoft Windows
multilayer system for the BK radiation platform for its XPert range of XRD based software can analyze a full range of
line. The system offers nearly double software. Based on the new universal elements without special accessories or
modifications.
Pro-Trace is the latest addition to the
companys XRF SuperQ software for cal-
culating net intensities in trace-element
analysis down to sub-part-per-million
levels. The system uses blank specimen
and optional calibration standards with a
calibrating power of more than 200 in-
ternational standard reference materials.
A Windows-based graphical user inter-
face and available on-line help featuring
worked examples simplify operation. The
software is designed for use with the
companys MagiX, MagiX PRO, and
MagiX FAST XRF systems.
Laser-X is a new industrial miniature
x-ray source for XRF analysis from
Photoelectron Corporation (Lexing-
ton, PA). The device is designed as an
OEM replacement for the radioactive
isotope sources used in hand-held XRF
units and process-control monitors. Ac-
cording to the company, advantages of
the new x-ray source over radioactive
materials include the following:
the radiation source can be turned
on and off at will
the intensity of the x-ray beam is
consistently high, and
the need for source replacement is
eliminated.
26 Spectroscopy 17(7) July 2002 Circle 22
w w w. s p e c t r o s c o p y o n l i n e . c o m
Conference Preview

Rigaku/MSC (The Woodlands, TX) is complete new line of microbeam XRF


introducing the ZSX series of x-ray spec- metrology tools. MicronX systems fea-
trometers designed for routine R&D. ture collimated XRF tools that provide
The ZSX101 is a full-size 4-kW produc- production-worthy film thickness and
tion XRF spectrometer. The ZSX100s composition measurement of metal
can be configured in a tube below or depositions for the microelectronics,
above, depending on the needs of the telecommunications, data storage, and
application. The ZSX mini is a benchtop metal finishing industries. The systems
WD-XRF system that operates at 110 V measure the thickness and composition
and does not require any cooling water of as many as six layers of deposited
for the tube or generator. It can be used metals simultaneously, from angstrom
for applications that dont require a full- to micrometer thickness ranges. These
powered analyzer, or as a back-up for a systems can also determine bulk alloys
larger WD-XRF unit. ZSX software in- for as many as 20 elements. The line in-
corporates the mostadvanced funda- cludes several options, allowing systems
mental parameters software available, to be optimized for specific applica-
according to the company. tions. Features include mechanical or
SPEX CertiPrep, Sample Prepara- optical collimation, several x-ray detec-
tion Division (Metuchen, NJ) intro- tor options, and vacuum options. Using
duces the 8000D and 8000M mixer/ a common system platform and stan-
mills. Both are two-clamp laboratory dardized subassemblies, each tool can be
mills featuring variable-range electronic configured to deliver the best perform-
timers, safety interlocks, and a wide ance for the set of applications meas-
choice of grinding and blending vials. ured by the tool.
The 8000D also offers forced-air cool- X-ray Instrumentation Associates
ing, has twice the capacity of its 8000 (XIA, Newark, CA) introduces the
DXP, a credit cardsized, low-power,
DXP/X-RAY INSTRUMENTATION

full-featured digital spectrometer for use


in portable or embedded x-ray, gamma
ray, and related applications. The system
uses the companys own gate array plus
DSP technology to achieve impressive
ASSOCIATES

performance in a small size and at rela-


tively low cost. Intended primarily for
OEM and other large-volume, low-cost
predecessor, and relies on two clamps applications, the analyzer is supported
that run in balance to provide smoother by a suite of development tools includ-
operation. Grinding/blending vials for ing the COM communication board
both mills are available in hardened and MANAGER board setup and data
steel, stainless steel, tungsten carbide, collection software. The onboard DSP
agate, alumina ceramic, zirconia ce- can support custom codes to carry out
ramic, acrylic, and polystyrene. The real-time spectral analysis and external
electronic timer range can be altered equipment control functions. The sys-
from 100 to 1000 min, making the mills tem offers a choice of 4-, 8-, or 16-MSPS
suitable for lengthy operations such as digitization rate, which sets basic power
mechanical alloying. The systems are consumption (500, 630, or 750 mW)
designed for pulverizing rocks, cement, and maximum throughput (50, 100, or
ceramics, and other materials in the 10- 200 kcps).
g range and for blending powders, XIA is also introducing the Saturn, a
paints, and emulsions in volumes as completely digital x-ray spectrometer
high as 60 mL or more. The high im- suitable for high-rate and high-
pact energy of the grinding balls has re- resolution applications. The newest
cently proved effective for mechanical member of the DXP line, the Saturn is
alloying in the production of supercon- intended for industrial and laboratory
ducting materials, the company reports.
Thermo NORAN, a Thermo Elec- Conference Preview continued on
tron business (Middleton, WI), offers a page 35
Circle 23 July 2002 17(7) Spectroscopy 27
TUTOR IAL

A Beginners Guide to ICP-MS


Part XI Peak Measurement Protocol
Robert Thomas
With its multielement capability, superb detection limits, wide dynamic the detection limits required
range, and high sample throughput, inductively coupled plasmamass the precision and accuracy
spectrometry (ICP-MS) is proving to be a compelling technique for more expected
and more diverse application areas. However, no two application areas the dynamic range needed
have the same analytical requirements. For example, environmental the integration time used
and clinical contract laboratories although requiring reasonably low the peak quantitation
detection limits do not really push the technique to its extreme routines.
detection capability. Their main requirement is usually high sample Before discussing these fac-
throughput because the number of samples these laboratories can tors in greater detail, and how
analyze in a day directly impacts their revenue. On the other hand, a they affect data quality, it is im-
semiconductor fabrication plant or a supplier of high-purity chemicals portant to remember how a
to the electronics industry is interested in the lowest detection limits the scanning device such as a
technique can offer because of the contamination problems associated quadrupole mass analyzer
with manufacturing high performance electronic devices. works (1). Although we will
focus on quadrupole technol-
ogy, the fundamental principles

Robert
Thomas
has more than
M odern ICP-MS must
be very flexible to
meet such diverse
application needs
and keep up with the increasing
demands of its users. Nowhere
techniques such as flow injec-
tion and laser ablation.

Measurement Variables
Many variables affect the quality
of the analytical signal in ICP-
of measurement protocol will be
very similar for all types of mass
spectrometers that use a scan-
ning approach for multielement
peak quantitation.

30 years of is this more important than in MS. The analytical requirements Measurement Protocol
experience in
the area of peak integration and of the application will often dic- Figure 1 shows the principles of
trace element
analysis. He is measurement protocol. The way tate this factor, but instrumental scanning with a quadrupole
the principal of the analytical signal is managed detection and measurement pa- mass analyzer. In this simplified
his own in ICP-MS directly impacts its rameters can have a significant example, the analyte ion (black)
freelance writing multielement capability, detec- impact on the quality of data in and four other ions (colored)
and consulting
tion limits, dynamic range, and ICP-MS. Some of the variables have arrived at the entrance to
company,
Scientific sample throughput the four that can affect the quality of the four rods of the quadrupole.
Solutions, based major strengths that attracted your data, particularly when car- When a particular rf/dc voltage
in Gaithersburg, the trace element community to rying out multielement analysis, is applied to the rods, the posi-
MD. He can be the technique almost 20 years include tive or negative bias on the rods
contacted by
e-mail at
ago. To understand signal man- whether the signal is continu- will electrostatically steer the
thomasrj@ agement and its implications on ous or transient analyte ion of interest down the
bellatlantic.net data quality in greater detail, the temporal length of the middle of the four rods to the
or via his web this installment of this series sampling event end, where it will emerge and be
site at www. will discuss how measurement the volume of sample converted to an electrical pulse
scientificsolutions1.
protocol is optimized based on available by the detector. The other ions
com.
the applications analytical re- the number of samples being of different mass-to-charge ra-
quirements. I will discuss its analyzed tios will pass through the spaces
impact on both continuous sig- the number of replicates per between the rods and be ejected
nals generated by traditional sample from the quadrupole. This scan-
nebulization devices and tran- the number of elements being ning process is then repeated for
sient signals produced by alter- determined another analyte at a completely
native sample introduction different mass-to-charge ratio
28 Spectroscopy 17(7) July 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m
Tutorial

until all the analytes in a multielement


analysis have been measured. +
Ion
The process for detecting one partic- + flow
Quadrupole rods
ular mass in a multielement run is rep- +
resented in Figure 2, which shows a 63Cu +
+
ion emerging from the quadrupole and
+
being converted to an electrical pulse by
the detector. As the rf/dc voltage of the
-
quadrupole corresponding to 63Cu
is repeatedly scanned, the ions as
electrical pulses are stored and counted +
+
by a multichannel analyzer. This multi-
channel data-acquisition system typi- Figure 1. Principles of mass selection with a quadrupole mass filter.
cally has 20 channels per mass and as
the electrical pulses are counted in each
channel, a profile of the mass is built- Detector
up over the 20 channels, corresponding
Quadrupole
to the spectral peak of 63Cu. In a multi-
element run, repeated scans are made
over the entire suite of analyte masses, Copper ion (m/z 63)
as opposed to just one mass represented Mass scan (amu)
in this example. 63
63
The principles of multielement peak Cu
acquisition are shown in Figure 3. In
this example (showing two masses), sig-
nal pulses are continually collected as
the quadrupole is swept across the mass
spectrum (in this case three times).
After a given number sweeps, the total 63
Cu scan Quadrupole
number of signal pulses in each channel mass scan
Multichannel data controller
are counted. acquisition
When it comes to quantifying an iso-
topic signal in ICP-MS, there are basi-
cally two approaches to consider (2).
One is the multichannel ramp scanning
approach, which uses a continuous Channel number 1 2 3 4 5 6 7 ............ n
smooth ramp of 1 to n channels (where
n is typically 20) per mass across the Figure 2. Detection and measurement protocol using a quadrupole mass analyzer.
peak profile. This approach is shown in
Figure 4.
The peak-hopping approach is where on the masses of interest. Full peak pro- mass stability is good enough to repro-
the quadrupole power supply is driven filing is not normally used for doing ducibly go to the same mass point every
to a discrete position on the peak (nor- rapid quantitative analysis because time. If good mass stability can be guar-
mally the peak point) and allowed to valuable analytical time is wasted taking anteed (usually by thermostating the
settle; a measurement is then taken for a data on the wings and valleys of the quadrupole power supply), measuring
fixed amount of time. This approach is peak, where the signal-to-noise ratio is the signal at the peak maximum will al-
represented in Figure 5. poorest. ways give the best detection limits for a
The multipoint scanning approach is When the best possible detection given integration time. It is well docu-
best for accumulating spectral and peak limits are required, the peak-hopping mented that there is no benefit to
shape information when doing mass approach is best. It is important to un- spreading the chosen integration time
scans. It is normally used for doing derstand that, to get the full benefit of over more than one measurement point
mass calibration and resolution checks, peak hopping, the best detection limits per mass. If time is a major considera-
and as a classical qualitative method de- are achieved when single-point peak tion in the analysis, then using multiple
velopment tool to find out what ele- hopping at the peak maximum is cho- points is wasting valuable time on the
ments are present in the sample, as well sen. However, to carry out single-point wings and valleys of the peak, which
as to assess their spectral implications peak hopping, it is essential that the contribute less to the analytical signal

July 2002 17(7) Spectroscopy 29


Tutorial

consideration with a single-shot laser


Number of ions pulse that lasts 510 s. Also with a con-
counted
tinuous signal produced by a concentric
nebulizer, we might have to accept a
compromise of detection limit based on
Sweep 1
the speed of analysis requirements or

Intensity (cps)
amount of sample available. What ana-
lytical precision is expected? If its iso-
1 60 tope ratio/
dilution work, how many ions do we
have to count to guarantee good preci-
sion? Does increasing the integration
Sweep 2 time of the measurement help the pre-
5 6 7 8 9 10 11 cision? Finally, is there a time constraint
Mass (amu) on the analysis? A high-throughput lab-
1 60 oratory might not be able to afford to
use the optimum sampling time to get
the ultimate in detection limit. In other
words, what compromises need to be
Sweep 3 made between detection limit, preci-
sion, and sample throughput? Clearly,
before the measurement protocol can
be optimized, the major analytical re-
Intensity (cps)

1 60
quirements of the application need to
Total
be defined. Lets take a look at this
process in greater detail.
Total of
n
sweeps Multielement Data Quality
Objectives
Because multielement detection capa-
1 60 5 6 7 8 9 10 11
bility is probably the major reason why
MCA channel number Mass (amu)

Figure 3 (above left). A profile of the peak is built up by continually sweeping the quadrupole Table I. Precision of Pb isotope
across the mass spectrum. ratio measurement as a function
Figure 4 (above right). Multichannel ramp scanning approach using 20 channels per amu. of dwell time using a total
Figure 5 (below right). Peak-hopping approach. integration time of 5.5 s.
Dwell
time %RSD, %RSD,
and more to the background noise. Fig- quirement of the application. When
(ms) 207
Pb/206Pb 208
Pb/206Pb
ure 6 shows the degradation in signal-to- multielement analysis is being carried
2 0.40 0.36
background noise ratio of 10 ppb Rh out by ICP-MS, a number of decisions
5 0.38 0.36
with an increase in the number of points need to be made. First, we need to know
10 0.23 0.22
per peak, spread over the same total inte- if we are dealing with a continuous sig-
25 0.24 0.25
gration time. Detection limit improve- nal from a nebulizer or a transient sig-
50 0.38 0.33
ment for a selected group of elements nal from an alternative sampling acces-
100 0.41 0.38
using 1 point/peak, rather than 20 sory. If it is a transient event, how long
points/peak, is shown in Figure 7. will the signal last? Another question
that needs to be addressed is, how many most laboratories invest in ICP-MS, it is
Optimization of Measurement elements are going to be determined? important to understand the impact of
Protocol With a continuous signal, this isnt such measurement criteria on detection lim-
Now that the fundamentals of the a major problem, but it could be an its. We know that in a multielement
quadrupole measuring electronics have issue if we are dealing with a transient analysis, the quadrupoles rf/dc ratio is
been described, let us now go into more signal that lasts a few seconds. We also scanned to mass regions or driven,
detail on the impact of optimizing the need to be aware of the level of detec- which represent the elements of inter-
measurement protocol based on the re- tion capability required. This is a major est. The electronics are allowed to settle

30 Spectroscopy 17(7) July 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m


Tutorial

3.5 4
Ratio of signal-to-background

Detection limit improvement


3.0
3
noise ( 105)

2.5
2
2.0

1.5 1

1.0
0
1 3 5 7 9 11 13 15 17 19
Be-9 Co-59 In-115 Tl-205 U-238
Number of points per peak
Element

Figure 6. Signal-to-background noise ratio degrades when more than Figure 7. Detection limit improvement using 1 point/peak rather than 20
one point, spread over the same integration time, is used for peak points/peak over the mass range.
quantitation.

and then dwell on the peak, or sit, and nificant amount of time is spent scan- Dwell Time  #Sweeps  #Elements  #Replicates  100

take measurements for a fixed period of ning and settling the quadrupole, which Dwell Time  #Sweeps  #Elements  #Replicates 

time. This step is usually performed a doesnt contribute to the quality of the Scanning / Settling Time  #Sweeps  #Elements  #Replicates
number of times until the total integra- analytical signal. Therefore, if the meas-
tion time is fulfilled. For example, if a urement routine is not optimized care- So to achieve the highest measure-
dwell time of 50 ms is selected for all fully, it can have a negative impact on ment efficiency, the nonanalytical time
masses and the total integration time is data quality. The dwell time can usually must be kept to an absolute minimum.
1 s, then the quadrupole will carry out be selected on an individual mass basis, This leads to more time being spent
20 complete sweeps per mass, per repli- but the scanning and settling times are counting ions and less time scanning
cate. It will then repeat the same rou- normally fixed because they are a func- and settling, which does not contribute
tine for as many replicates that have tion of the quadrupole and detector to the quality of the analytical signal.
been built into the method. This electronics. For this reason, it is essen- This factor becomes critically impor-
process is illustrated very simplistically tial that the dwell time which ulti- tant when a rapid transient peak is
in Figure 8, which shows the scanning mately affects detection limit and preci- being quantified, because the available
protocol of a multielement scan of sion must dominate the total measuring time is that much shorter
three different masses. measurement time, compared with the (3). Generally speaking, peak quantita-
In this example, the quadrupole is scanning and settling times. It follows, tion using multiple points per peak and
scanned to mass A. The electronics are therefore, that the measurement duty long settling times should be avoided in
allowed to settle (settling time) and left cycle (percentage of actual measuring ICP-MS because it ultimately degrades
to dwell for a fixed period of time at time compared with total integration the quality of the data for a given inte-
one or multiple points on the peak time) is maximized when the quadru- gration time.
(dwell time); intensity measurements pole and detector electronics settling Figure 9 also shows that shorter dwell
are then taken (based on the dwell times are kept to an absolute minimum. times translate into a lower measure-
time). The quadrupole is then scanned Figure 9 shows a plot of percentage of ment efficiency. For this reason, it is
to masses B and C and the measure- measurement efficiency against dwell probably desirable, for normal quanti-
ment protocol repeated. The complete time for four different quadrupole set- tative analysis work, to carry out multi-
multielement measurement cycle tling times 0.2, 1.0, 3.0, and 5.0 ms ple sweeps with longer dwell times
(sweep) is repeated as many times as for one replicate of a multielement scan (typically 50 ms) to get the best detec-
needed to make up the total integration of five masses, using one point per tion limits. So if an integration time of
per peak. It should be emphasized that peak. In this example, the total integra- 1 s is used for each element, this would
this example is a generalization of the tion time for each mass was 1 s, with translate into 20 sweeps of 50 ms dwell
measurement routine management the number of sweeps varying, depend- time per mass. Although 1 s is long
of peak integration by the software ing on the dwell time used. For this ex- enough to achieve reasonably good de-
will vary slightly, based on different ercise, the percentage of measurement tection limits, longer integration times
instrumentation. efficiency is defined by the following generally have to be used to reach the
It is clear from this information that, calculation:
during a multielement analysis, a sig-
32 Spectroscopy 17(7) July 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m
Tutorial

lowest possible detection limits. Figure Figure 8 (left).


10 shows detection limit improvement Scan back to mass Multielement scanning and
A and repeat measurement protocol of a
as a function of integration time for Scan
238 Scan Scan quadrupole.
U. As would be expected, there is a
fairly predictable improvement in the
detection limit as the integration time is Settle, dwell, Settle, dwell,
Settle, dwell,
increased because more ions are being measure measure measure
counted without an increase in the
background noise. However, this only
holds true up to the point where the
pulse-counting detection system be-
comes saturated and no more ions can Figure 9 (below). Percent
be counted. In the case of 238U, this oc- Mass A Mass B Mass C of measurement efficiency
curs around 25 s, because there is no as a function of dwell time
obvious improvement in detection limit with varying scanning/
Increasing mass
at a higher integration time. So from settling times.
these data, we can say that there appears
to be no real benefit in using an inte- 100
gration time longer than 7 s. When de-
ciding the length of the integration 90
% Measurement efficiency

time in ICP-MS, you have to weigh the Scanning/settling time


80
detection limit improvement against 0.2 ms
the time taken to achieve that improve- 1 ms
70
3 ms
ment. Is it worth spending 25 s measur-
5 ms
ing each mass to get a 0.02 ppt detec- 60
tion limit if 0.03 ppt can be achieved
50
using a 7-s integration time? Alterna-
tively, is it worth measuring for 7 s 40
when 1 s will only degrade the perform- 100 90 80 70 60 50 40 30 20 10 5
ance by a factor of 3? It really depends Dwell time (ms)
on your data quality objectives.
For applications such as isotope
dilution/ratio studies, high precision is
also a very important data quality ob- introduction pulsations or plasma fluc-
jective (4). However, to understand %RSD  N  100 tuations) dominate the signal, which
N
what is realistically achievable, we must leads to poorer standard deviation
be aware of the practical limitations of In practice this holds up very well, as values.
measuring a signal and counting ions in shown in Figure 11. In this plot of stan- So based on counting statistics, it is
ICP-MS. Counting statistics tells us that dard deviation as a function of signal logical to assume that the more ions
the standard deviation of the ion signal intensity for 208Pb, the dots represent the that are counted, the better the preci-
is proportional to the square root of the theoretical relationship as predicted by sion will be. To put this in perspective,
signal. It follows, therefore, that the rel- counting statistics. It can be seen that it means that at least 1 million ions
ative standard deviation (RSD), or pre- the measured standard deviation (bars) need to be counted to achieve an RSD
cision, should improve with an increase follows theory very well up to about of 0.1%. In practice, of course, these
in the number (N) of ions counted as 100,000 cps. At that point, additional kinds of precision values are very diffi-
shown by the following equation: sources of noise (for example, sample cult to achieve with a scanning quadru-

Table II. Impact of integration time on the overall analysis time for Pb isotope ratios.
Dwell time Number of Integration time %RSD, %RSD, Analysis time
(ms) sweeps (s)/mass 207
Pb/206Pb 207
Pb/206Pb for 9 reps
25 220 5.5 0.24 0.25 2 min 29 s
25 500 12.5 0.21 0.19 6 min 12 s
25 700 17.5 0.20 0.17 8 min 29 s

July 2002 17(7) Spectroscopy 33


Tutorial

quired. That is why integration times of


510 min are commonly used for deter-
0.09 mining isotope ratios with a quadru-
pole ICP-MS system (5, 6). For this type
0.08 of analysis, when two or more isotopes
0.07
are being measured and ratioed to each
other, it follows that the more simulta-
Detection limit (ng/L, ppt)

0.06 neous the measurement, the better the


0.05
precision becomes. Therefore, the abil-
ity to make the measurement as simul-
0.04 taneous as possible is considered more
desirable than any other aspect of the
0.03
measurement. This is supported by the
0.02 fact that the best isotope ratio precision
data are obtained with time-of-flight or
0.01
multicollector, magnetic sector ICP-MS
0.00 systems, which are both considered si-
1 3 5 7 9 25 50 multaneous in nature. So the best way
Integration time (s) to approximate simultaneous measure-
ment with a rapid scanning device, such
Figure 10. Plot of detection limit against integration time for 238U. as a quadrupole, is to use shorter dwell
and scanning/settling times, resulting in
more sweeps for a given integration
10,000 time. Table I shows precision of Pb iso-
Predicted tope ratios at different dwell times car-
Measured ried out by researchers at the Geological
1000 Survey of Israel (7). The data are based
Standard deviation (cps)

on nine replicates of a NIST SRM-981


(75 ppb Pb) solution, using 5.5 s of in-
tegration time per isotope.
100
From these data, the researchers con-
cluded that a dwell time of 10 or 25 ms
offered the best isotope ratio precision
10 measurement (quadrupole settling time
was fixed at 0.2 ms). They also found
that they could achieve slightly better
1 precision by using a 17.5-s integration
101 102 103 104 105 106 107 time (700 sweeps at a 25-ms dwell
Signal intensity (cps) time), but felt the marginal improve-
ment in precision for nine replicates
Figure 11. Comparison of measured standard deviation of a 208Pb signal against that predicted by was not worth spending the approxi-
counting statistics. mately 3.5-times-longer analysis time,
as shown in Table II.
This work shows the benefit of opti-
pole system because of the additional ones at or near the center of the peak. mizing the dwell time, settling time,
sources of noise. If this information is For this reason, if good precision is a and the number of sweeps to get the
combined with our knowledge of how fundamental requirement of your data best isotope ratio precision data. The
the quadrupole is scanned, we begin to quality objectives, it is best to use researchers were also very fortunate to
understand what is required to get the single-point peak hopping with integra- be dealing with relatively healthy signals
best precision. This is confirmed by the tion times in the order of 510 s. On for the three Pb isotopes, 206Pb, 207Pb,
spectral scan in Figure 12, which shows the other hand, if high-precision iso- and 208Pb (24.1%, 22.1%, and 52.4%
the predicted precision at all 20 chan- tope ratio or isotope dilution work is abundance, respectively). If the isotopic
nels of a 5 ppb 208Pb peak (2). being done in which analysts would signals were dramatically different like
Therefore, the best precision is ob- like to achieve precision values ap- in 235U to 238U (0.72 % and 99.2745%
tained at the channels where the signal proaching counting statistics then abundance, respectively), then the abil-
is highest, which as we can see are the much longer measuring times are re- ity to optimize the measurement proto-

34 Spectroscopy 17(7) July 2002 w w w. s p e c t r o s c o p y o n l i n e . c o m


Conference Preview

Conference Preview continued from and 50 W of power to generate an ex-


1.8 7
page 27 tremely high flux-density gain, the com-
pany reports. Increased spatial resolution

Predicted precision from counting


1.6 cps 6
1.4 %RSD applications and provides everything and beam stability are also promised. An
Sensitivity (cps  105)

5 necessary for a typical x-ray detector, integrated cooling system eliminates the
1.2
4
incorporating one DXP spectrometer need for a separate cooling unit.
1.0
channel, preamplifier power, and detec- Attendees can see many of these
0.8 3 tor HV bias in one compact chassis. Its products, along with others not men-
0.6
2 input is compatible with a wide range tioned, at the 2002 Denver X-ray Con-
0.4
1
of common detectors, including pulsed ference sponsored by the Interna-
0.2 optical reset, transistor reset, and RC tional Centre for Diffraction Data at
0 0 feedback preamplifiers. The Saturn of- Antlers Adams Mark Hotel (formerly
207.5
207.6
207.7
207.8
207.9
208
208.1
208.2
208.3
208.4
208.5

fers complete computer control over all Antlers Doubletree Hotel), Colorado
amplifier and spectrometer functions Springs, Colorado, July 29August 2,
Mass (m/z)
including gain, filter peaking time, and 2002. For more information, contact
Figure 12. Comparison of % RSD with signal pileup inspection criteria. Its DXP digi- Denise Flaherty, DXC Conference Co-
intensity across the mass profile of a 208Pb peak. tal filters significantly increase through- ordinator, 12 Campus Boulevard, New-
put compared to typical analog town Square, PA 19073-3273, (610)
systems. 325-9814, fax: (610) 325-9823,
col for individual isotopes becomes of The new X-Beam x-ray source from e-mail: flaherty@icdd.com, web site:
even greater importance to guarantee X-Ray Optical Systems (Albany, NY) www.dxcicdd.com.
precise data. delivers an intense, micrometer-sized
It is clear that the analytical demands focal spot. Designed for OEM use in
put on ICP-MS are probably higher micro-XRF instruments, the compact
than any other trace element technique unit uses polycapillary focusing optics
because it is continually being asked to
solve a wide variety of application
problems. However, by optimizing the
measurement protocol to fit the analyt-
ical requirement, ICP-MS has shown
that it has the capability to carry out
rapid trace element analysis, with su-
perb detection limits and good preci-
sion on both continuous and transient
signals, and still meet the most strin-
gent data quality objectives.

References
1. R. Thomas, Spectroscopy 16(10),
4448 (2001).
2. E.R. Denoyer, At. Spectroscopy 13(3),
9398 (1992).
3. E.R. Denoyer and Q.H. Lu, At. Spec-
troscopy 14(6), 162169 (1993).
4. T. Catterick, H. Handley, and S. Merson,
At. Spectroscopy 16(10), 229234
(1995).
5. T.A. Hinners, E.M. Heithmar, T.M. Spit-
tler, and J.M. Henshaw, Anal. Chem. 59,
26582662 (1987).
6. M. Janghorbani, B.T.G. Ting, and N.E.
Lynch, Microchemica Acta 3, 315328,
(1989).
7. L. Halicz, Y. Erel, and A. Veron, At. Spec-
troscopy 17(5), 186189 (1996).

Circle 25 July 2002 17(7) Spectroscopy 35

You might also like