Determination of wear metals in lubricating oil by electrothermal vaporization inductively coupled plasma mass spectrometry

MISSING IMAGE

Material Information

Title:
Determination of wear metals in lubricating oil by electrothermal vaporization inductively coupled plasma mass spectrometry
Physical Description:
xi, 175 leaves : ill. ; 29 cm.
Language:
English
Creator:
Escobar, Monica Paola, 1968-
Publication Date:

Subjects

Subjects / Keywords:
Inductively coupled plasma mass spectrometry   ( lcsh )
Lubricating oils -- Analysis   ( lcsh )
Chemistry thesis, Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 167-174).
Statement of Responsibility:
by Monica Paola Escobar.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002045654
oclc - 33399206
notis - AKN3583
System ID:
AA00009022:00001


This item is only available as the following downloads:


Full Text









DETERMINATION OF WEAR METALS IN LUBRICATING OIL
BY ELECTROTHERMAL VAPORIZATION
INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY










By

MONICA PAOLA ESCOBAR


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1995












ACKNOWLEDGEMENTS


I would like to sincerely thank my research advisor, Dr. Jim Winefordner, for

his guidance and support. I would also like to thank Dr. Ben Smith for his assistance,

friendship, and encouragement and the Winefordner research group for their support.

I am grateful to the service engineers at Finnigan MAT, George Henderson, Tom

Musselman, and Frank Trensch, for their assistance.

I would like to thank everyone who has made this a special and enjoyable

experience: Nikki Schultz, Nancy Mullins, Sue Ohorodnik, Andi Pless, Lena Yacoub,

Laura Cole, Denise Imbroisi, Stefanie Pagano, Taylor Hale, Eugene Wagner, and

Sherrie Hall.

I am grateful to my family for their support and understanding. I am especially

grateful to Bill Maehlmann for his patience and love and for being there whenever I

needed him. This last year would have been much harder without his constant support

and encouragement.

Finally, I would like to thank the Air Force Office of Scientific Research for

supporting my research.













TABLE OF CONTENTS


Page

ACKNOW LEDGMENTS ......................................................................................... ii

LIST OF TABLES ................................................................................................ v

LIST OF FIGURES .................................................................................................. vi

ABSTRACT ........................................................................................................... x

CHAPTERS

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

2 INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY
BACKGROUND .......................................................................... 5

History .................................................................................. 5
Recent Developments .......................................................... 7
Inductively Coupled Plasma ................................................ 10
ICP-M S Interface .................................................................. 18
Mass Spectrometer .............................................................. 27
Analytical Characteristics of ICP-MS ................................ 31
Sample Introduction Techniques ......................................... 37
Electrothermal Vaporization ................................................ 40

3 INSTRUMENTATION ................................................................... 61

Inductively Coupled Plasma Mass Spectrometer ............... 61
Electrothermal Vaporizer .................................................... 66







4 DETERMINATION OF METALS IN AQUEOUS
SOLU TIONS .................................................................................. 71

Instrum ent M modifications .................................................... 71
Experim ental ........................................................................... 77
Results and Discussion ........................................................ 84
Conclusions ............................................................................ 119

5 DETERMINATION OF METALLO-ORGANICS IN BASE
OIL .................................................................................................. 125

Instrum ent M modifications .................................................... 125
Experim ental ...........................................................................128
Results and Discussion ......................................................... 131
Conclusions ........................................................................... 161

6 CONCLU SION S ................................................................................163

Future Studies ................................................................................. 165

REFERENCES ..........................................................................................................167

BIOG RA PH ICA L SKETCH .................................................................................. 175












LIST OF TABLES


Table Page

1 Dominant species in the ICP ..................................................................... 16

2 Interfering ions in nebulization of 5% HNO3 ........................................... 34

3 Interfering ions in organic solution nebulization ...................................... 35

4 Calculated temperatures of the vapor-gas mixture required to obtain
S=10 for species generated under ETV conditions ................................... 52

5 Ion lens voltage ranges ............................................................................... 65

6 New graphite tube conditioning ................................................................. 82

7 ETV temperature ramp used for aqueous solutions .................................. 89

8 Summary of modifier performance .............................................................123

9 ETV temperature program for metallo-organic aluminum .......................... 134

10 Two-step atomization ETV temperature ramp ...........................................136

11 Results of NIST and Conostan check standards for the determination
of m etallo-organic A l .................................................................................. 146

12 Optimum ETV ramp for Fe ........................................................................ 148

13 Results of NIST and Conostan check standards for the determination
of m etallo-organic Fe .................................................................................. 151

14 Modified ETV ramp for Mg .......................................................................153

15 Optimum ETV ramp for Mg .................................................................... 155

16 Summary of metallo-organic Al, Mg, Fe, and Y results ...........................162













LIST OF FIGURES


Figure Page

1 Inductively coupled plasma. H designates the magnetic field................ 13

2 ICP torch...................................................................................................... 14

3 Calculated values for degree of ionization (%) for M' at T=7500 K,
ne=l xlO 5 cm-3 ................................................................................................. 19

4 Inductively coupled plasma mass spectrometer interface........................ 20

5 Supersonic jet............................................................................................... 24

6 Sampling interface....................................................................................... 26

7 Ion transfer optics........................................................................................ 28

8 Nebulizer, spray chamber and ICP torch.................................................... 39

9 M odified commercial graphite furnace....................................................... 42

10 Typical ETV temperature ramp................................................................... 44

11 Dome-type ETV............................................................................................ 46

12 SOLA ICP-M S vacuum system .................................................................. 63

13 ETV workhead............................................................................................. 67

14 Gas flows for ETV-ICP-M S........................................................................ 69

15 Sample introduction for ETV...................................................................... 70

16 ETV-tubing interface................................................................................... 74

17 Diagram of ETV-tubing interface and pinch valve.................................... 76

vi







18 Sample introduction glassware for ETV..................................................... 78

19 Calibration curve for Mg without carrier added......................................... 85

20 Comparison of peak intensity for ICP-MS and ICP-AES simultaneous
analyses for 27 consecutive injections of 1.5 ng Mg................................. 87

21 Effect of NaCI added to Mg. A) 0.02 ng Mg. B) 0.1 ng Mg.
C) 0.25 ng Mg. D) 0.5 ng Mg. E) 2 ng Mg. F) 4 ng Mg..................... 91

22 Mg temporal peak shapes (0.01 ng Mg). A) Without Carrier.
B) 4 pg N aC l added.................................................................................... 94

23 Calibration curves for Mg. A) 0.004 pg NaCl added. B) 0.04 ipg
NaCI added. C) 0.4 pig NaCl added. D) 4 pg NaCl added.
E) 10 pg N aCI added.................................................................................. 95

24 Effect of NaCl added to Al. A) 0.02 ng Al. B) 0.1 ng Al. C) 0.5 ng Al.
D ) 4 ng A l............................................................................................ ....... 97

25 Calibration curves for A l. ........................................................................... 100

26 Effect of NaCl added to Fe. A) 0.1 ng Fe. B) 0.25 ng Fe. C) 0.5 ng Fe.
D ) 2 ng F e..................................................................................................... 102

27 Calibration curves for Fe. A) No carrier added. B) 0.004 pg NaCl
added. C) 0.04 pg NaCl added. D) 0.4 pg NaCI added. E) 4 pg NaCI
added. F) 10 pg added................................................................................ 105

28 Effect of Pd(N03)2 added to Mg. A) 0.02 ng Mg. B) 0.1 ng Mg.
C) 0.5 ng M g. D) 4 ng M g......................................................................... 108

29 Calibration curves for Mg. A) No carrier added. B) 0.6 pig Pd(N03)2
added. C) 2 pig Pd(NO3)2 added. D) 12 pg Pd(N03)2, added. .................110

30 Effect of adding Pd(NO3)2 to Al. A) 0.02 ng Al. B) 0.25 ng Al.
C) 0.5 ng Al. D) 2 ng Al. E) 4 ng Al..................................................... Ill

31 Calibration curves for Al. A) No carrier added. B) 0.6 pig Pd(N03)2
added. C) 2 pg Pd(N03)2 added. D) 12 pg Pd(N03)2 added. .................115

32 Effect of Pd(N03)2 on Fe. A) 0.02 ng Fe. B) 0.1 ng Fe. C) 0.5 ng Fe.
D ) 4 ng Fe....................................................................................................1... 16







33 Calibration curves for Fe. A) No carrier added. B) 0.6 pig Pd(NO3),
added. C) 2 pg Pd(N03)2 added. D) 12 pg Pd(NO3), added. ..................118

34 Effect of SF6 addition on Y. A) 0.1 ng Y. B) 0.5 ng Y. C) 4 ng Y........120

35 Calibration curves for Y. A) 10 mL/min SF6. B) 20 mL/min SF6.
C) 30 mL/min SF6. D) 40 mL/min SF6......................................................122

36 Pneum atic plug holder.................................................................................. 126

37 Carrier flow through ETV. A) During drying and ashing steps.
B) During vaporization step......................................................................... 129

38 Effect of NaCI on Al metallo-organic standards. A) 1.29 ng Al.
B ) 3.30 ng ................................................................................................... 132

39 Temporal metallo-organic Al signal. A) From initial temperature
program of 50 ng. B) From two-step vaporization program of 52 ng........135

40 Analytical curve of metallo-organic Al using two-step vaporization
E T V ram p ...................................................................................................... 137

41 Effect of 20 C rest time on signal intensity. 0.525 ng metallo-organic
A l............................................................................................................ ...139

42 Effect of different sample volumes on metallo-organic Mg signal
intensity. 300 ppb M g................................................................................ 141

43 Calibration curve for metallo-organic Al. A) Peak intensity. B) Peak
area. Log-log plot. C) Peak intensity. D) Peak area.................................142

44 Calibration curve for metallo-organic Al, showing check standards..........145

45 Calibration curve for metallo-organic Fe in base oil. A) Calibration
curve. B) Log-log plot................................................................................. 149

46 Calibration curve for metallo-organic Fe, showing check standards..........150

47 Calibration curve for metallo-organic Mg in base oil. A) Analytical
curve. B) Log-log plot................................................................................. 156

48 Temporal signal of 3.5 ng Y. A) Before graphite tube conditioning
with SF6 B) After conditioning...................................................................157







49 Calibration curve for metallo-organic Y. A) Analytical curve.
B) Log-log plot............................................................................................. 159

50 Temporal profile of metallo-organic Y signal. A) Y blank.
B) 0.6 ng Y.................................................................................................. 160












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

DETERMINATION OF WEAR METALS IN LUBRICATING OIL BY
ELECTROTHERMAL VAPORIZATION INDUCTIVELY COUPLED PLASMA
MASS SPECTROMETRY

By

Monica Paola Escobar

May 1995

Chairperson: James D. Winefordner
Major Department: Chemistry


The inductively coupled plasma mass spectrometer (ICP-MS) is one of the few

techniques capable of multielement determinations with low detection limits. ICP-MS

is an almost ideal technique for the analysis of lubricating oils due to its many

advantages, which include broad linear dynamic range, multielemental analysis, rapid

acquisition of mass spectra, and low detection limits. However, there is a serious

problem when an organic matrix is nebulized into the ICP-MS. Carbon condenses on

the cooled interface cones, sooting and blocking the apertures. This is prevented by

adding a small amount of oxygen (2 10%) to the carrier gas so that all the carbon

is oxidized and volatilized. Too little oxygen will result in clogging of the interface

and too much will result in rapid deterioration of the orifice. A second disadvantage

of nebulizing organic into the ICP-MS is that carbide compounds formed in the ICP







may act as molecular interference for some elements, making it difficult or impossible

to detect these elements.

Electrothermal vaporization (ETV) was chosen as a sample introduction

technique because it eliminates several of the problems associated with conventional

nebulization. When using ETV with ICP-MS, the graphite furnace vaporizes the

sample, and it is then transported to the ICP where it is ionized. Each step can then

be optimized separately, adding several advantages to the technique: analysis of small

sample volumes, high sample transport efficiency, even higher sensitivity, and

separation of the analyte from the sample matrix, which leads to the ability to analyze

samples in organic, high acid, and high solid matrices. This makes it possible to

analyze oil samples while eliminating oxygen addition and reducing molecular

interference due to the organic matrix.

This dissertation describes the development of a method for the determination

of wear metals in lubricating oils by electrothermal vaporization inductively coupled

plasma mass spectrometry. Hardware modifications were made to the ETV to

improve transport efficiency to the ICP. These included the addition of a graphite

plug holder, a new ETV-tubing interface, a new sample transfer valve, and a

modification to the sample introduction glassware. The technique was then

characterized using aqueous solutions of the analytes of interest with sodium chloride

and palladium nitrate added as sample carriers. Finally, a method was developed for

the detection of Al, Mg, Fe, and Y metallo-organic compounds in an oil matrix.












CHAPTER 1
INTRODUCTION


There is a strong interest in replacing metallic ball bearings with those made

of ceramic, or silicon nitride (Si3N4), for a wide variety of applications. The

determination of wear-related trace elements in the lubricating fluids used with these

bearings is important in the development of models concerning the mechanisms of

wear for these materials. The routine determination of wear debris may also serve as

an indication for optimization of engine maintenance procedures.

Bearings consist of two rings spaced apart by a set of balls which are separated

by a cage or retainer, and are often lubricated. Conventionally, M50 steel is used to

make bearings; however advanced ceramics have recently been studied as materials

for bearing elements. Silicon nitride ball bearings have several advantages over steel

ball bearings currently being used in engines (1-6) Si3N4 bearings are lightweight,

chemically inert, and exhibit reduced wear. The high strength of the Si-N covalent

bond gives silicon nitride a high modulus of elasticity, hardness, and resistance to high

temperatures (550 800 oC). Ceramic ball bearings are electrically resistive and

nonmagnetic; therefore, damage from electrical arcing is eliminated and they can be

used at reduced temperatures in magnetic fields. It has been shown that they can have

an increased fatigue lifetime of fifty times over M50 bearing steel (6). Because of







2

their high performance capabilities, ceramic ball bearings have many potential

applications (6), including high-speed engines, space products, guidance systems, and

machine tool spindles. They may be used in applications that require dry operation

because of product contamination, such as chemical processing, semiconductor

manufacturing, and high vacuum systems. The disadvantages of using silicon nitride

as a bearing material include the high cost of fabrication and the variability in

manufacturing. These two problems should improve with continued research into the

manufacturing process.

Because the use of silicon nitride as a bearing material is relatively new, there

has not been much research studying its wear mechanisms (1). The wear mechanisms

of ball bearings must be well understood before they can be put into widespread use

to ensure catastrophic failure will not occur. One method used to study wear

mechanisms is to analyze the lubrication fluid for wear debris (7). The composition

of wear debris provides insight into how the bearings are wearing as well as

information about other engine components. The routine testing of lubricants used in

engines is also important for diagnostic purposes, including aiding in preventive

maintenance by indicating when oil changes, component replacement, and engine

overhauls are required. The wear metal debris can be metallic (from mechanical

wear), oxidized (from oxidative corrosion), or metallo-organic (from chemical

corrosion). Therefore it is important to use an analysis technique that will detect

dissolved species to mm-sized particles.







3

Silicon nitride bearings are manufactured by sintering or hot isostatically

pressing of the Si3N4 powder (8). Sintering has been the most widely used technique.

Because silicon nitride is difficult to sinter, sintering aids are added to the powder

mixture (9). These include mixtures of yttrium oxide (YO3), spinel (MgAl2O4),

aluminum oxide (A1203), aluminum nitride (AIN), or magnesium oxide (MgO). The

powder is molded into shape, then sinterred in a nitrogen or inert gas atmosphere at

1400 1650 C. A liquid phase is formed between the silicon nitride powder and the

sintering aids. The powder, or a-phase, dissolves in the liquid, then precipitates as

a fine crystalline phase, or P-phase. Crystal growth continues and contraction and

densification occur. As the temperature increases, crystal grain size increases. The

result is interlocking fiber-like P-grains with a residual secondary phase at the grain

boundaries. The resulting material is 89.0 93.0 % by weight Si3N4 powder and 7.0 -

11.0 % by weight sintering aids. The proposed mechanism of wear focuses on the

grain boundaries because they are the weakest part of the bearing (10). Therefore, this

dissertation describes the development of a method for the determination of Mg, Al,

Y, and Fe in lubricating oils by electrothermal vaporization (ETV) inductively coupled

plasma mass spectrometry (ICP-MS). The detection of Fe is important because it is

present in other engine components.

The inductively coupled plasma mass spectrometer is one of the few techniques

capable of multielement determinations with low detection limits. Electrothermal

vaporization was chosen as the sample introduction technique because it eliminates

several of the problems associated with conventional nebulization, which are discussed







4
in Chapter 2. However, the ETV contained several design flaws that affected its

performance. Therefore, modifications were made to its hardware to improve transport

efficiency to the ICP. The technique was then characterized using aqueous solutions

of the analytes of interest and these results are shown in Chapter 4. Finally, a method

was developed for the detection of Al, Mg, Fe, and Y metallo-organic compounds in

an oil matrix, discussed in Chapter 5.












CHAPTER 2
ICP-MS BACKGROUND


History


A detailed history of ICP-MS may be found in several reviews (11-14); only

a brief discussion is given here. In the 1970s it became apparent that inductively

coupled plasma atomic emission spectroscopy (ICP-AES) was inadequate for trace

element detection in geological samples. Because geological samples have high

matrix concentrations, even weak emission lines from matrix components can interfere

with elements of interest. This prompted the search for alternative methods of multi-

element trace analysis. Spark source mass spectrometry was already is use for the

determination of metals in geological samples, but the technique was inadequate for

trace analysis, and only a few samples could be analyzed per day.

Moruzzi, at the University of Liverpool, conducted a feasibility study under the

direction of Gray, using a capillary DC arc source mass spectrometer (15,16). The gas

temperature of 3500 K was too low for efficient ionization of some elements, it

experienced severe matrix effects, and produced nonlinear analytical curves. The

study did demonstrate that mass spectra could be produced from an atmospheric

pressure ionization source, although a hotter plasma was necessary.







6

The publication of the DC plasma results attracted attention in Canada and the

United States. Douglas and French, at the University of Toronto, started development

of a microwave induced plasma MS, which was later changed to an ICP-MS. Fassel

and Houk, at the Ames Laboratory at Iowa State University, started development of

an ICP source mass spectrometer. This latter group developed a close collaboration

with Gray's group in Surrey and obtained the first spectra of analyte ions produced

from an ICP source mass spectrometer (17). The main problem with this instrument

was in sampling ions from the bulk plasma. The apertures used for ion extraction

were too small and a cool boundary layer formed between the plasma and the interface

cone where molecular species formed. In 1980, by using larger diameter apertures,

continuum flow was first extracted from the bulk plasma, which led to improved

analytical performance (18).

During this time, Douglas and coworkers had developed the center-tapped load

coil design, which solved problems caused by the plasma offset potential (19). This

led to the release of the first commercial ICP-MS, the Sciex ELAN, in 1983 at the

Pittsburgh Conference. Soon after, the VG Plasma Quad, based on the instrument

developed in Surrey, was introduced in 1983 at the ASMS Meeting. Since that time,

the use of ICP-MS has greatly increased, and several different manufacturers now

market instruments.









Recent Developments


Now that ICP quadrupole MS instruments are in routine use, research has

begun on coupling different types of mass analyzers to the inductively coupled plasma.

The quadrupole mass analyzer has several limitations, including low mass resolution

and relatively slow mass scanning. In an attempt to overcome these disadvantages,

different mass analyzers have been used.

Finnigan MAT (20) recently marketed a high resolution magnetic sector ICP-

MS. The mass analyzer consists of a high resolution magnetic sector of reversed Nier-

Johnson geometry. Several scan modes are available, including magnetic, high

voltage, or their patented synchro-scan mode, in which counteracting and simultaneous

electric and magnetic scans are possible. The resolution is variable, with a maximum

resolving power of 7000. Since a resolution of 2700 is sufficient to separate spectral

interference associated with ICP-MS, this instrument allows for the detection of

elements previously impossible to detect with a quadrupole instrument, such as Fe, and

V. With the instrument in the low resolution mode (300), detection limits are in the

10 pg/ml range. The authors were able to detect V, Fe, Cu, Zn, and Ag in human

serum.

Myers and coworkers (21) constructed an ICP time-of-flight mass spectrometer.

An orthogonal TOFMS was used as the mass analyzer, with argon ions deflected away

from the detector to reduce detector dead time and space-charge effects. The

advantage of using a TOFMS is the ability to produce a spectrum of all ions sampled







8

within 30 to 100 Is. The orthogonal TOFMS was chosen because the velocity of the

ions in the flight tube are one to two orders of magnitude faster than those exiting the

sampling interface. This allows the extraction region to slowly fill with ions while a

previously sampled pulse is mass analyzed. When operated in the nonreflecting mode,

the resolving power of the instrument is 500. The sampling duty cycle of the mass

analyzer is 3%, limiting the sensitivity of the instrument, and the transmission

efficiency was approximately 20%. No analytical figures of merit were given.

Koppenaal et al. (22) designed, built, and tested an ICP ion trap mass

spectrometer (ITMS). The advantages of using an ion trap with an ICP source are the

same as when used with organic MS. These include its high sensitivity and resolution,

and its ability to select, store, and manipulate ions for dissociation and mass analysis.

The authors used an ICP-linear quadrupole-ITMS configuration, in which the

quadrupole could be used in the resolving (r.f./d.c.), non-resolving (r.f. only), or notch

(r.f. + r.f.ax) filtering modes. The ion trap was able to eliminate molecular

interference by collisions with He bath atoms in the ion trap chamber. The ion trap

almost completely neutralized Ar+ and two mechanisms were proposed for the

reaction. The LODs were in the 10 500 pg/ml range, and were expected to improve

upon optimization of instrument parameters. The same group is also currently

developing an ion cyclotron resonance mass spectrometer to couple to the ICP. This

will greatly enhance the resolving power of ICP-MS.

Other ionization sources have also been studied in an attempt to reduce spectral

interference, increase the versatility of the instrument and increase the ionization of







9

nonmetal species. Microwave plasma, as well as low pressure plasma mass

spectrometers have been constructed.

Evans and coworkers (23) modified a VG PlasmaQuad 2 instrument so that a

low pressure ICP could be formed in place of the atmospheric pressure ICP. The

advantage of this low pressure ICP is that it may be used for the ionization of

molecular or atomic species. A vacuum seal was formed between the low pressure

plasma torch and the sampling cone and the vacuum pumping rate in the interface

region was also increased. The He flow from a GC was directed into the center of a

special torch with argon as the plasma gas. By adjusting the forward power and the

plasma gas flow, the degree of fragmentation of the analyte was changed. The authors

were able to detect organotin, organolead, organoiron, and organohalide compounds

with detection limits ranging from 13 15 pg.

Fechner and Nagengast (24) modified the plasma interface and torch on a VG

PlasmaQuad 2 ICP-MS for use as a microwave induced plasma (MIP) MS. The

plasma was sustained in 3 to 5 L/min He gas flow with 380 W of power. An

advantage of using He over Ar is that helium has a higher ionization potential, making

it possible to detect nonmetals which are not well ionized in an Ar ICP. A lower

background spectrum was observed, but the polyatomic ion formation increased due

to the lower power applied to the MIP. Limits of detection for As, Se, Br, and I were

between 0.05 and 1.1 p.g/L, which is comparable to LODs found with ICP-MS.

Biological and fish reference samples were analyzed.







10

Wu (25) and coworkers constructed an He microwave plasma torch MS. The

advantages of this system are the same as for other He plasmas, including high

electron temperatures and moderate electron density. The 150 W atmospheric pressure

microwave plasma had a He flow rate of 4.0 L/min and produced fewer background

peaks than those found in ICP-MS. The limits of detection for F, Cl, Br, 1, S, C, and

P were 12 ng/ml to 1.0 jig/ml. These are comparable to or better than LODs found

for an ICP ion source on the same instrument.


Inductively Coupled Plasma


The inductively coupled plasma was first studied as an excitation source for

atomic emission spectroscopy (AES) in the 1960s, independently, by the research

groups of Greenfield and Fassel (26,27). However, at this time atomic absorption

spectroscopy (AAS) was in widespread use, and ten years passed before the ICP

gained acceptance as an analytical tool. During this time, technological advances were

made that improved detection limits with ICP-AES, and the first commercial ICP was

introduced in 1974 (28). Since this time the use of the ICP as an excitation source for

AES has grown tremendously (29,30) and in the 1980s it gained use as an ion source

for elemental mass spectrometry (17,31).

The use of the ICP as a source for mass spectrometry has several advantages.

Because the samples are introduced at atmospheric pressure, sample changeover is

simple and rapid. The 7,000 K temperature of the ICP ionizes most elements easily







11
and produces primarily singly charged ions. It efficiently dissociates sample

components and produces few molecular fragments.

There are disadvantages associated with the ICP ion source. The high gas

temperature and pressure of the plasma requires an interface that is fairly inefficient,

with only 1% of the ions that pass through the sampling cone passing through the

skimmer. Molecular ion formation in the plasma, especially from atmospheric gases

reacting with argon, is a primary disadvantage of ICP-MS. Molecular ions cause

isobaric interference over certain areas of the mass range and make it difficult to

detect trace levels of some elements.

A thorough review of ICP, including instrumentation, fundamental properties,

and applications may be found in Montaser and Golightly (30) and other references

(29). An ICP is an electrodeless discharge at atmospheric pressure maintained by

energy coupled to it from a radio frequency generator. A copper coil acts as the

primary of the RF transformer and the plasma is the secondary. Argon is normally

the main support gas, although other gases, such as helium, xenon, and nitrogen have

been used as additives (32-36) and helium has been used as the main support gas as

well (37-39). The plasma is generated in a quartz torch similar to those used in ICP

atomic emission spectroscopy (AES) with usually the only modification being the

horizontal instead of vertical mounting. It is surrounded by the load coil, which is

connected to an RF generator.

The RF current travelling through the coil produces a magnetic field in the

plasma which varies in time at the generator frequency of 27 or 40 MHz. Eddy







12

currents are induced within the gas which flow in circular paths that run concentrically

inside the torch. The magnetic lines of force are oriented axially inside the quartz

tube and follow elliptical paths outside the coil (Figure 1). The magnetic field is

coupled to the plasma gas containing ions and electrons. The charged particles meet

resistance to flow, resulting in Joule heating. These charged particles are accelerated

and collide inelastically with gas atoms. At atmospheric pressure, the mean free path

of the particles is approximately 10' mm, resulting in a high rate of collisions which

quickly heats the plasma.

Because an inert gas is a poor electrical conductor, the plasma is initiated by

"seeding" the argon gas with electrons from a Tesla coil. Electrons and argon ions are

formed which absorb energy from the alternating magnetic field. Collisions generate

an avalanche of charged particles that absorb the energy from the RF field and once

the electrons reach the ionization potential of the support gas, further ionization takes

place forming a stable, self-sustaining plasma.

A typical ICP torch is shown in Figure 2. It consists of an outer tube of inner

diameter of about 18 mm, which contains two concentric quartz tubes of 13 and 1.5

mm inner diameter. The two outer tubes have side arms supplying gas entering

tangentially to create a vorticular flow. The center tube carries the sample gas flow

and runs down the axis of the torch. The outer gas flow, or the coolant flow, is at 10

- 15 L/min, is the main support gas and also protects the quartz outer wall from

melting. The copper load coil, which surrounds the plasma torch, has 2 4 turns and





















H H










Argon
Coolant Flow

Argon
Intermediate Flow


Sample Inlet


Figure 1. Inductively coupled plasma. H designates the magnetic field.
















0
0


14














.-
Q
3'


O
0
0
0

~0
~0



8
U






15

may be gas or water cooled. The bottom turn of the coil is located 3 mm from the

top of the injector tube and the top turn of the coil is 13 mm from the sampling cone.

The coolant flow experiences a 'skin effect' because most of the energy from

the coil is coupled to this outer region. The temperature in this induction region is

10,000 K. The next tube contains the intermediate, or auxiliary gas flow, and keeps

the plasma away from the tip of the central quartz tube. The auxiliary flow rate is

typically 0-1.5 L/min. The central gas flow, or injector, nebulizer, or carrier gas flow,

transports the analyte from the sample introduction technique, and is normally about

1 L/min. Because of the small diameter of the injector tube, a high velocity gas jet

is formed in the injector tube which punches a cool channel through the center of the

plasma. The gas in the center channel is heated by radiation and conduction from the

induction region to a temperature of 5000 to 7000 K. The central channel is

physically distinct from the outer region of the plasma; therefore, sample composition

can vary without greatly affecting the electrical processes in the plasma.

The ICP contains many different species that interact with each other physically

and chemically. Table 1 is a list of the major species found in the ICP discharge (40).

A thorough understanding of all the plasma mechanisms has not yet been achieved;

the majority of fundamental research has focused on analyte excitation mechanisms.

Several different analyte excitation mechanisms have been proposed and have been

reviewed by Blades et al. (41). The most significant collisional and radiative













Table 1
Dominant species in the ICP


From Ar gas:
Argon atom: Ar
Argon ion: Ar'
Argon dimers: Ar2, Ar,
Argon molecules: ArO, ArX*
Free electrons: e

From solvent:
Aqueous solvent:
Hydroxyl radical: OH
Atomized solvent species: 0, H
Organic solvent:
Di-carbon and hydroxyl radical: C2, OH
Atomized solvent species: C, H, 0, S, etc.

From analyte and concomitant elements:
Analyte atom: M
Analyte ion: M*
Analyte molecular species: MO, MO*
Concomitant atoms, ions, and molecular species: X*, X'*, XO, etc.

From entrained air:
Air-N2, N2,, 02, O2, 0, N, CO, CO,, NO


*X represents any element
(40)







17

processes are listed below. The ground state analyte atom (M,) and ion (Mi,) may

undergo collisional excitation (-) by electrons (e-):

M, + e (high energy) = Mp + e- (low energy)

Mi+ + e (high energy) Mp+ + e (low energy)where Mp is the excited analyte

atom, and Myp is the excited and ionized analyte atom. Collisional ionization (-) and

three-body recombination (-) by electrons also occur:

Mi + e- Mp+ + 2e-

Excited and ionized analyte atoms may also be produced from charge transfer of the

ground state analyte with an argon ion:

Ar' + M, Mp + Ar + AE

Ar' + Mi, Mp + Ar + AE

The de-excitation of species may also involve radiative decay with photon emission

of energy hvlm,:

MP Mi + hvime

MP+ Mi+ + hvu,

Radiative recombination of the ionized analyte with an electron may occur, producing

background emission:

M+ + e Mp + hvcont

Penning ionization with excited argon (Ar*) may also occur:

Ar* + M, Mp+ + Ar + AE

Of overall importance in ICP-MS is the effective production of singly-charged

ions. The efficiency of the ICP in this regard varies among the elements, generally







18
in proportion to the ionization potential. Most elements are highly ionized (>90%),

which makes the ICP an almost ideal source for elemental mass spectrometry. The

degree of ionization is normally low for nonmetals, to near 100% for the alkali metals

(42). Figure 3 lists calculated values for the degree of ionization for most elements

of the periodic table, estimated using the Saha equation (43).


Inductively Coupled Plasma Mass Spectrometer Interface


Each instrument manufacturer has their own interface design, but most

instruments have a sampling and skimmer cone, and their functions are the same for

each interface. The instrument used in this research was the Finnigan MAT SOLA

ICP-MS. The difference between this instrument and others is discussed below.

Once the sample is ionized, the ions flow through a differentially pumped ion

extraction interface (Figure 4). Ions first pass through a sampling cone made of Ni

or Pt with an orifice diameter of 1.1 mm. While the plasma is at atmospheric

pressure, the area behind the sampling cone is kept at 2 3 torr by a 18 m3 per hour

single stage rotary pump, and this pressure is low enough for a supersonic jet to form

in this region. As the plasma interacts with the sampler, two processes take place(l 1).

First, the plasma is cooled and deflected. A boundary layer of gas forms between the

sampler and the plasma. The temperature of the boundary layer is inbetween that of

the plasma and the sampler and molecular species, such as oxides, readily form in this

region.











F -) C K4 3 00
E ^< ^ X6(


1!
0%


-U0'


en C-4 0 N
o0 C6N 4\ O 0





00 00 o O
me n a xQoo -ar

o\ as or -^ o


n I I l U
Cuorjo


N r-


3 0



Zo%




co o
bi oi





0 0


U (,
'- 00
QO\0


COm
3 tn







u 00






0
F-
0 00










Z O
00


a.,


po00



3-


0-


U 00
0 r


I
E-~o~ NojZo~


W.-
0% *
ON


Cu 0%
0%


0











0




-I -

0
*0


>0

Q 2


z


-o



Z


















u
co
E
U








0z

z.



Zu


19


I.-





.l








Eo
a-








1 .5





(U
~.






0

II











SCu





L.








0 0









I-





20













>-- |
II1I 1 I I
I8~






I "r"



I */I
E
E
E

_ _







21

The second interaction between the plasma and the sampler is electrical. The

plasma contains an equal number of positive ions and electrons, making it electrically

neutral. The metal sampling cone attracts electrons as well as positive ions. Because

the mobility of electrons is higher than that of ions, a negative charge collects on the

surface of the sampler. The sheath region that forms around the sampler is depleted

of electrons and contains an excess of positive ions. Because the sampler is grounded,

the plasma seems to float at a positive potential.

In addition to inductive coupling between the load coil and the torch, the load

coil is also electrostatically coupled through the torch wall by the capacitance between

the two. The capacitive coupling is dependent on the arrangement of electrical

connections to the coil. Conventionally one end of the coil is grounded and one end

is connected to the high voltage RF source. This results in a potential gradient along

the coil, except at the moment when the field polarity reverses.

When the plasma contacts the sampler and partially enters the orifice, the

plasma and cone are coupled through the thin sheath. The resistance of the sheath is

less than that of the capacitative impedance between the cone and plasma, and the

plasma's RF potential is determined by the ratio of the two impedances. As RF

current travels through the interface, the plasma flow to the sampler is modified by the

different mobilities of the positive ions and electrons in the sheath region. During

negative half cycles the current is carried by electrons, which move faster than positive

ions that carry the current during positive half cycles. This results in the plasma

carrying a net mean positive DC potential which may be larger than the floating







22

potential due to the sheath alone. Typically, the plasma may float at a positive

potential of 40 V from the grounded sampling cone (12).

When the plasma potential is high, there is an electrical discharge between the

plasma and the sampling orifice, called the secondary discharge. It is observed as a

bright white emission from the gas flowing into the orifice. This secondary discharge

is undesirable because it erodes the sampling orifice, produces multiply charged ions,

and induces high kinetic energies and a large spread of kinetic energy in the ion beam.

Minimizing the discharge was an important step in the early development of

ICP-MS. This was accomplished in several ways. Modifying the load coil by

grounding the end closest to the cone reduced the electric field close to the sampler.

The plasma potential was also reduced by using a low aerosol gas flow rate (0.5 0.9

L/min) or reducing the solvent load on the plasma. Using a two-turn instead of a

three-turn coil, or moving the coil closer to the cone also minimized the discharge.

It is thought that these last two modifications change the temperature and ion number

density of the plasma, which affects the sheath conditions, changing the plasma

potential.

One modification that almost totally eliminated the secondary discharge was

the center tapped coil, which was developed by Douglas and French (44). In this

arrangement, the center turn of the coil was grounded while voltages of equal

magnitude but opposite polarity were applied to each end. The positive gradient from

one end to the center was balanced by the opposite phase, negative gradient on the

other end. This resulted in a low bias potential in the plasma. The center tapped load







23
coil was first available on the early Sciex ELAN instruments. This arrangement

provided a great advantage for these instruments in the marketplace. However, the

other modifications discussed above reduced the discharge sufficiently, so that this

arrangement was not necessary for the development of ICP-MS.

As the ions pass through the sampling orifice, a large number of collisions

convert the random motion into directed mass flow along the beam axis (45). A shock

wave forms due to collisions with background gas molecules, and the leading edge is

called the Mach Disc (Figure 5).

The position of the Mach Disc is given by



(Eqn. 1) Xm=0.67Do(PJPi)/2



where Xm is the distance of the Mach disc from the sampling orifice, Do is the

diameter of the sampling orifice, Po is the pressure of the ICP, and P, is the

background pressure in the extraction chamber. The formation of the supersonic jet

allows for the extraction of an essentially undisturbed bulk plasma.

The skimmer cone, located 8 mm behind the sampler, has a 0.8 mm orifice

diameter. This cone extracts only the center of the supersonic jet, or less than 1 % of

the plasma, so that a minimum of oxides passes through into the mass spectrometer.

The skimmer is placed with its orifice inside the Mach disc to avoid losses of ions due

to collisions and scattering, and its optimum position from the sampler is about 2/3 of








24



























4v


s I-

Ad *r
8 5 Le







25
the distance to the onset of the Mach disc (2/3Xm). The center of the zone of silence

passes through the skimmer cone into the next vacuum stage.

The SOLA has an accelerator cone, which is unique to this instrument, located

10 mm behind the skimmer (46). This region is pumped by a 330 L/s turbo pump

which keeps the pressure at about 1 x 103 torr. This cone has a voltage of +2 kV

applied to it which focuses and accelerates the ion beam. A cross-over (Figure 6)

occurs in the center of the cone for a large range of energies and angles of ions

exiting the skimmer. It also acts as the differential pumping aperture between the

intermediate and high vacuum regions of the instrument.

The most important advantage of the accelerator cone is that it reduces "space

charge" effects that are commonly observed in ICP-MS. Space charge effects occur

behind the skimmer cone when the ions emerge as a column of positively charged

particles. Each is repelled by the coulomb force e2/r2, and the transverse acceleration

is proportional to e2/mr2. Therefore, ions with lighter masses will be repelled faster;

and if the time it takes for the ions to pass through the ion optics is long when

compared with the time it takes for the lighter ions to be repelled, the result is loss of

the lighter ions. With the accelerator cone in place behind the skimmer, the distance

travelled through the high pressure region behind the skimmer cone is minimized,

reducing the transit time, reducing space charge effects.
































5s









Mass Spectrometer


Ion Transfer Optics


In an ICP-MS system, it is important to remove neutrals because they interact

with ions, and photons need also be removed because they are detected as signal by

the detection system used. Conventional ICP-MS systems use a bessel box to remove

photons and neutrals from the ion beam emerging from the plasma. It contains a

central stop which physically removes these species. Ions are electrostatically steered

around the stop and refocussed into an exit slit. The SOLA ICP-MS employs the

accelerator cone to produce a high velocity stream of ions condensed into a tightly

focused beam. This beam is diverted off axis and later realigned to exit parallel to

the entry beam into the quadrupole mass filter (47). Because the need for the central

stop is eliminated, the SOLA system is reported to have a higher transmission and a

lower background than conventional ICP-MS instruments.

A diagram of the transfer optics is shown in Figure 7. The X-Y deflectors and

focus plate steer the ion beam off axis. The Y steer realigns the beam so that it is

parallel to the entrance beam. Because the SOLA ion optics are operated at potentials

higher than normally used in ICP-MS instruments, the energy of the ions emerging

from the ion optics is too high for the quadrupole filter. The quadrupole requires ions

of low energy so that they have enough residence time in the electric field to be

differentiated. In order to solve this problem, a retarding lens, or the phase match
















HI I 0
I -
U,



f- I U
I I I Imi







29

lens placed before the quadrupole, matches the velocity of the ions to that required by

the analyzer.

The important advantage of the SOLA ion optics is that mass bias effects are

reduced. Mass bias effects are electrostatic in nature and are observed as a loss in

the transmission of ions at both ends of the mass range. The action of the ion transfer

lenses is based on the energy of particles passing through the system. The system is

tuned so that one energy is efficiently transferred, with a reduced efficiency for

particles of different energies. However, in ICP-MS, when particles emerge from the

sampler and skimmer cones, they are at the same velocity; thus, their energies are

mass dependent. The SOLA interface reduces mass bias effects by means of a

transfer lens system operated at higher potentials. This eliminates mass discrimination

because a wide range of energies can be focused into the quadrupoles.


Quadrupole Mass Analyzer


A thorough review of quadrupole mass analyzers may be found in Dawson (48)

and Watson (49); only a brief description is given here. The quadrupole mass

analyzer consists of four parallel rods arranged symmetrically. Diagonally opposite

rods are connected together and RF and DC voltages are applied to each pair. The

DC voltage is positive for one pair and negative for the other, and the RF voltages are

of equal amplitude but opposite sign. Ions from the transfer lenses travel along the

longitudinal axis of the four rods. Ions of one mass are transmitted while ions of

other masses have unstable trajectories, collide with the rods, and do not reach the







30
detector. The mass spectrum is scanned by varying the amplitude of the voltages

while keeping the ratio of RF to DC voltage constant. The mass of the ions

transmitted increases linearly with the magnitude of the DC and RF voltages, and

results in a linear mass scale.

The mass resolution is set by the ratio of RF to DC voltages and in practice has

a limit:



(Eqn. 2) R=n2/c



where n is the number of RF cycles the ions spend in the RF field, and c is a constant,

normally 10 25. The resolution is normally a unit mass and there is a trade-off

between resolution and sensitivity. Because resolution is a function of the DC to RF

voltage ratio, as the ratio is adjusted for the maximum resolution, fewer ions are stable

in the quadrupole field and the sensitivity decreases.


Ion Detection


The Faraday plate or cup is the simplest, most inexpensive, and most reliable

ion detector. However, it does have some disadvantages, including high noise and

slow response time. As positive ions strike the collector they are neutralized by

electrons drawn from ground after being passed through a high ohmic resistor. The

voltage across the resistor is amplified and is a measure of the ion current. The

amplified signal is directly proportional to the number of ions and the number of







31
charges per ion. The detector response is independent of the energy, mass, and

chemical nature of the ions.

The electron multiplier is the most popular ion detector because it has a low

noise level and a fast response time. This detector typically improves the signal

sensitivity by three orders of magnitude. The electron multiplier consists of a glass

tube coated with a high resistance film. The ion beam is focused onto the entrance

of the detector, which emits electrons in direct proportion to the number of

bombarding ions. The secondary electrons are accelerated and strike another area of

the detector, which emits secondary electrons. This cascading effect continues and the

number of electrons ejected is greater than the number hitting it. The final collector,

the anode, is connected to a conventional amplifier.

The electron multiplier is not as stable as the Faraday plate because the gain

depends on the condition of the coating. There is also some mass discrimination

associated with using this detector. The coefficient of secondary emission is

dependent on the mass, charge, and nature of the incident particles, so that an ion with

mass-to-charge ratio 50 will produce more secondary electrons than one with mass-to-

charge ratio 500. Overloading and saturating of the detector occurs at an output

current greater than 10-8 A.


Analytical Characteristics of ICP-MS


ICP-MS is a technique which combines the efficient ionization abilities of the

ICP and the sensitivity and selectivity of MS detection. This combination results in







32

a very sensitive technique for elemental analysis. There have been several reviews of

ICP-MS covering applications and general studies of the technique (11,31,43,50-53).

When nebulizing aqueous solutions, ICP-MS has several advantages. It has a

broad linear dynamic range (up to 6 orders of magnitude), made possible, in some

cases, by the use of two ion detectors. It is capable of multielemental analysis due to

the broad range of elements ionized by the ICP and the rapid acquisition of mass

spectral data. Detection limits using ICP-MS are low (0.01 0.1 Vpg/L) for most

elements, and the technique produces simple mass spectra, with not more than 10 lines

per element. It has good signal stability ( 5% RSD for several hours) and is capable

of isotopic analysis.

The largest obstacle in ICP-MS is the presence of interference arising from

different sources (11,54). For this reason, several factors must be considered when

using ICP-MS. These include total solute content of the sample and potential

interference from the background spectrum and the matrix components. Interferences

may divided into two general groups: spectroscopic and non-spectroscopic, or matrix,

effects.

Spectroscopic effects are well understood and can further be divided into four

categories: isobaric overlap, polyatomic ions, refractory oxide ions, and doubly

charged ions. Isobaric overlap occurs when two elements have isotopes of similar

masses that cannot be resolved by the mass spectrometer. Most elements have 1 to

2 isotopes free from isobaric overlap, except indium, and this problem can generally

be corrected. An isotope of the interfering element that is free of overlap is measured.







33
The relative contribution of the overlapping isotope may then be calculated using

isotope abundance information. Many commercial instruments can incorporate these

corrections in their data collection software.

Polyatomic ions arise from the dominant species present in the plasma (Table

1). They only present significant problems at masses below 82. They can be

dependent on the specific instrument and are difficult to correct due to the instability

of the polyatomic ion signal. Because these interference are somewhat dependent on

sample composition, they will vary with type of sample; however, polyatomic ions

present in dilute aqueous solutions have been well studied (55). Table 2 lists those

arising from nebulization of a 5% nitric acid aqueous solution. The elements most

affected by interferents are Si, P, S, Cl, Ca, K, Cr and Fe. Nitric acid results in the

cleanest background spectrum as compared to hydrochloric and sulfuric acids, and is

the preservative of choice for ICP-MS. Lowe and Stahl (56) studied the polyatomic

ion interferents present when an organic solvent is nebulized into the ICP-MS. Their

results, including detection limits for affected analyte ions, are presented in Table 3.

Several of the analyte elements had poorer detection limits due to the interferents, and

24Mg could not be determined.

Refractory oxides may be present due to the incomplete dissociation of sample

matrix or from recombination reactions in the plasma. Generally, oxide formation is

less that 1.5%, but it is dependent on the plasma operating conditions and nebulizer

flow rate. The formation of doubly charged ions depends on the second ionization

energy of the element and the condition plasma equilibrium. Only elements with a













Table 2
Interfering ions in nebulization of 5% HNO3


Mass


Element (abundance)


Interferent


16 0(99.76) 160
19 F(100) 16OH
20 Ne(90.92) 16OH,
28 Si(92.21) 14N14N, 12C160
29 Si(4.7) 14N14NH, 2CI60OH
30 Si(3.09) "N16
31 P(100) 14N16OH
32 S(95.02) 160160
33 S(0.75) 1606OOH
34 S(4.21) 1601"O
35 Cl(75.77) 16018OH
40 Ar(99.6), Ca(96.97), K(0.01) 40Ar
41 K(6.91) 40ArH
52 Cr(83.76) 40Ar'2C, -ArO60
54 Fe(5.82) 40Ar14N
56 Fe(91.66) 40ArO60
80 Se(49.82), Kr(2.27) 4Ar40Ar


Br(49.82)


(55)


"4ArOArH













Table 3
Interfering ions in organic solution nebulization


Element


Detection Limit
(pg/mL)


Interferent


24Mg 12C,

2sMg 7.2 13"CC, 12CIH
27Al 0.033

4Ti 0.16 32S160, 12C4
siV 0.042 38Ar13C

5sCr 0.20
52Cr -* 4ArO60

6Fe 0.48 4ArO60

s5Ni 0.033 4ArlO
6Zn 0.27 4ArC12C, 32S., 32S160,

6Zn 0.21 4"Ar2C'4N, 40Ar13C,,
4160o,


68Zn


0.69


4OAr12C160


*These isotopes
(56)


were not able to be detected.







36
second ionization energy lower than the first ionization energy of argon will form any

doubly charged ions. These include the alkaline earth elements, some transition

metals, and rare earth elements. However, the number of doubly charged ions is

normally less than 1 %. The formation of doubly charged ions has two effects, it

decreases the number of singly charged ions, reducing sensitivity, and produces

isotopic overlaps at half the analyte mass.

Several methods have been attempted to reduce spectroscopic interference.

Instrument optimization is the most obvious method. Nebulizer flow rates, sampling

depth, and ICP forward power can be varied to reduce spectroscopic interference

(57). Mixed gas plasmas or He plasmas as well as different plasma types have been

used, as discussed previously in this chapter. Using He plasmas reduces the Ar

polyatomic ions, however it introduces a whole new set of He interferents. The

addition of molecular gases improves the transfer of energy within the plasma because

of their higher thermal conductivities. Sample introduction techniques other than

nebulization, such as ETV, laser ablation (LA), or cryogenic desolvation (58) often

reduce the level of spectroscopic interference by introducing a 'dry' sample. This

reduces polyatomic ions due to hydrogen and oxygen. Another solution, which was

also discussed earlier in this chapter, is to use a different type of mass discriminator

that can effectively resolve the interferents from the analytes of interest.

Matrix effects are not as well understood (59) and can be divided into two

categories: suppression and enhancement effects, and physical effects. Suppression

and enhancements effects occur when the presence of one species reduces or increases







37
the signal obtained from the analyte, which depends on the type of sample and its

concomitant species. In general, the lower the mass and the degree of ionization of

the analyte, the greater the effect of an added matrix element on the ion signal of the

analyte. There is also an increased effect on the analyte signal when the atomic

mass and degree of ionization of the added matrix element is higher. Normally these

effects are not severe, but dilution of the sample is one solution. The use of a closely

matched internal standard (60) and standard addition calibration techniques (11) have

also been attempted. Varying instrument parameters has also decreased matrix

interference (61). The most effective method of reducing matrix effects for some

samples is to separate the analyte from the matrix using chromatography or co-

precipitation (11).

Physical interference effects normally involve blocking of the ICP-MS interface

apertures. This may occur by sooting with carbon containing materials, which is

discussed later, or by a high concentration of total dissolved solids (TDS). This

causes large signal drift over a very short time (several minutes). To prevent this,

solutions should contain less than 0. 1% of TDS (62).


Sample Introduction Techniques


The various types of sample introduction techniques give ICP-MS its versatility,

and several reviews have been published on this topic (11,30,63,64). Solid samples

may be introduced using laser ablation (65-73), electrothermal vaporization (ETV)

(74,75), slurry nebulization (75,76) direct sample insertion (77), or arc nebulization







38

(78,79). Gas phase introduction techniques have been applied to hydride (80) and

osmium tetroxide (81) vapor generation and to supercritical fluid chromatography (82)

and gas chromatography (82,83). The introduction of liquids has been the most

widely researched area. Some techniques are used with a nebulizer, such as flow

injection (84) and liquid chromatography (82), while others, such as ETV, eliminate

the need for a nebulizer. For the direct determination of liquids, there are several

types of nebulizers. Ultrasonic (85), thermospray (86,87), and pneumatic nebulizers

(86) have been used.

Pneumatic nebulization is the most commonly used method of liquid sample

introduction (11). A concentric nebulizer and double pass spray chamber coupled to

the ICP are shown in Figure 8. For ICP-MS, a nebulizer gas flow of 0.75 1.1 L/min

is normally used, with 0.2 1 mL/min of liquid flow. A peristaltic pump is often used

to monitor the liquid uptake. The spray chamber acts as a low pass filter, allowing

only small droplets (< 10 ptm) to reach the plasma. This leads to the primary

disadvantage of pneumatic nebulization, which is that only 1 % of the sample solution

reaches the plasma.

There are special problems with the nebulization of organic solutions into the

ICP-MS (56,88,89). The large amount of hydrocarbon introduced causes loading of

the plasma, making it unstable and inefficient, because its energy is being used to

dissociate the organic matrix. The carbon also condenses rapidly on the cooled

sampling cone and quickly causes sooting and blocking of the cone aperture. To








39














O 0




0
U I -

&.
T3





















41
4) ---n
/ __
4)" "*







40
reduce these effects, a small percentage of oxygen is often added to the nebulizer

carrier gas to ash to organic matrix. This is a delicate procedure because if excess

oxygen is added, it quickly erodes the interface cones, but if too little oxygen is added,

clogging of the tiny orifices results. The addition of oxygen also increases the amount

of oxides present, raising the level of molecular interferents. The nebulizer suffers

from relatively slow sample changeover as well. However, operation of the nebulizer

and spray chamber is simple, and it provides continuous sample introduction. This

allows for long scanning times by the mass spectrometer and the use of an internal

standard.


Electrothermal Vaporization


The use of an electrothermal vaporizer (ETV) for sample introduction into an

ICP has several advantages over nebulization (63); small (p.L) sample volumes are

used; a higher transport efficiency (20 80%) results, and the solvent and matrix are

vaporized independently of the analyte, allowing the analysis of difficult samples, such

as organic samples or those containing high total dissolved solids. In addition,

samples in a variety of physical forms may be analyzed without much sample

pretreatment.

There are also disadvantages associated with the technique. A transient signal

is generated by the ETV, requiring rapid data acquisition software for multielement

analysis. The technique also has a poorer reproducibility (10% RSD) due to

inconsistent sample insertion and sample transport. Analyte reactions with the







41

furnace material are another problem, as well as the added complexity and cost as

compared to nebulization.

An ETV is an electrically heated cylindrical graphite tube that in this case

produces sample vapor for introduction into an ICP (Figure 9). It is open at both ends

and has a central hole for sample introduction. Most tubes are 1-5 cm in length with

3-8 mm diameter and can accommodate 5-50 iL sample volumes. A tube is

replaceable and fits between a pair of cylindrical graphite electrodes at the ends of the

tube. The electrodes are located within a water-cooled metal housing. Two gas flows

are provided in the furnace. The external flow surrounds the outside of the tube and

prevents entrance of outside air and incineration of the tube. The internal stream

flows through the center of the tube for sample transport into the ICP and is normally

about 1 L/min.

Most tubes are made of graphite, however, tantalum, tungsten, and rhenium

have been used (77,90-94). These metals provide faster heating rates, do not absorb

the sample solution, and do not form carbides with the analyte. These problems can

also be reduced by using pyrolitically coated graphite tubes; the coating seals the pores

of the graphite. The alternative metals form oxides and tend to vaporize more quickly

and contain more impurities than graphite. Tsukahara and coworkers (90) added a

small amount of hydrogen to the argon carrier flow to minimize oxidation of a

tungsten furnace. Different furnace shapes have also been used, such as filaments,

strips, and loops (81,90,93,95-97) The platform tube (98)is another useful design.

This tube has a platform placed on the bottom under the dosing hole which is heated






























j).-

C0
I

3'


0
bOO H







43

mainly by radiation from the tube walls. The temperature of the platform therefore

lags behind that of the tube walls and the analyte is vaporized after the wall has

reached a near steady-state temperature. This configuration produces an almost

constant-temperature furnace.

The furnace is normally heated in three steps (Figure 10). The temperature is

increased by increasing the current through the tube. The first step is the drying or

desolvation step. The tube is slowly ramped to about 110 C and held for several

seconds. In this step, the solvent is evaporated and a solid residue is left behind in

the furnace. The objective of this step is to evaporate the solvent as fast as possible

without splattering or vaporization of the analyte. This step may last 20 45 s.

The next step is the ashing step. The furnace is ramped to 350 1200 C and

held for several seconds. This step may also last 20 45 s. The purpose of the ashing

step is to ash any organic components in the sample and to vaporize any volatile

inorganic species present. The temperature must be high enough to remove volatile

components without loss of analyte.

The third step is atomization and/or vaporization. The furnace temperature

used in this step is in the range of 2000 3000 C. This step has a fast ramp rate

which vaporizes the analyte. The duration is 3 10 s.

A fourth step is occasionally used to clean the furnace of any remaining

sample. The temperature used for the cleaning step is normally higher than that used

for vaporization.





























0 10 20 30 40 ,
Time (Seconds)


Temperature (C)


Ramp Time (s)


Hold Time (s)


1 110 5 10

2 1200 10 10

3 2500 2 7


Figure 10. Typical ETV temperature ramp.


2500


2000

0
1500
I

a. 1000
E

500


0


Atomization







Ashing


Drying


Step







45

Two different types of furnaces have been used for sample introduction into an

ICP-MS. The first kind is a modification of commercially available graphite furnaces

(74,95,99-108), while the other is a furnace especially designed for sample

introduction into the ICP-MS (94,109-113). An example of each type is given below.

A typical modified graphite furnace is shown in Figure 9. Shen et al. (95)

modified a Perkin-Elmer HGA-300 graphite furnace for sample introduction into an

ICP-MS. They modified four areas of the furnace: coolant flow, front adaptor, rear

adaptor, and furnace tube. The main modification was the addition of a coolant flow

several millimeters downstream of the furnace tube. This flow condensed the sample

flow into particulate for better sample transport into the plasma. The side quartz

windows were removed and replaced with stainless-steel adapters. The front adaptor

was connected to a PTFE tube which transports the sample to the base of the ICP

torch. They also modified the furnace tube by inserting a tungsten loop that held the

sample solution into a tantalum tube that entered through the rear adaptor. Because

there is no means to vent the solvent during the drying step, a second drying step at

a higher temperature is required to remove condensed solvent from the tube walls.

Transport efficiency was improved by the additional coolant flow, and the detection

limit for lead was 10 fg.

Park and coworkers (109) designed an ETV taking into account the flow

dynamics and gas temperature distributions of the system (Figure 11). A glass dome

surrounded the rhenium filament and was sealed to its metallic base by an o-ring. The

glass dome has a 5 mL volume, which is large enough to minimize sample













To ICP




Glass Dot


Filament


Teflon

Al Base








Copper Wire to Power Supply
Copper Wire to Power Supply


Figure 11. Dome-type ETV.







47

condensation on its walls, but small enough to reduce the dilution of the sample in the

carrier gas. The ETV is oriented vertically and carrier gas is introduced tangentially

so that the sample is spiraled upwards, avoiding the dome walls. The sample is

carried through tygon tubing into the plasma and auxiliary argon flow is provided into

the plasma torch during drying and ashing of matrix and during sample deposition,

when the glass dome is removed. Using this system, better than 80% sample transport

efficiency was achieved.


Sample Transport in ETV-ICP-MS


Kantor (114) has developed a theory based on aerosol science to explain the

mechanism of sample transport from an electrothermal vaporizer to an ionization

source, such as an ICP or flame. He refers to the process as "thermal dispersion", or

the "production of a dry aerosol by high temperature processes," and

differentiates this from mechanical dispersion of solids or nebulization of liquids.

According to his theory, the change in analyte signal under constant furnace conditions

is the result of variations in the transport efficiency of the technique.

The theory attempts to explain the dependence of aerosol formation on (1)

properties of the analyte (single component sample), (2) concentration of the sample

in the vapor phase, which depends on the mass of the sample, the heating rate, and the

flow rate of the carrier gas, and (3) the properties and amount of matrix vaporized

with the analyte. All of these factors influence the sensitivity, linearity of calibration

curves, and matrix interference resulting from the technique.







48

Aerosol formation is a process that begins when the sample is heated and its

hot vapor is generated from the surface of the furnace. Physical condensation of the

analyte occurs when the vapor is cooled and supersaturation is achieved. In the case

of ETV, this cooling occurs when the hot vapor mixes with the turbulent argon carrier

gas stream. Initial condensation results in nuclei clusters, which are present even in

unsaturated vapor because of collisions. As the supersaturation increases, more

clusters form and their size passes through a critical stage by the attachment of single

molecules. The formation of stable nuclei relieves the supersaturation, and the critical

droplet diameter, dp*, can be calculated according to nucleation theory (115):



(Eqn. 3) do*=4yVJkT1nS



where y is the surface tension of the liquid droplet, V~ is the molecular volume of the

vapor species, k is the Boltzmann constant, T is the temperature (K) and S is the

saturation ratio dimensionlesss), which defines the magnitude of supersaturation. This

process is known as homogeneous nucleation or self-nucleation because the

condensation nuclei form from the vapor itself.

At a given S, droplets smaller than the critical diameter evaporate, while

droplets larger than the critical diameter continue to grow. The saturation ratio may

be calculated (114):


S=Pvap/Pe(T)


(Eqn. 4)







49

where pvap is the partial vapor pressure and p,(T) is the equilibrium vapor pressure at

the temperature of nucleation.

For ETV methods where a carrier gas is used, and the sample is being heated,

the partial vapor pressure may be calculated to an acceptable accuracy using (114):



(Eqn. 5) Pvap=NsTqR/Vgtv



where the units of pvp are in atm, N, is the amount of vapor, in moles, from the

evaporated sample, T, is the temperature of the gas without heating in K, R is the gas

constant (82.05 cm-atm K-' mol-'), V, is the flow rate of the gas (cm3s-), and t, is the

mean evaporation time (s) of the majority of sample.

From the above equations, it can be deduced that, under the same conditions,

a higher supersaturation is observed for a less volatile substance because of the lower

value of p,(T). An increase in S results in an increase in the concentration of clusters

and, from Eqn 3, a decrease in the critical diameter, which increases the concentration

of stable nuclei. Below a certain value for S, there is negligible formation of stable

nuclei.

When there is a high concentration of stable nuclei and the supersaturation has

decreased, heterogenous condensation takes place. This is characterized by

condensation of vapor on existing particles. This condensational growth (g/s) is

proportional to the degree of supersaturation. When a high concentration of fine

particles is reached, growth continues by Brownian coagulation. Coagulation occurs







50

when two particles collide and adhere to form one particle. When the velocity is

determined by random motion, it is termed Brownian or thermal. Because the smaller

particles move faster, the rate of the process increases with smaller particles. This

leads to an increase in the mean particle size.

Flagan and Friedlander (116) and Kantor (114) took all of the above processes

into consideration and derived the following equation for the growth of particles:



(Eqn. 6) dp=klc,0.4



where dp is the particle diameter (cm), c, is the initial vapor concentration at the

temperature of nucleation (g/cm3), and k, is a constant specific for the vapor species

and conditions used in aerosol formation. It can be determined from Eqn. 6 that a

large change in the initial vapor concentration is needed to change the particle

diameter significantly. Therefore, when the initial analyte concentration is increased,

the number of particles will increase rather than the particle diameters. This means

that the transport efficiency does not change with sample mass, which leads to a broad

linear dynamic range.

If the vapor and particle concentrations continue to increase, particle growth

continues by the coalescence of finer particles and sintering of larger particles formed

by coagulation and coalescence. This is thought to occur when the major component

has a mass above 50 100 Vg. When this occurs, the particles become large enough

so that they gravimetrically separate out of the argon carrier flow and sample is lost.







51

In graphite furnace ETV methods, the carrier gas flow is normally about 1

L/min, and the sample is vaporized in about 1 second due to the fast heating ramp

used. Using these data and equations 4 and 5, the temperature dependence of the

saturation ratio was calculated (114) for various analyte vapor species and different

masses. It is assumed that when S approaches 10, stable nuclei are formed. Table 4

shows the results of these calculations. The temperatures in Table 4 are those to

which the vapor-gas mixture must be cooled to achieve a saturation ratio of 10. The

results show that with increased volatility of the species (species lower in Table 4),

stable nucleation occurs at lower temperatures, or further downstream from the hot

furnace. Because of the longer distance travelled by the vapor in this case, deposition

of the vapor species on the cool walls of the transport tubing is more likely to take

place, resulting in the loss of analyte for more volatile species. This analyte loss has

in fact been observed for zinc (117), arsenic (118), and cadmium (119), confirming

the data in Table 4.

Vapor condensation on foreign particles. Kantor's theory also explains the

"carrier effect", or heterogeneous condensation of vapor on foreign particles added to

the sample (116). This occurs when vapor from a volatile analyte condenses on the

stable nuclei formed from the foreign species in the furnace and is how cloud

formation takes place in the atmosphere. This can happen several different ways. The

first is when the two species have similar volatilities. Partial covolatization takes

place and stable nuclei are formed from the most abundant species according to

equations 3 and 4.













Table 4
Calculated temperatures of the vapor-gas mixture required to obtain S=10 for
species generated under ETV conditions.


Sample mass


Vapor


1 ng


(metal)


100 ng


NaCI 747 K 850 K

PbO 735 838

Pb 688 805

Ca 648 760

MoO3 612 700
Mg 540 583
CdCI2 500 572

Zn 455 530
ZnCl, 443 510
As4 410 464


390


458


(114)







53

A second process occurs when the volatilization of sample components is

different. Even in this case, there is partial mixing of the vapor from the volatile

element and the stable nuclei from the less volatile element because of the fast heating

rates used in ETV methods. Another example of the carrier effect may be due to the

electrothermally heated metal sample holders used in ETV methods. Metals such as

platinum, tungsten, and rhenium release vapor at the temperatures normally used for

atomization. It is thought that a surface oxide layer forms which is then vaporized and

forms stable nuclei (120).

Another possible mechanism occurs when the analyte vapor condenses on

stable nuclei produced from organic compounds present as solid sample matrices or

additives (114). It is known that during the pyrolysis of several organic materials,

aerosol particles are formed from the decomposition products, such as soot.

Stable nuclei may also form from gaseous organic substances added to the carrier gas.

Under ETV conditions, stable aerosol particles may form from the pyrolysis of

hydrocarbon or halogenated hydrocarbon vapors.

These carrier mechanisms may improve transport efficiency, but may also result

in loss of analyte if the foreign species is present in excess and particle formation is

controlled by coalescence and sintering.

Chemical condensation of vapor. Chemical condensation occurs when the

analyte vapor reacts to form a compound that is less volatile. An example of this is

the formation of a metal oxide or hydroxide when a volatile metal halide is mixed

with water or oxygen in the carrier gas. When the hot analyte vapor mixes with the







54
cool argon carrier gas, it is likely that oxidation takes place due to impurity oxygen

in the carrier gas or water vapor from the solvent. This results in chemical

condensation because the metal oxide is less volatile than the metal halide.

Effect of operating parameters on sample transport. In the case of ETV-ICP-

MS, the carrier gas flow rate from the furnace to the plasma must be about 1 L/min

in order to puncture the plasma and create the center channel. Any flow rates much

below this value will produce little or no signal; therefore, carrier flow is not a

parameter that can be widely varied.

When self-nucleation is the dominant process in aerosol formation, an increase

in the heating rate will improve transport efficiency. When the heating rate is

increased, the vaporization time of the analyte decreases, and according to equations

4 and 5, this will increase the supersaturation, improving transport efficiency. When

there is a large mass of sample, and coalescence and sintering are the main processes,

increasing the vaporization rate up to a certain point will improve transport efficiency.

However, above this rate, the transport efficiency decreases because the

supersaturation of the sample vapor is too high. Decreasing the heating rate reduces

supersaturation, reducing coalescence and sintering, and thus sample loss.

The final temperature of the furnace also affects sample transport. When the

final temperature of the furnace is increased, the temperature of the vapor-gas mixture

is also increased. Therefore, it takes longer for the mixture to cool down enough for

stable nuclei to form. This takes place further down the length of the transfer tubing

and allows for the sample vapor to condense onto the walls, resulting in sample loss.







55

The length of the transport tubing, therefore, also has an affect on sample

transport efficiency. It has been shown that with longer tubing, the signal area

remains the same, but peak height decreases. This is expected when ideal conditions

exist for stable nuclei to form. However, when large particles formed from

coalescence and sintering are transported through longer tubing, loss to gravitational

forces is more likely to occur. In addition, any volatile species that do not form stable

nuclei will have a greater chance of condensing on the tube walls.


Matrix Modification


Matrix modifiers are commonly used in graphite furnace atomic absorption

spectroscopy (GFAAS). Modifiers retain the analyte in the furnace at higher

temperatures, making it possible to ash at higher temperatures to remove matrix

compounds (121). When used with ETV-ICP-MS, matrix modifiers have a different

purpose. They act as physical carriers of the analyte from the furnace to the plasma,

and also may change the chemical composition of the analyte (114). Modifiers

improve transport efficiency, precision, and linearity of the technique (122).

Few researchers have used modifiers with ETV-ICP-MS. Ediger and Beres

conducted a comprehensive study concerning the effect of several modifiers on the

analysis of 20 different elements (122). The modifiers studied included sodium

chloride, palladium nitrate, magnesium nitrate, and tellurium nitrate. Before the

addition of the modifiers, the peak intensity increased nonlinearly when plotted versus

sample mass. The use of modifiers improved linearity and sensitivity for most







56
elements examined. They also studied the use of citric acid as a modifier for Co, V,

and In and found that it also enhanced the sensitivity of the technique. The use of

methane as a modifier was studied by adding 0.1% methane to the argon carrier,

however no improvement in transport efficiency was observed.

Hyamine hydroxide has been used as a modifier by two different research

groups. Forster et al. (101) used hyamine hydroxide in the determination of Pt in

leukemia cells. Newman and coworkers (102) used the same modifier for the

determination of Te in biological fluids. The researchers did not fully understand the

role of this modifier. It is thought that hyamine hydroxide disrupts the surface tension

of blood plasma, allowing it to dry over a larger area of the furnace. The detection

limit for Te in plasma was 5.7 ng/ml, which was a significant improvement over other

methods. Both groups used a post-furnace addition of oxygen to ash the organic

matrix present.

Gregoire (81) studied the effect of several modifiers on the determination of

Os by ETV-ICP-MS. A small enhancement in signal was found using NaCl, thiourea,

and 8-hydroxyquinoline. Ni and Se increased the sensitivity 5-fold, and Te had the

greatest effect, with a 20-fold increase in Os signal.

The same group examined the effect of mixed Mg/Pd modifiers on the

determination of volatile elements: Pb, Sn, and Ag (97). With the use of these

modifiers they observed a 5-fold increase in sensitivity, and an improvement in the

precision and linearity of the technique. Without the use of the matrix modifiers, the

analytical curves for these elements were similar to those obtained previously (122).







57
They found that the Mg(N03)2 and Pd(N03)2 mix increased the analyte signals two

times over the Mg(N03), and PdC1, mix. It was also found that different proportions

of each mixture had different enhancement effects on the signals and that each

modifier alone did not improve the analyte sensitivity. This does not agree with the

results obtained by Ediger and Beres (122), who found an enhancement in signal

using Pd(N03)2 as well as Mg(N03)2 individually.

Hub and Amphlett (123) used potassium iodide as a matrix modifier when

determining metal contaminants for semiconductor process control. The instrument

used was an ELAN Perkin Elmer Sciex ICP-MS with an HGA 600 MS ETV.

Concentrated hydrofluoric acid and buffered oxide etch (BOE), which is a mixture of

ammonium fluoride and hydrogen fluoride, were analyzed for eight metal contaminants

with detection limits below 0.2 ng/g. The surfaces of silicon wafers were studied by

vapor phase decomposition or droplet surface etching combined with ETV-ICP-MS.

Detection limits for the simultaneous analysis of the eight elements for this analysis

were between 0.2 to 2 x 109 atoms/cm2 per 6-inch wafer. By using the KI modifier,

the sensitivity was improved by a factor of 3-4, and the reproducibility and linearity

of the analytical curves were substantially improved. The optimum amount of

modifier was found to be approximately 1 [Lg. For the analysis of BOE, the relative

standard deviation (RSD) was about 15% for Na, Al, and Fe, which were found in

concentrations of 0.2 0.8 ng/g. Zn, which was also in this concentration range, had

about a 25 % RSD. The reproducibility was about 30% RSD for Ni, Cu, and Pb,

which were found in concentrations less than 0.1 ng/g. The authors attempted to







58
measure 5Cr as well, but were unable to detect it at the 1 ng/g level because of the

"Ar 1C molecular interference.


Analysis of Organics By ETV-ICP-MS


The ETV is an excellent sample introduction technique for organic samples.

It requires only a very small sample volume, and has rapid sample changeover. Most

importantly, the organic matrix may be removed and vented to atmosphere by using

a carefully programmed ETV temperature ramp, reducing the level of carbide

interferents. For this reason, oxygen addition for organic matrix ashing is normally

not required. However, not much research has been performed in this area. A brief

discussion of work in this field follows.

Hall and coworkers (77) were the first to use ETV-ICP-MS for the

determination of a metal in organic material. They determined gold in isobutyl methyl

ketone (IBMK), a solvent commonly used for the extraction of geological samples.

The authors used a laboratory-built glass domed ETV, similar to that described

previously in this chapter. They vented the organic matrix during the drying step

before atomizing the Au into the ICP. They had favorable results, with a 3% RSD

and a detection limit for gold of 0.008 ng/mL.

Osborne (106) determined elemental as well as organic mercury in petroleum

samples. He used a modified Perkin Elmer HGA 500 furnace coupled to a Sciex 250

ICP-MS. A post-furnace addition of oxygen to the carrier gas was used to ash the

organic matrix. The ETV parameters used were a vaporization step at 2650 C and







59
a cleanout step at 2700 C. No statistical difference was found between the

determination of elemental and organic Hg. A comparison was made of the organic

Hg signal obtained when dissolved in different length alkanes (C, C32), and it was

found that the Hg signal decreased as the carbon chain length was increased. The best

results were obtained with a carbon length of 12 or lower. The precision was between

5.9 13 %RSD for the shorter chains and 70 and 55 %RSD for CG and C32,

respectively, and the detection limit was 3 ppb Hg in pentane. The decrease in Hg

signal with increasing chain length was attributed to the increased plasma loading

because of the increased energy demands for longer carbon chains and the spatial

distribution of the sample as it combusts not being well confined to the plasma axis.

The author also noted an increase in the wear of the interface cones due to the

additional oxygen.

Richner and Wunderli (124) used ETV-ICP-MS to differentiate between organic

and inorganic chlorine. A Sciex Elan 5000 and ETV-600 were used, along with

platform graphite tubes. Oxygen gas flow at 20 mL/min was added to the furnace

carrier argon to ash the organic matrix. The ETV ramp temperature program consisted

of a drying step at 140 C, an initial vaporization step at 400 C to vaporize the

organic Cl, and a second vaporization step at 2650 C to vaporize the inorganic Cl.

The main application of this method was in the determination of polychlorinated

biphenyls (PCBs) in waste oils. The precision of the method for Cl was 10 20

%RSD, with detection limits of 0.5 9 Vig/g for different PCBs. However, PCBs

containing less than three Cl atoms could not be detected, even at 50 p.g/g. The







60
authors attributed this to the high volatilities of these compounds. They concluded

that the Cl compounds were being vaporized during the drying step, accounting for the

loss of signal. Because the method could not distinguish between PCBs and other

organic Cl compounds, it was suggested that it may be used only as a rapid screening

technique for the determination of PCBs in waste oils.

Manninen (105) used ETV-ICP-MS for the determination of extractable organic

chlorine (EOCI). EOCl is the portion of total organic chlorine which is separated from

a sample by using an organic solvent extraction procedure. A Fisons PlasmaQuad PQ

II with a Fisons ETV Mk III instrument was used for the analysis. The furnace

heating was as follows: the drying step was at 100 C, the ashing temperature was

1000 C, and the vaporiztion temperature was 2600 C. The furnace was not vented

during drying or ashing, and no oxygen was added to ash the organic matrix; carbon

build-up was observed on the interface cones. Several chlorinated compounds were

studied, including p-chloroanisol, pentachlorophenol, trichloroethene, and o-

chlorobenzoic acid. Chlorine was determined in ethyl acetate with a detection limit

of 10 ng. The precision was 5 %RSD for standards and from 5 to 30 %RSD for real

samples.












CHAPTER 3
INSTRUMENTATION


Inductively Coupled Plasma Mass Spectrometer



The research in this dissertation was performed on a Finnigan MAT SOLA

inductively coupled plasma quadrupole mass spectrometer. The instrument is

controlled by a DELL Dimension 486 (Dell, Austin, TX, Ultrascan 15FS) computer

interfaced to the instrument by a Programmable Logic Controller (PLC) (Mitsubishi,

Himeji, Japan, model FX-48MR). The state of all vacuum, cooling, gas flows, and

electronic control units is constantly monitored by the PLC. In case of failure, the

instrument undergoes the shut-down procedure and vents to argon.

The ICP torch and spray chamber are shown in Chapter 2 (Figure 8); the torch

is similar to that discussed earlier. The torch and spray chamber material were

constructed of quartz by Precision Glassblowing (Englewood, CO). The spray

chamber has a water cooling jacket and its own chiller that circulates 50% ethylene

glycol in the temperature range from -30 to 100 C (model RTE-100, Neslab,

Newington, NH). Argon is obtained from a 180 L dewar of liquid argon supplied

from Liquid Air Corporation (Walnut Creek, CA). The coolant gas flow through the

ICP torch is argon at 15 L/min, with the auxiliary gas flow at 1.2 L/min. The carrier







62

gas flow is also argon and is normally 1.0 L/min, depending on the sampling

accessory used. The three-turn copper coupling coil is water cooled. The RF power

is 1.3 kW at a frequency of 27.12 MHz produced by a radio crystal controlled

generator (Henry Electronics, Los Angeles, CA), and an automatic tuning circuit is

used for impedance matching.

The peristaltic pump (model M312, Gilson, France) used for sample

introduction may also be controlled through the computer.

A second, larger chiller (model CTF-75, Neslab, Newington, NH) supplies the

copper load coil, matching box, sampling interface, ETV, and all turbo vacuum pumps

with water cooled to 19 C.

The sampling interface is a three cone system through which the pressure drops

from atmospheric to 1 x 10-5 torr. The sampling cone is located 13 mm from the load

coil and has an orifice diameter of 1.1 mm at its tip. It is made of nickel with copper

sandwiched in the middle. The space behind it is pumped by an 18 m3 per hour single

stage rotary pump (model EMI18, Edwards, West Sussex, UK) and the pressure in

this region is normally 2-3 torr. The overall layout of the vacuum system is shown

in Figure 12. The skimmer cone is located 8 mm behind the sampler, has an aperture

of 0.8 mm, and is also made of nickel.

Located 10 mm behind the skimmer is the nickel accelerator cone. The region

between the skimmer and accelerator is pumped by a 330 L/s turbo pump ( model

TPH 330, Balzers, Germany) (Figure 12) which keeps the pressure at 1 x 10-3 torr.







63
















E

E


CNC
C/3
2

C.)
O
&,
N

00


s







64
The quadrupole mass filters are pumped by a 240 L/s turbo pump (model TPH

240, Balzers, Germany), and the pressure in this analyzer housing is maintained at

about 2 x 105 torr. The detector housing is pumped by a 60 L/s turbo pump (model

TPH 062, Balzers, Germany) which keeps the pressure in this region less than 5 x 10-6

torr. The turbo pumps are backed by a 12 m3/h rotary pump (model E2M 12, Edwards,

West Sussex, UK).

There are three vacuum gauges on the system controlled by an Edwards

vacuum gauge controller (model 1105, West Sussex, UK). Pirani gauges (model PRM

10, Edwards, West Sussex, UK) are located in the expansion housing (#2), and in the

intermediate housing (#1) (Figure 12). A Penning ionization gauge (#3) (model

CP25-K, Edwards, West Sussex, UK) monitors the pressure in the analyzer housing.

Pirani gauge 1 is interlocked to the Penning gauge so that the Penning is not turned

on until the pressure in the intermediate housing is below 1 x 102 torr. This protects

the Penning gauge from damage when running at high pressures.

The optics lenses used to transfer the extracted ions into the quadrupole mass

analyzer have been discussed earlier (Chapter 2) and are shown in Figure 7. The

voltage range for each lens is shown in Table 5.

The quadrupole mass analyzer is a square array of four accurately matched

cylindrical rods (model QMS 511, Balzers, Germany). The specifications are as

follows: rod diameter is 16.00 mm, rod separation is 21.16 mm, rod length is 300

mm, with RF at 1.6 MHz. The mass range of the quadrupole is 1 255 a.m.u. Peak













Table 5
Ion lens voltage ranges


Min. Voltage


Max. Voltage


Extraction Cone 0 +5 kV

X Deflection -100 V +100 V

Y Deflection -330 V +330 V
Focus 0 +160 V

Y Steer -330 V +330 V

Phase Match -512 V 0


Lens


Quadrupole Bias


-12 V


12V







66
width is adjusted by the resolution control on the front of the mass spectrometer, and

abundance sensitivity is controlled by the pole bias voltage. The quadrupole may be

used in the "remote" or "local" modes. In the remote mode, the computer controls the

mass setting, while in the local mode, the operator has manual control via a knob

located on the front panel.

The collector system consists of an analog Faraday amplifier and a pulse

counting electron multiplier (model 4870, Galileo, Sturbridge, MA). The multiplier

is used for signals smaller that 106 ions per second, while the Faraday plate may be

used for signals up to 6 x 1010 ions per second. A pair of deflector electrodes controls

which detector the ion signal enters.


Electrothermal Vaporizer


The electrothermal vaporizer is a GBC graphite furnace (model GF3000, GBC

Scientific Equipment, Australia) that has been modified by Finnigan for use as a

sample introduction device for the ICP. It consists of a control unit, interfaced to the

computer, and the graphite furnace workhead (Figure 13). The left window of the

original furnace has been removed and replaced by a titanium vapor extraction

interface which connects to flexible tubing for sample transport (Figure 13). The

carrier argon and auxiliary gas flow through the center of the tube are controlled by

two mass flow controllers (MFC) on the ETV workhead unit. The carrier argon flow

has a range of 0 2 L/min and the auxiliary flow has a range of 0 0.2 L/min. They










Graphite
Plug


Rubber
Bellows Titanium Graphite
Chimney Shroud G te
Left Side Electrode
Block


Window
Plastic
Tubing
Fitting


Right Side Block
\ Furnace Tube
Center Block


Figure 13. ETV workhead.







68
enter the tube tangentially, so the flow is turbulent. The shroud gas flow is controlled

by a flow meter on the front of the control unit.

The carrier argon exits the ETV through the Ti interface and is vented to

atmosphere until atomization of the sample begins, when the sample valve directs the

flow into the plasma (Figure 14). The carrier flow then enters through a "Y"

connector located between the spray chamber and ICP torch (Figure 15). The

computer controls the ETV temperature ramp as well as the time when the sample

valve is actuated and when the mass spectrometer starts scanning. A graphite plug

(P/N 4019D, Spectrographic Services, Sussex, NJ) is placed in the dosing hole of the

graphite tube before heating to prevent sample loss. Hollow as well as integrated

platform pyrolytically coated graphite tubes (P/N 4090-73 and 4090-75, CPI, Santa

Rosa, CA) may be used in the furnace.

















0



E-r,2


zo

























/


-,,r-












CHAPTER 4
DETERMINATION OF METALS IN AQUEOUS SOLUTIONS


Because the instrument was new to the laboratory and was one of the first

ETVs that Finnigan had delivered to a customer, it was decided to characterize the

system using aqueous standards of volatile elements and the elements of interest

before research was begun on the metallo-organic standards. It was thought that the

determination of Cu, Zn, Mg, Al, Fe and Y would be simple and straightforward, and

would be a good start for familiarizing ourselves with the new instrument.


Instrument Modifications


When the ETV was used as received from Finnigan MAT, poor results were

obtained. Precisions of 100 to 150% were not uncommon for aqueous standards of

Cu, Zn, Al, and Mg and calibration curves were nonlinear. Therefore, several

modifications were made to the ETV to improve the efficiency and precision of

sample vapor transport to the ICP.


Graphite Plug Holder


The first and most obvious change was with the graphite plug. Finnigan

provided no means of securing the plug in the dosing hole, and it often fell out during







72
ETV heating, resulting in sample loss through the opening and contributing to the poor

precision. A laboratory-constructed spring loaded graphite holder made of ceramic

with a bakelite support was adapted for the SOLA ETV from another graphite furnace

in the lab. It was anchored to the ETV by a magnetic mount which held the bakelite

support. The bakelite support held a phenolic sleeve that held a ceramic holder within

a spring. The graphite plug was pressed into the ceramic holder, then placed in the

dosing hole. The holder was tightened with a set screw going through the bakelite.

After correspondence with Finnigan MAT (San Jose, CA), this graphite plug

holder was soon replaced with one supplied by Finnigan MAT. It consisted of a metal

arm that screwed into the ETV workhead. It had a hinge that allowed it to be raised

or lowered, to plug or unplug the dosing hole using the mass of the arm to hold the

graphite plug into the dosing hole. The area of the arm that contacted the plug was

hollowed out and replaced with a small piece of insulating ceramic so that the current

traveling through the graphite tube would not travel through the arm.


ETV Tubing Interface


The interface between the graphite tube and the flexible transport tubing

supplied by Finnigan is shown in Figure 13. The titanium tube, however, was several

millimeters away from the graphite tube, leaving a dead volume that may have caused

sample loss. This interface was replaced with one constructed in the laboratory. It

consisted of a solid teflon cylinder constructed to fit the ETV that was drilled out in







73
the middle to provide a path for the argon flow and sample to travel. The teflon piece

also had an o-ring groove to seal it to the window opening. The transfer tubing fit

snugly within the tunnel drilled through the teflon, and it was pushed through the

teflon fitting into the ETV workhead. This provided less dead volume in that the

teflon piece was flush with the hole, while the other interface was hollowed out

around the titanium tube.

After further correspondence with Finnigan MAT, they provided a longer

titanium tube for the interface (Figure 16). It was not in contact with the graphite

tube, but was much closer than the previous Ti tube. The titanium tube was used for

all further work.


Sample Valve


The original sample valve was a solenoid microvalve with a teflon interior that

was actuated by the ICP-MS software according to the ETV temperature ramp. It was

a three-way valve so that the ETV flow could be vented to air or be transported into

the ICP. After experimenting with metallo-organic oil compounds in the ETV, oil

could be seen oozing out the vent port of the valve and a small puddle had formed

outside the valve. It was apparent that the tiny passages in the valve were too small

and that sample was condensing on the valve walls. We replaced the valve with a

manual three-way valve (Jax Valve & Fitting Co., Jacksonville, FL) with a much








74

















[.I







Cu
6t
0




oo
D *5














.5T
)**I
o <







75
larger internal diameter that was actuated manually with the aide of a stopwatch. The

transport tubing was attached to the valve with 1/4 inch Swagelock fittings.

After further communication with Finnigan, they provided a 'pinch' valve in

which the transport tubing from the ETV to the ICP passed through a valve (Figure

17). A plastic tee was placed before the valve with silicon tubing connected to each

arm of the tee. Both tubes exiting from the tee passed through the valve, with one

connected to the ICP and the other venting to the atmosphere. A lever between the

two tubes pinched one of the two silicon tubes to constrict, or block, passage of the

ETV carrier gas. The carrier argon was then forced to travel through the other tube.

This valve provided a path with a larger diameter for sample transport, and was used

for all further work. The tubing is about two feet long, with an 1/8 inch i.d. and 1/4

inch o.d.


Sample Introduction Glassware


Finnigan MAT recommended that the ETV be operated with the spray chamber

and nebulizer in place with water being pumped continuously through them. They

claimed that hydrogen ions in the plasma contributed to a higher ionization efficiency.

However, this eliminates one of the advantages of ETV, which is the use of a 'dry'

plasma to reduce molecular interferents due to water in the plasma. Since one of the

analytes of interest in this research is iron, it is important to limit the amount of water

in the plasma to reduce the amount of the ArO molecular interferent at mass 56. The































*U
C-(



a)





A
I


a)s
-







77
sample introduction interface at the ICP was modified to eliminate the use of the spray

chamber (Figure 18). The nebulizer gas flow was clamped to one arm of a "Y"

connector, with another arm connected to the ETV gas flow with a nylon Swagelock

fitting, and the third arm connected to the torch injector tube by the glass ball joint.

The nebulizer flow, before the "Y" connector, was passed through a three-way

solenoid valve that was wired to the ETV sample valve. It was hooked up oppositely,

so that nebulizer argon flowed to the ICP when the ETV valve was venting to

atmosphere and the nebulizer flow was vented to atmosphere when the ETV flow was

directed to the ICP. This provides a constant flow of injector gas into the ICP,

reducing the chance of plasma 'blow-out'.


Experimental


Instrument Optimization


A difficulty with using a 'dry' plasma is with optimizing instrumental

parameters. A constant signal is required to tune the instrument, or adjust the voltage

settings on the optics lenses to obtain a maximum signal. Constant signal is also

required to optimize the placement of the plasma torch position over the sampling

orifice. Many researchers have tuned the optics lenses using a wet plasma, or

nebulizing a standard solution, then removing the nebulizer and spray chamber after

optimization and hooking up the ETV interface (77,91,99,100,107,108,125,126). A

few researchers have tuned with a wet plasma, then continued to nebulize blank water






78







l) o
0
0

E





CO




00
Ifl





4)2







79
throughout the ETV runs in order to minimize changes in the ion lens settings

(90,127). Other researchers have tuned on a signal from an element slowly vaporized

from the furnace, such as mercury (96,128,129). One group tuned on the "2C signal

generated from a blank ETV firing or from CO, present as an impurity in the argon

gas supply, then tuned on the Cd signal generated from its slow vaporization off of

the furnace (111,112). Other researchers have tuned on the signal present from a

spectral feature in the plasma, such as Ar2 (128) or ArH (92). Tuning on the signal

from an element vaporized from the furnace is difficult because a stable signal is

difficult to generate, and adjustment of the torch position using a spectral feature is

impossible because the species, such as ArE, is present throughout the plasma.

In our laboratory, several different approaches were attempted. 12 crystals were

placed into the graphite furnace and heated to produce a signal. A Pb shot was also

introduced into the furnace and vaporized slowly at a low temperature, and a

concentrated solution of Hg was slowly heated. However, none of these methods

worked well because a constant signal was impossible to obtain for more than a few

minutes.

The last method attempted gave the best results. House nitrogen was connected

to the auxiliary mass flow controller on the ETV. Using this controller, N, gas was

slowly bled into the main ETV argon gas flow to the plasma. A nitrogen gas flow of

10 30 mL/min added to the 1 L/min argon flow produced enough signal at mass 14

or 28 to detect on the Faraday plate. Without the added nitrogen flow, no signal was







80
observed on the Faraday detector. Nitrogen is present throughout the plasma.

However, during the tuning procedure, the nitrogen concentration was higher in the

central channel than throughout the rest of the plasma. Therefore, the torch was able

to be finely adjusted over the sampling cone orifice. Initial adjustments were made

using a wet plasma, nebulizing a standard tuning solution of 2 ppm Al, Mg, Cd, Pb,

and In, once every one to two weeks. The ion optic lenses were then tuned using

nitrogen at mass 14 (N') or 28 (N,2), or argon at mass 40, depending upon the mass

of the analyte of interest. This method worked well because a constant signal could

be maintained for as long as necessary.

The ETV parameters were optimized for each element, except carrier flow,

which was kept constant at 1 L/min, as this is the flow needed to puncture the plasma

and form the central channel. The vaporization temperature was optimized by running

three injections of the same concentration for a series of temperatures. The peak

intensities and precisions for each set of runs were compared, and the temperature that

gave the highest intensity, with good precision (<15% RSD), was chosen as the

optimum atomization temperature. There generally was a definite optimum

temperature, with peak intensity decreasing below and above the optimum. The mass

monitored for each element was optimized in a similar way. Each element was run

three times over a 1 amu mass range, for example, for Mg at mass 24, mass 23.8 was

studied, and increased by 0.1 until 24.7. The optimum mass was then chosen based

on the best peak intensity and precision. Because temporal profiles were taken, it was







81

important to monitor the correct mass, or precision and signal intensity suffered

because the tail of the peak was being monitored. The intensity may drop by 2 times

just 0.5 amu from the peak optimum. A hollow graphite tube was used for the

determination of aqueous elements.


Standard Preparation


Atomic absorption single metal standard solutions (Fisher Scientific) of Al, Mg,

Fe, and Y at 1000 ppm were diluted to prepare calibration standards. NaCI solutions

were prepared from solid, purified NaCl (Puratronic grade, Johnson Matthey, Ward

Hill, MA), and Pd(N03)2 solutions were prepared from solid, purified Pd(N03),

(99.9%, Johnson Matthey, Ward Hill, MA). Each solution contained 1% trace metal

grade HNO3 (Optima grade, Fisher Scientific) and was prepared using deionized water

purified in the laboratory (Milli-Q Plus, Millipore, Bedford, MA) with a specific

resistance of 18 MQ cm.


Instrumental


Optical emission from the plasma was detected by placing a fused silica lens

(2 in focal length, 1 inch diameter) so as to image 1:1 light from a region between the

ICP load coil and the sampling cone upon a fused silica fiber optic bundle of 5 mm

diameter. The other end of the fiber optic cable was configured as a slit and

positioned 6 mm in front of the entrance slit of a 0.34 meter focal length







82
monochromator (Spex model 340E, Spex Industries, Edison, NJ). The signal was

detected using a photomultiplier tube, current-to-voltage converter and an analog strip

chart recorder. The spectrometer slits were typically set to 100 pm producing a

spectral resolution of 0.25 nm.

Sulfur hexafluoride was introduced into the furnace using the auxiliary mass

flow controller equipped on the ETV workhead.


Procedure


A 10 pL aliquot of aqueous standard was introduced into the graphite tube

through the dosing hole with a fixed 10 pL Eppendorf pipette (model 4800,

Brinkmann Instruments, Westbury, NY). Modifier, if used, was added first in 10 ipL

aliquots, using a 0 10 pL Eppendorf pipette (model 4810, Brinkmann Instruments,

Westbury, NY ). The dosing hole was then plugged using a graphite plug that had a

few mm of its tip sanded off. Each plug was conditioned twice by consecutive blank

firings to remove any impurities in the graphite, and the plug was replaced daily.

Each graphite tube lasted for several hundred firings, usually it split in half after this.

A new graphite tube was conditioned using the ETV ramp procedure recommended

by Perkin Elmer (Table 6). The ETV run was then initiated on the computer. At the

appropriate time set by the operator in the software program, the sample valve was

actuated by the computer so that the ETV carrier argon flow was directed from

venting to the ICP.













Table 6
New graphite tube conditioning


Temperature
(oC)


Ramp Time
(s)


Hold Time
(s)


1 2650 60 2
2 20 1 20
3 2650 10 10
4 20 1 20
5 2650 10 10
6 20 1 20
7 2650 10 10


Number







84

The mass spectrometer could only be set to scan over a preselected mass range,

but there was a "bug" in the software that prevented it from scanning over more than

one mass range. The time resolved software that allows the mass spectrometer to scan

very rapidly has not been released by Finnigan, and without it, it is difficult to scan

over more than a few elements. In order to monitor only one mass throughout the

length of the ETV ramp, the mass analyzer had to be manually switched to the "local"

mode using the switch on the front control panel of the mass spectrometer and the

mass of interest was selected. Time profiles of all signal peaks were obtained in this

research. The Faraday detector was used for all runs because the peak intensities

produced from the ETV were too large for the electron multiplier detector. Limits of

detection were calculated based on the 3a criteria using the standard deviation of a

series of blank injections as the measure of blank noise.


Results and Discussion


After the previously discussed instrument modifications were made, sample

precision and analytical curve linearity were still poor. Calibration curves like the one

for Mg shown in Figure 19 were typical for all elements studied. Reproducibilities

for this data ranged from 17.5 86.0 % RSD. Several different parameters were

varied, including ETV ramp temperatures, ETV carrier gas flow, and volume of

sample introduced into the furnace. Because the precisions were so poor, it was

impossible to optimize these parameters.









II I I I I 85










04

















-oE
(0
co
--
I ,.--

















o0










00









o






o 00 0 0


(S/lunoo) AI!sulul 18eed ueeiO







86
In an attempt to find the cause of the poor precision, several experiments were

carried out. To ensure that the problem was not due to a "bug" in the software, a

chart recorder was wired to the electronics on the Faraday detector. The signal

measured on the chart recorder correlated rather well with the signal measured by the

Finnigan SOLA software. Therefore, it was concluded that the software data

acquisition was not the problem.

An optical emission detector was set up to measure the emission from the ICP

to ensure that the problem was not with the mass spectrometer. A 1.5 ng aliquot of

Mg was introduced into the furnace and the signal was monitored using the mass

spectrometer and optical emission detector simultaneously. Figure 20 shows the

results of this experiment. The peak intensities from each detection method followed

the same trend. The precision for the emission peak heights was 21 % RSD and it was

16% RSD for the mass spectral peak intensities. The results show that the poor

precision was not due to the mass spectrometer; therefore, it was concluded that the

problem was being caused by the ETV.

The reproducibility of the furnace heating was also measured using a

photodiode equipped with a red filter to measure the continuum emission of the

graphite tube and found to be acceptable. According to Planck's radiation law (130)

the emission is proportional to the temperature of the object. Most graphite furnace

systems are equipped with a photodiode to constantly measure the emission, and this

photodiode was adapted from another graphite furnace in the laboratory.













. 1 I I I 'I I I I II I I I ,


/






- =

I


Ci I
_ I


I I I A I I I I I I I I I I \ I I I I I I
o- r- 1'- I- o c 04o
0 0 0 0 0 0 0 0
+ + + + + + + +
w w w w w w (UwO)
0 0 0 0 0 0 0 0 (Lilo)
T-T -1-(00(0 *^ (


saI(jsuI U|I 8leead


V.



C.





ou











0



E,5





04
Lu









Cu




















0

tu
t














.0o
CM <.


(S/slunoo)







88
The precision of the Eppendorf pipette was measured by depositing 10 pL of

Millipore water into a beaker on the mass balance. It was found to have 1.58 % RSD,

which agrees with the manufacturer's claim for precision.

The argon flow was then measured using a mass flow controller (Teledyne)

to check if the graphite plug was sealing well after each injection. The dosing hole

was unplugged, then plugged, and the gas flow was monitored using a chart recorder.

The signal was measured and found to have 4.0 % RSD. Although this is rather high,

it is not high enough to account for all the reproducibility problems.

These experiments did not find the main cause of the problem, which was

inconsistent sample transport from the ETV to the ICP. The next section describes the

solution to the reproducibility problem.


Matrix Modification


The use of matrix modifiers, or carriers, with the ETV was then investigated.

Two different modifiers, NaCl and Pd(N03)2, were chosen because of their previous

success in ETV-ICP-MS (122) and because they did not contain any of the elements

of interest. They can also be found in highly purified forms, which is important in

reducing contamination of the sample. Solutions of NaCl and Pd(N03), of various

concentrations were prepared and added to several concentrations of each element

(Mg, Al, Fe, and Y). The general ETV temperature ramp for the aqueous solutions

is shown in Table 7. The vaporization temperature (no. 3) changed with the element













Table 7
ETV temperature ramp used for aqueous solutions


Temperature
(C)


Ramp Time
(s)


Hold Time
(s)


1 90 10 20
2 250 10 20
3 1 6


* This temperature varied with element.


Number




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 EJ1M2TTZX_MJIMCI INGEST_TIME 2012-02-20T23:12:14Z PACKAGE AA00009022_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES