NBS
U.S. Department
of Commerce
National
Bureau ot
Standards
Office of Environmental
Measurements
Washington, DC 20234
EPA
United States
Environmental Protection
Agency
Office of Environmental Engineering
and Technology
Washington DC 20460
EPA-600 7-80-122
June 1980
Research and Development
Microanalysis of
Individual Layered
Particles by
Secondary Ion Mass
Spectrometry
Interagency
Energy/Environment
R&D Program
Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
MICROANALYSIS OF INDIVIDUAL LAYERED PARTICLES BY
SECONDARY ION MASS SPECTROMETRY
by
Dale E. Newbury
Gas and Particulate Science Division
National Bureau of Standards
Washington, DC 20234
Interagency Energy/Environment R and D Program Report
-------
DISCLAIMER
This document is a preliminary draft. It has not been formally released
by the U. S. Environmental Protection Agency and should not at this stage be
construed to represent Agency policy. It is being circulated for comments
on its technical merit and policy implications.
ii
-------
FOREWORD
The role of the National Bureau of Standards in the Interagency
Energy/Environment R&D Program, coordinated by the Office of Research
and Development of the U. S. Environmental Protection Agency, is to
provide those services necessary to assure data quality in measurements
being made by the Federal, state, local, and industrial laboratories
participating in the interagency program. The work at NBS is coordinated
by the Office of Environmental Measurements and is conducted in the
Center for Analytical Chemistry, the Center for Radiation Research,
and the Center for Thermodynamics and Molecular Science. NBS activities
form part of the Characterization, Measurement, and Monitoring Program
Category and address the data quality assurance needs of air and water
monitoring programs. NBS efforts in support of data quality assurance
include:
0 Studies of the feasibility of production of
Standard Reference Materials which could be
used for the verification of performance audit
samples for quality control programs or used
for the calibration of field and laboratory
instruments.
0 The development and demonstration of new or
improved measurement methods, particularly when
needed for the certification of Standard Reference
Materials.
0 The evaluation and dissemination of data on the
physical and chemical properties of effluents,
products and raw materials of environmental
significance in energy production.
0 The provision of reference materials for the
evaluation and validation of monitoring methods.
This report is one of the Interagency Energy/Environment Research and
Development Series Reports prepared to provide detailed information
on the development of an NBS measurement standard or method.
WILLIAM H. KIRCHHOFF, Crfief
Office of Environmental Measurements
National Bureau of Standards
-------
ABSTRACT
Secondary ion mass spectrometry is evaluated for application to the
determination of the composition and structure of individual particles.
Analyses of many elemental constituents at the ppm level can be obtained
in individual particles as small as 2 ym in diameter. Molecular signals,
both organic and inorganic, can be detected from particles above 20 ym in
diameter. Quantitative analyses of elemental constituents can be made with
a relative accuracy of 25 percent by means of empirical relative sensitivity
factors and within a factor of two by means of a physical model. Multi-
element depth profiles can be obtained from individual particles as small
as 4 ym in diameter. Depth profiles of individual particles from SRM 1648
Urban Air Particulate reveal pronounced surface concentrations of lead and
barium. Implementation of SIMS depth profiling requires automation to make
use of the full spectral information and to eliminate matrix effects by
normalization.
IV
-------
CONTENTS
Abstract iv
Figures vi
Tables viii
1. Introduction 1
2. Conclusions 3
3. Methods for Quantitative Analysis in Secondary Ion
Mass Spectrometry 4
4. Experimental Methods 23
5. Detection of Molecular Signals by SIMS 30
6, Application of SIMS to Environmental Samples 38
References 57
v
-------
FIGURES
Number
1 Schematic illustration of the sputtering process,
illustrating secondary ion emission and neutraliza-
tion processes 5
2 Secondary ion mass spectrum of NBS glass K 326 showing
secondary ion of light elements, and associated
molecular ions 6
3 Secondary ion mass spectrum of NBS glass K 309 showing
secondary ions of an intermediate mass species, barium,
and the associated oxide and hydroxide ions 7
4 (a) Normalized positive secondary ion yields under 0
bombardment (5); 4. • • • 8
(b) Normalized negative secondary ion yields under Cs
bombardment (5) 9
5 (a) Working curve for niobium in an iron matrix (NBS
SRM 660 steels) 12
(b) Working curve for chromium in an iron matrix ...... 13
6 Relative sensitivity factors for selected elements
compared to silicon as measured in a glass matrix and
as calculated from pure element data (11) 15
7 Error factor histogram for quantitative analysis of
glasses by the relative sensitivity factor method with
average values of sensitivity factors (4) . . . . 19
8 Error factor histogram for quantitative analysis by
the relative sensitivity factor method. Blind round
robin results on unknown glasses (19) 20
9 Error factor histogram for quantitative analysis of
glasses by the local thermal equilibrium model method
(4) 21
10 Spectrum of bulk NBS glass K 227 obtained with
negative primary ions 22
vi
-------
11 Spectrum of a single particle obtained with a
positive primary ion 27
12 Secondary ion spectrum of pure silicon showing
inorganic molecular signals 31
13 Secondary ion spectrum of glycine on a nickel
sheet substrate showing numerous molecular
fragments 33
14 Secondary ion spectrum of atropine on a
nickel sheet substrate 34
15 Secondary ion spectrum of fluorodopamine on gold-
alloy substrate 35
16 Portion of a secondary ion mass spectrum from glycine
on a 20 ym nickel ball 36
17 Secondary ion mass spectrum of glass K 309 in fiber
form 39
18 Secondary ion mass spectrum of glass K 309 in bulk
form 40
19 Secondary ion spectrum of K 309 spherical glass
particle collected from wall of remelt furnace 41
20 Secondary ion mass spectra of an individual particle
from SRM 1648 (Urban Air Particulate) (a) Positive
secondary ions; (b) Negative Secondary ions 43
21 Multi-element depth profile obtained by manual scanning.
Sample: individual particle of SRM 1648 (Urban Air
Particulate) with a diameter of approximately 10
micrometers 46
22 Multi-element depth profiles of individual particles
from SRM 1648 obtained with automated ion microscope
(a) Particle size 5 micrometers; (b) Particle size
4 micrometers 47
23 Near-surface secondary ion spectra obtained from single
10 micrometer sized sediment particles (a) New York
Harbor; (b) Ohio River; (c) Indiana Harbor Canal 49
24 Single element depth profiles obtained from different
individual particles from New York Harbor (a) Iron;
(b) Carbon, (c) Aluminum and (d) Silicon 52
vii
-------
TABLES
Number
1 Comparison of Measured and Calculated Relative
Sensitivity Factors 16
2 Detection Limits in Silicon-Lead-Oxygen Glass 29
3 Analysis of the Surface Layer of An Urban Environmental
Particulate with Relative Sensitivity Factors Derived
from NBS Glasses 42
viii
-------
1. INTRODUCTION
When a particulate sample is examined by an analytical technique which
is capable of characterizing individual particles, we frequently observe
that the sample is inhomogeneous on the scale of the individual particles,
even when major elements are considered. It is often of considerable value
in identifying the origin(s) of a particulate sample to have knowledge of
the types of individual particles which make up the sample and their fre-
quency. The highly developed and widely used technique of electron probe
microanalysis is capable of characterizing major and minor elements at a
concentration greater than 0.1% and with an atomic number of 11 or more in
individual particles. Once we are capable of assessing variations in com-
position between individual particles, the next level of interest becomes
the study of structure which may exist within an individual particle. Elec-
tron probe microanalysis can characterize local inhomogeneities, but the
spatial resolution is limited by the finite size of the electron interaction
region, which has a volume of about 1 ym3. Studies with the electron probe
frequently reveal that individual particles have complex variations in com-
position on this scale. However, because of the penetration of the electrons
into the solid, to a range of 1 ym or more, it is not possible to study
compositional variations along the depth axis of individual particles.
In recent years, the technique of secondary ion mass spectrometry (SIMS)
has been developed, offering a number of interesting advantages to the study
of particles (1,2,3). In the SIMS technique, a beam of energetic ions is
used to sputter atoms from a target. During the sputtering process a frac-
tion of the atoms knocked from the target is emitted in a charged state, the
so-called secondary ions. These secondary ions are analyzed within a mass
spectrometer, producing a mass spectrum of a small volume of the target.
Because it is a mass spectrometric technique, SIMS is capable of detecting
all elements in a target. Molecular species can also be detected in SIMS
spectra, including complex organic compounds. Isotopic ratios can also be
measured.
The lateral resolution of the ion microprobe, a focussed beam SIMS
instrument, is of the order of 1 ym, which is similar to that of the electron
microprobe. However, the depth resolution of a SIMS instrument is much
better than that of the electron microprobe. Secondary ions are emitted from
a region of the sample which lies with 1 - 2 nm of the sample surface, which
is about one percent of the sampling depth of x-rays in electron probe
microanalysis. The sputtering process which forms the basis of SIMS provides
a means by which the sample can be eroded in a controlled fashion. Thus, in
a layered sample, the structure perpendicular to the surface can be exposed
in a sequential fashion. Coupled with the shallow sampling depth of the
-------
secondary ions, SIMS can provide an elegant technique for the characteriza-
tion of particles both laterally and in depth.
In order to successfully apply SIMS to the study of particles, this
study addressed three problems. (1) Quantitative analysis by SIMS is com-
plicated by the existence of strong matrix effects on secondary ion signals.
In order to establish a quantitative basis for SIMS analysis, an extensive
study was conducted on empirical and theoretical models for the interpreta-
tion of the SIMS spectra. This study is reported in Section 2 of this
report. (2) Particles of a non-conducting nature represent a difficult
target for charged-beam analysis techniques. Methods were developed for the
mounting and analysis of individual particles (Section 3). (3) SIMS spectra
frequently contain signals which are related to inorganic and organic molec-
ular fragments. A study was conducted to assess the utility of the ion
microprobe, a high current density SIMS instrument, for the detection of
characteristic signals from various compounds when the target is in the form
of small particles (Section 4). Finally, based upon these studies, a number
of particles taken from samples of environmental interest have been examined
by SIMS (Section 5).
-------
2. CONCLUSIONS
This study has demonstrated that:
(1) High sensitivity mass spectrometry can be successfully carried out on
individual environmental particulates with a size below 10 ym, even if these
particles have an insulating character.
(2) Molecular signals, both inorganic and organic, can be detected at inter-
mediate ion beam current densities appropriate to the measurement of particles
as small as 20 ym.
(3) Quantitative analyses can be performed with relative sensitivity factors
with a relative accuracy of 25 percent in a majority of cases. Physical
models of ion emission can yield relative accuracies of a factor of two or
better in a majority of cases.
(4) Sensitivities in the ppm range can be obtained for many elements in a
silicate matrix when the sample is in the form of a particle of 2 ym or
larger.
(5) Multi-element depth profiles can be successfully obtained from individual
particles as small as 4 ym in size. Depth resolutions of 100 nm or better
are possible on particles.
(6) In SRM 1648 Urban Air Particulate, depth profiling reveals that heavy
metals such as lead and barium are predominantly concentrated at the surface
of the particles, with the surface concentration as much as 100 times greater
than the interior.
(7) Practical implementation of the SIMS depth profiling technique to indi-
vidual particles requires an automated instrument to make full use of the
multi-element characterization capability and to maximize the depth resolu-
tion.
-------
3. METHODS FOR QUANTITATIVE ANALYSIS IN SECONDARY ION MASS SPECTROMETRY
INTRODUCTION
In performing chemical analysis for the elemental constituents of a
sample by SIMS, it is necessary to relate the measured secondary ion inten-
sities detected from the sputtered assemblage to the atom concentrations in
the original, undisturbed material. The analytical methods by which the
composition can be calculated from the spectral data include a variety of
empirical and theoretical approaches (4). To appreciate the magnitude of the
calculational problem, it is useful to consider the various factors which
influence the measured secondary ion intensities: (1) Matrix effects—the
sputtering process is illustrated schematically in Figure 1. The secondary
ion fraction of the sputtered atoms is small, typically in the range 10 3 -
10 . Atomic ions, M~, are the predominant species observed in the secondary
ion fraction for light elements (Z<40); cluster ions, M~, and molecular ions,
M 0~, are also observed (Figure 2 and Figure 3). In the case of bombardment
x y
of a heavy element with a reactive primary ion such as oxygen, the molecular
ions tend to dominate the atomic ions by a factor of 10 or more.
The normalized positive secondary ion yield observed from the pure
chemical elements, Figure 4, shows a variation of more than four orders of
magnitude across the periodic table, approximately related to the reactivity
of the individual elements with oxygen (5). When an electropositive primary
ion such as cesium is used and negative secondary ions are measured, a
similar strong variation in the normalized secondary ion intensity is observed,
with the sensitivity related to the electron affinity (4). These extraor-
dinary variations in the sensitivity are believed to occur as a result of
differences in the availability of electrons in the surface region from which
the secondary ions are ejected. Free electrons act to modulate the processes
of neutralization or electron capture (6). Thus, variations in the elec-
tronic character of the surface can lead to pronounced matrix effects. The
presence of adsorbed reactive gas species, the accumulation of the implanted
primary ion species, as well as the basic characteristics of the bulk matrix
specier, can influence the surface electronic character. Moreover, the nature
of the surface region can change during the course of a SIMS experiment due
to sputter induced variations in the adsorbed oxygen on the surface and the
time dependent build-up of the implanted primary ion species. Thus, the SIMS
signal is subject to time-varying effects. (2) Instrumental effects—the
extremely strong matrix effects noted above have tended to overshadow the
-------
Primary ions
(n*; /x>
Secondary ions
Neutral atoms
>-
/
Vacuum
Surface
Solid
Figure 1. Schematic illustration of the sputtering process, illustrating secondary ion emission and
neutralization processes.
-------
Figure 2. Secondary ion mass spectrum of NBS glass K 326 showing secondary ions of light elements, and
associated molecular ions.
-------
Figure 3. Secondary ion mass spectrum of NBS glass K 309 showing secondary ions of an intermediate mass
species, barium, and the associated oxide and hydroxide ions.
-------
13 S K«V O-
• PURE ElEMENT
A COMPOUND
10
40 50 60 70
ATOMIC NUMBER (Z)
100
Figure 4. (a) Normalized positive secondary ion yields under 0 bombardment
(5).
-------
108 =
N.D.? tB.D. I Inl B.D.t
I I
16.5 K«V C»+
• PURE ELEMENT
A COMPOUND
B.O. = BARELY DETECTABLE
N.O. = NOT DETECTED
102
Ao ~
ThA
10 20 30 40 SO 60 70 80 90
ATOMIC NUMBER (Z)
Figure 4. (b) Normalized negative secondary ion yields under Cs bombardment
(5).
-------
existence of instrumental variations in the relative sensitivity. The rela-
tive sensitivity, S , , is defined as:
SX/M -
-------
Empirical Methods
Absolute Sensitivity Factors—
An absolute sensitivity factor, S , can be defined as:
A.
SY = iy/CY (2)
A A A
where i is the intensity measured for an element X in a standard of known
composition C. Analysis of an unknown with absolute sensitivity factors
requires a set of standards containing all elements of interest and a defined
set of operating conditions, since the measured intensity is a function of
the primary ion species, the beam current and energy, the secondary ion
extraction efficiency, and the detector gain. The absolute sensitivity factor
method contains no inherent correction for matrix effects. This method can
only be applied to SIMS analysis when standards are available which are close
in composition to the unknown. By measuring the absolute sensitivity factors
on the same instrument used for the analysis of the unknowns and operating
under defined conditions, instrumental discrimination effects can be almost
completely eliminated.
Working Curves—
A "working curve" establishes the instrumental signal response as a
function of concentration in a given matrix from a series of standards.
Examples of SIMS working curves are shown in Figures 5 (a) and (b). A work-
ing curve is an elaboration of the absolute sensitivity factor method. A
suite of standards is employed which spans the composition range of interest
for the unknowns. Analysis of the unknown is carried out directly by compar-
ing the signal measured on the unknown with the working curve to yield the
predicted concentration. SIMS working curves are usually simple linear func-
tions for those samples in which the solute element is present at low con-
centrations (e.g., 10 atom percent or less) and situated in only one phase.
Non-linearities in the working curves are frequently observed at high concen-
trations or in cases in which an element is partitioned in two or more phases.
Analytical accuracies of five percent relative can be achieved with the work-
ing curve approach, even in the case of non-linear response, provided suffi-
cient standards are available to define the working curve. Working curves
inherently incorporate instrumental discrimination effects, and hence must
be iocally determined. No matrix corrections are contained in the working
curve function, so the method does not have flexibility in the analysis of
unknowns.
Relative Sensitivity Factors—
The relative sensitivity factor, S , , defined by equation (1), can also
be employed for quantitative analysis. By algebraic manipulation of equation
(1), the measured ratio of secondary ion intensities for two elements X and M
can be related to the ratio of concentrations, providing the appropriate
relative sensitivity factor is available from measurements on known standards:
• < VV
11
-------
CO
o
Q_
NIOBIUM IN STEEL (93)
0663
0,0
0,1
PERCENT NIOBIUM
0,2
0,3
Figure 5. (a) Working curve for niobium in an iron matrix (NBS SRM 660 steels)
12
-------
CHROMIUM IN STEEL (52)
ce
<
cc
\-
t—i
m
en
o:
o
o
Q_
661
662
0-5 1-0
PERCENT CHROMIUM (ATOM)
Figure 5. (b) Working curve for chromium in an iron matrix.
-------
where I = i/f Is the total secondary ion signal for an element considering all
isotopes. By measuring (I /I ) for all elements in a sample,
iCX + CM = 1
Equation (5) is an equality in one unknown, and therefore the value of CM can
be determined. This value of C can then be used in equation (3) to yield
values of Cv for all elements in the sample, resulting in a complete analysis
X
of the sample.
The relative sensitivity factor method has a number of advantages for
SIMS analysis. Because one element is measured relative to another, a degree
of compensation for matrix effects is often obtained, since matrix factors,
such as the amount of oxygen, will often similarly affect all elements pre-
sent. An example of this compensation can be observed in highly oxidized
systems such as glasses. In Figure 6, relative sensitivity factors for
elements with 20 <_ Z <^ 32 compared to silicon are shown as measured in a
glass matrix and as calculated from the pure element secondary ion yield
values of Andersen (11) . The range of relative sensitivity for these elements
is substantially reduced in the glass matrix compared to what might be
expected from the pure element behavior. Moreover, a systematic trend in the
relative sensitivity factor with atomic number is observed in this period.
Similar results have been reported for an alumina matrix by McHugh (2) . In a
matrix rich in oxygen, such as a glass which contains 60 atom percent oxygen,
and which is further enriched by implantation of the oxygen primary ion beam,
the elements behave in a similar fashion with respect to secondary ion
production.
The relative sensitivity factor method has another useful feature when
the analyst is confronted with an unknown which differs from the standards
previously used to determine the suite of factors. If the analyst has deter-
mined a suite of factors for a series of elements X. relative to a matrix
element M, a set of sensitivity factors for a new matrix N can be calculated
from the equation:
SX/N = SX/M * SM/N (6)
An example of the application of this equation for generating new sensitivity
factors is given in TABLE 1, where calculated and measured sensitivity factors
are compared for a germanium — oxygen system derived from a silicon — oxygen
system. In this case, the errors introduced by calculating the sensitivity
factors are generally less than a factor of two. This procedure works most
satisfactorily for highly oxygenated systems such as glasses, ceramics,
minerals, etc., in which a reactive species provides a common matrix element.
14
-------
10
1
SX/Si
0-1
001
• GLASS MATRIX
o PURE ELEMENTS
(ANDERSEN)
I I I I I I I I I
20 22 24 26 28 30 32
figure 6. Relative sensitivity factors for selected elements compared to
silicon as measured in a glass matrix and as calculated from
pure element data (11) .
15
-------
TABLE 1. COMPARISON OF MEASURED AND CALCULATED RELATIVE SENSITIVITY FACTORS
Element Measured S , Calculated Svlr, Error
^^ _ X/Ge
B 0-77 0.78 +1.3%
Al 9.98 7.51 -25%
Si 1.66 2.56 +54%
Ti 9.71 7.12 -27%
Fe 4.52 3.02 -33%
Zr 6.97 5.00 -28%
Ce 5.75 7.00 +30%
Ta 0.49 0.60 +22%
Pb 1.69 1.18 -30%
Relative sensitivity factors inherently contain local instrumental bias
and therefore should not be transferred from one instrument to another. As
such, however, analysis with relative sensitivity factors tends to eliminate
errors resulting from instrumental effects, providing care is taken to
establish the same standardized operating conditions for the analysis of the
unknowns as is used for the determination of the factors.
Theoretical Models for SIMS Analysis —
A number of models for the production of secondary ions have been pro-
posed, including thermodynamic models (12,13), quantum mechanical models (14),
valence models (15), and electrostatic models (16,17). The validity of these
models as a description of the secondary ion emission process is a matter of
considerable debate. Most of the models have only been developed to the
point of considering secondary ion emission during noble ion sputtering of
elements or simple compound targets in the absence of oxygen. The local
thermal equilibrium (LTE) model has been one of the most successful models
in describing secondary ion spectra of multi-element systems under reactive
ion bombardment (12). In brief, the LTE model describes the secondary ion
emission as originating in a dense plasma at the surface of the target. The
plasma is assumed to be in thermal equilibrium, allowing application of the
Saha-Eggert equation to relate the sputtered ion/neutral ratio in the plasma
to several parameters.
log (n+/n°) = 15.4 + l.Slog T + log(2B+/B°) -
^ (I - AE) -
-------
where n and n are the numbers per unit volume of ions and neutrals for a
given element, B is the partition function, I is the ionization potential,
AE is_the depression of the ionization potential, T is the plasma temperature,
and n is the density of free electrons. With the exception of T and n , the
other parameters in equation (7) can be obtained from reference data. In
applying the LTE model to an unknown sample, the temperature and electron
density are, in general, not known in advance. In some cases, these param-
eters can be obtained from previous experiments on similar samples of known
composition. Alternatively, in the "internal standard" LTE method, the known
compositions of two elements in the sample and the measured secondary ion"
intensities for those elements are used to solve equation (7) as a system of
simultaneous equations in the variables T and n . The values of T and n thus
determined are used to relate the measured secondary ion intensities n for
the remaining elements in the sample to the numbers of neutral atoms n. in the
sputtered assemblage, which yields the composition. In further elaborations
of the LTE model, negative secondary ions, molecular ions, and multiply-
charged ions can be considered (12).
It is important to note that the LTE model, as well as the other theo-
retical models, calculate the ion-to-neutral ratio as emitted from the sample.
The ion signal actually measured differs significantly from the emitted ion
signal, since the spectrometer transmission is only of the order of 0.1 - 10
percent. If the transmitted signals were proportional to the emitted signals
with the same constant of proportionality for the ions of all elements, the
loss due to transmission could easily be compensated. However, the trans-
mission is usually mass dependent, and the ion detection system can also give
a mass dependent response, and the ion detection system can also give a mass
dependent response, so that the overall system response is a complicated
function of mass. Analysis with a theoretical model which does not incor-
porate any corrections for the instrumental response can be in error by as
much as an order of magnitude. Often, these errors in analysis are taken to
be evidence of the failure of a model to describe secondary ion emission
when, in fact, the errors are associated with the measurement process.
COMPARISON OF RESULTS OF QUANTITATIVE ANALYSIS METHODS
To compare the accuracy which could be achieved with the relative sensi-
tivity factor method and with the LTE theoretical matrix correction method,
Newbury and Heinrich studied a system of multi-element glasses as a model
system (18). Glasses offer a number of advantages as model system for SIMS
analysis: (1) Glasses can be prepared with many combinations of elements in
a single matrix which is homogeneous at the micrometer scale of spatial
resolution. (2) The oxygen content of glasses is typically 60 atom percent,
which is further increased by implantation of the primary oxygen ion beam.
Secondary ion emission from such an oxygen-rich system is less susceptible
to artifacts introduced by gases adsorbed from the specimen vacuum environ-
ment. (3) A common element, such as silicon, can be included in the composi-
tions of the various glasses to provide a reference element for the relative
sensitivity factors.
17
-------
For assessing the accuracy of analysis, an error factor, F, was defined
as,
F = C(true)/C(calculated) (8)
where C is the atomic concentration.
Analysis of Glasses by Relative Sensitivity Factors
In the analysis of the glasses by the relative sensitivity factor method,
average values of the relative sensitivity factors were first determined from
a group of glasses, and these average values were then employed for the anal-
ysis of a large suite of glasses. The resulting histogram of error factors
for more than 100 elemental determinations is shown in Figure 7. The error
distribution is such that 83 percent of the analyses fall within an error
factor of two, and 99 percent fall within an error factor of five. No error
factors greater than seven were observed. Closer examination of the error
distribution near unity reveals that 53 percent of the analyses fall within
an error factor of 1.2.
This study was extended to simulate the analysis of unknowns by conduct-
ing a "blind sample round robin" (19). In this study, four laboratories which
had previously been supplied with a suite of glasses for the determination of
relative sensitivity factors were asked to analyze an additional group of
glasses as unknowns. In this case, each laboratory selected relative sensi-
tivity factors from the reference standards which most closely matched the
unknowns. The analytical performance for this situation is summarized in the
error histogram of Figure 8. The error distribution is such that 91 percent
of the analyses fall within an error factor of two, and 100 percent within a
factor of five.
LTE Analysis of Glasses
Analysis of the suite of glasses was also carried out with the LTE model
in the method devised by Andersen and Hinthorne (12). Two elements were
selected as the internal standards, and the true composition values for these
elements were used as input to the calculation. The sum of the elemental com-
position other than oxygen was constrained to a total of 40 atom percent.
Analyses were made with and without corrections for the molecular ions in the
spectrum. No corrections were included for instrumental discrimination. The
resulting error factor distribution, Figure 9, reveals that 53 percent of the
analyses fell within a factor of two, and 84 percent within a factor of five.
Error factors greater than five were observed in 16 percent of the analyses,
with error factors in excess of 10 in about 5 percent of the cases. Error
factors as high as 40 were observed for heavy elements such as lead and
tantalum.
Comparison
When the error factor distributions in Figures 7, 8, and 9 are compared,
the relative sensitivity factor method is clearly seen to produce superior
overall results. In particular, analysis with the relative sensitivity factor
18
-------
SIMS ANALYSIS
OF GLASSES WITH
AVERAGED RELATIVE
SENSITIVITY FACTORS
j45
40--i
(1/F)> 5:1 CASE
1/FT-r
30
2!
15
to
0-5 < F < 2 83%
0-2 < F < 5 99%
F > 5: 0 CASES
Figure 7.
543212345
OVERESTIMATE UNDERESTIMATE
Error factor histogram for quantitative analysis of glasses by the
relative sensitivity factor method with average values of sensi-
tivity factors (4).
19
-------
Blind Sample Round
Robin
-t-40
1 case
57%
^F5 2
3 cases
VF
Figure 8
Error factor histogram for quantitative analysis by the relative
sensitivity factor method. Blind round robin results on unknown
glasses (19). 20
-------
(1/F) > 5
6 DETERMINATIONS
1/F FT
I 4
••
••
:
••
i
••
»
£.
-
i
mm
4
m
r7
•
0
^
4
-flT
2345
NBS GLASSES
LTE ANALYSIS
O/O2, Si/SiO
AND NO OXIDE
F=C (TRUE)/C(LTE)
311 DETERMINATIONS
05 < F < 2 53%
0-2 5
43 DETERMINATIONS
=UF
OVERESTIMATE UNDERESTIMATE
Figure 9. Error factor histogram for quantitative analysis of glasses by the local thermal equilibrium
model method (4).
-------
method greatly reduces the number of large errors exceeding a factor of five.
While the amount of standards data required for the relative sensitivity
factor method exceeds the LTE method, it must be noted that application of
the LTE method to obtain even the accuracy of Figure 9 requires some method
of independent analysis to obtain the compositional values for the internal
standards. In many cases of practical interest, the relative sensitivity
factor method will give virtually the same flexibility in analysis as the LTE
method with its constraints.
FUTURE DEVELOPMENT OF ANALYTICAL METHODS FOR SIMS
While the relative sensitivity factor method seems at present to be the
most practical approach to the analysis of an unknown, the difficulties of
obtaining enough standards to analyze matrices of widely varying electronic
character are quite formidable. The theoretical methods have the analytical
flexibility to solve, in principle, any matrix. However, their accuracy in
an analysis is frequently quite suspect, a condition which arises in part due
to their susceptibility to instrument artifacts which affect measured ion
intensities. A promising approach to the problem of developing a flexible,
quantitative analysis method is that originally suggested by McHugh (20). In
this method, relative sensitivity factors measured in a matrix M are modified
to factors appropriate to a different matrix M' by the use of a series of
multiplicative matrix correction factors, S,:
o = c . y • v . y
SX/Mf SX/M Ll L2 L3 . . . (9)
These matrix correction factors could be determined either experimentally
from the actual spectra of the reference sample and the unknown through the
I i _i_ i I
measurement of a matrix sensitive signal, such as M /M or MO /M , or the
Z factors could be calculated theoretically with the LTE model or an alterna-
tive physical model. A combination of experimental and theoretical E factors
could be used. The advantage of such an approach is that the locally deter-
mined relative sensitivity factors, S / > measured for example from a suite
of glasses, automatically incorporate local bias in the instrumental response.
The matrix correction factors provide the capability to shift from one matrix
to another. The approach illustrated in equation (9) should offer the
required analytical flexibility to successfully deal with a wide variety
of specimens while eliminating the large instrumental effects on analysis.
22
-------
4. EXPERIMENTAL METHODS
SPECIMEN CHARGING
Most particles of interest derived from environmental sources are found
in the form of chemical compounds, such as oxides or sulfides, rather than as
pure elements. This is an important distinction, since compounds tend to be
electrically insulating, while many pure elements have a conducting or semi-
conducting nature. Under charged ion (or electron) beam bombardment, insulat-
ing particles tend to accumulate electrical charge. When a sufficient charge
is developed, the particles often repel each other, causing a change in posi-
tion of the target of interest. In extreme cases, the particles may be
ejected from the substrate and lost.
In SIMS, the analyst has a choice of either a positive primary ion beam
(usually 16C*2 , ll*N2 , 40Ar , or 133Cs ) or a negative primary ion beam
(160~, or N20~). The source brightness (in A/cm2) is typically 10 to 100
times brighter for a positive primary ion compared to a negative primary ion
(e.g., 16C>2 compared to 160 ). Thus, in terms of the beam current obtained
in the final probe, which is one of the determining factors in establishing
the limit of sensitivity, it is more desirable to employ a positive primary
ion beam. However, bombardment with positive primary ions tends to lead to
significant problems with specimen charging.
Under the impact of a primary ion, regardless of its charge, secondary
electrons are emitted from the target. The secondary electron emission
coefficient under ion bombardment is frequently greater than unity. When a
positive primary ion beam such as 16C>2 strikes the particle injecting posi-
tive charges, and secondary electrons are emitted, removing negative charges,
the particle tends to acquire a positive charge. This positive charge can
cause the positional instability mentioned above and can also affect the
trajectories of the sputtered secondary ions, causing unstable signals or
reducing the sensitivity. With negative primary ion bombardment, negative
charges are injected into the target, while the emission of secondary elec-
trons effectively removes negative charges. As a result of this flow of
negative charge both to and from the target, a charge balance can be obtained.
While the surface potential may not be at earth, a potential of the order of
a few volts is not exceeded, and moreover, the potential is stable. Strong,
stable secondary ion signals can be obtained. Thus, bulk insulating samples,
such as oxidic materials, can be studied, see for example Figure 10. Second-
ary ion spectra from bulk insulators could not be obtained under positive
23
-------
Figure 10. Spectrum of bulk NBS glass K 227 obtained with negative primary ions.
-------
to
Ul
Figure W continued. Spectrum of bulk NBS glass K 227 obtained with negative primary ions.
-------
primary ion bombardment. Negative primary ions were initially employed for
the study of individual particles.
PARTICLE DISPERSION AND MOUNTING
The ideal substrate for SIMS study of individual particles is one which
produces very few secondary ions. When positive secondary ion spectra are
considered, several elements stand out as candidate materials. Of these,
carbon and gold are most feasible. A second desirable property of the sub-
strate is high reflectivity to aid in location of the particle via the optical
microscope. While vitreous carbon can partially satisfy this requirement,
gold is ideal in all ways. In addition, by choosing a high atomic number
material with a single isotope, the possibility of introducing extraneous
peaks in the spectrum is reduced. Gold substrates were prepared either from
electroplating onto polished brass blocks or by pressing gold wire into thin
foils. Particles were dispersed onto these substrates by ultrasonification
in ethyl alcohol, followed by rapid drying on the substrates. For negative
primary ion bombardment, no further preparation was needed.
The substrate typically contributed only sodium and gold signals to the
spectrum of a particle. When measuring a spectrum on a particle, the gold
signal from the substrate was found to be higher than when the substrate was
measured directly. This was attributed to the resputtering of components of
the particle from the adjacent substrate surface. For example, the silicon
sputtered from the particle contaminated the surface of the gold, increasing
the ionization efficiency of the gold compared to the clean substrate. In
addition to the elemental gold secondary ion signal, a molecular signal for
213(AuO ) is observed which is frequently higher than the parent ig7Au
signal. Again, this is attributed to the presence of sputtered material on
the substrate near the particle under analysis.
USE OF POSITIVE PRIMARY ION BEAMS
In order to maximize sensitivity from small particles, experiments were
carried out to explore the possibility of employing high brightness positive
primary ion beams. Particles were dispersed onto gold substrates as described
above. Coatings of thermally-evaporated high purity carbon of various thick-
nesses were applied. With carbon coatings of the order of 20 - 40 nm thick,
stable secondary ion signals could be obtained from individual particles less
than 10 vim in diameter, Figure 11. Since positive ion beams with a diameter
of 1 ytn could be generated, the use of positive ions allows the study of
inhomogeneities in heterogeneous particles with sub-structure on the order of
the beam size.
DEPTH PROFILING
In order to study layered particles, the depth profiling mode of opera-
tion of the ion microprobe has been employed extensively. Since the sample is
continuously sputtered in order to generate the secondary ion signal, control
of the sputtering process can lead to a known rate of sample erosion. Thus,
the sample can be effectively "peeled" atom layer by atom layer. By monitoring
the secondary ion signals as a function of erosion time, elemental depth
26
-------
Figure 11. Spectrum of a single particle obtained with a positive primary ion.
-------
profiles can be obtained, providing the rate of erosion can be determined.
In the present work, the particles examined typically contained silica as a
major matrix constituent, combined with aluminum oxide, calcium oxide and
iron oxide at significant levels. In order to calibrate the erosion rate on
environmental particles, the erosion rate was measured on a number of NBS
Research Material glasses, which had similar compositions to the environ-
mental particles.
Ion erosion can be carefully controlled in the case of a flat, semi-
infinite sample to yield a depth resolution of the order of 5 nm. Unfor-
tunately, the irregular nature of particles deviates substantially from this
ideal case, which results in a degraded depth resolution. The extent of the
degradation could be measured if a suitable particle with a known layered
structure were available. At present, this factor remains to be fully
evaluated.
In preparing elemental depth profiles, another artifact must be taken
into account. The absolute value of the secondary ion signal is strongly
dependent on the amount of oxygen present in the matrix. The oxygen content
can vary during the initial stages of the ion erosion due to the implantation
of the primary oxygen ions until a dynamic equilibrium is established. During
this period, absolute secondary ion signals are not a good indication of com-
positional variations. Moreover, at interfaces between regions of different
composition, absolute signals can also change due to variations in the
sputtering rate, and hence, in the oxygen concentration. In order to minimize
these dynamic effects, the technique of measuring signals relative to a major
matrix peak has been employed. Silicon has been chosen as a suitable matrix
peak, because of the frequency with which it occurs in environmental particles.
By measuring the signals of interest relative to a matrix peak, the varia-
tions in the signal of interest due to a varying oxygen content are minimized
since the matrix reference element can be expected to vary in a similar
fashion.
QUANTITATIVE ANALYSES
Based on the results of Section 2, quantitative analyses were carried
out by the relative sensitivity factor method. Relative sensitivity factors
appropriate to the situation of an oxidized matrix were obtained from the NBS
RM 30 and 31 series research material glasses (21). These glasses contain
most elements of environmental interest in a silicate matrix. Sensitivity
factors were measured relative to silicon and the procedures described in
Section 2 were employed for quantitative analyses.
SENSITIVITY LIMITS
Sensitivity limits were estimated from the count rates per nanoamp of
primary ion beam current on selected RM 30 glasses. The detection limit was
estimated as the concentration which was expected to produce three times the
background count rate (background approximately 1 c/s in the absence of
interferences). The detection limit for a variety of elements is listed in
TABLE 2.
28
-------
TABLE 2. DETECTION LIMITS IN SILICON-LEAD-OXYGEN GLASS
(NBS K493 and K523 Glasses)
1 nA
detection limit
, (1 sec. integration)
Element c/s/nA/Atom Percent ppm
Li
B
Mg
Al
P
Ti
Cr
Fe
Ni
Ge
Zr
Ba
Ce
Eu
Ta
Th
U
7.22 x 104
4.69 x 103
4.66 x 104
6.32 x 101*
1.32 x 103
5.37 x 10U
3.01 x 104
1.77 x 104
1.13 x 101*
7.45 x 103
4.20 x Wk
7.41 x Wk
1.85 x 104
6.45 x I0k
2.51 x 103
1.38 x 104
1.42 x 10^
0.42
6.4
0.64
0.47
23
0.56
1
1.7
2.7
4
0.71
4.0
1.6
4.6
12
2.2
2.1
Q
All elements as dilute solutes in the indicated matrix.
Corrected for mass abundance.
c
Concentration which is expected to produce three times
tne background count rate (1 c/s).
29
-------
5. DETECTION OF MOLECULAR SIGNALS BY SIMS
The secondary ion spectrum frequently contains ion species which do not
correspond to elemental ions. Figure 12 shows a positive secondary ion spec-
trum of silicon bombarded with oxygen primary ions. The Si elemental peaks
are observed. In addition, a cascade of molecular peaks are observed, includ-
ing SiO , Si2 , Si20 , Si02 , Si202 , etc. In general, an oxidized structure
containing metallic ions M and N can produce a wide variety of molecular ion
± ± ± ±
signals, including MO , N 0 , M N , and M N Oj . These signals can be used
x y x y xx x y
to identify the presence of compounds within the original sample. However,
the act of sputtering the sample can cause the breaking of chemical bonds and
the creation of new molecular species within the ion-bombarded-region. To
properly interpret the inorganic molecular spectrum, it is necessary to study
model systems in order to identify the characteristic molecular ions. In a
complicated, multi-component environmental particulate, the multiplicity of
possible molecular fragment ions can make spectral interpretation difficult.
Only the molecular signals associated with the major constituents can be
studied usefully.
Considerable interest has arisen in the past several years in the detec-
tion of organic compounds by means of SIMS (22,23). Benninghoven and his
co-workers have successfully obtained SIMS spectra from a large number of
organic compounds with parent ions in the mass range 50 - 300 AMU.
Benninghoven determined that to optimize the detection of organic molecules,
which might be decomposed by the action of the bombarding ions, it would be
necessary to employ conditions of "static SIMS". In static SIMS, the damage
cross section for the incident ion is minimized by employing a low primary
ion beam energy, typically 1-5 keV, and also by using an extremely low
primary ion current density, of the order of 1 x 10 9 A/cm2. In order to
produce a detectable signal, a primary beam current of 1 nA must be used, but
the beam is defocused to approximately 1 cm in diameter. The total dose
during the recording of a secondary ion mass spectrum (e.g., for 1000 s) is
of the order of lO1^ ions/cm2 or less. The probability of a given lattice
location being struck more than once by an incident ion is therefore very
low, and the measured secondary ions are generated from undamaged material
throughout the period of the experiment. While organic mass spectrometry by
static SIMS is very promising, the study of particles by static SIMS is not
feasible. Because of the low current density, useful signals could not be
obtained from individual microscopic particles. Large aggregates of particles
could be studied by static SIMS, but the results would only be valid if the
sample was homogeneous on a particle by particle basis.
30
-------
tQQQi
10
20 30 /.O 50 60 70 80 90
Figure 12. Secondary ion spectrum of pure silicon showing
inorganic molecular signals.
31
-------
The primary ion bombardment conditions which are employed in the ion
microprobe deviate greatly from the condition of static SIMS and are consid-
ered to constitute "dynamic SIMS". In dynamic SIMS, the sample is effectively
undergoing constant change during the measurement of a spectrum. The com-
bination of a high bombardment energy (10 - 20 keV), which results_in a large
damage cross section, and a high current density (typically 1 x 10 * to
1 x 10 3 A/cm2), leads to the situation in which the dose is so high that all
surface atom sites are hit by primary ions in the first few seconds of bom-
bardment. Thus, if a molecule undergoes ion-induced damage, the opportunity
for measuring that species is quickly lost. On the basis of these arguments,
most investigators have concluded that the detection of organic molecules
with the ion microprobe is an impossibility.
The ion microprobe does, however, have an extremely high specific sensi-
tivity (i.e., sensitivity/unit bombarding ion) so that useful signals (100 -
1000 c/s) can be obtained from the major constituents of a sample even with
a greatly reduced primary beam current. Investigations were carried out to
determine if secondary ion spectra of organics could be obtained in a micro-
probe mode with a primary beam current density of 10 ^ to 10 6 A/cm2, which
lies between the conditions of static SIMS (10~9 A/cm2) and dynamic SIMS
(10 1 A/cm2). In these experiments, dilute water solutions of several organic
compounds, including glycine, atropine, and fluorodopamine, were applied to
metal film substrates, including nickel and gold, following the methods of
Benninghoven et al., (22,23). The primary ion beam consisted of 1602 ions at
18.5 keV (effectively 160 at 9.25 keV at interaction) or 8.5 keV (4.25 keV
at interaction) with beam sizes of 10 ym - 50 ym and beam currents of 0.1
nA - 1 nA, producing current densities in the range 10 6 - 10 A/cm2.
Complex mass spectra were obtained on a nickel substrate, Figures 13 and
14. Signals were obtained for molecular fragments as well as for nickel com-
plexed with organic molecular fragments.
Fluorodopamine dispersed on gold films yielded secondary ion spectra with
strong signals for the parent ion, Figure 15. In parallel with Benninghoven's
observations (22), a stronger signal was observed for the parent molecule plus
a hydrogen. A strong fragment peak corresponding to the parent minus an
NH or CH3 group was also observed. This experiment marks the first reported
observation of the sputtering of intact large molecules under conditions which
deviate substantially from the static SIMS case.
To evaluate the utility of this experiment for measuring organic layers
on particles, a glycine solution was applied to spherical nickel particles
with a diameter of approximately 20 urn. A portion of the mass spectrum
obtained from a single nickel particle is shown in Figure 16. Note the sim-
ilarity of this spectrum with the same portion of the spectrum obtained with
glycine dispersed on a bulk nickel substrate. This experiment indicates the
feasibility of obtaining useful organic mass spectra from layers on individual
small particles, at least down to the size of 20 ym.
The ion microprobe employed in this experiment actually has a significant
discrimination against molecular species due to the fact that its transmission
32
-------
TGlycine
r Nickel substrate
100
Figure 13. Secondary ion spectrum of glycine on a nickel sheet substrate showing numerous molecular
fragments.
-------
OJ
^ Positive secondary ions
10 20 30 £0 50 60 70 80 90 100 110 120 130
Figure 14. Secondary ion spectrum of atropine on a nickel sheet substrate.
-------
U1
5-fluorodoparmne
100 200
Figure 15. Secondary ion spectrum of fluorodopamine on gold-alloy substrate.
-------
Atropine
Nickel sheet substrate
160 + t 8-5 keV
Positive secondary ions
Figure 16. Portion of a secondary ion mass spectrum from glycine on a 20
nickel ball.
36
-------
is maximized for high energy secondary ions, i.e., those sputtered with a
kinetic energy above 25 eV. Molecular secondary ions have a probable energy
of the order of a few eV, with a narrow energy distribution. Thus, the
detection of organic molecular ions in Figures 13 - 16 by ion microprobe
suggests that if an instrument were used with enhanced sensitivity in the low
kinetic energy region, organic secondary ion mass spectra could be obtained
from particles on a routine basis. The ion microscope, a SIMS instrument
with adjustable ion optics has the capability to select an energy window to
maximize molecular ion collection. This instrument probably offers the best
approach to organic mass spectrometry of individual particles.
37
-------
6. APPLICATIONS OF SIMS TO ENVIRONMENTAL SAMPLES
To demonstrate the feasibility of SIMS for the analysis of environmental
samples, several problems have been examined: (1) a study of the capability
of SIMS to obtain surface information on particles; (2) a study of elemental
depth distributions in particles in SRM 1648 (Urban Particulate Matter); and
(3) a study of particles obtained from river sediments.
SURFACE INFORMATION
An example of the capability of obtaining surface information on particles
was obtained. In the process of preparing microparticles of known composition
to serve as NBS Research Materials in this project, experiments were carried
out to convert bulk glasses of known composition into particle form. Fila-
mentary material was created by rapidly pulling threads of molten glass at
speeds up to 10 m/s, producing fibers whose diameter ranged as small as 3 urn.
Mass spectra were obtained from the fibers under beam conditions chosen to
restrict the analysis to a layer within 50 nm of the surface. The spectrum
thus obtained (Figure 17) is quite similar to the spectrum of the bulk material
(Figure 18), indicating that no surface contamination occurred during the
fabrication. Particles were also prepared in spherical form by remelting
ground glass particles in an air stream passing through a muffle furnace. In
an early experiment, the air stream velocity was too low to prevent the
particles from contacting the walls of the furnace. As a result, the surface
layer of the .particles showed considerable contamination from the boron and
sodium components of the ceramic furnace tube, as revealed in a SIMS spectrum
(Figure 19). When these particles underwent ion erosion in order to study
the interior, spectra similar to the bulk material were obtained. The obser-
vation of a thin surface contaminant layer on these artificial particles
indicates that useful surface information can be obtained from individual
particles.
As an example of a near surface analysis of a single urban particulate,
the measured peak intensities in a secondary ion spectrum were converted to
compositional values by the sensitivity factor method. The results for this
particle analysis are listed in TABLE 3. This particle consisted principally
of calcium, sodium, aluminum, silicon, and iron. Minor elements include
boron, magnesium, potassium, titanium, chromium, nickel and lead. Lithium
and barium were observed at trace levels.
38
-------
0000
u>
—-K309 Fiber, 10pm -j
Figure 17. Secondary ion mass spectrum of glass K 309 in fiber form.
-------
Figure 18. Secondary ion mass spectrum of glass K 309 in bulk form.
-------
1000
Heat Treated Particle
Furnace Wall Sample
Figure 19. Secondary ion spectrum of K 309 spherical glass particle collected from wall of remelt
furnace.
-------
TABLE 3. ANALYSIS OF THE SURFACE LAYER OF AN URBAN ENVIRONMENTAL
PARTICULATE WITH RELATIVE SENSITIVITY FACTORS DERIVED
FROM NBS GLASSES
Element Concentration (atomic percent)
LI 0.0067
B 0.62
Na 3.3
Mg 1.3
Al 4.0
Si 6.1
P 2.1
K 1.7
Ca 12.8
Ti 0.25
Cr 0.21
Fe 6.8
Ni 0.42
Ba 0.012
Pb 0.28
The sum of all the metallic and methalloid
elements was assumed to be 40 atomic per-
cent; balance - oxygen.
DEPTH PROFILING OF LAYERED ATMOSPHERIC PARTICLES
Individual particles from NBS SRM 1648 (Urban Particulate Matter) were
examined by SIMS. Spectra obtained from the near-surface region show a range
of compositions (e.g., Figures 20a and 20b) in which silicon, aluminum,
calcium, sodium, chlorine, and iron are found at high levels which very from
particle to particle. This result is in accordance with our observations of
individual particles in this material by electron probe microanalysis. Minor
elements observed in individual particles include lithium, fluorine, boron,
barium, and lead.
42
-------
10000
A
1000
mass/charge '•':
Figure 20. Secondary ion mass spectra of an individual particle from SRM 1648
(Urban Air Particulate). (a) Positive secondary ions.
-------
H
-I—
C2H2
CH
Cl
Cl
m
Urban Dust Particle
16CT. 21-5 keV
Negative ions
111 in ; i ^11 linn u
I r i 11 iri -» i ci .il
ill) ,1 I llif
liJ till: 111 II III L;**
••.!! ' !
.! 111 i
Figure 20. Secondary ion mass spectra of an individual particle from SRM 1648 (Urban Air Participate)
(b) Negative secondary ions.
-------
Depth profiles were recorded on selected individual particles in the size
range 5-10 ym. Figure 21 shows a multielement depth profile obtained by
manual tuning of the secondary spectrometer to record the intensities of seven
elements as a function of depth. The intensities of aluminum, iron, barium,
lithium, lead, and boron were normalized with the signal for silicon to mini-
mize matrix effects. A complex behavior is observed in this specimen for a
number of elements. Initially, all elements except aluminum increase with
depth. After approximately 0.5 vim of the particle was removed, the barium and
lead levels dropped significantly, before recovering to a nearly stable level
at a depth of 2 ym. The range of the barium and lead levels over the depth
scale studied is more than two orders of magnitude, with the interior depleted
in these heavy elements relative to the surface. This result was observed
frequently in a collection of about 20 particles which were studied in this
manner (24). The result was recently confirmed on a state-of-the-art ion
microprobe which had a computer-controlled secondary ion spectrometer for
rapid peak switching for the recording of multi-element depth profiles. Two
particles studied Figures 22a and 22b showed decreasing levels of barium and
lead toward the interior while the silicon signal remained remarkably constant.
In the UAP sample as a whole, bulk chemical analysis reveals a lead
level of about 0.65 weight percent. The result that at least a fraction of
the particles contain enhanced surface levels of lead suggests that the
biological significance of the lead content of the sample may be substantially
greater than the bulk analysis suggests. Thus, when these urban air pollu-
tion particles are ingested, the outer layers of the particles will be the
first to interact with body tissues. It is specifically these outer layers
in which the heavy metals in the sample are concentrated. Thus, the amount
of heavy metals which can be absorbed into the tissue locally near a particle
may be much greater than the average analysis value for the whole sample
would indicate.
The practicality of using SIMS depth profiling to study particles is
chiefly limited by the degree of automation available on the instrument, par-
ticularly with regard to multiplexing secondary ion signals. In the NBS ARL
ion microprobe mass analyzer, peaks of interest must be tuned manually, making
the recording of multi-element depth profiles tedious. Moreover, the time
factor causes a loss of depth resolution. In the new CAMECA IMS-3F ion micro-
scope, the state of automation is such that depth profiles such as Figures 22a
and 22b are rapidly obtained in a straightforward manner.
STUDIES OF STREAM SEDIMENTS
Individual particles from a series of stream sediments were examined from
three locations: New York Harbor, Indiana Harbor Canal, and the Ohio River.
Surface spectra from representative individual particles in these samples are
shown in Figures 23a, 23b, and 23c. Again, complex particle compositions are
observed, with high levels of silicon, carbon, aluminum, potassium, calcium,
and iron. Minor levels of lithium, boron, sodium, titanium, manganese, stron-
tium, barium, and lead have also been observed.
Elemental depth profiles for iron, carbon, aluminum, and silicon in
separate particles from New York Harbor are shown in Figures 24a, 24b, 24c,
45
-------
Urban Dust Particle
Pb/28Si
10
Figure 21.
0 1 2
Depth, pm
Multi-element depth profile obtained by manual scanning. Sample:
individual particle of SRM 1648 (Urban Air Particulate) with a
diameter of approximately 10 micrometers.
46
-------
Counl*
10E
ims 3f
10£ I
Depth (pm)
1-0
Figure 22. Multi-element depth profiles of individual particles from SRM 1648
obtained with automated ion microscope (a) Particle size 5 micro-
meters.
47
-------
10E
18E
IK.
10E
C/l
CD
"c
10E
1BE
i me
3f
2/1/80
108
Depth (pm)
1-0
Figure 22. Multi-element depth profiles of individual particles from SRM 1648
obtained with automated ion microscope (b) Particle size 4 micro-
meters.
48
-------
10000
Harbor jSediment
Surface Spectrum
p. e 23 Near-surface secondary ion spectra obtained from single 10
Figur • ^^ ^ ^^ York Harbor>
-------
10000
Ul
o
sif Ohio River Sediment
i Surface Spectrum
Figure 23. Near-surface secondary ion spectra obtained from single 10 micrometer sized sediment parti-
cles (b) Ohio River.
-------
10000
1000
c
100
Indiana Harbor Cana Sediment
Figure 23. Near-surface secondary ion spectra obtained from single 10 micrometer sized sediment parti-
cles (c) Indiana Harbor Canal.
-------
xio
1C
8-
3-
2-
1-
0-
New York Harbor Sediment
Iron Profile
Depth, pm
Figure 24. Single element depth profiles obtained from different individual
particles from New York Harbor (a) Iron.
52
-------
x10
151
1V
c/s 10H
5-
0
New York Harbor Sediment
Carbon Profile
0
1
Depth, jjm
Figure 24. Single element depth profiles obtained from different individual
particles from New York Harbor (b) Carbon.
53
-------
x10*
11
10
1*
8
7
5
3-
1-
New York Harbor Sediment
Aluminum Profile
1
Depth, pm
-Figure 24. Single element depth profiles obtained from different individual
particles from New York Harbor (c) Aluminum.
54
-------
and 24d. In this case, the secondary mass spectrometer vas tuned to a
specific peak and the signals were measured at a constant time interval. In
the resulting profiles, features where the composition varies with a depth
resolution of 0.1 ym or less can be observed, demonstrating the depth resolu-
tion which can be obtained by the SIMS technique. In a modern ion microscope
with rapid peak switching, this depth resolution could be achieved even when
multi-element depth profiles are measured.
55
-------
X105-
20H
15-
c/s 10-
0
New York Harbor Sediment
Silicon Profile
0
Depth,
Figure 24. Single element depth profiles obtained from different individual
particles from New York Harbor (d) Silicon.
56
-------
REFERENCES
1. K. F. J., Heinrich, and D. E. Newbury, eds. Secondary Ion Mass Spectrom-
etry, National Bureau of Standards Special Publication 427, Washington,
1975.
2. McHugh, J. A. Secondary Ion Mass Spectrometry. In: Methods of Surface
Analysis, A. W. Czanderna, ed., Elsevier, Amsterdam, 1975. pp. 223-278.
3. Andersen C. A., and Hinthorne, J. R. Ion Microprobe Mass Analyzer.
Science, 175:853-860, 1972.
4. Newbury, D. E. Quantitative Analysis by Secondary Ion Mass Spectrometry.
In: Quantitative Surface Analysis of Materials, N. S. Mclntyre, ed.,
American Society for Testing and Materials STP 643, ASTM, Philadelphia,
1978. pp. 127-149.
5. Storms, H. A., Brown, K. F., and Stein, J. D. Evaluation of a Cesium
Positive Ion Source for Secondary Ion Mass Spectrometry. Analyt. Chem.,
49:2023-2030, 1977.
6. Andersen, C. A. A Critical Discussion of the Local Thermal Equilibrium
Model for the Quantitative Correction of Sputtered Ion Intensities. In:
NBS SP 427, ibid. pp. 79-119.
7. Newbury, D. E. Report on the United States - Japan Cooperative Analysis
by Secondary Ion Mass Spectrometry. In: Secondary Ion Mass Spectrometry,
Fundamentals and Applications, M. Someno and D. Wittry, eds., Japan
Society for the Promotion of Science, Osaka, 1978. pp. 52-74.
8. Newbury, D. E. The Influence of Instrumental Sensitivity Variations on
Quantitative Analysis by Secondary Ion Mass Spectrometry. In: Micro-
beam Analysis - 1979, D. E. Newbury, ed., San Francisco Press, San
Francisco, 1979. pp. 335-337.
9. Newbury, D. E. Instrumental Effects in Secondary Ion Mass Spectrometry.
Proc. 2nd Int'l SIMS Conf., Stanford, 1979. in press.
10. Schroeer, J. M. An Outline of Secondary Ion Emission Models. In: NBS
SP 427, ibid. pp. 121-127.
11. Andersen, C. A. Analytical Methods and Applications of the Ion Micro-
probe Mass Analyzer. In: Microprobe Analysis, C. A. Andersen, ed., Wiley,
New York, 1973. pp. 531-553.
57
-------
12. Andersen, C. A., and Hinthorne, J. R. Thermodynamic Approach to the
Quantitative Interpretation of Sputtered Ion Mass Spectra. Analyt. Chem.,
45:1421-1438, 1973.
13. Jurela, Z. The Application of Nonequilibrium lonization to the Emission
of Secondary Ions. Int'l J. Mass Spec, and Ion Phys., 12:33-51, 1973.
14. Schroeer, J. M., Rhodin, T. N., Bradley, R. C. A Quantum-Mechanical Model
for the lonization and Excitation of Atoms During Sputtering. Surf. Sci.,
34:571-580, 1973.
15. Plog, C., Wiedmann, L., and Benninghoven, A. Empirical Formula for the
Calculation of Secondary Ion Yields from Oxidized Metal Surfaces and
Metal Oxides. Surf. Sci., 67:565-580, 1977.
16. Joyes, P. Evaluation Theorique de la Pulverisation Cathodique Isotrope.
J. de Physique, 29:774-790, 1968.
17. Williams, P. Mechanism of Oxygen Enhancement of Sputtered Ion Yields.
In: Proc. 13th Conf. Microbeam Analysis Society, 1978. pp. 1A-1I.
18. Newbury, D. E., and Heinrich, K. F. J. Quantitative Procedures in Ion
Probe Microanalysis. Mikrochimica Acta. in press.
19. Newbury, D. E. On the Accuracy of Quantitative Analysis in Secondary Ion
Mass Spectrometry - Round Robin Results. In: Proc. 13th Conf., Micro-
beam Analysis Society, 1978. pp. 6A-6I.
20. McHugh, J. A. Empirical Quantitation Procedures in Secondary Ion Mass
Spectrometry. In: NBS SP 427, ibid. pp. 129-134.
21. National Bureau of Standards Research Materials 30 and 31—Glasses for
Microanalysis.
22. Benninghoven, A., and Sichtermann, W. Secondary Ion Mass Spectrometry:
A New Analytical Technique for Biologically Important Compounds. Organic
Mass Spectr., 12:595-597, 1977.
23. Benninghoven, A., Jaspers, D., and Sichtermann, W. Secondary-Ion Emission
of Amino Acids. Appl. Phys., 11:1-5, 1976.
24. Newbury, D. E. Secondary Ion Mass Spectrometry for Particulate Analysis.
In: Environmental Pollutants—Detection and Measurement, T. Y. Toribara,
J. R. Coleman, B. E. Dahneke and I. Feldman, eds., New York, Plenum, 1978.
pp. 317-348.
58
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
600/7-80-122
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Microanalysis of Individual Layered Particles by
Secondary Ion Mass Spectrometry
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Dale E. Newbury
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Gas and Participate Science Division
National Bureau of Standards
Washington, DC 20234
10. PROGRAM ELEMENT NO.
625 BE
11. CONTRACT/GRANT NO.
EPA-IAG-D8-E684
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Office of Research and Development
Office of Environmental Engineering and Technology
Jjashington, DC 20460
13, TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/ORD/17
15. SUPPLEMENTARY NOTES
This project is part of the EPA-planned and coordinated Federal Interagency Energy/
Environment Research and Development Program
16. ABSTRACT
Secondary ion mass spectrometry is evaluated for application to the determination of
the composition and structure of individual particles. Analyses of many elemental
constituents at the ppm level can be obtained in individual particles as small as
2ym in diameter. Molecular signals, both organic and inorganic, can be detected from
particles above ZOyim in diameter. Quantitative analyses of elemental constituents
can be made with a relative accuracy of 25 percent by means of empirical relative
sensitivity factors and within a factor of two by means of a physical model. Multi-
element depth profiles can be obtained from individual particles as small as 4um
in diameter. Depth profiles of individual particles from SRM 1648 Urban Air
Particulate reveal pronounced surface concentrations of lead and barium. Imple-
mentation of SIMS depth profiling requires automation to make use of the full spectra
information and to eliminate matrix effects by normalization.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Depth Profiling
Ion Microprobe
Layered Particles
Microanalysis
Particulate Analysis
Secondary Mass Spectrometry
Air Pollution Control
7B
7C
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS {This Report/
Unclassified
21. NO. OF PAGES
L7
20. SECURITY CLASS /Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (R«v. 4-771 PREVIOUS EDITION is OBSOLETE
------- |