ALS Methods Super Trace Analysis

Loading...

Taking too long?

Reload document

| Open in new tab

ICP OPTICAL EMISSION SPECTROSCOPY

TECHNICAL NOTE 12

Introduction to Atomic Emission Spectrometry

Lisa Goldstone, Odile Hirsch*
Jobin Yvon, Inc., 3880 Park Avenue, Edison, NJ 08820
* Jobin Yvon SAS, 16-18, rue du Canal, 91165 Longjumeau, Cedex, France

Keywords: atomic emission, sample preparation, analysis

1 Principle of atomic emission

For a liquid sample, the atomization process can
be illustrated as follows.

1.1 General information

A given atom has a large number of possible
energy levels. An emission spectrum is produced
by an electronic transition from a high energy
level En to a lower energy level Em. The accept-
able transitions are given by the quantum
mechanics selection rules. A quantity of energy
Q is transferred to an atom by collision with
another particle, resulting in excitation of the
atom. An electron from an outer layer of the
atom is excited to a higher energy level.
Following this electron excitation, the electron
returns, in one or several stages, to its original
energy level.

The atomic emission technique measures the
energy lost by an atom passing from an excited
state to a lower energy state. The energy is
released in the form of light rays with a wave-
length l , or more specifically, in the form of a
photon with a frequency v carrying energy h x v.

Figure 1: Energy transition

The atomic emission spectrum is composed of
discrete spectral lines. The number of photons
emitted is proportional to the number of atoms of
the element present. To be excited, the sample
must be atomized, meaning dissociated into free
ions or atoms.

Figure 2: Transformation of sample

Depending on the species excited, the lines
have different types:
* emission from an atom: line I
* emission from an ion once ionized: line II
* emission from an ion twice ionized : line III

Lines I and II are frequently observed; lines III are
rarely observed in the plasma and lines of a high-
er degree are not observed. This is due to the
energies involved.

1.2 Plasma

The emission phenomena takes place in a plas-
ma. A plasma is an electrically neutral highly ion-
ize ionized gas. The gas used is typically argon.

1.2.1 Plasma creation

An argon flow travels through a crystal tube
inside a solenoid. The lines of force, generated by
the magnetic fields, are directed along the axis of
the solenoid inside the tube and take the form of
an ellipse on the outside.

ICP OPTICAL EMISSION SPECTROSCOPY

TECHNICAL NOTE 12

2.1 Introduction

As discussed previously, an atom subjected to a
plasma emits characteristic photons. This property
makes
it possible to perform a qualitative
analysis. The number of photons emitted is propor-
tional to the number of atoms of the considered ele-
ment. This is the basis of the quantitative analysis.

2.2 Qualitative analysis

A qualitative analysis consists of searching for the
elements in an unknown sample.JY. The lines
recorded reveal the presence of an element. An
approximate quantification can also be performed.
To select the lines, see the wavelength tables. Note
that experience will play a major role in interpreting
the spectra. This semi-quantitative analysis can be
performed very rapidly using the JY Win-IMAGE
option.

2.3 Quantitative analysis

The quantitative analysis links the energy emitted to
the number of atoms contained in the sample.
Atomic emission is not an absolute method. The
relation existing between the intensity emitted by a
line and the concentration of the associated element
must be calculated: this is called the line calibration
curve. The advantage of ICP-OES is that these cali-
bration curves are, in most cases, linear to several
orders of concentration. Great care must be taken in
calculating these curves as they will determine the
ultimate accuracy of the analysis.

2

An electrical discharge is created to arc the plasma
by partially ionising the gas in the torch.

2 Analysis

The electrons produced are subjected to the mag-
netic field induced and circulate along the axis of
the crystal tube describing annular circuits. Induced
or eddy currents are thus produced. The electron
path is stopped by collision, resulting in heating and
ionisation of the other gas atoms. The plasma is self
maintaining and continous.

Figure 3 : Magnetic field

1.2.2 Plasma gas

The gas generally used is argon, which like all rare
gases, is monatomic, chemically inert and has a high
ionisation energy (15.6 eV). A certain number of
advantages are provided by argon:
– Emission of a relatively simple spectrum producing
little spectral interference in emission spectrometry,
– Capacity to atomise, ionise and excite most of the
elements of the periodic table,
– Absence of formation of stable composites
between argon and elements,
– Lower cost than that of other rare gases as it is the
most widely available (1 % in air).

Its only limitation is its low thermal conductivity
compared to that of molecular gases such as nitro-
gen or oxygen. During heating, the argon ions trans-
fer energy to the atoms present in the sample solu-
tions.

ICP OPTICAL EMISSION SPECTROSCOPY

TECHNICAL NOTE 12

Repeatability: Represents the fluctuations of the
signal during a single reading under the same meas-
urement conditions.

Reproducibility: Represents the fluctuations of the
signal when one of the measurement parameters
has varied.

Robustness: Capacity of plasma to accept different
changes without significant variation in element
concentration. These changes can come from the
matrix, the analytical parameters or the environ-
ment.

Sensitivity: Slope of the calibration line; signal
intensity versus concentration.

Uncertainty: Estimation within which the true value
is located.

2.4.1 Repeatability, reproducibility and accuracy

These three terms are often understood to mean
the same thing. Repeatability and reproducibility are
expressed as a relative standard deviation. The only
difference is the fact that for reproducibility, one of
the parameters has varied. Accuracy is a principle
that can be difficult to measure. It is evaluated
using certified reference materials. It mainly
depends on the calibration curve and the prepara-
tion of the samples.
The repeatability and accuracy concepts can be
summarised using a target. The true value being the
center of the target.

Figure 4: Calibration line

2.4 Analytical performance

An emission spectrometer provides a large number
of analytical performance features: accuracy,
repeatability, reproducibility, selectivity, robustness,
sensitivity, detection limit, linearity, dynamic range
…

The main terms are defined below:
Accuracy: The accuracy of an instrument is its
capacity to give results that are free of systematic
error, meaning having a good degree of exactitude.
The accuracy is evaluated by the difference
between the measured mean value and the true
value of element concentration.

Blank: Matrix without analytes

Detection limit: Smallest concentration which can
be detected with certainty with respect to a blank

Error: Deviation between the mean value and the
true value

Fidelity: Capacity of an instrument to give good
repeatability.

Long term stability: Equivalent to repeatability over
a long period of time (several hours).

Figure 5: Repeatability and Accuracy

Precision: Corresponds to repeatability and repro-
ducibility.

Quantification limit: Concentration corresponding to
a given repeatability (for example 5%).

In figure 5 the second case is better than the third,
even though the result seems bad (high relative
standard deviation).

3

ICP OPTICAL EMISSION SPECTROSCOPY

TECHNICAL NOTE 12

2.4.2 Detection limits (LOD), Quantification limit
(LOQ)

expressed in multiples of standard deviation.

Figure 8 shows the influence of the coefficient k on
the percentage of error.

Here, we assume that the fluctuation of the spec-
tral background follows a Gaussian distribution
defined by its standard deviation.

Figure 6: Gaussian distribution of spectral back-
ground distribution fluctuations

The detection limit is derived from the signal that
will be statistically possible to extract from the
background. This comes down to stating the fol-
lowing question: does a measurement point belong
to the background or to the signal?

k corresponds to a safety coefficient. The higher
the k, the easier it will be to differentiate a signal
point from a background point.

Figure 8: Smallest detectable signal XL

IUPAC recommends applying k = 3. The percent-
age of error is then 13%.

Three equivalent formulas are used to express the
detection limit:
(CL., concentration limit) :
(1) CL = 3 C SB /S
(2) CL = 3 C RSDB/SBR
(3) CL = 3 BEC RSDB

Where,
S = net signal of solution
C = concentration of solution measured.
sB = standard deviation of blank
RSDB = standard deviation relative to blank
SBR = signal to background ratio
BEC = Background Equivalent Concentration

Formula (2) is mainly used for plasma optimisation.
Formula (3) is used to calculate the detection limits
for a given matrix following a calibration.

The Background Equivalent Concentration (BEC) is
represented on the figure 9. It corresponds to the
concentration for which the signal is equal to the
background (S = B or SBR = 1).

Figure 7: Detection of a point among the back-
ground noise

The Figure 7 represents the overlap of the fluctua-
tions of background B and signal S as a function of
the deviation between the two mean values,

4

ICP OPTICAL EMISSION SPECTROSCOPY

TECHNICAL NOTE 12

I

2.4.3 Robustness

Depending on the composition of the matrix, the
same concentration of an element will not give the
same signal.For example, 5 ppm of Mg in water or
in NaCl 10 g/L does not give the same net intensi-
ty.

Figure 9 : Background Equivalent Concentration
(BEC)

BEC is the absolute value of the ordinate of the ori-
gin of the calibration curve when no background
correction is applied. If a background correction is
applied, use formula (2).

For the detection limit, RSD is expressed:

RSD =

/1(

k

2)

Therefore, for k = 3, RSD = 47%.

To define the quantification limit, you define the
maximum desired RSD, for example 5 or 10%.
RSD = 10%, for LOQ = 4.5 LOD
RSD = 5%, for LOQ = 10 LOD
RSD = 2% for LOQ = 30 LOD
Generally, the minimum LOQ used = 3 LOD.

The detection limit is an estimation.
The quantification limit is a measurement.

The detection limit can be calculated in water or in
a special matrix. It is very important to know the
detection and quantification limits for each of the
applications.

Figure 10: Influence of matrix

To minimize this difference, the most robust plasma
possible must be used. The robustness can be eval-
uated by the ratio between two lines (an ionic line
and an atomic line). For this purpose, we use the
ratio provided by two Magnesium lines:
Mg II 280 nm/ Mg I 285 nm.

The higher the ratio, the more robust is the plasma.
A plasma is robust when the ratio of magnesium is
greater than 6. In this case, the plasma is in ther-
modynamic equilibrium.

3 Sample Preparation

The samples are usually introduced into the plasma
in solution form but solids, finely divided, can also
be used. Care should be taken in obtaining a solu-
tion of a solid sample since there is a risk of loss of
element, or contamination. The final solution
obtained by the analyst depends on the nature of
the sample and the concentration of elements to be
determined.
Two main types of sample preparation are used for
ICP analysis:
– Acid Digestion.
– Dry attack.

5

ICP OPTICAL EMISSION SPECTROSCOPY

TECHNICAL NOTE 12

3.1 Acid Digestion

3.3 Remarks

Acids, whether singly or in a mixture, use their oxi-
dising or reducing properties. Care must be taken
with respect to possible loss of volatile elements.
For example, loss of As, Se, Sn in the form of chlo-
ride in the presence of HCl. Creation of precipitates
of Ca, Ba, Pb also represents a possible error in the
presence of H2SO4. Open and closed microwave
systems are increasingly used in laboratories.

3.2 Dry attack

Alkaline fusion is often used, as well as high-tem-
perature calcination (450 – 600 °C) with acid
recovery of ashes. Care must be taken with respect
to contaminants introduced by the reagents. Losses
by volatilisation and insolubilisation can be more
than negligible.

– Whatever the type of solution, the quantity of acid
or flux must be as low as possible to ensure mini-
mum perturbation of the plasma. Certain acids are
preferable to others as they perturb the plasma
less.

Order of preference (from more favorable to least
favourable): HNO3, HCl, HCLO4, H2SO4, H3PO4.

– Caution, HF requires a specific sample introduc-
tion system.

– Na has a high depressive effect on sensitivity.

– The presence of hydrofluoric acid can result in the
formation of Ca, Mg Mn… fluoride precipitates.
This occurs frequently during mineralization of geo-
logical products, even after evaporation of the
hydrofluoric acid. This can be avoided by adding a
few drops of boric acid.

In the USA:
Jobin Yvon Inc.
3880 Park Avenue
Edison, NJ 08820
Tel: 1-732-494-8660
Fax: 1-732-494-8796
E-mail:
emission@jyhoriba.com

1-877-JYHORIBA

In France:
Jobin Yvon S.A.S.
16-18, rue du Canal
91165 Longjumeau Cedex
Tel: (33) 1/64 54 13 00
Fax: (33) 1/69 09 90 88

Germany: (49) 89/46 23 17-0
Italy: (39) 2/57 60 56 90
U.K.: (44) 20/82 04 81 42

In Japan:
Horiba Ltd.
2 Miyanohigashi, Kisshoin
Minami-ku, Kyoto 601-8510
TEL: (81) 75 313 8121
FAX: (81) 75 321 5725
www.horiba.com

China: (86) 10/68 49 2216
Spain: (34) 91/724 16 57
Other Countries: Contact JY
S.A.S.

6