NASA
N
M
v?,
“;5
n
ZL
c
TECHNICAL
NOTE
QUANTITATIVE DIRECT-CURRENT ARC
ANALYSIS
OF
RANDOM COMPOSITIONS
OF
MICROGRAM RESIDUES IN SILVER
CHLORIDE COMMON MATRIX
by William
A.
Gordon
and
~ilbert
B.
Chapman
Lewis
Research Center
Cleveland,
Ohio
NATIONAL AERONA~TICS AND SPACE ADMINISTRATION
0
WASHINGTON, D.
C.
0
NOVEMBER
1969
1.
Report
No.
NASA TN
D-5532
2.
Government Accession No.
3.
Recipient's Catalog No.
4. Title and Subtitle
QUANTITATIVE DImCT-CURmNT ARC
ANALYSIS
OF
RANDOM COMPOSITIONS
OF
MICROGRAM
RESIDUES
IN
SILVER CHLORIDE COMMON MATRIX
5.
Supplementary Notes
5.
Report Date
November
1969
'
6.
Performing
Code
6.
Abstract
A procedure
is
described for spectrochemical analysis of random compositions of
20
ele-
ments in the concentration range from
0.1
to
100
percent. Samples
are
microgram
resi-
dues derived from dissolution of metal samples. Calibration curves, one for each
ele-
ment,
are
prepared over the range from
0.01
to
10
micrograms.
Absolute amounts of
analyte elements and their percentage concentration
are
determined from these curves
for
all
compositions. Analyses of some alloy standards with and without use of compari-
son standards
are
reported
as
examples of results attainable. No matrix
effects
attribu-
table
to excitation suppressions
or
enhancements were found.
7.
Authods)
William A. Gordon and
Gilbert
B.
Chapman
9.
Performing Organization Name and Address
Lewis Research Center
National Aeronautics and Space Administration
Cleveland, Ohio
44135
2.
Sponsoring Agency Name and Address
National Aeronautics and Space Administration
Washington,
D.
C.
20546
17.
Key Words
(Suggested
by
Author(s))
Chemical analysis
Emission spectroscopy
Spectrochemical analysis
8.
Performing Organization Report No.
E-527%
10.
Work Unit No.
129-03
11. Contract
or
Grant No.
13.
Type of Report and Period Covered
Technical Note
14.
Sponsoring
Agency Code
18.
Distribution Statement
Unclassified
-
unlimited
19.
Security Classif. (of this
report)
20.
Security Classif.
(of
this page)
Unclassified Unclassified
21.
No.
of
Pages
22.
Price*
23 $3.00
~ ~ ~
~~ ~
*For
sale
by the Clearinghouse for Federal Scientific and Technical Information
Springfield, Virginia
22151
QUANTITATIVE DIRECT-CURRENT
ARC
ANALYSIS
OF
RANDOM COMPOS
,
CROGRAM
RES
LVER CHLORIDE COMMON MATRIX
by William
A.
Gordon and Gilbert
B.
Chapman
Lewis Research Center
SUMMARY
A
spectrochemical procedure
is
described for quantitatively determining random
combinations of
20
programmed elements.
The programmed elements
are
Al, Co,
Cr,
Fe,
Hf,
Mn,
Mo,
Nb,
Ni,
Pd, Re, Si, Ta, Ti, Th,
U,
V,
W,
Y,
and
Zr.
The response
of
a
direct-reading spectrometer
is
calibrated for absolute amounts of each of these
ele-
ments alone over the range from
0.01
to
10
micrograms. Absolute amounts of elements
in the sample
are
determined from these calibration curves, with correction made for
background and spectral line interferences. Percentage compositions
are
calculated
from the absolute amounts of analyte elements determined.
The analytical samples
are
dissolved and microliter amounts of the solutions are de-
posited onto carbon electrodes. The carbon electrodes are specially prepared to contain
4
milligrams of silver chloride (AgC1) in the carbon matrix. The silver chloride serves
as
a
buffer
material and also
as
a
common matrix. The electrodes
are
arced in agron,
which results in the vaporization and excitation of the metal constituents in the residue.
With
a
minimum AgCl-to-sample weight ratio of
400,
no matrix effects attributable to
excitation suppressions or enhancements
were
found.
Analyses of some alloy standards with and without use of
a
comparison standard
are
reported
as
examples of results attainable by the procedure. With automation of sample
arcings and data reduction, the analysis time
is
about
4
minutes per sample, including
preparation of electrodes through data recording but not including sample dissolution.
The need for appropriate comparison standards in emission spectrochemical analysis
is
the single most limiting characteristic of the method. The root cause of the problem
is
the complex effects of sample composition on atomic emission.
In
this report,
as
is
customary in the field, the term matrix effect
is
used to denote any mechanism which
either increases
or
decreases the atomic emission from
a
given amount of analyte
ele-
ment in the sample. The origins of matrix effects are discussed
in
references
1
and
2.
Because of these effects, the samples must accurately simulate the chemical and physi-
cal forms of the comparison standards. Much effort
is
required to prepare standards and
validate procedures, often for relatively minor changes in sample composition.
As
a
re-
sult, the emission method
is
inefficient and uneconomical in many analytical applications,
particularly those of
a
nonroutine nature.
The ideal solution to this problem
is
to calculate chemical compositions theoretically
from measurements of emitted light intensities. Some progress has been made in this
approach in recent years (ref.
3).
However, because of the complexities and instabilities
of the vaporization and excitation phenomena in the
arc
discharge, this approach has not
resulted in
a
practical method for spectrochemical analysis.
Another approach to the problem
is
to reduce the unknown samples to
a
common
chemical and physical form. Calibration standards of the same form
as
the samples can
then be used for
all
sample compositions. Many practical spectrochemical methods have
been described which use this principle. The most effective procedure
is
to dilute the
sample
in
a
relatively larger mass
of
a
pure material, and then excite the spectra with
a
direct-current arc. The dilution can be achieved by either mechanical blending of pow-
ders (ref.
4),
or
by fusing the sample with the diluent (ref.
5).
With proper selection of
the diluent,
a
common environment
is
provided for
all
sample compositions.
The effectiveness of this approach
is
dependent on
a
high diluent-to-sample ratio.
Ratios higher than
100
parts diluent to
1
part sample
are
desirable but, because of the
insufficient detectability of the
arc,
the dilution ratio
is
limited to the order of
10
parts
diluent to
1
part sample. Because of the low dilution ratios, these methods
are
best
ap-
plied to sample types having relatively small variations
in
composition, rather than to
random compositions of elements.
Furthermore, due to errors associated with arc exci-
tation, the methods have been semiquantitative, with errors ranging from about
30
per-
cent to
100
percent of the amount present.
The method described
in
this report
is
applicable for quantitative analysis of random
compositions of
20
programmed elements. Analytical results
are
derived from single-
element calibrations'for each of the programmed elements. Calibration curves of abso-
lute amounts of each element, ranging from
0.01
to
100.0
micrograms,
are
permanently
filed. Percentage compositions
are
calculated from the absolute amounts determined.
No
internal standards
are
used and
a
single analytical line for each element
is
used for
at
least three decades of concentration. Corrections for spectral line interferences can be
made with no foreknowledge of them. The method
is
applicable to samples
as
small
as
10
micrograms and the sample weight need not be known.
If
the sample weight
is
known,
a
mass balance with respect to metal constituents can
be
calculated.
The advantages of this procedure derive from refinements of the arc source (refs.
6
2
to
8).
In
reference
6,
conditions
are
reported
for
detecting nanogram amounts of
ele-
ments in the argon arc in the presence
of
milligram quantities of silver chloride (AgCl).
Because
of
the good detectability
of
analyte elements in the presence of larger quantities
of
AgC1, dilution ratios of
1
to 400
are
feasible with AgCl serving
as
a
common matrix.
The use of
a
tantalum (Ta)-tipped cathode for eliminating errors caused by arc wander
is
described in reference
7.
In
reference
8,
a
method
is
described for controlling either
arc
light intensities
or
arc current.
In
the present work, the
arc
current was controlled
according to
a
time-dependent program. Recently, the authors developed
a
carbon anode
which provides quantitative vaporization of microgram samples into the argon arc. The
results of this work are discussed in the appendix. All of these developments
were
com-
bined in the analytical procedure described herein.
The most important characteristic sought in the development of this procedure
was
freedom from matrix effects, particularly those associated with excitation
in
the
arc
column.
A
secondary, but
still
important, consideration
was
the economy and speed of
the determinations. The procedural steps of the method
were
established
so
that they
are
simple, rapid, and amenable to automation, insofar
as
possible.
After
sample dissolu-
tion, about
4
minutes per sample
are
required through recording of data. The excitation
sequence
of
11
samples in the arc chamber
is
completely automated,
as
is
the reduction
of data. However, detailed description of the automation
is
not within the scope of this
report. Reference to certain features of the automation
is
made only to
facilitate
under-
standing of the total analytical concept.
INSTRUMENTS AND PROCEDURES
The instruments used in this work consisted
of
a
specially developed
arc
chamber,
a
conventional direct-reading spectrometer, and
a
Lewis developed paper tape data
recording system. The primary characteristics of the spectrometer
are
summar-
ized in
table
I,
along with pertinent operational information. The
arc
chamber, illus-
trated in figure
1,
was
described in detail in references
7
and
8.
A
description of the
data recording system and computer programs
is
in preparation.
The operation cycles of the arc chamber, spectrometer, and recording system
are
programmed
so
that
11
samples are arced in sequence and the data recorded without
at-
tendance by an operator. The sample electrodes
are
specially prepared to contain
4
mil-
ligrams of AgCl intimately mixed with
a
residue containing
10
micrograms of sample.
These electrodes
are
loaded into the arc chamber and arced according to the sequence
shown in figure
2.
The time-variable current program
is
shown in figure
3.
After
an
initial arcing time of about
5
seconds
at
a
current of
11
amperes, the current
is
gradually
increased to the maximum current of about
36
amperes.
Also shown in figure
3
is
the
3