Multielement analysis in whole blood using a capacitively coupled microwave plasma atomic emission spectrometer


Material Information

Multielement analysis in whole blood using a capacitively coupled microwave plasma atomic emission spectrometer
Physical Description:
ix, 166 leaves : ill. ; 29 cm.
Besteman, Arthur David, 1971-
Publication Date:


Subjects / Keywords:
Trace elements -- Analysis   ( lcsh )
Blood -- Analysis   ( lcsh )
Atomic emission spectroscopy   ( lcsh )
Chemistry thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1997.
Includes bibliographical references (leaves 154-165).
Statement of Responsibility:
by Arthur David Besteman.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 028623269
oclc - 39476611
System ID:

This item is only available as the following downloads:

Full Text







I would like to dedicate this dissertation to my

parents, Arthur and Audrey Besteman. Their love and support

has meant so much to me. I am very blessed to have them as

my parents.


I would like to thank Dr. Jim Winefordner for allowing

me to be a member of his group and for teaching me so much

about analytical chemistry. Working for him has truly been

a pleasure. I would also like to thank Dr. Ben Smith for

the great deal of help that he gave me in my research.

The whole Winefordner group has contributed to this

research, whether it be help with the project or help in

making working here a better experience. I would especially

like to thank Bryan Castle for all his help with the

computers and Dr. Kobus Visser for his help in making

modifications to my project. I must also thank Jeanne

Karably for her help with everything.

Over the course of this research project I was

fortunate enough to have three undergraduate assistants,

Don-Yuan Liu, Nancy Lau, and Gail Bryan. They all

contributed a great deal to this project and I appreciate

all their hard work.

I am grateful to the Chemistry Department Machine Shop

for their work in constructing the electrodes. I am also


grateful to the University of Florida Infirmary for drawing

my blood without causing too much pain.

I must thank the Centers for Disease Control and

Prevention for the initial funding of this work. I also

thank Texaco Company and the University of Florida Division

of Sponsored Research for providing my support for the past

three years.

There are also several people I must thank for their

contributions before I entered graduate school. My high

school chemistry teacher, Mr. Roger Bratt, helped cultivate

my interest and appreciation for chemistry. During my

undergraduate education I worked for three summers with Drs.

Mark and Karen Muyskens. Through this experience I learned

much about doing research and how to work independently. I

am very grateful for all that they taught me.

Finally, I must thank my family and friends for all the

support they have given me while I have been so far from

home. I could not have done it without them.



ABSTRACT . .. viii


1 INTRODUCTION ................ 1

2 BACKGROUND .. ... . 4

Atomic Emission Spectrometry .
Excitation Sources .
Choice of Emission Lines .
Microwave Plasmas in Atomic Emission .
Capacitively Coupled Microwave Plasma
Microwave Induced Plasma .
Microwave Plasma Torch .
Comparison to the Inductively Coupled
Conclusion . .

. 4
. 7
. 8
. 10
. 12
. 13
. 15
Plasma 16
. 16


Introduction . .
Medical Significance . .
Trace Elements . .
Lead . .
Manganese . .
Lithium . .
Zinc . .
Magnesium . .
Major Elements . .
Sodium . .
Potassium . .
Methods of Analysis . .
Lead . .
Screening methods .
Clinical methods .
Primary and Trace Elements .
Atomic absorption spectrometry
Atomic emission spectrometry .
Spectrophotometry .


. 19
. 20
. 20
. 21
. 24
. 24
. 25
. 26
. 27
. 27
. 27
. 28
. 28
. 28
. 32
. 38
. 39
. 40
. 41

Inductively coupled plasma mass
spectrometry . 42
X-ray fluorescence . 44
Electrochemical techniques .. .44
Conclusion . . .. 47


Setup .. . . .48
Microwave Plasma Electronics .. .48
Waveguide . . .. .50
Torch . . ... .. .50
Plasma Gases . ... .52
Electrode . ... .53
Lens Setup . ... .54
Detector . . .. .56
Photodiode array . .. .57
Charge coupled device . 57
Computer Software . ... .60
Materials . . 61
Aqueous Standards . 61
Blood Standards . ... .61


Introduction . . 65
Methods of Sample Introduction into a CMP ... .65
Nebulization . .. .66
Thermal Vaporization . ... .68
Cup electrode . .. .68
Filament electrode . .. .72
Hydride Generation . .. 74
Development of Electrode for Blood Analysis 75
Cup Holder Electrode . ... .75
Platform Electrode . ... .79
Suspension Method . ... .87
Spiral Filament Electrode . ... .89
Conclusion . . ... .. .93


Introduction . .. 94
Optimization of Parameters . .. .94
Helium Flow Rate . 94
Drying and Ashing Conditions . 95
Cleaning . . ... 98
Sample Size . . .99
Sources of Noise . 100
Analysis . . ... 100
Aqueous Standards .. . 100
Bovine Blood Standards . 101
NIST Standards . .. 107


Human Blood Standard . 107
Blood and Aqueous Standards .. .109
Conclusion . .... 114


Introduction . . ... 115
Trace Elements . ... 117
Zinc . .... 117
Lithium . . ... 118
Magnesium .. . 125
Manganese . .. 130
Primary Elements . ... 130
Sodium . . ... 130
Potassium . ... 134
Comparison to Literature Values .. .138
Conclusion . . 140


Conclusion . . 141
CMP-AES as a Lead in Blood Screening Method 141
CMP-AES as a Multielement Clinical Technique 141
Future Work . . 142
Analysis of Other Health Related Elements 143
Using Aqueous Standards for Blood Analysis .145
Commercially Made Filaments .. .146
Simultaneous Multielement Analysis ... .148
Other Biological Fluids . .. .149
Plasma . ... 149
Serum . . 150
Urine and spinal fluid .. 150
Miniaturize System . .. 151
Acousto-optic tunable filter 151
Miniature detector . 152
Atomization Method for Other Techniques 153



Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Arthur David Besteman

December 1997

Chairperson: James D. Winefordner
Major Department: Chemistry

A capacitively coupled microwave plasma atomic emission

spectrometer (CMP-AES) has been evaluated as a clinical

method for the direct analysis of several of the primary and

trace elements in whole blood. A tungsten filament spiral

electrode was used with the CMP, and whole blood samples

were deposited on the electrode and subsequently dried,

ashed, and atomized. The emission was measured with a

spectrometer and either a photo diode array or a charge

coupled device detector. A sample size of only 2 4L was

required and the time for each sample run was under 4

minutes. This method has a wide dynamic range, allowing the

determination of both the primary elements in blood and the

elements present in trace quantities.


Much of the initial work focused on measuring the

levels of lead in blood. A detection limit of 30 ppb for

lead in whole blood was obtained and good accuracy was

obtained in the analysis of whole blood standards from the

National Institute of Standards and Technology.

The research then focused on applying the CMP-AES to

other elements in blood. The elements studied were

potassium, sodium, lithium, magnesium, manganese, and zinc.

Good linearity was obtained for these elements and the

concentration levels obtained for these elements were

consistent with literature values.

The primary advantages of this method are that no

sample pretreatment or dilution is required, it is easy to

run, has a low instrument cost, and is capable of doing

multielement analysis.


The research project involved developing a capacitively

coupled microwave plasma atomic emission spectrometer (CMP-

AES) as a clinical method for multi-element analysis in

whole blood. Microwave supported plasmas are an excellent

source for atomic emission spectrometry. They produce a

high degree of excitation of atomic and polyatomic species,

have a relatively low cost, and are simple to operate.

CMP's have shown the ability to accomplish direct elemental

analysis in complex matrices [1-3]. Direct analysis is

desirable because it eliminates the use of hazardous

chemicals and the dilution of the sample in trying to

minimize matrix effects [1]. Direct analysis also

eliminates any contamination of the sample or interference

with the plasma that could be introduced by the solvent.

A sample introduction method has been developed that

enables the CMP-AES system to directly determine the

concentration of several elements in whole blood without any

sample dilution or pretreatment. A tungsten electrode is

used which has a spiral loop at the top. The samples of

blood are placed on this loop and dried by inductively

heating the electrode using microwave power. A flow of

helium gas is introduced through a quartz torch that

supports the electrode, and a low power plasma is formed at

the top of the electrode to ash the blood sample. The power

of the plasma is then increased for atomization and

excitation of the sample. The resulting emission is

measured using a spectrometer and either a photodiode array

or a charge coupled device detector. The peak area of the

atomic emission line of the analyte is then compared to an

analytical curve of standards to determine the concentration

of the analyte in the blood.

Initially this research focused on developing a

screening method for lead in whole blood. The design of the

filament was optimized, as well as the conditions for

drying, ashing, and atomizing the sample. All optimizations

were done using lead as the analyte. The method worked well

with whole blood, and gave excellent linearity and good

precision. The accuracy was tested by analyzing blood lead

Standard Reference Materials (SRM's). Good agreement was

obtained with SRM's with concentrations greater than 100


The CMP-AES method was then used for the analysis of

several of the medically significant primary and trace

elements in blood. The following elements were chosen for

analysis: sodium, potassium, magnesium, manganese, lithium,

and zinc. For each element an atomic emission line was

chosen that was free from interference and that produced a

linear analytical curve over the concentration range of

interest. The operating conditions had to be modified to

some extent for several of the elements. Analysis was

performed on human blood standards for each element

sequentially by the method of standard additions. The CMP-

AES gave good linearity and precision for these elements.

The determined blood levels for most of the elements studied

were consistent with those found in literature.

In this dissertation, the development of the CMP-AES

for use as a method for elemental analysis in whole blood

will be discussed. A brief overview of atomic emission and

the microwave plasma as an analytical method will be given.

The clinical importance and the methods currently used to

analyze the elements studied will also be presented. The

main portion of the dissertation will discuss the

experimental setup used, the development of the sample

introduction system, and the results of the analysis of the

selected elements in whole blood. The last chapter of the

dissertation will discuss the conclusions made from the

research and the possibilities for future work.


Atomic Emission Spectrometry

Atomic emission spectrometry is a useful method for

elemental analysis. It is very specific, has a wide dynamic

range, and has the capability of measuring many elements

simultaneously. Typically, its disadvantages include poor

sensitivities and serious matrix effects [4].

Atomic emission is the process of an atom being brought

to an excited state. The relaxation of the atom from the

excited state results in the emission of radiation. The

outer shell (valence) electrons are the components of the

atom that are excited. The electrons can be excited to a

number of different levels. The photons emitted from the

electrons as they relax from the different energy levels

have characteristic frequencies (v) giving rise to many

wavelengths for each element. The energy levels of each

element are different which results in a distinct emission

spectrum for each element. The energy (E) associated with

each emitted photon is determined by the product of Planck's

constant (h = 6.63 x 103" Js) and the frequency,

E = hv = hc/X

where c = the speed of light (3.00 x 108 m/s in a vacuum).

Figure 2.1 gives a simple example of the energies associated

with various transitions. The dotted lines represent the

excitation of electrons to two different energy levels, and

the solid lines indicate the various modes of relaxation

with their corresponding energies.

Not all transitions have the same probability of

occurrence. In general, the strongest emission is observed

from transitions which terminate at the ground electronic

level. This is called a resonance transition. If the

condition of thermal equilibrium is maintained, then the

number density (atoms per cm3) of analyte atoms in a given

excited state (nj) can be related to the total number

density of analyte atoms (nt) by the Boltzmann


Sneg -E l/kT

The temperature (T) is the absolute temperature (K). E, is

the excitation energy (J) relative to the ground state, and

gi is the statistical weight of state i. Z represents the

electronic partition function:

Z(T) = Y 0 ge-Ei/kT
_j 0 i

Level 2

Level 1

Ground State

<- E21=hv21=hc&1

<- E2=hv2=hc/X2

<- E =hv1=hc/k

Figure 2-1. Energy diagram for excitation and emission [5].


The radiant power of emission (<,) between two states (from

state i to state j) is given by the product of the

population density of the excited atoms (n,), the transition

probability (A j, s ') that an excited atom will undergo the

transition from j to i, the energy of the emitted photon

(hvji), and the volume element observed (V in cm3):

,E = nhv A 1V

This value, as well as the number density of excited atoms,

is proportional to the analyte concentration in the sample.

This relationship is good only for low concentrations. By

measuring the intensity of emission from standards of

various concentrations of the analyte, the exact

relationship between analyte concentration and 0, can be


Excitation Sources

Many types of excitation sources are used for atomic

emission. Generally the excitation source is also the

atomization source. In atomic emission, the sample must

first be atomized. This is the process of forming free

atoms. When the free atoms are formed, they can then be

collisionally excited to produce the atomic emission lines.

For many years, flames were the most commonly used

atomic emission source because they are simple,

reproducible, and inexpensive. The flames used in atomic


emission are formed by the combustion of an oxidant gas and

a fuel gas. There are several disadvantages to using a

flame as an atomic emission source. The energy of the flame

is difficult to control, and it does not generate enough

energy to atomize all the elements or to populate high

excited states of some transitions.

The introduction of plasma sources for atomic emission

spectrometry has significantly improved the detection

limits, accuracy and precision for atomic emission

spectroscopy. A plasma is a partially ionized gas sustained

through an electrical discharge or through a microwave or

radiofrequency field [5-6]. Plasmas are advantageous

because they have a higher temperature and a less reactive

environment than flames. Plasmas also produce a higher

degree of excitation generating more atomic emission lines

for use in analysis. Inductively coupled plasmas (ICP) are

currently the most widely used. Microwave plasmas are also

effective atomic emission sources. Two types of microwave

plasmas are the microwave induced plasma (MIP) and the

capacitively coupled microwave plasma (CMP).

Choice of Emission Lines

The atomic emission lines are spectrally separated by

using an optical dispersion device, typically a grating or a

prism. The emission lines must be chosen carefully for

optimum signal-to-noise (S/N). The most intense line for

the element of interest is not always the best line for

analysis. The spectral line chosen must be free from

spectral interference. Interferences may come from the

emission of the inert gas used, impurities in the gas, other

concomitants in the sample, or the plasma support materials.

For analysis in complex matrices, concomitants in the sample

can be a significant problem. The resolution of the

spectrometer plays an important role in determining how well

the spectral line can be distinguished from nearby spectral


Another factor in choosing an emission line is self-

absorption. As discussed previously, the emission intensity

is proportional to the number density of the excited atoms.

At high concentrations, the number density of atoms in the

various energy levels can be very high. The atoms present

in the lower energy levels can absorb the energy emitted

from the relaxing excited states. At high number densities

in the lower energy levels, a significant fraction of the

emitted energy can be absorbed instead of being detected as

emission signal. This is a significant problem if a

resonance line is used because the majority of the atoms are

present in the ground state. When self absorption begins to

occur, the slope of a log-log plot of emission intensity vs


concentration will deviate from the desired value of one and

approach a limiting value of one half [5]. In conventional

flames, the linear concentration range is often no more than

two orders of magnitude because of self absorption [7].

When analyzing samples of such a high concentration that

self absorption occurs, there are two basic strategies. The

first is to dilute the sample to a concentration where self

absorption does not occur. The second is to choose a weaker

spectral line that gives a linear response over the

concentration range of interest.

Microwave Plasmas in Atomic Emission Spectrometrv

Microwaves are radio waves in the frequency range of

1.0 GHz and upward [8]. Microwaves have been very useful

for applications in radar and communications because of

their high frequency and short wavelength. The high

frequency of microwaves provides wide bandwidth capability.

The wavelength is long enough to penetrate materials, but

short enough to allow microwave energy to be concentrated in

a small area. This feature has been taken advantage of in

microwave ovens [9].

Two methods for transmitting microwave energy from one

point to another are the coaxial cable, and the waveguide.

The coaxial cable consists of two cylindrical conductors


separated by a continuous solid dielectric. The microwaves

travel through the dielectric [10]. Coaxial cables are

capable of a large bandwidth and are small in size, but have

the disadvantages of high attenuation and cannot handle high

powers. The waveguide can be either a circular or

rectangular hollow pipe. It is capable of handling high

powers with low loss, but is large in size and only has a

narrow bandwidth.

The first microwave discharge was observed in the

1940's by electrical engineers and physicists working on

radar equipment [11]. It was viewed as a nuisance instead

of a potential technological advancement. In 1951, Cobine

and Wilbur described some of the features of a microwave

plasma [12]. They described the plasma using helium, argon,

air, oxygen, and nitrogen as the support gases.

In 1958, Broida and Chapman used a microwave-induced plasma

(MIP) to analyze nitrogen isotopes [13]. Kessler and

Gebhardt used a capacitively coupled microwave plasma (CMP)

to analyze limestone in 1967 [14]. Mavrodineanu and Hughes

used a microwave plasma torch in 1967 to view the emission

spectra of several elements by introducing solutions into

the crater of a graphite discharge tip [15]. Fallgatter et

al., examined an argon microwave plasma as an excitation

source for atomic emission spectrometry in 1971 [16]. The

development of microwave plasmas has grown over the years

because of their high excitation efficiency for both

metallic and non-metallic elements, their low background

emission, and their low cost [17]. In recent years,

microwave plasmas have been applied to many different

analytical applications including analysis of solids,

biological fluids, and oil [1-3]. Several authors have

written extensive reviews of the use of microwave plasmas in

spectrochemical analysis [17-22].

Capacitively Coupled Microwave Plasma (CMP)

For a CMP, a magnetron (microwave power tube) generates

the microwaves which are conducted through a coaxial wave

guide. Within the waveguide, a standing wave is produced

which builds up microwave energy that is transferred to the

tip of a central single electrode. By oscillating in the

microwave field, the electrons gain enough kinetic energy to

collisionally ionize the support gas. This produces a

flame-like plasma at the tip of the electrode. The plasma

that is produced is capable of atomizing and exciting the

analyte in a sample. The signal is measured by focusing the

emission on the entrance slit of a spectrometer. The

multielement emission is usually measured with a photodiode

array (PDA) or a charge coupled device (CCD).


Several parameters must be optimized in order to obtain

satisfactory results with a CMP. These parameters include

the microwave power, the plasma gas flow rate, and the

position of the electrode with respect to the detector.

Optimum microwave powers differ depending on the type of

samples and the method of sample introduction. Helium is

used most often as the support gas with flow rates ranging

from 3 to 10 L/min [23].

Spencer et al. studied various parameters for a high flow

rate (>6 L/min) CMP [24]. The temperature measurements were

made with the following plasma conditions: 10 L/min helium,

150 cm3/min hydrogen and 700 W of applied power. The

following results were obtained for the analysis of aqueous

solutions: excitation temperature = 3430 K; and electron

number density = 4.4 x 1014 cm3. They determined that the

values of Tec and n. are not statistically different for the

introduction of aqueous and organic solutions into the


Microwave Induced Plasma (MIP)

Microwave induced plasmas (MIP) are created by using an

external resonant cavity or some other structure to couple

microwave energy to a stream of gas in a quartz tube. MIP's

are sustained at low powers (25 to 200 W) with argon or


helium as the support gas [18]. A microwave power supply is

attached to an antenna or circuit loop by a coaxial cable.

The energy goes through the antenna or loop and is

introduced into the resonant cavity generating a standing

wave. A quartz tube is placed in the cavity in such a way

that its axis is parallel to the line of electric field

oscillation. MIP's that use an electrothermal type of

atomizer have resulted in the best detection limits [18].

Microwave induced plasmas are more widely used than

capacitively coupled microwave plasmas because MIP's require

lower power and can be operated at atmospheric pressure. In

addition, CMP's involve the use of an electrode which can

cause spectroscopic contamination and memory effects if it

erodes [23]. However, CMP's do have several advantages over

MIP's. MIP's can only be operated at low powers while CMP's

are stable over a wide range of power levels (50-2000 W).

At higher powers, there are fewer matrix effects and more

intense signals. Also, a wide range of gases can be used to

sustain CMP's and CMP's are more tolerant to the

introduction of foreign materials than MIP's [25]. Sample

introduction problems have hindered the development of

commercial MIP instruments [5]. Memory effects are also a

problem in MIP atomic emission spectroscopy [18]. The

memory effects are probably a result of etching of the

quartz tube by the plasma providing a region where analyte


atoms can collect. An MIP is most useful as an excitation

source when it is combined with a separate sample atomizer.

Microwave Plasma Torch

Jin et al. developed a new type of microwave plasma

called the microwave plasma torch (MPT) [26-32]. A MPT

contains three concentric tubes, with the outer tube made of

brass and the inner tubes made of copper. The outer tube

serves as the microwave cavity which couples the microwave

energy to the torch forming a plasma at the top of the

torch. The carrier gas containing the sample aerosol enters

the inner tube and the plasma gas (helium or argon) flows

through the middle tube. This microwave plasma is very

stable and has a high tolerance to the introduction of

foreign materials [30]. The linear dynamic range for the

MPT was generally more than three orders of magnitude and

the detection limits for 15 rare earth elements were in the

part-per-billion (ppb) range [32]. This is a significant

advancement over the MIP because the MPT can withstand the

introduction of wet aerosols. Solutions are nebulized by an

ultrasonic nebulizer and the resulting aerosol is introduced

through a desolvation-dessicator system. The MPT does

however suffer from matrix effects and air entrainment in

the torch.

Comparison to the Inductively Coupled Plasma (ICP)

The inductively coupled plasma (ICP) is widely used in

industry and research. The ICP has a somewhat higher

temperature than the microwave plasma and produces a high

degree of excitation. The ICP consists of several

components. A gas, typically argon, flows through a torch

made out of three concentric quartz tubes. The top of the

torch is immersed in a high energy induction coil which

carries radiofrequency power (at 27 or 40 MHz) in the range

of one to three kilowatts. This causes the oscillation of

the argon atoms, and the high energy collisions that result

produce a plasma at the top of the torch with a temperature

more than 6000 K [6,33]. The sample is generally introduced

by nebulization [34]. A fine mist of sample is generated by

pumping the sample through a pneumatic nebulizer and spray



Atomic emission spectrometry is a very selective

analytical method that can be used for many types of

samples. Plasma sources have further increased the

usefulness of atomic emission spectrometry. Table 2-1 shows

a comparison of the plasma sources discussed. Although the

ICP and the MIP are currently in wider use than the CMP, the


CMP is better able to analyze complex matrices. This

feature of the CMP can be used for direct elemental analysis

in blood.

Table 2-1. Comparison of inductively coupled plasma (ICP),
capacitively coupled microwave plasma (CMP),
microwave induced plasma (ICP) and microwave
plasma torch (MPT) for atomic emission
spectrometry [21].


Argon Argon
Gas Argon Helium or or
Helium Helium

Power (W) 500-1500 70-1000 10-150 40-500

temperature 2000-6000 2000-3500 500-2000 1000-6000

standard 0.5-2% 2-10% 0.5-2% 1-5%

dynamic ~105 ~104 ~103 -104

Limit of
detection 0.1-100 0.1-1000 0.1-100 0.1-100



The human body requires a delicate balance of the

levels of various elements. Too much or too little of a

particular element can have devastating physiological

effects. Some typical ailments are a result of an imbalance

of elements in the body. High levels of sodium and low

levels of potassium, magnesium, and calcium all lead to

hypertension (high blood pressure) [35]. Hypertension is

the most common disease in industrialized societies and

contributes to the development of cardiovascular disease,

stroke, and renal failure [35]. The reduction of the levels

of certain elements caused by medical treatment with some

drugs can also cause serious problems [36].

Elements are transported by the blood and taken up in

varying amounts by organs and tissues [4]. The significance

of the levels of various elements in health makes it

important to have readily available techniques to monitor

these elements. The biological importance of each element

studied during the course of this research and the methods

used to measure them will be discussed.

Medical Significance

Trace Elements

An element is classified as a trace element if the

concentration is below 250 ppm [37]. Trace elements can be

classified into two groups, essential and nonessential. An

element is considered essential if lack of that element

causes problems. Essential trace elements include

manganese, copper, zinc, tin, and nickel. Nonessential

elements are those that are present in biological organisms

but have not been determined to play an important role [4].

The role of trace elements in the body has received

much attention by the scientific community. Although they

are present in low concentrations, they can play essential

roles in biological functions, and can also be detrimental

to biological activity if present in too great an amount

[14]. Trace elements are very important in the structure of

enzymes and are needed in the production of proteins. Trace

elements are also essential for the normal growth and

development of the human skeleton [36].

The metabolism of certain trace elements is involved in

various diseases. Therefore, the measurement of the


concentration of trace elements in biological fluids can be

used as a test for certain diseases. Poor health can also

be caused by environmental exposure to some elements. The

proper levels of trace elements are especially important

during pregnancy to insure a healthy child [36], and in the

elderly for their immune response [38]. Trace element

losses must be monitored closely in patients receiving

radiation therapy or chemotherapy. These patients may

require additional supplements of certain elements to

compensate for losses due to significant weight loss from

their illness or treatment [39].


Lead (Pb) poisoning is the leading environmental threat

to children in the United States [40-41]. The primary

sources of lead exposure are lead based paints and lead-

contaminated dust and soils. The Department of Housing and

Urban Development estimates that 4 million homes containing

young children have lead-based paint hazards [42]. Children

can also be exposed to lead though air, water, and food.

Lead poisoning affects virtually every system in the

body. It is especially harmful to the developing brain and

nervous system of unborn and young children [40, 43]. Lead

causes health problems because the body is unable to

distinguish between lead and calcium. When a person

consumes lead it is assimilated into the blood stream the

same way calcium is. Young children and pregnant women

absorb calcium more efficiently to meet their added

requirement so they are particularly at risk. Typically,

adults absorb 10 to 15 percent of the lead that reaches

their digestive tract. Pregnant women and young children

can absorb as much as 50% of the lead [44].

In blood, 94% of the lead is bound to the hemoglobin.

Within the past five years, it has been found that lead

concentrations as low as 100 ppb in the blood can be

detrimental to the health and intellectual development of a

child [40, 45]. Figure 3-1 shows the health effects of

various levels in the blood [40]. Acute lead poisoning can

result in anorexia, dyspepsia and constipation followed by

abdominal pain [46]. The detrimental effect of lead

poisoning on young children has led the Centers for Disease

Control and Prevention (CDC) to lower the acceptable level

of lead concentrations in the blood of children to 100 ppb,

compared to a level of 250 ppb considered acceptable from

1985 to 1991. The symptoms of lead poisoning are often

invisible at first, preventing the diagnosis and treatment

of most cases. The number of lead poisoning cases can be

greatly reduced if a large scale screening program is

implemented. This would require an inexpensive, easy-to-use

method to detect trace amounts of lead in blood.

of lead in blood

Health effect in


1000 Brain and kidney damage (adults)

Brain and kidney damage (children)

Increased blood pressure
(middle-aged men)
Decreased IQ and growth in
young children

Pre-term birth, reduced birth weight,
and decreased mental ability in
infants from mother's exposure during
100 pregnancy

Figure 3-1. Health effects of lead poisoning [40].


Manganese (Mn) is a important as a constituent of

metalloenzymes and as an enzyme activator [37, 47].

Research with animals has shown that Mn deficiency can lead

to impaired growth, skeletal abnormalities, disturbed

reproductive function, and problems with lipid and

carbohydrate metabolism. A deficiency of manganese leads to

a decreased level of blood clotting proteins and has also

been observed in several diseases including epilepsy [47-


Toxic levels of manganese can be the result of chronic

inhalation of airborne particulates containing high

concentrations of Mn from mines, steel mills, or some

chemical industries. Patients with liver disease are also

at risk from Mn toxicity because their liver may not

adequately clear the Mn absorbed from a normal diet. The

main signs of Mn toxicity include depressed growth and

appetite, impaired iron metabolism, and altered brain

function. Severe psychiatric abnormalities including

hyperirritability, violent acts, and hallucinations can be

caused by Mn toxicity [37].


Lithium (Li) is used in the treatment of manic

depressive psychosis [49]. Lithium is administered in the

form of lithium carbonate or another lithium salt with as

much as 1800 mg taken daily [50]. Blood normally only

contains lithium at a level of low ppb, but for therapeutic

lithium levels, a range of 0.5-1.5 mM is maintained in the

blood [51]. This is close to the toxic level, and a level

of 5 mM can be lethal. This necessitates the monitoring of

lithium levels in patients receiving this type of treatment.


Zinc (Zn) is the second most plentiful trace element in

the body. Zinc is important in the metabolic functions of

the body and is essential for the production and functioning

of over 40 enzymes that contain zinc [36]. Zinc is also

vital in the synthesis of DNA and RNA in every living cell.

Zinc plays a role in immune functions, in growth and

development, and in the synthesis and release of

testosterone. Zinc is especially important in expectant

mothers and in the growth of young children [52-53].

Zinc plays a major role in fighting infections and in

the healing process [47]. Those at risk for zinc deficiency

include women of child bearing age, young children, and the

elderly [54]. If the zinc level is too low it can cause

congenital malformations, including spina bifida and central

nervous system abnormalities. It can also cause severe

growth retardation, arrested sexual maturity and a loss of

appetite [36]. The level of zinc is especially critical in

the elderly because of the deterioration of immune function

with age. Low zinc levels can indicate diabetes because

zinc is important in the storage and release of insulin. A

zinc imbalance may be involved in hypertension.

Most cases of zinc toxicity have been related to food

poisoning incidents and to industrial pollution [53]. Too

much zinc causes anemia (reduced hemoglobin production),

elevated white blood cell count, muscular problems,

exhaustion, diarrhea, nausea and dizziness [52-53]. Very

high levels of zinc can impair metabolic functions that are

dependent on other trace elements [54]. High levels of zinc

can also interfere with the absorption of copper which can

provoke iron deficiency and anemia [47].


Magnesium (Mg) is essential in the transfer, storage,

and utilization of energy. Mg regulates and catalyzes over

300 enzyme systems in mammals [55-56]. Mg also maintains

the cardiovascular system, regulates DNA and RNA synthesis

and structure, and is important in cell growth,

reproduction, and membrane structure. Mg controls many

processes in the body including neuronal activity,

neuromuscular transmission, cardiac excitability, muscular

contraction, blood pressure, and peripheral blood flow [47,


A deficiency of Mg promotes hyper coagulability of

blood, atherogenesis, vasoconstriction, cardiac arrhythmias


and also damage to the cardiac muscles. A Mg deficiency may

also be related to cardiovascular disease, hypertension,

diabetes, depression, and atherosclerosis [35, 52].

Major Elements


Sodium is the principal cation of the extracellular

fluid. It is essential in maintaining the pH balance of the

fluid and is also important in nerve transmissions and

muscle contraction. Sodium levels can be depleted by

vomiting, diarrhea, or heavy sweating. If depletion occurs

it is critical to take measures to bring the sodium level

back up to a healthy level. If the sodium level is too high

it can result in hypertension, kidney disease, or heart

disease [47].


Potassium is important in the body in its maintenance

of fluid and electrolyte balance and cell integrity.

Diabetic acidosis, dehydration, or prolonged vomiting or

diarrhea can cause low potassium levels. Symptoms of low

potassium levels are muscular weakness, paralysis, and

mental confusion. Too much potassium can also cause

muscular weakness, confusion, as well as numbness, slowed

heart rate, vomiting, and eventually cardiac arrest [47].

Methods of Analysis


The Centers for Disease Control and Prevention has set

forth a number of desirable characteristics for an improved

blood lead measurement system. These characteristics

include an accuracy and precision of 10 ppb at 100 ppb, a

detection limit of 10-20 ppb, a sample volume of less than

200 yL, a low cost-per-test, an analysis time under five

minutes, portability, and minimal operator training required

to perform the method.

Screening methods

Currently used screening methods for lead in blood

which measure the level of either erythrocyte protoporphyrin

(EP) [57] or zinc protoporphyrin [45] in blood as an

indication of lead poisoning are not sensitive enough to

measure blood lead levels below 250 ppb. The EP test is

based on the increase in the amount of EP caused by an

increase in Pb. Porphyrins are the metabolic intermediates

in the biosynthetic process that produces heme [52]. Lead

impairs heme synthesis, preventing the incorporation of iron

into the protoporphyrin. This allows free protoporphyrin to

chelate cytosolic zinc. The amount of free protoporphyrin

can be measured because it fluoresces deep red [58]. The

whole blood is diluted and matrix modifiers are added. The


porphyrins are then separated from the blood and measured by

molecular fluorometry. This test has been recommended by

the CDC since 1978 [57]. Hematofluorometers have also been

used to screen children for lead [59]. These are portable

instruments that measure the zinc protoporphyrin directly in

a single drop of blood.

Two current methods being developed as portable

screening methods are anodic stripping voltammetry (ASV)

[60-65] and potentiometric stripping analysis (PSA)[66-67].

In anodic stripping voltammetry, a decomplexing agent is

added to the blood sample to free up the lead for

electrolysis. The Pb is reduced at a controlled potential

causing it to plate out on the surface of a mercury

electrode. A voltage sweep of the electrode releases the

lead and produces a current between the working and

reference electrode. By measuring this current, the amount

of lead can be determined [64]. This method has the

problems of instrument instability, slow speed, and

variations in response due to other elements present in the

blood. ASV also requires a plating solution. The use of

ASV with microelectrode arrays and indium as an internal

standard has improved the detection limit and precision for

the analysis of lead in blood [61].

In potentiometric stripping analysis, the lead analyte

is preconcentrated in a mercury film on a glassy carbon


electrode. This occurs by potentiostatic deposition where

electrons are added to the metal. The stripping step is

then achieved chemically by adding an oxidant. During the

stripping, the potential of the working electrode as a

function of time is closely monitored. This will produce a

well-defined stripping plateau which can be used for the

analysis of lead. Whole blood has to be diluted by a factor

of ten for analysis by the method of standard additions

[67]. The total time for analysis is about 5 minutes.

Both ASV and PSA possess the required accuracy and

precision to detect low blood lead levels. Electro-chemical

methods are advantageous as a screening method because they

are both portable and inexpensive; however, they have the

disadvantage that they require the use of reagents and

sample pretreatment when analyzing whole blood.

Exeter Analytical (North Chelmsford, MA) has developed

a commercial atomic absorption spectrometry (AAS) instrument

that can be used for blood lead screening [68]. The lead

absorption line at 283.31 nm is used with near line

background correction using the non-absorbing 287.33 nm lead

line. This instrument used a 150 W tungsten coil filament

in an enclosed chamber. Tungsten coils are excellent

atomization sources because of their high heating rate and

their commercial availability. Tungsten coils that are made

for halogen projector lamps can be used so they are


relatively inexpensive. One coil can last for approximately

70 runs. The blood samples were diluted by a factor of ten

with 0.2 % nitric acid, 0.5% Triton x-100, and 0.2% NH4H2PO4.

Calibration is done with aqueous standards, and a detection

limit of 30 ppb with a RSD of 9.0% at 100 ppb is obtained.

This method produces results in less than 3 minutes, has a

low cost-per-test, and is easy to operate [68].

Recently, a portable, battery powered AAS was developed

by Jones and coworkers for lead in blood screening [69]. A

tungsten coil was used as the atomizer and a miniature fiber

optic spectrometer with a charge coupled device (CCD)

mounted on a input card of a personal computer was used as

the detector. The blood samples were digested in nitric

acid by microwave heating and then diluted with distilled

deionized water. A 20 pL sample was placed on the coil and

then dried for 2 minutes at 3.0 A and then atomized at a

current of 6.0 A. The absorption signal was collected using

a hollow cathode lamp and a fiber optic. The spectrometer

and multichannel detector allowed near-line background

correction technique to be used. The lead absorption line

at 283.3 nm was used for analysis and the average of the

nonabsorbing lead lines at 280.2 and 287.3 nm was used for

background correction. The total cost of this entire system

was below $6000. A detection limit of 1 ppb for lead was

determined. The linear dynamic range was 2 orders of


magnitude and the precision was 5%. The method was proven to

be accurate by analyzing NIST blood standards. The coil was

used in the analysis of up to 400 samples [69].

Clinical methods

Research is being done by many different government

agencies and universities to improve blood lead measurement.

Isotope dilution inductively coupled mass spectrometry (ID-

ICP-MS) [70] and graphite furnace atomic absorption

spectrometry (GFAAS) [71-72] are methods that are able to

detect trace amounts of lead in blood below the level of

concern (100 ppb). Both of these methods are very accurate

and precise, but have the disadvantages of requiring sample

pretreatment and expensive instrumentation. The expense of

testing is a major consideration since millions of children

would need to be tested in a large-scale public health

screening program.

Atomic absorption spectrometry (AAS). Graphite furnace

atomic absorption spectrometry (GFAAS) is one of the most

popular methods for lead in blood analysis [71-90]. GFAAS

has excellent sensitivity and selectivity, large throughput,

and is capable of analyzing very small volumes. Many GFAAS

methods use a L'vov platform which is a small platform

placed in the graphite tube to hold the sample and ensure

that the tube and sample come to the same temperature at the

same time. In 1991, the Centers for Disease Control (CDC)


surveyed the methods being used by clinical laboratories for

blood lead analysis. Of the laboratories surveyed, 61% used

GFAAS, 5% of the labs used Delves cup AAS, 7% used

extraction AAS, 1% used carbon rod AAS, and 26% used ASV


The methods used to analyze lead in blood by GFAAS

include: direct introduction of blood into the furnace;

dilution with either water, Triton X-100, or acid;

deproteinization with nitric acid; matrix modification;

solvent extraction; or a combination of several methods

[82]. The direct injection of blood samples into a graphite

furnace has many problems associated with it. The blood can

seep into the graphite and produce major memory effects [83-

84], and during drying and atomization, the blood residue

can cloud the viewing windows [85]. Also, a carbonaceous

residue from the proteins in the blood builds up in the

furnace and is unable to be vaporized even at high

temperatures [76, 83]. Diluting the blood samples with

water alone is not sufficient to reduce adequately the

amount of carbonaceous residue [86]. The presence of water

in the blood sample also gives rise to a slow precipitation

of the red cell membranes, reducing the homogeneity of the

sample [87]. Diluting the blood samples with a 0.5 to 2%

solution of Triton X-100, a surfactant, causes complete


lysis of the blood cells and produces a clear solution that

minimizes the negative effects of the blood matrix [88].

The problem of carbonaceous residue build up can be

virtually eliminated by deproteinization of the blood with

30-50% nitric acid. The supernatant of the resulting sample

can then be injected into the graphite tube. This procedure

destroys the bulk of the organic matter in the blood.

However, the use of nitric acid shortens the life of the

graphite tube because of the oxidation of the tube's

pyrolytic coating [89]. The blood could be deproteinized at

lower concentrations of acid, but the inorganic salts

present were removed, necessitating the use of standard

additions [90].

Adding matrix modifiers to the blood can help in

retaining the analyte while volatilizing away most of the

matrix. The most common matrix modifiers used in blood lead

analysis are diammonium hydrogen phosphate, ammonium

dihydrogenphosphate, and phosphoric acid [75, 82]. By

adding these matrix modifiers, higher furnace temperatures

can be used to ash away the matrix without significant loss

of the analyte. The method of solvent extraction can also

minimize matrix effects, but it is very tedious, prone to

contamination, and does not completely remove interference



A GFAAS method has been developed which allows aqueous

standards to be used for blood lead analysis [71]. Prior to

analysis, the blood is deproteinized with a 5% nitric acid

solution containing 0.1% Triton X-100. The supernatant is

collected and the concentration of lead is measured using

Zeeman GFAAS. Parsons and coworkers have also developed a

method capable of calibrating with aqueous standards [45,

75]. A transversely heated graphite tube/platform called a

stabilized temperature platform furnace (STPF) was used.

This method produced a nearly isothermal system which

reduced the time of analysis, increased the precision, and

eliminated many of the chemical and matrix interference.

Blood samples preserved in EDTA were diluted by a factor of

10 with a solution containing ammonium dihydrogen phosphate,

triton X-100 and nitric acid. The samples were directly

introduced into an autosampler where the mixing with the

solution occurs. Twelve micro-liter aliquots were injected

into the furnace and atomized. Each analysis took 90 s, and

the system was able to run approximately 100 samples per day

with duplicate injection. The precision was better than 5%.

While both of these methods are advantageous because aqueous

standards can be used for calibration, they have the

disadvantage of requiring appreciable sample treatment.

A flame AAS method has been developed that used 20 yL

of blood samples spotted on filter paper and then analyzed

in a Delves cup [93]. A Delves cup is a small nickel cup

that is positioned in the flame for analysis. The blood

sample must be allowed to dry on the filter paper and is

then ashed. The ashing step burned away the paper and then

the sample was introduced into the flame to be analyzed for

lead by measuring the absorption at a wavelength of 283.3

nm. The entire analysis time was 15 s per sample and a

limit of quantitation of 40 ppb was obtained. This method

gave excellent reproducibility and accuracy [93]. It has

the disadvantage that there was considerable variability in

the adsorptiveness of the papers which was detrimental to

the accuracy. Also, this method's requirement of allowing

the blood to dry on the filter paper resulted in the sample

being susceptible to contamination from airborne particles.

As a clinical method, flame AAS has the disadvantage that

the equipment is expensive and cumbersome and requires a

combustible gas source.

Inductively coupled plasma atomic emission spectrometry

(ICP-AES). A carbon rod atomizer has been used to analyze

blood samples with a ICP atomic emission spectrometer [94].

Blood samples were diluted by a factor of five with

distilled water. The samples were placed on the carbon rod

atomizer and then dried and volatilized. The resulting

vapor was carried into the plasma by the plasma gas. This

method of sample introduction was more efficient than

nebulization. An aqueous detection limit of 7 ppb was

reported for lead with a relative standard deviation (RSD)

of 0.2%.

Inductively coupled plasma mass spectrometry (ICP-MS).

ICP mass spectrometry is a very sensitive method for the

measurement of lead in blood [70, 95-96]. The main method

of sample introduction in an ICP-MS is a nebulizer. Aqueous

samples are transferred to a nebulizer by a peristaltic

pump. The aerosol produced by the nebulizer is carried to

the plasma by a flow of gas, typically argon. The high

temperature of the plasma vaporizes and ionizes the sample

and the ions are then detected in a mass spectrometer

according to their mass to charge ratio [97]. ICP-MS with

isotope dilution, is the method with lowest bias for

determining lead in whole blood and serum [70, 95]. Isotope

dilution mass spectrometry involves measuring the change in

the relative abundance of two isotopes of an analyte after

adding a known amount of one of the isotopes to the sample.

The CDC uses isotope dilution (ID) ICP-MS for the analysis

of its certified reference material, lead in bovine blood,

from its Blood Lead Laboratory Reference System. An aliquot

of the whole blood sample is spiked with a radiogenic lead

isotopic standard. This aliquot along with an unspiked

aliquot is then digested with ultrapure nitric acid in a

microwave oven. After cooling, both samples are diluted and

then aspirated into an ICP-MS. The isotope ratios of lead

at mass 206 and mass 208 are then measured. While this

method is very accurate and precise for determining lead in

blood, it is more suitable for determining reference values

than being used as a clinical method because of it's high

cost and low throughput (10 samples per day) [70, 95].

Primary and Trace Elements

The main methods for trace elemental analysis in the

clinical laboratory are absorption or emission spectro-

photometry. Typically, the blood is separated, and the

plasma or serum is used for analysis [4]. Methods capable

of performing trace elemental analysis include AAS, ICP-AES,

and ICP-MS. Other methods include electrochemical, neutron

activation, flame atomic fluorescence spectrometry,

molecular absorption spectrometry, X-ray fluorescence,

particle-induced X-ray emission and radiochemical techniques

[46]. However, many of these methods are not suitable for

routine use in a clinical setting. Neutron activation, for

example, is a very sensitive technique but requires the use

of a nuclear reactor and requires a very long time for

analysis [98]. Currently, AAS is the most widely used

method in clinical laboratories, usually employing

electrothermal sample introduction [46]. Recent reviews of

clinical methods of analysis have appeared in Analytical

Chemistry [99] and in the Journal of Analytical Atomic

Spectrometry [100].

Atomic absorption spectrometry (AAS)

Sodium, potassium, zinc, magnesium, and iron blood

levels can be determined by flame atomic absorption

spectrophotometry (FAAS) [47, 101]. The samples are diluted

and introduced into the flame. The analysis of each element

requires a hollow cathode lamp that produces light at a

wavelength specific for that element. The fraction of

absorbed light is used to determine the concentration of the

element present. Shang and Hong have used a microvolume

injection technique to measure the levels of Cu, Zn, Ca, Mg,

and Fe by FAAS [102]. The blood samples were treated with

triton x-100 and then diluted with a mixture of 0.18 M HC1,

0.003 M La203, and 0.013 M KC1. The injection volume used

was 10 pL. Atomic absorption has greater sensitivity than

either flame atomic emission spectrometry (FAES) or ion

selective electrodes (ISE), but it is less precise and not

as suitable for routine clinical analysis. It has a high

initial cost and the necessity for compressed gases and

flames are undesirable in the clinical laboratory.

GFAAS is a very popular method for elemental analysis

in blood. The various methods used for lead analysis are

also used for many other elements and have the same


advantages and disadvantages [90]. The levels of magnesium,

manganese, lithium and iron have all been determined by

GFAAS [47, 50, 52]. GFAAS has achieved a detection limit of

2 ppb for manganese in blood and is the most common method

for analysis of lithium in blood [50, 103]. The main

disadvantage of GFAAS as a clinical technique is its

limitation as a single element technique. Some researchers

have developed complex methods of determining two or three

elements simultaneously, but it is difficult and expensive,

requiring a complicated optical setup [74].

Atomic emission spectrometry (AES)

Sodium and potassium in serum are usually analyzed by

either flame atomic emission spectrometry (FAES) or by ion-

selective electrode potentiometry (ISE) [101]. FAES

requires a dilution of the sample by 100 to 200 times, often

adding lithium or cesium to the sample as an internal

standard and ionization suppressant. An air-propane flame

is used, and the sodium emission is monitored at 589 nm and

the potassium emission at 766 nm. Only 1 to 5% of the atoms

in the flame are excited to emission, but the concentration

of the elements is sufficient for accurate and precise

measurements [101]. Lithium levels can also be reliably

measured using flame emission spectrometry [104].

Flame photometric flow-injection analysis has been

successfully used to simultaneously measure the levels of

lithium, sodium and potassium in blood serum [105]. The

serum samples were diluted ten-fold with doubly-distilled

deionized water. The sample was then injected and split into

three portions so that each portion reached the detector at

a different time. Between the analysis of each sample

portion, the filter on the detector was changed to be

specific for each analyte. This method allowed the analysis

of 108 samples per hour [105].

ICP-AES has been used to measure the levels of Fe, K,

Mg, Na, Li and Zn in human serum and blood [106-108]. Serum

samples were digested in nitric acid or diluted with

deionized water. A microsampling system has been developed

for ICP-AES which uses <0.1 mL of sample [107]. By

digesting the blood or serum sample with acid, aqueous

analytical curves could be used for calibration.


Spectrophotometry involves selectively completing and

separating an analyte using either an inorganic or organic

colorimetric reagent. Various organic reagents have been

used as spectrophotometric agents for the analysis of

lithium, magnesium, and iron in blood and serum [51-52].

Calmagite, methylthymol blue and formazam dye, are some

examples of chromophores that have been used for the

analysis of magnesium. The level of iron in blood is

analyzed by exposing the blood sample to strong acids to


dissociate the iron from its binding proteins. A chromogen

is then added to the sample to produce a iron chromogen

complex that has an absorbance maximum in the visible

region. The concentration of lithium in serum can be

measured by observing shifts in the spectrum of a reagent

caused by the presence of lithium. The reagents must be

very specific for lithium because sodium, which is present

at high concentrations in blood, is generally an

interferent. Crown ethers can be used for lithium analysis.

By using different cage sizes, conformational flexibility,

and various side groups, crown ethers can be made to form a

complex selectively with the several analytes of interest.

The complex formed can be extracted into an organic solvent

with an anionic reagent that is colored allowing

spectrophotometric analysis [51].

A major disadvantage of spectrophotometry is the

limited selectivity due to overlapping absorption bands. It

is, however, easy to use, rapid, and can be readily

automated [4].

Inductively coupled plasma mass spectrometry (ICP-MS)

ICP-MS has been used for the measurement of trace

elements in whole blood and serum [33, 109-112]. The

advantages of using ICP-MS include high throughput (40

samples/hour), possibility of simultaneous analysis, and

good detection limits. Over 50 elements have detection

limits in the range of 0.01 to 0.1 ppb [33]. Adding an

internal standard can often correct for matrix effects and

instrument drift.

Blood and serum samples for ICP-MS are usually digested

with acid or diluted. The sample pretreatment often

includes a separation step. The amount of time needed for

sample preparation has been reported as 25 minutes for 50

specimens [33]. Barany and Bergdahl reported on a method

for ICP-MS of trace analysis in blood where whole blood was

diluted 50 to 100 times with an alkaline solution. Each

analysis required only 75 seconds. Even with dilution, some

problems were encountered with the buildup of denatured

proteins from the blood so the torch required occasional

cleaning. This method was used for the determination of 7

trace elements in blood. It was not suitable however for

Mn, Se, Hg, or Cr [113].

The major disadvantage of using ICP-MS is the high cost

of the instrument and the operator expertise needed. Also,

the analysis of lighter elements is difficult because of

more interference. Interferences arise from mass overlap

from either polyatomic ions, doubly charged ions, or

elements with the same isotopic masses. Currently, it is

not possible to analyze chromium, manganese, or iron by ICP-

MS in biological samples due to the presence of

interference [33, 109-110, 113].

X-ray fluorescence

X-ray spectrometry involves bombarding the sample with

radiation of distinct energy. This removes electrons from

the inner shells forming atoms in an excited state. The

electrons from the outer shells fall into the shells vacated

by the removed electrons according to specific transition

rules. The radiation emitted by this process is very

characteristic. The method of x-ray fluorescence can be

used for simultaneous multielement analysis on a very small

sample of blood (2-3 pL) without destroying the sample [114-

115]. The blood levels of potassium, calcium, chromium,

iron, nickel, zinc, selenium and lead can all be determined

in one measurement. Detection limits ranged from 21 ppm for

phosphorus to 30 ppb for lead in blood [114-116]. The

method of X-ray fluorescence has the disadvantage that it is

very difficult to match the composition of the calibration

standards to the matrix of the sample [4].

Electrochemical techniques

Voltammetry, an electrochemical method, is also capable

of measuring trace elements in blood. In voltammetry, the

measurements are based on the potential-current behavior of

a small electrode that is easily polarized [4]. Voltage is

applied to a microelectrode and the diffusion current is

measured as a function of the voltage. This allows both

quantitative and qualitative analysis of the trace element.

For this method, it is necessary to digest completely the

samples prior to analysis [4].

Ion selective electrodes (ISE) are capable of

determining the level of potassium, sodium, magnesium and

lithium in blood or serum by measuring the potentiometric

charge as a function of ion concentration [51, 55, 117-120].

The membranes of ISE's are ideally sensitive to only one

ion. Most membranes, however, respond to ions other than

the one for which they are designed. Polymer-bound liquid

membranes use a membrane that contains a sensing material

dissolved in the polymer support matrix. If the sensing

material is neutral in charge, then it must complex with the

analyte in some way to transfer it across the membrane or it

must be able to facilitate ion exchange. Neutral sensing

materials are called ionophores and are often some type of

crown ether. Crown ethers can be made in such a way that

they can selectively complex a given ion. The polymer

matrix containing the sensing material is often polyvinyl

chloride (PVC) [121]. Bulky crown ethers used in a PVC

membrane ISE can exhibit a selectivity up to 2000:1 for

lithium [51]. A glass ion-exchange membrane is used for the

analysis of sodium, and a valinomycin neutral-carrier

membrane is used for potassium [101].

The use of ISE's to analyze clinical samples involves

either the direct analysis of undiluted samples or the

indirect analysis of pre-diluted samples. Direct ISE

methods are subject to bias because of the difference in the

serum matrix and the aqueous samples used for calibration.

Indirect ISE is susceptible to error introduced by the


ISE's to monitor Mg can yield rapid results on blood,

plasma, serum and aqueous solutions with sample sizes

ranging form 100 to 200 pL [55]. The Mg ISE's employ

ionophores using neutral carrier based membranes with

excellent precision reported at 2 to 4%. However, this

method does experience problems with very low levels of

magnesium because the analytical response is not linear at

these low concentrations [52].

ISE's compare favorably to the methods of atomic

absorption spectrometry and flame emission spectrometry for

the analysis of several elements. ISE's have the advantages

that they function in turbid solutions, have a wide dynamic

range, have a rapid response, are inexpensive, and are very

portable with current instruments weighing between 7 to 12

kg [117]. The rapid response is very beneficial in

monitoring dosages and compliance with medical treatment

such as lithium treatment in psychiatric patients. ISE's

have the disadvantages that they have limited sensitivity,

are subject to interference from other ions and memory

effects, and require frequent calibration.


The analysis of the primary and trace elements in blood

is very important in maintaining and monitoring the health

of individuals. Although there are many methods capable of

doing multi-element analysis in blood, there is still much

room for improvement. The most accurate and precise methods

all require some sort of sample pretreatment. Sample

treatment requires time and is a possible source of

contamination. Ideally a clinical method for blood analysis

would be able to analyze whole blood directly, without

sample pretreatment, and would be able to use simple

standards (i.e. aqueous) for calibration.



The experimental setup is shown in figure 4-1. Each

component of the experimental setup will be described.

Microwave Plasma Electronics

The microwave plasma was generated by an 870 W

magnetron (Samsung OM75A) at 2450 MHz. This type of

magnetron is commonly found in domestic microwave ovens.

Magnetrons are capable of high power with low cost and high

efficiency. Magnetrons produce microwaves through the

combination of an anode, cathode, and magnet. Electrons are

emitted from the cathode and are introduced into a

combination of electric and magnetic fields which cause the

electrons to move around the cathode. The electrons then

move toward the anode and exchange potential energy,

building up the microwave field. When the electrons hit the

anode, the power is coupled directly to the output. The

output allows the microwaves to be taken out via an external

Removable Quartz

Helium Flow in --

Teflon Tape

Figure 4-1. CMP-AES experimental setup.


transmission line [9]. A diagram of the magnetron is shown

in figure 4-2.

The magnetron was powered by a current regulated

analog-programmable power supply (Model 106-05R, Bertran

High Voltage, Hicksville, NY, USA). An A.C. power

transformer (Magnetek Triad, model F-28U, Newark

Electronics, Chicago, IL) was used to provide a high

current, low AC voltage for the magnetron filament.


The rectangular waveguide was made out of aluminum and

constructed in the laboratory. The waveguide had the

following dimensions: height = 47 mm, width = 98 mm, length

= 277 mm. The waveguide had a hole near one end on the top

allowing the output of the magnetron to be inserted, and

holes on the top and bottom near the other end allowing the

torch to be suspended within the waveguide. The hole

diameter for the torch was 44 mm, and the center of the hole

was 58 mm from the end.


The torch consisted of four concentric quartz tubes:

an outer quartz tube (outer diameter (o.d.) = 19 mm)

directed the flow of helium; a removable quartz tube (o.d. =

15 mm) reduced the dead volume of the torch; and an inner

Heater leads
and cathode




Output antenna

Figure 4-2. Magnetron

quartz tube (o.d. = 5 mm) that supports a short piece of

quartz tubing (o.d. = 2 mm) in which the filament rests.

The inner quartz tube is used to decrease the volume of the

torch in order to reduce the amount of helium gas required.

The inner quartz tube is a separate piece of quartz tubing

held in place in the torch with teflon tape. In some

experiments, the inner tube is brought up around the plasma

so that it shields the plasma. The inner tube as a shield

is better than using the torch itself because the inner

quartz tube is easily replaced if the plasma attacks it and

makes it optically unclear. A quartz chimney surrounds the

top of the torch to reduce instabilities caused by air


Plasma Gases

Helium (BOC Gases, The BOC Group, Inc., Murray Hill,

N.J.) was used as the plasma gas. Helium was an excellent

plasma gas for atomic emission spectroscopy because of its

high ionization energy [122]. The ionization energy of

helium is 24.6 eV compared to 15.8 eV for argon [123]. The

high ionization energy enhanced the possibility of energy

transfer to the analyte. A helium plasma is able to excite

efficiently elements introduced into the plasma, and has low

background characteristics. Hydrogen (BOC Gases, The BOC

Group, Inc., Murray Hill, N.J.) was introduced into the

plasma at a flow rate of 250 cm3/min for the cleaning step.

The presence of hydrogen in the plasma helped to create a

reducing environment and increased the temperature of the

plasma [124]. The higher temperature and reducing

environment helped in the removal of the carbonaceous

residue left over from the blood sample.


The graphite cup holder electrodes were made out of

spectroscopic grade carbon (Union Carbide, Carbon Products

Division, Cleveland, OH). The metals used for the cups and

the electrodes were obtained from Alfa Aesar/Johnson

Matthey, Ward Hill MA. The following metals were obtained

as rods and machined to make the various electrodes: nickel

(99.5% pure), titanium (99.99% pure), and tungsten (99.95%

pure). The tungsten screen used was obtained from Newark

Wire Cloth Co., Newark, NJ.

The tungsten wire (99.95% pure) used was also obtained

from Alfa Aesar/Johnson Matthey, Ward Hill, MA. Three

diameters of wire were used: 0.25 mm, 0.5 mm, and 0.75 mm.

The final filament used was made out of the 0.5 mm tungsten

wire. The top of the filament was a tight 2.5 turn spiral

with a diameter of 3 mm. The total length of the filament

was 6.5 cm.

Lens Setup

The initial lens setup (figure 4-3a) used two

planoconvex lenses. The first lens (diameter = 50.8 mm,

focal length = 125 mm) was placed 125 mm from the plasma to

collimate the emission from the plasma. The second lens

(diameter = 25.4 mm, focal length = 50.8 mm) was placed so

that the emission was focused onto the entrance slit of the


In an attempt to improve the precision of the CCMP-AES,

the lens setup was changed after the lead-in-blood work was

completed. The lenses (figure 4-3b) were set up so that the

emission from the plasma filled the collimating mirror of

the spectrometer. Two lenses had to be used because a

single lens could not be placed close enough to the plasma

for the desired focusing. A first lens (focal length = 38.1

mm, diameter = 38.1 mm, Esco Products/Precision, Oak Ridge,

New Jersey) was used to form a one-to-one image of the

plasma at a proper distance away from the plasma. It is

placed 76 mm from the plasma. The second lens (focal length

= 25.4 mm, diameter = 25.4 mm, Esco Products/Precision, Oak

Ridge, New Jersey) was then used to magnify the image in

such a way that the collimating mirror of the spectrometer

was completely filled with emission. The second lens was

placed 102.78 mm from the first lens. The distance for each


.. .. .


Lens 1

Lens 2


To collimating
mirror of

1:1 Image
of Plasma

s I

Figure 4-3.

Lens setup (not to scale): a) lead in
blood work, b) multielement work.



lens was calculated from the equations:

1/f = 1/s + 1/s' and m = s'/s

where f is the focal length of the final lens, s is the

distance between the emission source and the final lens, s'

is the distance between the lens and the mirror, and m is

the resulting magnification of the image. In the

modification of this lens setup, s became the distance from

the second lens to the one-to-one image formed by the first

lens. This lens setup resulted in a magnification of the

plasma image of approximately 25 times.


The detector consisted of a 0.5 m spectrometer (Spex

1870, Edison, NJ, USA) and either a photodiode array (PDA)

or a charge coupled device (CCD). The spectrometer grating

contained 1200 grooves/mm with a blaze wavelength of 300 nm.

The preliminary work and the lead in blood research was done

using the PDA. The multielement work was done with the CCD.

The spectrometer slit width was adjusted for each

element. If greater sensitivity was needed, the slit width

was opened to as much as 40 gm. For elements requiring less

sensitivity and higher resolution, a slit width as small as

10 lm was used. The slit height was kept constant at 2 cm.

Both the PDA and the CCD gave a spectral window of 40 nm.

Photodiode array

The intensified photodiode array (Tracor Northern TN-

6122A, Middleton, WI, USA) consisted of 1024 silicon

photodiodes arranged linearly, each spaced 25.4 pm apart.

Each photodiode consisted of a layer of silicon doped with

atoms containing extra electrons (p-type semiconductors) on

top of a layer of silicon doped with atoms with one valence

electron less than silicon (n-type semiconductor). This

allows the current to flow in only one direction. A reverse

biased potential is applied across the diode so that when

exposed to light, electron hole pairs are created producing

a current that is proportional to the amount of light [125-


Charge coupled device

The detector was changed from the photodiode array

(PDA) that was used for much of the lead-in-blood work, to a

charge coupled device (CCD) for all of the multielement

work. This change was necessary because of problems that

developed with the hardware and software that controlled the

PDA. The CCD detector has the advantage that it was two

dimensional and was cryogenically cooled to reduce the dark


The CCD contained 296 x 1152 picture elements (pixels).

Each pixel was 20 pm square and consisted of a metal-oxide-

silicon (MOS) capacitor. The pixels were made out of an

insulating silicon dioxide layer over a p-type silicon

substrate. This was topped by a thin metal electrode [125,

127-128]. When a photon struck a pixel, it penetrated the

lattice breaking the covalent bonds between adjacent silicon

atoms. This created electron-hole pairs which were measured

as an electric charge. The radiation striking each pixel

was proportional to the resulting charge and was measured by

transferring the charge to a single point. The covalent

bonds could also be broken by thermal agitation. The

thermal generation of charge was reduced by cooling the CCD.

Figure 4-4 shows the effect of cooling on the CCD background

counts. The temperature was maintained constant by a

heating element in the CCD dewar. The temperature was

maintained at -110 C even though there was not much change

in the dark counts below a temperature of -40 "C. At

temperatures higher than -90 C the liquid nitrogen

evaporates too quickly. At temperatures lower than -140 C,

the charge transfer efficiency from pixel to pixel may be

lowered, degrading the CCD performance [126].

If the light levels reaching the CCD were too high,

blooming could occur. Blooming is the spillage of charge

from an over-illuminated pixel to an adjacent pixel [126,

128]. The signal to noise ratio and the dynamic range could

be improved by the process of inning. Binning combines the

charge from adjacent pixels during readout. The charge read



1. Ox1o


i 8.0x10 6



S4.0x10 6

2.0x10 -

0.0 -



-120-100 -80 -60 -40 -20 0 20
Temperature (OC)

Figure 4-4. CCD background dependence on temperature.


will increase by the number of pixels binned, but the noise

will stay the same. Binning has the disadvantage of

reducing the spatial resolution [126].

Originally, the CCD detector was not sensitive to

emission below 400 nm. The camera was sent to Spectral

Instruments (Tucson, Arizona) so that a UV enhancement

coating could be applied to the CCD element. The coating

was lumogen yellow, an organic phosphor. The phosphor

absorbs light in the UV range and re-emits it in the visible


Computer Software

The programmable power supply and the triggering of the

detector were controlled by a computer (PC's Limited, model

28608L, PC's Limited, Austin, TX) and a computer interface

(Model SR 245, Stanford Research Systems, Palo Alto, CA,

USA) using a program written in Microsoft QuikBasic 4.50

(Copyright Microsoft Corporation, 1985).

The emission spectra were collected using CCD9000"

spectral acquisition software, version 2.2.2 (copyright

1990-1992, Photometrics, Ltd.). The peak areas were

determined using the program LabCalc" (copyright 1987-1992,

Galactic Industries Corporation). Analytical curves were

constructed using Origin" version 4.0 (copyright 1995,

Microcal" Software, Inc.).


Aqueous Standards

Aqueous standards were prepared by sequentially

diluting 1000 ppm reference standards for each element

(Fisher Chemical, Fisher Scientific, Fair Lawn, New Jersey).

The standards used in the standard additions of the blood

analysis were prepared from the salts of the elements being

analyzed. This was necessary because the standards needed

to be non-acidic to prevent denaturing of the blood. Also,

the concentrations required for some of the elements were

larger than the available aqueous standards. All the

aqueous standards were prepared using deionized water

(specific resistivity 18 MQ/cm) from a Milli-Q Plus water

system (Millipore Corporation, Bedford, MA). The aqueous

standards were introduced for analysis using a 2 kL air

displacement pipetter (Eppendorf, Brinkman Instruments Inc.,

Westbury, NY).

Blood Standards

The lead bovine blood standards used were Quality

Control Materials (QCM) produced and distributed by the CDC

Blood Lead Laboratory Reference System (BLLRS). The samples

were collected by the CDC from two cows kept at the CDC

livestock facility (Lawrenceville, GA) that were given


dosages of lead nitrate in gelatin capsules. The blood was

collected from the cows and the initial concentration was

determined using atomic absorption spectrometry. Varying

amounts of the two blood samples were then blended to give a

range of lead concentrations. The final concentrations of

the samples were determined using ID-ICP-MS [70].

Human whole blood was collected by venipuncture into a

Vacutainer (Becton Dickinson Vacutainer Systems, Franklin

Lakes, NJ, USA) coated with KEDTA as an anticoagulant. The

standard addition samples were made by adding varying

amounts of an aqueous standard to a 0.75 mL portion of whole

blood. Deionized water was added to the sample to produce a

final volume of 1.0 mL. This resulted in a sample that was

75% whole blood. The samples were gently rolled and then

sonicated for 5 minutes to thoroughly mix the aqueous

standard and the blood.

A lead-in-blood Standard Reference Material (SRM 955a)

was purchased from the National Institute of Standards and

Technology (NIST) (Gaithersburg, MD). This SRM consisted of

four vials of frozen bovine blood each containing a

different concentration of lead. The concentration of lead

in each SRM was determined by NIST using ID-ICP-MS and

confirmed using GFAAS and laser-excited atomic fluorescence

spectrometry. The concentrations are shown in Table 4-1.

The uncertainty reflects a confidence level of 95%.


It was necessary to keep the blood samples frozen when

not in use. By freezing the blood, the bacterial and

chemical interaction of the blood sample were greatly

reduced. If not frozen, the various elements can bind to

proteins in the blood and settle out. Although the content

of the element in the vial remains the same, the

concentration in the liquid portion will be less than the

target value for the standard. The proteins could also

denature leading to a change in the homogeneity of the blood

[129]. Prior to use, the blood samples were allowed to thaw

to room temperature, homogenized by gently rolling, and then

sonicated for 10 minutes.

The blood samples were introduced for analysis by a

positive displacement micropipetter (Drummond model 525,

Drummond Scientific Co., Broomall, PA). Before depositing

the sample, the outside of the glass capillary tip was wiped

with a KimwipeM to remove any blood that had adhered to the

tip. The pipetter was cleaned between sample runs by

repeatedly depressing the plunger first in a solution of 5%

nitric acid solution and then in deionized water.

Table 4-1. Concentration of lead in SRM 955a at 22C.

Vial Number Concentration (ppb)

955a-1 50.1 0.90

955a-2 135.3 1.3

955a-3 306.3 3.2

955a-4 544.3 3.8



Capacitively coupled microwave plasma atomic emission

spectrometry (CMP-AES) has been used to analyze various

types of samples directly. Discrete sample introduction in

a CMP is easier than in a microwave induced plasma or an

inductively coupled plasma because a CMP is the only one

that uses an electrode to support the plasma.

Investigations have involved the analysis of various

matrices including dry tomato leaf samples, coal fly ash,

steel, oil, and biological materials [1-3, 130-132]. A

variety of methods have been used for sample introduction

into a CMP. They have included nebulization, thermal

vaporization, and hydride generation.

Methods of Sample Introduction into a CMP

In the earliest work with a CMP as an atomic emission

source, solid electrodes were used to support the plasma.

The analyte solution was vaporized and carried into the

plasma by premixing the analyte carrier gas and the plasma


gas. Hanamura et al. used a platinum clad tungsten

electrode with this method of sample introduction [133].

The platinum coating was used because the platinum is

thermally stable, chemically inert, and has a low thermionic

emission rate. These properties of platinum increased the

electrode lifetime and reduced the contamination of the

plasma by elements present in the electrode. Interfering

emission lines from the electrode is one of the major

drawbacks of the single electrode CMP. Hanamura and

coworkers use this type of electrode to analyze hydrogen and

oxygen in metals and also mercury in water [133-134].


Several researchers have used a nebulizer to introduce

aqueous samples into a CMP [124, 135-137]. A nebulizer is

an easy and inexpensive way to introduce a solution into a

plasma. Nebulization is a process where the sample to be

analyzed is transferred by a peristaltic pump to a nebulizer

which converts the sample into an aerosol in a spray

chamber. The aerosol is then swept by a carrier gas through

the center of the electrode into the plasma. A disadvantage

of nebulization is that much of the sample is lost in the

spray chamber.


Patel et al. used pneumatic nebulization with a CMP to

analyze aqueous samples for 15 elements [135]. Sample

solutions were nebulized with a Meinhard nebulizer and a

laboratory-constructed spray chamber and desolvation system.

A tubular electrode made out of tantalum was used to support

the plasma. The analyte carrier gas passed through the

center of the hollow electrode and entered into the plasma

at the top of the electrode [25]. By introducing the sample

directly into the core of the plasma, the interactions

between the sample and plasma were improved. Also, the

concentration of the analyte in the plasma viewing region is

increased, improving the detection limits, signal to noise,

and signal to background. It was determined that this

method gave low detection limits with a wide linear dynamic

range for a number of different elements. Several other

tube materials and forms of electrodes were evaluated. The

electrode types included a platinum tube, a copper tube, a

platinum coated tungsten wire (0.5 mm o.d.) a molybdenum

rod, and a tungsten rod with platinum cladding [135].

Hwang et al. used graphite as the electrode material

[137]. This electrode had a lower emission background and

did not significantly contaminate the plasma in comparison

to the metal rod electrode. Excellent detection limits for

several elements in aqueous solutions were obtained [137].

Thermal Vaporization

Two previous methods of sample introduction by thermal

vaporization (TV) include a tungsten filament electrode

(figure 5-1a) [23, 131-132] and a cup holder electrode

(figure 5-1b) [1-3, 130, 138-139]. The sample was

introduced into the CMP by directly depositing it on the

filament or in the cup held by the electrode. Thermal

vaporization is advantageous over nebulization in that a

greater percentage of the sample is introduced into the

plasma. TV has the disadvantages of poorer precision,

greater interference effects, and a lower throughput of


Hanamura et al. used a another method of thermal

vaporization with a CMP to measure carbon, hydrogen,

nitrogen, oxygen and mercury in orchard leaves and tuna fish

[140]. A separate furnace vaporizer was used. The sample

was held in a quartz crucible which was heated. The carrier

gas was flown through the sample chamber to carry the

volatile constituents through the center tube of the torch

and into the plasma for analysis.

Cup electrode

A cup can be used to introduce both liquid and solid

samples into a CMP. In order to use a cup electrode, an

electrode must be fabricated such that the top of the


Electrode designs: a) filament electrode,
b) cup holder electrode, c) platform electrode,
d) titanium electrode with nickel cap,
e) titanium electrode with titanium cap.


Figure 5-1.

electrode has a hollowed out portion that will snugly hold

the cup. The electrode must be made out of a material that

is conductive and has a higher melting point than the

thermal temperature of the plasma. Materials that can be

used for the electrode are graphite or various metals.

Graphite electrodes are cheaper and more resistive than

metal electrodes, and emission from the metal electrode can

also cause interference in analysis. Electrodes made out of

metal have several advantages over those made out of

graphite. Metal electrodes are more durable and last longer

than graphite electrodes. Also, graphite electrodes form

refractory carbides and produce gaseous molecular carbon

species which cause interference in emission measurements.

Cups made out of both graphite and tungsten have been

used to hold the sample. The cup was placed on the top of

the electrode and the electrode containing the cup was

placed into the central tube of the torch. The plasma was

ignited at a low power (around 100 W) to ash the sample and

was then raised to 400 700 W to atomize and excite the

analyte, enabling the measurement of the emission [1]. By

using a cup instead of a wire filament, higher powers could

be used so that there were fewer matrix effects and the

signal was larger. The disadvantage of using a cup was that

the atoms were dispersed over a wider volume so the number

density of excited atoms was smaller.


Ali et al. used CMP-AES with a cup electrode with both

the electrode and the cup made out of graphite. Detection

limits ranging between 10 and 210 pg were obtained for 12

elements with a precision better than 12% [138]. A sample

volume of only 5 gL was used. The graphite cup was coated

with tantalum carbide to reduce memory effects. Because

graphite was quite porous, memory effects were observed for

all the elements analyzed. The cups lasted 30-40 firings

and then had to be replaced due to etching of the cup rim by

the plasma. Multielement analysis with this system was

performed on coal fly ash and tomato leaves [1]. Spencer et

al. used a tungsten cup to analyze silicon in oils [3].

Tungsten was found to be an excellent cup material because

of its tolerance to high temperatures, long lifetime, low

emission background, and low memory effects.

More recently Pless et al. used a tungsten cup in a

graphite electrode for multielement analysis [139] The

cup had a total volume of 30 pL. Detection limits in the

low picogram range were obtained for 10 pL samples of

cadmium, magnesium, and zinc in aqueous solutions. Cadmium

in solids was also analyzed obtaining a detection limit in

the picogram range [130]. Various matrices were analyzed by

this system including coal fly ash, tomato leaves, soil,

bovine liver, and oyster tissue. The results achieved good

agreement with the certified values of the reference


Filament electrode

Ali and Winefordner evaluated a tungsten filament

electrode for multielement analysis in aqueous solutions

[23]. Filaments have the advantage that they are simple and

inexpensive. The sample was introduced into the plasma by

placing a few microliters of sample in a loop in the

filament. The sample was then dried at low microwave

power. The filament heated up rapidly creating a high rate

of volatilization. After the sample was dry, the plasma was

ignited and the sample was ashed if necessary by a low power

(30 W) microwave plasma The power of the plasma was

increased until the sample was atomized and excited so that

the emission could be measured. It was found that adding a

low flow rate (100 mL/min) of hydrogen gas with the plasma

gas reduced the background emission from the tungsten

filament. The absolute detection limits of 12 elements were

in the range of 1 to 100 pg and this compared favorably to

the method of graphite furnace atomic absorption

spectrometry (GFAAS). A linear dynamic range of 3 to 4

orders of magnitude was obtained and the precision was

better than 10%. Reported lifetimes for the filaments were

500-1000 runs [23].

Wensing et al. evaluated a CMP-AES for a lead blood

screening method using a tungsten loop (figure 5-la) as the

electrode [131-132]. A tungsten wire of 0.25 mm thickness

was tied in a knot, leaving a small loop in the center, and

the remaining ends of the wire were bent so that they could

be inserted into a piece of quartz tubing which was then

inserted into the plasma torch. The blood samples were held

in the loop by adhesion to the wire.

A 5 pL blood sample was placed in the filament loop and

subsequently dried, ashed, and atomized. Drying was

accomplished using microwave power to inductively heat the

electrode for 90 seconds. After drying, the helium gas flow

was turned on and a small plasma was ignited, ashing the

sample at a power of 55 W for two minutes. The sample was

then atomized in a helium plasma at a power of 170 W. The

lead emission at 405.8 nm was measured using a photodiode

array (PDA). A cleaning step was necessary in order to

remove the carbonaceous residue from the left over blood

sample. Cleaning was performed by increasing the power to

200 W and adding a flow of hydrogen gas to the helium

plasma. The cleaning procedure lasted for one minute and

effectively removed all blood residue from the filament.

The filament electrode CMP-AES method was able to meet

two of the criteria set forth by the Centers for Disease

Control (CDC) for a lead in blood screening method. The

detection limit for lead in blood was 7 ppb and the analysis

time was under five minutes. However, the filament

electrode was not accurate for blood lead concentrations

unless matrix effects were reduced by diluting the blood

with deionized water by a factor of approximately one half.

Even with dilution, the method was not sufficiently accurate

for blood lead concentrations below approximately 200 ppb

and the precision did not meet the requirements set forth by

the CDC for lead in blood screening methods.

Using the filament as the electrode had several drawbacks.

The filaments were handmade and so were difficult to make

reproducibly. It was also difficult to deposit the sample

in the filament loop with adhesion to the wire as the only

source of support for the sample. Finally, the lifetime of

the filament electrode was greatly shortened if the

microwave plasma power was raised above a certain point.

Hydride Generation

Hydride generation involved introducing the elemental

analytes to the CMP as a gas. An acidified aqueous solution

of the sample was added to a small volume of 1% sodium

borohydride in a reaction cell. After a certain amount of

time had passed, the resulting hydride of the element was

carried to the CMP by a flow of the plasma gas. Akatsuka

and Atsuya used a CMP with hydride generation to analyze

arsenic in sewage sludge, and iron in steels. They obtained

a detection limit of 0.25 ppb for arsenic in solution [141].

Uchida et al. used the method of hydride generation with a

CMP to analyze inorganic tin [142].

Development of Electrode for Blood Analysis

Cup Holder Electrode

A cup holder electrode (figure 5-1b) was investigated

for the analysis of lead in blood. A cup holder electrode

had several advantages over the filament electrode used by

Wensing et al. for lead blood analysis. Using a cup holder

electrode allowed the introduction of larger sample volumes

and made sample deposition easier and more reproducible.

Also, a cup holder electrode is a more robust electrode

allowing the atomization power to be increased, which could

lead to increased emission intensity.

Initially the electrode material chosen was graphite.

Graphite is a good material because it can sustain very high

temperatures, it is inexpensive, and it is easy to machine

to make modifications to the electrode. The easy

machinability of graphite allowed various parameters of the

electrode (length and penetration) to be optimized before


switching to a metal cup holder. The metal holder would be

more durable but not as easy to machine as graphite.

The graphite electrode contained a hole in the top

which held a nickel cup. A graphite cup would not be

suitable for blood analysis because the graphite could form

refractory compounds which would interfere with the lead

signal. Also, the blood could seep into the graphite

causing memory effects. A cup made out of metal would be a

better choice because there would be less of a chance of

interfering species and memory effects. Initially, nickel

was chosen as the cup material because it did not oxidize

easily, making it very durable. The cup had a sample

capacity of 20 yL.

Several parameters for the cup electrode were studied.

The change in electrode from the filament to the cup

electrode required reoptimizing the conditions of the

plasma. The coupling of the microwave energy and the

stability of the plasma were affected by the length of the

electrode, and the electrode's penetration into the

waveguide. The optimum coupling position of the electrode

was determined by varying the electrode length penetration

into the microwave field to find the parameters where the

minimum microwave power was needed to sustain the plasma.

The first parameter studied was the length of the

electrode. Initially the height of the electrode above the

waveguide was kept constant at 8 mm. This was the height

used in the work with the tungsten filament electrode.

Different lengths of the electrode were then used

determining the range of powers over which a stable plasma

would form. The power supplied to the magnetron was

gradually increased until a plasma would form. The power

was increased until the plasma was no longer stable and then

decreased until the plasma could no longer be sustained.

The lengths of the electrodes ranged from 3.5 cm to 7.0 cm

which corresponded to a depth of penetration into the

waveguide of 2.7 cm to 5.7 cm. This experiment was repeated

keeping the penetration depth of the electrode into the

waveguide constant at 3.2 cm and evaluating different

lengths of the electrodes. As the penetration depth was

increased, the plasma could be maintained at higher powers.

For some positions of the electrode, the maximum power of

700 W could not be achieved because the plasma started to

make a very loud whining noise. The length and penetration

that produced a stable plasma over the widest range of

powers was chosen as the optimum. Figure 5-2 shows some

examples of the data collected for the optimization of

electrode length and penetration.

Using the optimum conditions found for the electrode,

the analysis of aqueous lead solutions was studied. At an

atomization power of 370 W, the lead emission took

S 3.5 cm Electrode

Penetration (cm)

Peneraton (c)
Penetration (cm)

" 4.5 cm Electrode


Penetration (cm)

S 7.0 cm Electrode

0 on

Penetration (cm)

Figure 5-2. Optimization of electrode length and
penetration depth of graphite electrode.


approximately 20 seconds to appear and then remained for 90

seconds for a 10 pL sample of a 50 ppm lead standard. At

such a high concentration and power, the signal should have

been much greater. Tungsten and titanium cups were

evaluated with the graphite cup holder, but these cups also

gave poor results.

The next electrode evaluated was a cup holder made out

of titanium with a nickel cup. Titanium metal instead of

graphite could improve the coupling of the microwaves to the

electrode, improving the efficiency of atomization and

excitation. After atomization with this system, the cup

remained very hot and required a long time to cool down. If

an aqueous sample was deposited in the cup before it was

sufficiently cooled, the sample would vaporize. No

improvement of the lead signal intensity was observed.

Platform Electrode

A cup holder electrode made out of titanium that had

not been drilled to hold a cup was then studied as the

method of sample introduction. This type of electrode was

labelled the platform electrode because the sample was

placed on the flat top portion of the electrode. The top

portion was approximately 6.5 mm long and 6 mm in diameter

and the post was ~ 3 mm in diameter. A 10 pL aqueous lead

sample was placed on the flat top of the electrode, dried

for 90 seconds at approximately 150 W and then atomized with

a plasma power of 300 W. The titanium platform electrode

gave a larger signal than had previously been obtained. The

signal still lasted a long time (figure 5-3a) but not as

long as with the graphite electrode and the nickel cup. The

signal increased (figure 5-3b) when the bulky top part of

the electrode (that was intended to hold the cup) was

removed, yielding a thin metal rod (Figure 5-2c). With this

electrode design, the plasma formed on the same surface that

the sample was on, increasing the interaction of the sample

and the plasma, yielding more efficient atomization and


A significant problem was experienced with the titanium

electrode. After several runs, the lead signal would begin

to decrease in intensity and take longer to appear. It was

necessary to sand off the top of the electrode to regain the

larger signals. Each time the electrode was sanded, a

titanium atomic emission line appeared that interfered with

the lead line being used for measurement. The electrode had

to be tempered by igniting a plasma and then slowly taking

it to higher powers to eliminate the interference before

conducting another run after cleaning. The more the

electrode was used, the sooner in between runs it had to be

cleaned off. The interfering titanium line also limited the

amount of power that could be applied to the plasma. At


0 10 20 30 40
Time (seconds)

Time (s)

Temporal profiles for lead signal for the
platform electrodes: a) platform with cup
holder portion, b) thin titanium rod platform.

Figure 5-3.


powers above 400 W, the interfering line appeared. It is

desirable to have an electrode that could last for an

indefinite period of time and would not have any emission

lines which would limit the powers used. For these reasons,

several other materials were tried for the platform


Tungsten, which had been used for the filament

electrode, has a higher melting point and lower background

emission than titanium and so it was evaluated as the

electrode material. A 10 yL aqueous sample on the tungsten

electrode took approximately 5 minutes to dry at a microwave

power of 150 W. The lead signal, upon atomization, took

several seconds to appear and then lasted for about 30

seconds. The signal was small compared to the signal

obtained using the titanium electrode, even at higher

powers. Nickel was studied next, but it also produced

results similar to those obtained with the tungsten


A titanium electrode with a nickel cap which screwed

into the top was then evaluated (figure 5-2d). This design

was used in an attempt to obtain a similar signal as that

obtained with the titanium electrode, but with a longer

lifetime because of the nickel cap. The signal obtained for

this electrode was similar to that obtained when the whole

electrode was nickel; a low intensity signal was delayed in

appearing and had a long temporal profile.

From the results obtained, the electrode made out of

pure titanium was the best platform electrode even though it

would have to be changed on a regular basis. The titanium

platform electrode lasted approximately 130 firings for

aqueous samples. The titanium platform gave good results

for aqueous lead samples (figure 5-4) achieving a detection

limit of 30 ppb for a 5 pL sample volume. However, the

precision was poor for concentrations of 100 ppb and below.

Analysis of lead in whole blood was performed using

whole blood quality control materials (QCMs). The

analytical curve for these standards (figure 5-5) was linear

(R = 0.997) and produced a detection limit of 50 ppb. After

running an individual analysis, it was necessary to clean

the electrode by scraping off the remaining blood residue.

This was not difficult, but added approximately one minute

to the analysis time. Occasionally, the blood sample would

interfere with the plasma yielding poor precision. The

interference of the blood samples with the plasma might be

due to a problem in depositing the sample. Since the

surface of the electrode was flat, it was difficult to

deposit the sample in the same way each time. The maximum


c) 400



0 100 200 300 400 500 600

Concentration (ppb)

Figure 5-4. Analytical curve for aqueous lead standards on
the titanium platform electrode.






0 100 200 300 400 500 600 700

Concentration (ppb)

Figure 5-5. Analytical curve for whole blood lead standards
on titanium platform electrode.

capacity of the titanium rod platform was 5 pL, and the

sample would sometimes run over the side.

In order to achieve reproducible sample deposition,

electrodes containing a depression in the top were used. It

was found that the size and shape of the depression was very

important in measuring the lead signal. If the depression

was too deep, the signal was small. If the volume of the

blood sample was too large, the blood would interfere with

the formation of the plasma. When blood samples were run

with the depression electrode, the inside of the depression

became dirty and was difficult to clean. Even for shallow

depressions, the lead signal was approximately one half of

the signal obtained by the electrode without the depression.

The electrode with the depression had a short lifetime of

only 30 firings for blood samples.

To reduce the amount of time required to clean the

electrode, a titanium electrode with a cap was used (figure

5-2e). This electrode design allowed one cap to be cleaned

while another cap was being used for analysis. Samples

could also be dried separately, and then placed on the cap

holder to use the microwave plasma for the ashing and

analysis. This shortened the analysis time by ninety

seconds and could be beneficial for the storage and

transport of the blood samples in the clinical setting.

Caps with various diameters and various sized depressions

were used. Some problems were experienced with the

uniformity of the plasma on the titanium cap electrode. At

lower microwave powers, the plasma would sometimes form on

one side of the cap and either stay at that side or flicker

around the edge of the cap. At higher microwave powers,

some of the caps yielded good signal, but it was difficult

to clean them when analyzing blood samples. The caps were

also difficult to reproducibly construct. Caps with the

same design did not produce the same signal.

Suspension Method

The results with the platform electrode demonstrated

that it is necessary for the plasma to interact directly

with the blood sample. Anytime the blood sample was below

the plasma in some sort of depression, the signal was

drastically reduced. However, when the blood sample was on

the surface of the electrode where the plasma formed, the

blood would often interfere with the stability of the

plasma. A method for which the sample was suspended above

the electrode was used to try to account for keeping the

sample in the plasma without being on the surface where the

plasma forms. A titanium rod electrode was used to support

the plasma and a macor holder was used to support a screen

or a wire mesh above the electrode (figure 5-6). Initially

stainless steel screens (10-20 and 40 mesh) were used.

Top View

Macor Holdr Tungsten Mesh
Macor Holder

Titanium Electrode

Quartz Torch

Helium gas

Figure 5-6. Suspension method of sample introduction.

Platinum screens (40 mesh), tungsten screens (20 and 40

mesh), and a four-squared cross made out of tungsten wire

were also tried.

The sample would not dry with microwave power alone, so

a very low power plasma was ignited below the screen. For

blood samples, the drying caused some problems because if

the plasma was too close to the sample or too high in power,

it would cause the sample to bubble and spatter. The

titanium electrode was changed to a pointed tungsten

electrode which could sustain a very low power plasma which

dried the blood more effectively. The power of the plasma

was increased for ashing, and then further increased for

atomization. This method worked well for aqueous standards

and greatly decreased the background during atomization

because the macor holder shielded most of the emission from

the plasma. However, this method of sample introduction did

not work well for blood samples. The mesh became very

brittle at higher plasma powers and broke very easily.

Spiral Filament Electrode

Each method of sample introduction tried had various

advantages to it. The cup holder electrode held the sample

the best, the filament electrode was easiest to clean, the

platform electrode gave the best signal, and the macor

holder resulted in the lowest background. Various features

of several of these methods were combined to design an

improved filament electrode. A thicker (0.5 mm in diameter)

tungsten wire was used which was more durable and could

sustain higher powers than the original filament (0.25 mm in

diameter) could. Initially a single loop was made at the

top of the wire to hold the sample, but it was difficult to

deposit the sample in the loop. A two and a half turn

spiral was then made at the top of the electrode The spiral

served as sort of a platform and held a 2 pL blood sample

very well. The spiral filament electrode (figure 5-7)

performed well for both aqueous and blood samples (chapter

6) and was easy to clean. The filament was, however,

difficult to make because the tungsten wire was very brittle

and would often split during the construction of the spiral.

An attempt was made to use commercial tungsten light

bulb filaments to hold the sample. A 20 turn, rectangular

light bulb filament was placed over the loop of a filament

electrode. This method would remove the necessity of having

the spiral at the top of the electrode and could also help

make the electrodes more reproducible. The method worked

well for aqueous samples (figure 5-8) giving a detection

limit of 44 ppb, but was very hard to clean after the

analysis of blood samples. It also eroded quickly under the

high plasma powers used to clean the blood from the


~3 mm

65 mm
Wire Diameter = 0.5 mm

Figure 5-7. Tungsten spiral filament.

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EQBYJZLR1_7U40W7 INGEST_TIME 2014-04-18T23:39:07Z PACKAGE AA00014247_00001