Citation
Applications of laser ablation inductively coupled plasma mass spectrometry for the analysis of solids

Material Information

Title:
Applications of laser ablation inductively coupled plasma mass spectrometry for the analysis of solids
Creator:
Bi, Melody, 1968-
Publication Date:
Language:
English
Physical Description:
x, 159 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Calibration ( jstor )
Ions ( jstor )
Laser ablation ( jstor )
Laser beams ( jstor )
Laser power ( jstor )
Lasers ( jstor )
Plasmas ( jstor )
Signals ( jstor )
Silicon ( jstor )
Soil samples ( jstor )
Chemistry thesis, Ph. D ( lcsh )
Dissertations, Academic -- Chemistry -- UF ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2000.
Bibliography:
Includes bibliographical references (leaves 151-158).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Melody (Xiangqing) Bi.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
024876365 ( ALEPH )
45286902 ( OCLC )

Downloads

This item has the following downloads:


Full Text










APPLICATIONS OF LASER ABLATION INDUCTIVELY COUPLED PLASMA
MASS SPECTROMETRY FOR THE ANALYSIS OF SOLIDS














By

MELODY (XIANGQING) BI


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


2000
































To My Family














ACKNOWLEDGMENTS


I would like to express my immense gratitude to my advisor, Dr. James D. Winefordner who has offered not only his intellectual guidance for the project but also his patience for my mistakes and stubbornness in pursuit of the project. Dr. Benjamin W. Smith is the next person to whom I would like to express my gratitude for invaluable suggestions and comments on this research. He has also given me a lot of help on other issues.

I would like to thank Dr. Antonio M. Ruiz and Dr. Igor Gornushkin who have contributed directly to my experiments. I would also like to thank Dr. Chenglong Yang for many useful discussions concerning my research.

I would like to thank Dr. Stephen J. Pearton and Dr. Rajiv K. Singh (from the Materials Science & Engineering department) who provided semiconductor samples and useful information for my research. I would also like to thank Dr. Luisa Amelia Dempere and Mr. Wayne Acree from the Major Analytical Instrumentation Center, who helped me with the SEM-EDS study. I would like to thank Mr. Gill Brubaker from Energy Research Center for letting me use the ICP-AES instrument.

My gratitude also goes to Drs. Willard W. Harrison, Daniel R. Talham, Martin Vala and Paul H. Holloway for their kindness in serving on my committee. They were








also supportive of my research effort. I especially thank Dr. Harrison who attended our regular research group meeting and gave me useful advice every now and then.


I would like to thank Dr. Burtron H. Davis from the Center for Applied Energy Research, University of Kentucky for recommending me to the University of Florida and for his help with my oral project. I feel very lucky to know him.

I want to thank my brother Sean for his encouragement. I want to thank my mom Yuante Bi and my dad Junke Bi who live on the other side of the globe, and who have given me the love and support that I needed the most to pursue my life goal and dreams. I also thank my friends and colleagues, the entire Winefordner group, Merry Zhang, Lisa Lang, Detong Sun, Caijun Sun, Junqiang Sun, Johnathan Bishop, Qianrong Ma, Dale Cheng, Bob and Gay Greninger etc. who have made my time at the University of Florida a pleasant and interesting experience.

Finally I would like to acknowledge financial support from a Texaco Fellowship for my research.














TABLE OF CONTENTS


ACKN OW LEDGM EN TS ................................................................................................. iii

LIST OF ACRON YM S ................................................................................................... viii

ABSTRACT ....................................................................................................................... ix

CHAPTERS

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

2 BACKGROUND ...................................................................................................... 7

2.1. Principles of Laser Ablation ............................................................................... 7
2.1.1. Principles of Lasers .................................................................................... 7
2.1.2. Q-switch Lasers ........................................................................................ 9
2.1.3. Laser M aterial Interaction ........................................................................ 9
2.2. Principles of Inductively Coupled Plasma Mass Spectrometry ....................... 12
2.2.1. Inductively coupled plasm as .................................................................... 12
2.2.2. Sam ple Introduction ............................................................................... 15
2.2.3. Sam pling Interface ................................................................................. 15
2.2.4. Ion Lens in ICP-M S ................................................................................. 17
2.2.5. Space Charge Effect ............................................................................... 17
2.2.6. Quadruple M ass Spectrom eter ............................................................... 18
2.2.7. M ass Detectors ........................................................................................ 19
2.3. Principles of LA-ICP-M S ................................................................................. 20
2.3.1. Calibration Strategis .............................................................................. 21

3 INSTRUMENT DIAGNOSTICS AND OPTIMIZATION .................................... 36

3.1. Introduction ...................................................................................................... 36
3.2. Experim ental .................................................................................................... 36
3.2.1. Instrum entation ...................................................................................... 36
3.2.2. Sam ples .................................................................................................... 39
3.3. Results and Discussion ................................................................................... 39
3.3.1. Laser Output Energy ............................................................................... 39
3.3.2. Laser Focusing Positions ........................................................................ 41
3.3.3. Analysis of SiC ...................................................................................... 41
3.3.4. Plasm a Operating Conditions ................................................................. 42








4 STUDY OF SAMPLING STRATEGY FOR LA-ICP-MS ANALYSIS ................ 52

4.1. Introduction ...................................................................................................... 52
4.2. Experim ental .................................................................................................... 53
4.3. Results and Discussion ................................................................................... 54
4.4. Conclusions ...................................................................................................... 56

5 STUDY OF SOLUTION CALIBRATION OF NIST SOIL AND GLASS
SAMPLES BY LASER ABLATION INDUCTIVELY COUPLED PLASMA
M ASS SPECTROM ETRY ...................................................................................... 64

5.1. Introduction ...................................................................................................... 64
5.2. Experim ental .................................................................................................... 66
5.2.1. Instrum entation ...................................................................................... 66
5.2.2. Samples .................................................................................................... 68
5.2.3. Solution-based Calibration ...................................................................... 68
5.3. Results and Discussion ................................................................................... 71
5.3.1. Selection of an Internal Standard ............................................................. 71
5.3.2. Study of Fractionation Effect .................................................................. 73
5.3.3. Analysis of NIST Standard M aterials ...................................................... 74
5.4. Conclusions ...................................................................................................... 76

6 ELEMENTAL ANALYSIS OF SPANISH MOSS USING LASER
ABLATION INDUCTIVELY COUPLED PLASMA MASS
SPECTROM ETRY .................................................................................................. 87

6.1. Introduction ...................................................................................................... 87
6.2. Experim ental .................................................................................................... 89
6.2.1. Instrum entation ........................................................................................ 89
6.2.2. Samples .................................................................................................... 90
6.3. Results and Discussion ................................................................................... 91
6.3.1. Elemental analysis of Spanish moss using NIST leaf SRMs as matrix
m atched standards by LA-ICP-M S .................................................................... 91
6.3.2. Study of Fractionation ............................................................................. 93
6.3.3. Analysis of M oss Samples by LIBS ........................................................ 94
6.3.4. Standard Addition M ethod ...................................................................... 94
6.4. Conclusions ...................................................................................................... 96

7 LOW LEVEL COPPER CONCFNTRATION MEASUREMENTS ON
SILICON-WAFER SURFACE USING DIRECT LA-ICP-MS AND
SOLUTION SAM PLING ICP-M S ............................................................................ 110

7.1. Introduction ......................................................................................................... 110
7.2. Experim ental ....................................................................................................... 112
7.2.1. Instrum entation: ......................................................................................... 112








7.2.2. Sam ples ...................................................................................................... 113
7.3. Results and D iscussion ....................................................................................... 114
7.3.1. D irect Analysis of W afer Standards ........................................................... 114
7.3.2. Analysis of Solution Sam ples .................................................................... 115
7.4. Conclusions ......................................................................................................... 116

8 PROFILING OF PATTERNED METAL LAYERS BY LASER ABLATION
INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY ...................... 120

8.1. Introduction ......................................................................................................... 120
8.2. Experim ental ....................................................................................................... 122
8.2.1. Instrum entation .......................................................................................... 122
8.2.2. Sam ples ...................................................................................................... 123
8.3. Results and Discussion ....................................................................................... 124
8.3.1. Effect of Ablation Cham ber ....................................................................... 124
8.3.2. Optim ization of Argon Flowrates .............................................................. 124
8.3.3. Optim ization of Laser Focus ...................................................................... 125
8.3.4. Com parison of LA-ICP-M S w ith LA -OES ............................................... 126
8.3.5. Profiling of Patterned Metal Layers on Silicon Wafers ............................. 128
8.4. Conclusions ......................................................................................................... 129

9 CON CLU SION S AND FU TURE W ORK ................................................................ 142

9.1. Conclusions ......................................................................................................... 142
9.2. Future W ork ........................................................................................................ 145

REFEREN CES ................................................................................................................ 150

BIO GRAPH ICAL SKETCH ........................................................................................... 159














LIST OF ACRONYMS


LA-ICP-AES LA-ICP-MS


LA-ICP-OES LIBS MD-ICP-AES/MS


RSF RSD SEM SIMS SRM


Laser Ablation Inductively Coupled Plasma Atomic Emission Spectrometry Laser Ablation Inductively Coupled Plasma Mass Spectrometry

Laser Ablation Inductively Coupled Plasma Optical Emission Spectrometry Laser Induced Breakdown Spectrometry Microwave Digestion Inductively Coupled Plasma Atomic Emission Spectrometry/Mass Spectrometry Relatively Sensitivity Factor Relatively Standard Deviation Secondary Electron Microscopy Secondary Ion Mass Spectrometry Standard Reference Material














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

APPLICATIONS OF LASER ABLATION INDUCTIVELY COUPLED PLASMA
MASS SPECTROMETRY FOR THE ANALYSIS OF SOLIDS


By

Melody Bi

August 2000


Chairman: Dr. James D. Winefordner
Major Department: Chemistry

Laser-sampling techniques, such as laser ablation (LA) inductively coupled plasma (ICP) mass spectrometry (MS), have gained popularity for analyzing solid materials because little or no sample preparation is required. Practical concerns such as the difficulty of obtaining or making matrix-matched standards prevent such techniques from becoming analytical tools. This work was largely concerned with studies of calibration strategies for trace analysis in solids and the feasibility of surface profiling by LA-ICP-MS. The capabilities of surface profiling by LA-ICP-MS were also explored.

By analyzing NIST soil and glass samples, a method for the determination of trace element concentrations by LA-ICP-MS using solution calibration and an internal standard has been studied and evaluated. In most cases, the measured element concentrations were within +15% of the certified values. The choice of an internal








standard and the fractionation effect caused by laser irradiance on the surface of the sample were studied and discussed.

NIST archival leaf standards were used for reliable quantitative elemental analysis of Spanish moss samples by LA-ICP-MS. The results were compared with that obtained from microwave digestion (MD) ICP atomic emission spectrometry (AES) analysis. For most of the elements studied, the results for the two techniques agreed. A standard addition method was also studied and the results showed that it is an effective method when matrix-matched standards are not available.

LA-ICP-MS was applied to profiling of patterned thin metal layers on a polymer/silicon substrate. The parameters of the laser and ICP-MS operating conditions have been studied and optimized for this purpose. LA-ICP-MS has good sensitivity and was able to profile thin metal layers on the order of a few nm on the silicon surface. A lateral spatial resolution of 45 ptm was achieved.

Finally, LA-ICP-MS was used to measure the amount of copper contamination on a silicon wafer surface. A detection limit of 6.1 x 10"3 atoms/cm2 was obtained.














CHAPTER 1
INTRODUCTION

Inductively coupled plasma mass spectrometry (ICP-MS) has gained wide acceptance over the last 15 years as a technique for quantitative analysis and trace element detection.1-3 ICP-MS is most commonly used for the analysis of solution phase samples because accurate, multi-element calibration standards are readily available and matrix matching of the sample and standard is generally straightforward. Although ICPMS is a well-established analytical tool for solution analysis, an active area of interest in ICP-MS is the development, characterization and application of alternative methods for sample introduction. Popular methods that have been studied extensively and evaluated to date include: Electrothermal Vaporization (ETV).4-6 Flow Injection (FI),7,8 Direct Injection Nebulization (DIN) 9 and Laser Ablation (LA).10-12

Precise and accurate measurement of multiple elements at low levels is required in a variety of applications. 13-19 The presence of trace elements in soils, for example, can be the identification of contaminants and levels of toxic elements present. In another case, the trace elements in Spanish moss can be an indication of pollution source and levels. Direct analysis of these types of solid samples could provide numerous advantages over solution sampling including,20








" Allows the sample to be analyzed directly and therefore introduces little dilution of

the sample. The dissolution of the sample normally dilutes the sample and reduces

the actual detection limits available.

� Reduced sample preparation time. The time required for some solution preparation

schemes can be extensive, typically several hours to one day. This preparation time may be reduced considerably for solid sample introduction where samples are typically ground to a fine powder and suspended in a fluid (slurry) medium, mixed

with a binder and pelletized or analyzed directly.

* Reduced risk of sample contamination. In general, contamination can be minimized

for solids analysis because sample handling is minimal, and the agents added as

binders, for example, tend to have an organic matrix.

" When using laser ablation as a solid sampling technique, both bulk sampling and

surface profiling can be carried out. In this way, detailed elemental or isotopic

distribution through a sample can be investigated.

Among the most popular direct solid sampling methods are laser ablation, glow discharge (GD),21 secondary ion mass spectrometry (SIMS)22 and spark/arc ablation

(SA).23 GD converts a solid sample directly into an atomic phase by sputtering processes. Because sputtering is a primary sampling mechanism, heating is minimal and preferential vaporization is minimized. GD is best suited for conductive samples although non-conductive samples can be analyzed with a radio-frequency discharge.24-26 A GD source coupled to a high-resolution mass spectrometer is a powerful technology for achieving excellent depth profiling and detection characteristics with limits of detection








(LODs) of 10-100 ng/g. For arc/spark ablation, mass is eroded from the sample in the form of atoms, molecules, vapor, droplets, solid flakes, and particles. Normally, only conductive samples can be used with spark discharges. However, non-conductive materials can be analyzed by mixing the sample with conductive matrix. SIMS has been acknowledged as one of the most important techniques for surface analysis and for elemental imaging.27,28 The strong points of SIMS are its outstanding sensitivity, which is in the ppb range (in the solid); and its extremely fine lateral-profiling resolution, which is on the order of 0.1 to 1 jim.29

The use of pulsed laser ablation sample introduction for ICP-MS has attracted particular interest as a direct solids analysis technique.30-36 The unique characteristics of LA-ICP-MS include the following:

" No vacuum is required in the sample chamber; however, an airtight seal is necessary; " LA-ICP-MS, unlike LA-ICP-AES, separates the ionization step from the sampling

step - the laser is used to ablate the sample only and the material is transported to the secondary plasma source in the torch of the ICP. Therefore, both steps can be

independently controlled and optimized.

" LA-ICP-MS reduces the amount of H20 and other polyatomic species in the plasma. " The high sensitivity of the ICP-MS allows small samples to be quantified, which is

ideal for LA-ICP-MS in that spatial resolution can be used to investigate compositional gradients across a sample. If the sample of interest is the microscopic inclusion in a bulk material, or particles on a filter paper, it is not necessary to analyze








the entire sample. Most of the mass in such cases is not of interest and leads to large

background signals in the analytical source.

Although the relative sensitivity of LA-ICP-MS is poorer compared to solution ICP-MS, because much less material is injected into the ICP, trace detection with submicrogram per gram (ppm) sensitivity in the bulk is routinely achieved. Becker et aL 37 determined Zn, B, Si, Ge, Sn, Sb, P, S, Se and Te in a synthetic GaAs standard using secondary ion mass spectrometry (SIMS), SN-ICP-MS, ICP-OES, spark source mass spectrometry (SSMS), radiofrequency glow discharge mass spectrometry (GDMS) and LA-ICP-MS. For the LA-ICP-MS study, ablation at 20 Hz using 10 ns pulses (266 nm) was used with detection by a double-focusing spectrometer. Inspection of Table 1-1 reveals that LA-ICP-MS competes favorably with other state-of-the-art solid state techniques for these analyses.

A number of limitations of solid sampling include the following:

" Sample inhomogeneity, which can to some extent be overcome when a sample is

brought into solution.

* Particles of solid material can fractionate because the sampling process is remote

from the ICP,

" The most serious problem is calibration and quantitative measurement. This usually

requires that the sample and standard be of similar chemical matrix and physical form. Even when this is achieved, a mismatch of elemental response between the

sample and standard may occur.








" Deposition on the sampling cone can also occur if plasma loading is unacceptable or

if the particles of the solid are too large for the dwell time provided by the sampling

extraction position and gas flow rate.

" LA has poor precision, partly because of fluctuation in the arrival and size of

individual ablated particles transported to the ICP.

Among these, the major obstacle that prevents the widespread applicability of LAICP-MS is the difficulty of obtaining matrix-matched standards. The use of relative sensitivity factors (RSF) obtained from solution to determine the elemental concentrations in solids has been studied for both LA-ICP-OES36 and LA-ICP-MS.38,39 The method has obtained an accuracy of � 20% of the certified values for metals, soils and glass materials.

The present work applied LA-ICP-MS to the elemental analysis of solid samples including soils, glasses, Spanish moss, silicon wafers etc. A solution-based calibration methods for soils and glasses was studied and evaluated. We examined the utility of NIST leaf standards for calibration of Spanish moss samples. We also measured the surface contamination of Cu on silicon wafers by LA-ICP-MS. We studied sampling strategies for powdered materials for LA-ICP-MS analysis. In addition, we explored the surface profiling of thin metal layers on silicon wafer. Optimum experimental parameters were determined for these studies. Future research directions are suggested based on these studies.



















Table 1-1. Elemental concentrations in GaAs measured by different techniques.


ICP-AES 910 � 50 18+-6 <15
<40
<40
45 � 12 (650 + 1001) 390 + 100 420 � 60 110 � 30


Internal standard element. b Possible inhomogeneity


SSMS 870+ 160a
8.2+ 1.5 11+2 1144
4.2+ 1.2 14+4 1290 � 260 720 + 140 315 � 48 108 � 15


rf GDMS 870 i 00a
8.2 � 2.9 7.8 � 2.9

<10

850+ 100 475 + 62 120 22 43 � 14


LA-ICP-MS 870 � 80a 8+-1
7.9 � 3.6 36+ 1 23 � 1 132 � 4

74+� 1 48+� 1 62+� 1


Dopant
Zn
B Si Ge Sn Sb
P
S
Se Te


SIMS 1208 +90 17.8 +1.2 11.7+0.7

13.5 + 2b


450 � 80 400 + 75 113 �27


ICP-MS 827 + 22 19.5 - 0.7 11.5 - 0.8
20.5 + 0.6
6.0 � 0.2 49+� 1 328 � 30 316 � 20 395 � 12 97 � 3














CHAPTER 2
BACKGROUND

2.1. Principles of Laser Ablation

The use of a laser to extract material from a solid sample for analysis can be traced back to the early 1960s. Lasers also have been used an ion source, e.g. laser ionization mass spectrometry (LIMS) and laser microscope mass spectrometry (LMMS).40 Although LMMS is an established technique, its primary advantage of <25 rn spatial resolution is offset by the major disadvantage that it is a qualitative technique. Laser ablation for micro-sampling has been used as a sample introduction technique with ICP-AES for a number of years.41,42 The early work by Thompson et al.41 showed good signal reproducibility for SRM steels. The first application of laser ablation for solid sample introduction into ICP-MS was published by Gray in 1985 43 using a ruby laser with the Surrey research ICP-MS system.

2.1.1. Principles of the Lasers

Laser is an acronym for light amplification of stimulated emission of radiation. It was first achieved by Maiman44 of the Hughes Research Laboratory. The general principles of the laser are illustrated by Figure 2-1. According to the resonance condition (AE = hy) and to the rules of quantum mechanics, an atom can change its energy level, which leads to the absorption or emission of a photon. Stimulated emission is the basis of laser behavior. Stimulated emission leads to the emission of radiation that is coherent








with the incoming radiation. In order to have light amplification in a laser, the number of photons produced by stimulated emission must exceed the number of photons lost by absorption. Light amplification is only achieved when a population inversion from the normal distribution of energy states exists. Population inversion (activation of a laser material) is created by an external pumping source, so that a few photons of proper energy will trigger the formation of a cascade of photons of the same energy. As a consequence of its light-amplifying property, a laser produces spatially narrow and extremely intense beams of radiation with identical frequency, phase, direction and polarization properties.

Figure 2-2 shows the four energy levels of the Nd:YAG (Y3Al5012) laser, which is used in the Finigan Mat SOLA LA-ICP-MS system. In this system, lasing occurs between the meta-stable levels. As the terminal level is essentially empty at room temperature, the population of El can be increased by a relatively small pump power above that of the E3 level. This is a significant advantage over other systems such as the ruby laser. Many of the characteristics that make the Nd:YAG laser attractive in terms of output power, pulse repetition rate, and pulse energy result from the properties of the yttrium aluminum garnet (YAG) host. Large YAG crystals of high optical quality are readily available and the relatively large thermal conductivity together with four-level operation make high average output powers feasible. Nd:YAG systems are used widely because they are relatively simple and cheap, and they are incorporated into several commercial LA systems. Ablation using pulses, usually of 10 to 300 mJ at 1064 nm has become very common. Over the last 5 years, frequency doubling, tripling and quadrupling have been used to produce pulses of wavelengths of 532, 355 or 266 nm and








employed for analysis, especially in geological applications, where laser-sample coupling is often very poor for matrices.

2.1.2. Q-switch Lasers

Stimulated emission depletes the upper lasing level much faster than the pumping rate of the flash lamp. The light output thus consists of many intense spikes. The normal or free-running mode is illustrated in Figure 2-3. Also illustrated is a Q-switched (Qmode) pulse, which is produced by rapidly changing the quality (Q) factor, a measure of the energy storage capacity of the device, by, for example rotating one of the end mirrors. During most of the rotation cycle, the mirror is tilted at an angle, so the two mirrors are not facing each other, and the laser is pumped very strongly with a pulsed flashlamp. Near the end of the pump pulse, a very large population inversion will have been built up in the laser material, but no lasing action can begin because the mirrors are not parallel. At this point, the rotation mirror is stopped parallel to the fixed mirror. Because of the large population inversion, the gain and the available stored energy in the laser rod are very high, and the round trip gain in the laser cavity is much greater than one. Consequently, the laser oscillation builds up much more rapidly than in normal operation and rises very rapidly to a peak power level much above the normal. Thus, a very short, very high power pulse is produced. Laser pulses of 10 to 100 ns and peak powers from 10' to >10' W are so obtained.

2.1.3. Laser Material Interaction

Figure 2-4 shows a conceptual interpretation of laser ablation. Laser-material interactions involve coupling of optical energy into a solid, resulting in vaporization;








ejection of atoms, ions, molecular species, and fragments, shock waves, plasma initiation and expansion; and a hybrid of these and other processes. The interaction between the incoming radiation and the solid sample depend on numerous variables related to the laser, the sample and the atmosphere above the sample. Among these are the wavelength, energy, spatial and temporal profile of the laser beam and the heat capacity, heat of vaporization and thermal conductivity of the sample.45,46 Many models have been developed to describe these processes, but no models completely describe the explosive laser ablation process. Two general descriptions are described on the basis of irradiance, namely, vaporization and ablation.47-52

When the laser pulse duration is microseconds or longer and the irradiance is less than approximately 106 W/cm', vaporization is likely a dominant process influencing material removal from a target. Phonon relaxation rates are on the order of 0.1 ps, and absorbed optical energy is rapidly converted into heat. Heat dissipation and vaporization are fast in comparison to the laser pulse duration. The thermal and optical properties of the sample influence the amount of material removed during the laser pulse. The interaction is predominantly thermal. Melting is common and elements of higher vapor pressure will be enriched in the vapor relative to the concentration in the solid.

At higher irradiances, above 10' W/cm2 with nanosecond and shorter laser pulses focused onto any material, an explosion occurs. Phenomenologically, the surface temperature is instantaneously heated past its vaporization temperature through linear one-photon absorption and multi-photon absorption of the laser pulse duration; energy dissipation through vaporization from the surface is slow relative to the laser pulse width.








Before the surface layer can vaporize, the underlying material reaches its vaporization temperature. Temperature and pressure of the underlying material are raised beyond their critical values, causing the surface to explode. The pressure over the irradiated surface from the recoil of vaporized material can be as high as 10' Mpa.53

In this explosive interaction, melting often is not observed around the crater. Fractional vaporization is normally negligible. Power densities in the 106_ 109 W/cm2 range cause vaporization, ablation, both of these processes simultaneously, or additional mechanisms that have not yet been identified. The incident beam is partially reflected by the sample surface to a degree that depends on the nature of the surface and that decreases as the temperature of the sample surface is increased. An advantage of using giant pulses is the very high rate at which energy is deposited at the surface, which decreases reflection to the extent it can be neglected. Absorption of incoming photons produces photoelectrons and ions. The conversion of incoming energy to heat is very rapid, leading first to melting, then boiling, over the area of laser impact. Evolution of this process depends on the parameters indicated above. Specifically, the time required for the sample to be raised to its vaporization temperature is given by: t,; = KpC (T, - T0)2/4P2,

where K is the thermal conductivity (W/cm K), p is the mass density of the sample (g/cm3), C is the heat capacity (J/g K), To is the initial temperature (K), Tv is the vaporization temperature of the sample (K) and P is the laser power density (W/cm2). As K, C and T, depend on the composition of the sample, tv varies with composition. Hence different elements will vaporize at different rates. Moreover, at the periphery of the








crater, high temperature gradients exist, allowing segregation of elements of high and low boiling point. Consequently, the ablated material, in the form of droplets and vapor, may not have a composition wholly representative of the original sample. Table 2-1 lists some of the mechanisms that have been studied.

2.2. Principles of Inductively Coupled Plasma Mass Spectrometry


In 1980, Houk et al.54 at Iowa State University first demonstrated the combination of an argon inductively coupled plasma (ICP) and a quadruple mass spectrometer for elemental analysis of aqueous sample solutions. The technique, now known as inductively coupled plasma mass spectrometry (ICP-MS), developed rapidly, especially after the launch of the commercial instruments in 1983-1984. It is now a standard method for multi-elemental and isotope ratio analysis of diverse biological and geological samples.55 Recognized advantages of ICP-MS include direct analysis of solutions, calibration against aqueous standards which are readily available, pg mL-' detection limits for many elements, a wide elemental coverage and a linear dynamic range of up to 10 orders of magnitude. A typical ICP-MS is composed of an ICP, an interface system, ion lenses, a mass analyzer, and a detector. These are introduced individually.

2.2.1. Inductively coupled plasmas

The inductively coupled plasma is an electrodeless discharge in a gas at atmospheric pressure, maintained by energy coupled to it from a radio frequency generator. The gas used is commonly argon. A schematic of the torch arrangement in ICP is shown in Figure 2-5. The plasma is generated inside and at the open end of an








assembly of quartz tubes known as the torch. The torch has an outer tube within which there are two concentric tubes that terminate short of the torch mouth. Each annular region formed by the tubes is supplied with argon by a side tube entering tangentially so that it creates a vorticular flow. The center tube, through which the sample is introduced to the plasma, is brought out along the axis. The outer gas flow, i.e., coolant flow, protects the tube walls and acts as the main plasma support gas. The auxiliary flow is introduced to the inner annular space. It is used to ensure that the hot plasma is kept clear of the tip of the central capillary injector tube, to prevent its being melted. The central gas flow or nebulizer flow conveys the aerosol from the sample introduction system producing a high velocity jet of gas that punches a cooler hole through the center of the plasma.

The coupling coil of 2-4 turns of fine copper tube, cooled by a water or gas flow, is located with its outer turn a few millimeters below the mouth of the torch. The RF current supplied from the generator produces a magnetic field that varies in time at the generator frequency so that within the torch, the field lies along the axis. The discharge is initiated in a cold torch by a spark from a Tesla coil, which provides free electrons to couple with the magnetic field. Electrons in the plasma travel around the magnetic field lines in circular orbits and the electrical energy supplied to the coil is converted into kinetic energy of the electrons. At atmospheric pressure, the free electron path before collision with an argon atom, to which its energy is transferred, is only about 1 to 3 mm, and thus the plasma is heated, forming a bright discharge. At the frequency used, the skin effect occurring in RF induction heating ensures that most of the energy is coupled into the outer region of the plasma. The cool injector gas flow punches a channel through








the center of the plasma, carrying most of the sample aerosol, so that little appears in the outer annular part of the plasma. Gas in the center channel is heated mainly by radiation and conduction from the annulus and, while the temperature in the induction region of the plasma may be as high as 10,000 K, in the central channel, the gas kinetic temperature is between 5000 K and 7000 K. Power is coupled mainly into the outer region, which is physically distinct from the central channel through which the sample aerosol travels. Thus the chemical composition of the sample solution can vary substantially without greatly affecting the electrical processes that sustain the plasma. Physical separation between the region where the electrical energy is added and the region containing the sample is one reason for the mildness of physical and chemical interference in the ICP compared to that seen in most other spectro-chemical sources.

As an ion source, the ICP has several valuable properties.

" Samples are introduced at atmospheric pressure, and may be interchanged readily.

* The degree of ionization across the periodic table is relatively uniform and is mainly

singly charged ions.56

" Sample dissociation is very efficient at the gas temperatures experienced and few

sample molecular fragments remain.

" High ion populations of trace concentrations are produced and therefore potential

sensitivity is high.

The main disadvantages are the high gas temperature and the pressure at which the ions are produced which require an appropriate interface design to transfer the ions without significant distortion of their relative populations to a mass analyzer.








2.2.2. Sample Introduction

The ICP requires any sample to be introduced into the central channel gas flow as a gas, vapor or aerosol of fine droplets or solid particles. A wide variety of methods may be used to produce these such as pneumatic or ultrasonic nebulisation of a solution, electrothermal volatilization of microsamples from a hot surface, laser or spark ablation from a solid, and generation of volatile hydrides or oxides from a reaction vessel among others. Most ICP-MS systems are equipped with a pneumatic nebulizer. In a pneumatic nebulizer, a high velocity gas stream produces a fine droplet dispersion of the analyte solution. The larger droplets are removed by a spray chamber which allows only those below about 8 jtm diameter to pass on to the plasma. These small droplets carry only about 1% of the solution, which is usually metered to the nebulizer by a peristaltic pump. Although it is not an efficient sample introduction system, the pneumatic nebuliser is convenient, reasonably stable and allows for multiple sample changes. Other introduction methods are used to meet more specialized requirements.57

2.2.3. Sampling Interface

After formation in the plasma, the analyte ions must be transported into the analyzing mass spectrometer through an appropriate interface system. Figure 2-6 shows the layout of the sampling interface, which is used in the Finnigan Mat SOLA ICP-MS system. It is a three-cone system through which the pressure drops from atmospheric to approximately 10- torr. The first cone, the sampling cone, is located 14 mm from the end of the load coil and has an aperture of 1.1 mm at its tip. The inter-space behind it is evacuated by means of a 3 x 10' mL/s single-stage rotary pump, and with the plasma lit, a








pressure of 2 to 3 torr is maintained in this region of the system. The dynamics of the gas flow through the aperture are such that all the gas within a cylinder of some 4 mm diameter in the plasma flow through the aperture. The effective diameter D of the aperture is given by the relationship:

D = (GoGT)1/2DT

GT is the torch gas flow across a diameter DT and Go is the gas flow through the sampling orifice. Located 8 mm behind the sampling cone is the skimmer cone, which is used to select the center portion of the jet of plasma passing through the first aperture. This cone is located within the zone of silence behind the first cone where the components in the jet are moving at supersonic speed before the position in which the so called Mach Disc is formed. This is a shock wave where the supersonic jet meets the residual gas in the expansion chamber. The aperture at the end of this cone is 0.8 mm, which is sufficient to allow most of the ions sampled from the plasma to pass through to the next stage of the vacuum system while removing the cooler edge of the jet which are subsequently pumped away by the expansion chamber rotary pump.

Behind the skimmer cone is located the accelerator cone. The inter-space between the skimmer cone and the accelerator cone is pumped by a 330 L/s turbo pump which maintains this section of the vacuum system at a pressure of 10-3 torr while the plasma is being sampled. A voltage of -2 kV is maintained on this electrode, and this has the effect of focusing all ions passing through the skimmer cone to a fine cross-over in the 1 mm aperture at the end of this cone. This aperture at the end of the cone also acts as the differential pumping aperture between the intermediate vacuum section and the high








vacuum section of the instrument in which is located the quadruple mass analyzer. The operating pressure in this high vacuum housing is typically 5 x 106 torr for the multiplier.

2.2.4. Ion Lens in ICP-MS

The ion lenses are also shown in Fig 2-6. In each lens, several electrodes are stung together to confine the ions on their way to the mass analyzer. Each lens incorporates a central disc to prevent photons from the plasma from reaching the detector. The juxtaposition of these non-ideal conditions in ICP-MS means that different ion optical conditions are required to transmit ions of different m/z and that the sensitivity for different elements is not as even across the mass range as the high ionization efficiencies of the different elements would indicate. Furthermore, the extent, and possibly even the direction, of the mass discrimination effect depend on ion lens settings and ion energy, the latter of which can be influenced by plasma potential and plasma operating conditions.

2.2.5. Space Charge Effect

Few ions are lost to recombination during the extraction process. Thus, the ion current through the sampler is quite high (-0.1 ptA), and the ion current through the skimmer is normally 1 mA. In the plasma and in the supersonic jet, this ion current is balanced by an equal electron current, so the beam acts more or less as if it were neutral. However, as the beam leaves the skimmer, the electric field established by the lens collects ions and repels electrons. The electrons are no longer present to keep the ions confined in a narrow beam, the beam suddenly is not quasi-neutral, and the ion density is still very high. The mutual repulsion of ions of like charge limits the total number of ions








that can be compressed into a beam of a given size. This is called the space charge effect. The space charge effect should be substantial in the ICP-MS at total beam currents on the order of 1 ptA. The ion current causes space-charge effects that are reasons for the nonideal behavior of ion optics in ICP-MS. Also, even a small change in the total ion current caused by the addition of just a modest amount of matrix element can change the fraction of analyte ions that get through the lens. Heavy matrix ions are themselves deflected less and stay closer to the center of the ion beam where they can do the most damage. Therefore, the space charge effects are a major cause of matrix interference in ICP-MS.

2.2.6. Quadrupole Mass Spectrometer

A diagram of quadruple mass filter is shown in Figure 2-7. Four straight metal rods or metallised surfaces are suspended parallel to and equidistant from the axis. Opposite pairs are connected together. DC and RF voltages of amplitude U and V, respectively, are applied to each pair. The DC voltage is positive for one pair and negative for the other pair. The RF voltages on each pair have the same amplitude but are opposite in sign. The ions to be separated are introduced along the axis into one end of the quadruple structure at the velocities determined by their energy and mass. The applied RF voltages deflect all the ions into oscillatory paths through the rods. If the RF and DC voltages are selected properly, only ions of a given m/z ratio have stable paths through the rods and will emerge from the other end. Other ions will be deflected too much and will strike the rods and be neutralized and lost there.








2.2.7. Mass Detectors

The magnitude of the ion beam emerging from the quadruple mass analyzer is measured using either a DC Faraday amplifier or a pulse-counting electron multiplier. The arrangement is shown in Figure 2-8. Choice of collector depends on the size of the beam to be measured, and the appropriate collector can be selected electronically by deflecting the beam onto either the Faraday plate or the entrance aperture of the channeltron. The Faraday detector is based on electron counting. A 1-count-per-second (cps) signal on the Faraday cup corresponds to 5 x 10"7 A, which is equivalent to 6 x 10"0 ions per second arriving the collector.

Figure 2-9 shows a diagram of a channeltron electron multiplier. An open glass tube with a cone at one end is used. The interior of the tube and cone are coated with a lead oxide semiconducting material. The cone is biased at a high negative potential and the back of the tube near the collector is held near ground. Relative to either end, the resistance of the interior coating varies continuously with position. When a voltage is applied across the tube, a continuous gradient of potential exists with position inside the tube. Suppose a positive ion leaves the mass analyzer and is attracted to the high negative voltage potential at the cone. When the ion hits the surface, one or more secondary electrons are ejected. These secondary electrons hit another section of the coating and more secondary electrons are emitted. Thus, a discrete pulse containing as many as 108 electrons at the collector after an ion strikes the mouth of the detector.

The largest current that can be measured on the system is normally 108 A, which is equivalent to 6 x 10" ions per second arriving at the collector. For signals below 10"5








A, the multiplier is the preferred detector device. Each ion arriving at the front aperture of the multiplying element creates an electron cascade down the curved multiplying tube, resulting in a pulse of 10' to 108 electrons being released at its output. The large flux (106 ions per second) should not be allowed to reach the multiplier as these large flux beams can rapidly destroy the multiplier

2.3. Principles of LA-ICP-MS

The roots of LA-ICP-MS lie in part with the logical extension of previous experiences with LA-ICP-OES58,59 and in part with the anticipated analytical advantages mentioned in the introduction section. Figure 2-10 shows a typical schematic of the LA-ICP-MS system. When the laser beam is focused on a sample, interaction between the laser beam and the sample allows the conversion of photon energy into thermal energy, which is responsible for the vaporization of most of the exposed solid surface. The material ablated is swept away with an argon stream to an inductively coupled plasma mass spectrometer (ICP-MS) and is analyzed.

The laser may be mounted vertically or horizontally. In the latter, the beam is deflected onto the sample cell by a 450 mirror, typically incorporated into a viewing microscope. Optical and visual focusing may or may not be adjusted to coincide. Often the sample may be viewed remotely via a camera.

The sample sits within a glass ablation cell, the upper face of which may be slanted at 45' to the vertical to reduce back reflection. For use with ultraviolet lasers, a fused silica window must be installed in the cell.60 A platform or turntable, which is usually under computer control, allows positioning of the sample in the X, Y, Z








directions. Displacements as small as a few micrometers may be employed and the laser firing arranged to occur only between sample movements. According to the pattern of laser shots, individual features such as mineral grains in rocks may be analyzed, or depth profiling, line profiling, area or bulk analyses are possible.

The cell is fed an inert gas to ntrain the ablated material; this is the equivalent of the nebulizer flow in conventional solution nebulization ICP-MS. Argon or occasionally helium is employed; the latter is reported to improve scavenging.61 The sample remains at atmospheric pressure and this is an advantage over many other analytical techniques, such as X-ray photoelectron spectroscopy, where the sample must be under vacuum to permit analysis. As air must not enter the ICP, an arrangement for purging the cell to atmosphere must be included, usually by a three-way valve in the cell to ICP transport line.

2.3.1. Calibration Strategies

Various strategies are explored to obtain elemental concentrations in solids by LA-ICP-MS. Quantification using a single-point calibration, based on the sensitivity obtained by analysis of a sample containing an element of known concentration, is simple but generally inaccurate. A more sophisticated approach was developed by Hagger et al. using elemental response factors determined by solution ICP-MS and modified based on element-dependent volatilization efficiencies.62 This method, derived from a thermal model of the ablation process, requires either knowledge of one analyte concentration and the ablation temperature TA or two or more analyte concentrations from which TA can be








determined. Although external standards are not necessary, the obvious disadvantage is that the concentration data are not fully quantitative.

From sensitivities obtained from analysis of a single multi-elemental standard, a response-mass curve may be constructed, allowing elements other than those of known concentration in the standard to be determined. The curve needs to be corrected using a Saha factor which accounts for differences in the degree of ionization in the ICP. The agreement between the determined and accepted concentrations of Mg, Mn, Fe, Cu, Zn, Sr, Ba and U in two limestone reference materials is within � 10%.31

Element-for-element calibration against sensitivities obtained by analysis of an external well-characterized matrix has been used for the multi-elemental analysis of diverse biological 63,64 and geological samples.65 Numerous difficulties, however, are associated with this approach: (i) the samples and standard need to be matrix-matched,

(ii) the standard must be well characterized and must contain all the elements of interest at concentrations sufficient to give good sensitivity and (iii) the standards must be readily available and inexpensive.

Much effort has been expended in finding ways to produce multi-element solid standards cheaply and rapidly. Approaches include the addition of individual compounds to a powdered matrix, mixing and pressing with (or without) a binder,66 the addition of liquid standard solutions to a powdered matrix, drying and pressing,67,68 the production of glass fusions69 and the production of sintered compacts.70

One area where calibration is easier is metal analysis, for which more standards exist. Moreover, complete matrix matching of metals is sometimes unnecessary. For








example, Fe, Ni, Cu, Zn, Pd, Pt, Pt, Pb and Bi have been determined in gold, silver and gold-silver alloys using either gold or silver reference materials as standards.71 The ease of ablation and the relatively homogeneous distribution of elements in metals simplify metals analysis.

An alternative to solid standards is liquid-based calibration in which a dual flow system allows simultaneous introduction of laser ablated solids and nebulized aqueous solutions. As both forms of material are present in the ICP, calibration against the sensitivities obtained by analyzing aqueous standard solutions is possible.72 An obvious disadvantage, is the presence of oxide and hydroxide interference associated with the "wet" plasma.

Owing to variations, for example in laser energy output between shots, sequential responses obtained under the same conditions are not identical. To compensate for such fluctuations, internal standardization (IS) is almost always necessary in LA-ICP-MS analysis. This is normally conveniently done using the responses of a minor isotope of a major element, which may be expected to be relatively uniformly distributed in the matrix. Improved precision results from IS on a set of replicate analyses; IS from standard to sample should also improve accuracy and is possible when the concentration of at least one suitable element is known. In addition to being homogeneously distributed, an IS element should behave in the ablation process and in the ICP in a manner identical with that of the analytes.

An alternative to IS is to normalize elemental responses to the amplitude of the acoustic wave generated by laser ablation. This improves the precision of concentration








data obtained by LA-ICP-AES.73 Compensation for variation in transported mass may also be made using light-scattering measurements obtained using a photomultiplier tube (PMT) mounted above the cell to ICP transport line perpendicular to the beam of a He-Ne laser.74 Although not as effective in improving precision as the use of IS, this method requires no previous knowledge of the sample homogeneity or elemental composition. A similar system denoted an in-line mass transport measurement (MTM) cell and employed a UV/VIS spectrophotometer.75 Variations in the integrated total scattering effect of the ablated particles, obtained by summing the sub-peak areas of the intensities produced by this scattering, were used to normalize the analytical results. Accuracy was improved and precision (RSDs) lowered from 25 to 5% using this normalization in determination of multielements in arsenopyrite.

















Table 2-1. Laser ablation mechanisms. " Absorption (single, multiphoton, defect initiated, ...) " Reflection (time-dependent)
* Thermodynamics (melting, latency, phase change,...) " Plasma ignition
* Shock waves (gas) " Stress waves (solid)
* Laser-plasma interaction (inverse bremsstrahlung,...) " Plasma radiation/heating " Gas-dynamic expansion " Hydrodynamic expansion
























Stimulated Emission


Energy state 2 ',, Energy state I


Light Attenuation By absorption


NON-INVERTED POPULATION


Light Amplification By Stimulated Emission


INVERTED POPULATION


Figure 2-1. Principles of laser action.


4






























Radiation by stimulated emission Metastable level


Figure 2-2 Energy levels in Nd:YAG laser















Partially reflecting mirrow


Lfl


Laser rod


I


Q-swtiched (Q-Mode)


Free-running (N-Mode) pulse Time


Figure 2-3. Q-switch laser


Laser output


100% Mirror


























Atoms, Ions, Molecules, Pulsed Laser Beam Clusters, Particles












Target


Figure 2-4. Conceptual drawing of laser ablation

















analytical zone


/ induction region load coil








/II
~outer gas flow torch /







aerosol gas flow (into axial channel) Figure 2-5. Torch layout in ICP.

















Focus Accelerator cone Y steer
Phase match lens X-Y deflectors Skimmer cone Samplecone





plasma

















Analyser turbo Intermediate turbo Rotary pump


Figure 2-6. Interface in LA-ICP-MS.
















Detector


resonant ion nonresonant ,on


Ad


dc and ac voltages


Figure 2-7. Quadrupole mass analyzer.


Source


-"-















Faraday collector and amplifier


Amplifier and discrimnator for ion counting


Figure 2-8. Detectors in ICP-MS.


Bea n deflectors



















Secondary Electrons
Preamplifier











-3 kV







Figure 2-9. Channeltron detector.













CCD camera


Lens


mirror


ICP torch


sample cone


solution


nebulizer


Figure 2-10. Schematic of LA-ICP-MS system.


Laser beam


Argon in














CHAPTER 3
INSTRUMENT DIAGNOSTICS AND OPTIMIZATION

3.1. Introduction

The purpose of this chapter is to provide a better understanding of the Finigan LA-ICP-MS system. Information that are additional to the knowledge from the manual and other descriptions are provided. The study was focused on the laser ablation system since a detailed study on the ICP-MS operating conditions were described in the dissertation by Baker.76 In the present work, the influence of laser output energy, flashlamp voltage, laser focusing conditions or mass removal etc. were studied and discussed. The effect of intermediate flow rate and forward RF power on the LA-ICP-MS analysis of solids and solutions were also studied and discussed. This information was provided for general purpose; however, more specific information or the effect of modification of the system on particular measurements can be found in the subsequent chapters. The parameters were studied based on measurements performed on NIST glass and High Purity Solution (Charleston, SC) standards. The goal is to provide information that is applicable to all matrices.

3.2. Experimental

3.2.1. Instrumentation

An illustration of the LA-ICP-MS setup is show in Figure 2-10. It consists of a Finnigan MAT SOLA ICP-MS (Hemel Hempstead, UK) and System 266 laser ablation








accessory. The ablation module consists of a Spectron SL 401 Nd: YAG laser which has been frequency quadrupled to produce an output beam of 266 nm. A CCD camera is used for remote viewing of the sample. The CCD camera is mounted in parallel with the laser; however, the focusing of the CCD camera is different from the focusing of the laser beam which will be discussed in this chapter. The system also includes an x-y-z translation stage for adjusting the focus of the laser beam and for selecting the area of the sample for analysis. This system was modified to allow for x-y-z translation of the sample while the laser is fired repetitively.

Figure 3-1 shows the schematic of Finnigan 266 laser ablation system. The general components and their functions in the laser are described as follows:

A single flashlamp is used in the cavity and amplifier. It is located in the lamp housing above the rod together with its ceramic reflector. The lamp has leads which terminate at fixing posts mounted on the outside of the housing together with the lamp supply leads. The pockels cell compromises a cube polariser, a wedged KD*P crystal and a glass compensating block. The telescopic oscillator is used to obtain a low beam divergence by discriminating against the higher order modes within the cavity together with efficient energy extraction. It also has a dual purpose in compensating for thermal lensing of the laser rod at different repetition rates. The fundamental wavelength output from a Nd:YAG laser is 1064 nm (Infrared). It is quadrupled by propagation through suitable nonlinear harmonic generating crystals to produce wavelength of 266 nm (Ultra Violet). The harmonic wavelength generated propagates collinearly with the fundamental output. Therefore, as the conversion efficiency is not 100%, a beam separator is used to separate the harmonic output from the residual fundamental output. In order to maximize








the conversion efficiency of harmonic generation, the input beam must propagate through the crystal along a unique axis with respect to the crystalline axis. This phase matching angle is also dependent upon the crystal temperature. Consequently, all crystals are mounted in temperature stabilized ovens, held at elevated temperature.

The 266 laser system has repetition rates of up to 5 Hz with pulse energies of 1-10 mJ and pulse widths around 8 ns in duration. When focused to a 50 pm spot, laser irradiances of -8x108 - 8x 10 1 W/cm2 are achieved. The efficiency of detection for all analytes is a function of the ablation process, transport process, ionization in the ICP, and mass spectrometer response. In general, sample translation during ablation provided more representative sampling of the surface; however, it also resulted in a larger mass removal rate and subsequently higher sensitivities.39

For analyses, the sample is placed in an ablation cell consisting of a 6 cm glass tube with a quartz window for transmission of the UV laser beam. The cell has a total volume of -80 cm3. The ablated material is transported into the ICP through -1.0 m of 3/16' i.d. plastic tygon tubing with a stream of argon flow. In addition, the aerosol output from a concentric nebulizer and cooled Scott-type spray chamber can also be continuously introduced to the ICP-MS during ablation with a glass Y-connector. The ICP is typically operated between 950 and 1300 W with coolant and intermediate flows of 15 L/min and 0.9 L/min, respectively. Sample is carried to the plasma, whether from the ablation chamber, nebulizer, or a combination of both, with a total argon flow of -1 L/min.









3.2.2. Samples

The samples for this study include National Institute of Standards and Technology (NIST) (Gaithersburg, MD) glass 611 and 612. Solution samples were diluted from High Purity standard solutions.

3.3. Results and Discussion

3.3.1. Laser Power

A single laser shot produces a transient response;77 repetitive laser shots of a few Hz produce quasi-stable responses, which are more useful for quantification.39 The current work was done with multiple laser shots only. Typically, greater laser power produce greater ablated masses, greater transport to the ICP and hence greater responses for a constant analyte concentration. Sufficient response for good quantification (accuracy and precision) is required without saturation of mass peaks at m/z values of interest. Saturation of individual peaks may be tolerable if other elements present at trace concentrations are the only analytical interest. Excess ablated material, however, will have several deleterious results, including memory effects, cone blockage and plasma disequlibrium. Therefore, good control of laser power is necessary for obtaining the necessary sensitivity.

The laser power can be adjusted by changing the flashlamp voltage. However, the change of laser power is not linearly corresponding to the change of flashlamp voltage. The laser power was measured using a power meter (Scientech 361). The correlation plot is shown in Figure 3-2. The mass removal of Zn corresponding to the change of flashlamp voltage is shown in Figure 3-3. The plot shows an increase in mass removal with increasing flashlamp voltage.









The laser power at wavelength 1064 and 266 is measured by an oscilloscope (Tektronix, TDS 620A). It was found the voltage reading from the oscilloscope is directly proportional to the laser power, however, no effort was made to calibrate the reading for the actual laser power. It was found at the primary wavelength, 1064 nm, the laser power was stable when averaging 4 laser shots over a 2 hour period of time (Figure 3-4). However, when the laser frequency quadrupled to 266 nm, the laser power showed significant fluctuation when averaging 32 laser shots (Figure 3-5). A more stable laser power versus time plot was obtained when averaging 128 laser shots (Figure 3-6). This has indicated that it is necessary to apply over 128 laser shots for any analysis to obtain a reasonably stable signal. An apparent flicker noise is seen from Figure 3-6. It was found that the laser power dropped steadily over a period of time, e.g., -8% in half an hour time period. This corresponds with the signal decrease when measuring pure copper materials on ICP-MS. The depth and width of the laser track was measured using profilimetery. The mass removal was therefore estimated. The ICP-MS signal of copper was measured at the same time. Figure 3-7 shows the spectral intensity of Cu signal vs. mass removal over a 30 min period. The results show that the signal directly corresponds to the mass removal by laser ablation for pure copper samples. This phenomenon of power decreasing along with the time is most likely due to the temperature increase of the thermostat for the flashlamp, however, no definite reason is known. Because of the decay of the laser power, an internal standard is necessary to correct for the mass ablated for any quantitative analysis. In case an internal standard is not available, the laser power needs to be monitored and corrected.









3.3.2. Laser Focusing Positions

The focusing of the laser is crucial to the sampling spatial resolution, and affects sampling sensitivity. Defocusing the laser in order to sample larger areas of a target was critical for either enhancing the signal or obtaining more representative sampling of a heterogeneous sample. The Finnigan Mat LA-ICP-MS system has an optical system to focus the laser beam (UV light) on the sample surface and an optical system to focus the visible light on the sample so that the movement of the sample can be monitored. A digital displacement of the laser focusing position is designated as "d". The two optical systems share the same lens which can be adjusted; however, the focusing point of the UV light and visible light did not coincide. The difference was measured by firing the laser beam at a piece of carbon paper. The focusing lens was adjusted to produce different focusing distance according to the digital display. The size of the laser beam was measured by measuring the spot size in the carbon paper left by laser beam. At the smallest beam size, the laser is focused on the surface of the object. A plot of focusing distance vs. spot size was obtained (Figure 3-8). When the digital display is 0.0, a sharp image of the sample is seen on the monitor. The plot has indicated that there is a 2 mm difference of the focusing distance of the visual image and the UV light.

3.3.3. Analysis of SiC

A series SiC particles of different sizes were used to study the smallest individual particles feasible for the LA-ICP-MS analysis. The SiC particles were spread out on a foam tape so that single particle could be seen under the microscope on the monitor. The laser was repetitively fired at individual SiC particle of different mesh. The response of









the Si signals on mass spectra were plotted vs. the particle size (Figure 3-9). This work demonstrates that particles larger than -20 um (approximately focus laser beam diameter) should be used to maintain a constant mass signal. For smaller particles (<16 jim) mass signal intensities gradually decreased indicating that only a fraction of the laser radiation effectively vaporizes the particle and the rest of the laser radiation interacts only with the substrate material. Also, the transport of the ablated materials might not be efficient when the particle size is below 16 jtm.

3.3.4. Plasma Operating Conditions

A detailed study on the plasma operating was described by Scott Baker.76 The present study is meant to provide additional information about the operating conditions of the plasma for LA-ICP-MS analysis. The effects of intermediate flow rate and forward RF power on solution and solid analysis were studied. It was found that in multi-element analysis a compromise is inevitable since maximum responses do not necessarily occur at the same flow rate and RF power. This is illustrated in Figure 3-10 and 3-11, which shows the sensitivity of different elements at different intermediate flow rate and forward RF power for glass 611, 612 and 20 ppb multi-element solutions. In general, RF power at 1300 V produces the best sensitivity for a glass sample, and RF power at 1200 produces the best sensitivity for solutions standards. The sensitivity of each element reaches its maximum at different intermediate flow rates. These results indicate that elements in solids and in solution do not necessarily have their best sensitivity under the same plasma conditions. Tuning with a standard solution or solid materials is necessary if maximum sensitivity is required for a certain element in solid or solutions samples.













Pockels cell


Rear mirrow assembly


utput coupling mirrow and shutter Mode controlling ape e


Intra cavity telescope


Oscillator pumping chamber


450 comer beam steering mirrow


-U.. .... ... ..... ..


J
~ I


H




- Harmonic separator (prism)


Dichroic mirror


Beam dump


U


beam rotator Second harmonic generator

-Third/Forth Harmonic generator


Figure 3-1 Finigan SOLA laser ablation system layout.







44




0.18

0.16 x I

0.14 1

0.12

0.10

0.08 o 4 .. 0.06

CO 0.04
0.02
.4

0.00

-0.02
9 10 11 12 13 flash lamp voltage (kV)



Figure 3-2. Laser power vs. flash lamp voltage.




0.014 0.012 o 0.010 0)
E 0.008
o o 0.006


-) 0.004


0.002 x


0.000 , I
8.4 8.6 8.8 9.0 9.2 9.4 9.6 flashlamp voltage (kV)



Figure 3-3. Weight loss of Zn metal vs. flashlamp voltage.




















150140

- - 1 3 0 >, 120- 110. 100N 90
0
- 80
C )
-2 70-


0 U


U N U.N


0 U


Figure 3-4. Laser power measured by oscilloscope vs. time at 1064 nm.


I I I I .I I I I I . I . I . I 0 10 20 30 40 50 60 70 80 time (min)
























0 0


9 0


0 M N


"I * I * I 1I I * t
0 2 4 6 8 10 12 14 time(m in)


Figure 3-5. Laser power measured by oscilloscope vs. time at 266 nm averaging 32 laser shots.





-1.5 1 E


-2.0


-2.5


-3.0


-3.5


-4.0


-4.5-


U


U
U
E.g


U U
U
U


0 E 0


U
U
U


-5.0 1
I *** I I 0 2 4 6 8 10 12 time (min)

Figure 3-6. Laser power measured by oscilloscope vs. time at 266 nm averaging 128 laser shots.
























0 Cu signal



0 mass removal


2.OOE+07 1.80E+07 1,60E+07


1.40E+07 1.20E+07 1.OOE+07 8.OOE+06 6,OOE+06


4.OOE+06 2.OOE+06 0.GOE+00


0 2 4 6


10
time (min)


12 14 16 18 20


Figure 3-7. Weight loss of copper and signal intensity vs. time.


e


2000


1800 1600


1400 1200
E

1000 E 800 600













0.30 0.25


0.20


0.15


0.10


0.05-


M U 0


M M


II * I * I * I * I * I *
-4 -3 -2 -1 0 1 2 3
d (- toward sample), (+ away from sample)


Figure 3-8. Laser focusing position vs. diameter of laser beam size. At d = 0, a sharp image is seen on the monitor. At d = -2, the laser (266 un) is focused on the surface of the sample.



















































0 200 400 600

micrometer


800 1000 1200


Figure 3-9. Intensity of Si in LA-ICP-MS measurement vs. particle size.


3.5x10' 3.0x10 2.5x105 2.0x105 1.5x105 1.0xlO 5.0x104

0.0x

-5.0xl 0'


1~

















-0 0 CUA



M n -----BN0
V _ V V V Pb


1100


1200


13 00


1400


RF power (v)


Figure 3-10. Normalized intensity of 20 ppb multi-element solution vs. forward RF power.


1400


RF power (v)


Figure 3-11. Normalized signal intensity vs. forward RF power for glass 611.


1 000


Mn Pb


0(l


900


0 ...







51





Co
2.5



( 2.00 . - 1.5 Cu




z Mg


0.5- Ni=I�I
NN i





0.4 0.6 0.8 1.0 1.2 1.4 Intermediate flow rate (mL/min) Figure 3-12. Normalized intensity vs. intermediate flow rate for 611.


70
C:









60




~Co 5I9 N M = 30
0







Z 202
Cu 63


10 i 600.4 0.6 0.8 1.0 1.2 1.4 Intermediate flow rate (mL/min) Figure 3-13. Normalized intensity vs. intermediate flow rate for 20 ppb multielement solution.














CHAPTER 4
STUDY OF SAMPLING STRATEGY FOR LA-ICP-MS ANALYSIS

4.1. Introduction

The preparation of solid samples can be time consuming, but much less so than many commonly used dissolution procedures, which yield good quantification in the LAICP-MS analysis. Contamination during the production of solid samples is not usually observed, although problems occasionally arise. For example, tungsten from the lining of a tungsten mill has been observed to contaminate sediments during ball-milling.65 Sample preparations for laser ablation normally involves bringing the sample into powder form so as to obtain homogeneity. The powders are then pressed into pellets under a pressure of 35 Mpa (-5000 Psi) with or without binder. The pellets need to be pressed one by one in laboratory press and care needs to be taken when washing and cleaning the molds to reduce contamination. This can be troublesome when many samples are needed to be prepared. Also, pressing pellets involve the risk of contamination from the container. Pressure need to be carefully controlled when pressing each pallet in a set of samples. The differences in pressure may result in differences in the surface properties and thus can cause inaccuracy of the measurements. Adding binders for materials not easily pressed into pellets also involves the risk of contamination.

An alternative method of sample preparation for laser ablation has been used in LIBS (laser induced breakdown spectroscopy) study.78 The powdered materials were








placed on a double-sided tape stuck to a microscope slide. Good precision and accuracy were obtained. No significant background signals were found for the elements of interests. The IR laser used for LIBS study showed no sign of coupling with the tape or glass slides as substrates. Using tape as a sample holder for powdered samples is more convenient compare with pressing pellets provided no problems arise with the background from the tape. Less time for sample preparation is required and no binder is necessary for any powered materials or pressure control.

In this study, the feasibility of using tapes as sample holder for LA-ICP-MS study was studied by analyzing soil and leaf standard materials. Three types of tapes were studied for this purpose.

4.2. Experimental

The LA-ICP-MS system has been described previously (Chapter 3). Table 4-1 lists the typical ICP-MS operating conditions used in this work. A JSM-6400 Scanning Electron Microscope was also used in this study. It is a high-resolution scanning electron microscope with a modem digital image processing system.

NIST leaf samples (1515, 1570, 1573 and 1547) and soils samples (2709, 2711 and 2710) were used in these studies. Three tapes were tested for their feasibility of being used as sample holders. They are double sided transparent tapes from supermarket

(I) doubled sided foam tape (II), and specially-made double sided transparent tape (III) designed for research. They are all products of 3M company. A small -2 cm x 2 cm tape sample holder was used to contain loose powders of leaf or soil samples. One side of the transparent tape was stuck to a microscope glass slide. The powdered samples were








spread out on the tape using a glass slide so that only a mono-layer of the samples was left on the tape. Excess powders were tapped off the tapes. All the samples for the SEM study were coated with Au-Pd.

4.3. Results and Discussion

The general considerations in this study were,

* The laser output energy must be high enough to produce sufficient sensitivity for

trace element detection; however, the laser beam can not penetrate the tape, in which

case, the glass slides will be sampled and included in the background,

* If the laser beam couples with the tape, it is necessary that the tape does not produce

significant background signal for the isotopes of analytical interest. In other words, the tape should not contain that element, or the signal from the tape is constant and can be corrected as background, in which case, the element should homogeneously

present in the tape.

For a laser output energy of 7 mJ as in analyzing pellet samples, the laser track can be seen breaking the transparent tapes and coupling with the glass substrate beneath. High intensity signals of Si, Mg, Ca were observed in mass spectra indicating that the UV laser coupled efficiently with the polymer (tape) and glass materials. A series of laser energies were tested; if the laser output energy was adjusted to -2 mJ, there were no sign of the laser beam breaking through the tapes into the glass substrates.

Figure 4-1 and 4-2 are the SEM images of the powdered leaf SRM 1547 on tape (III) before and after the ablation. Figures 4-3 - 4-8 show tapes (I), (II) and (III) before and after laser ablation. A 2 mJ laser output energy was used. Comparison of the structures of the laser track in Figure 4-2 and Figure 4-8 reveals that at this output energy,









the laser beam ablated the tape to some extent. However there were no signs from the ICP-MS measurement that the glass materials were ablated.

The elemental content in the tapes was studied by measuring a few chosen elements; their intensities were normalized to the background. Figure 4-9, 4-10, 4-11 show the results obtained from the three tapes. It was found that for tape (I), all the elements chosen in the study showed higher intensity compare to the background. Also, the relative standard deviations (RSDs) were high (>40% in most cases), which indicated that the foam tape had serious elemental contamination in the matrix which could not be corrected since these elements were not evenly distributed in the matrix of the tape. The results for tape (II) indicate even more significant elemental contamination compare to tape (I). In both cases, elemental contaminants include Mg, Ca, Ni, Cu Zn and Pb. The results from tape (III) showed that the elemental contamination were very low and the RSDs were low for all the elements measured. The RSDs were relatively higher for Ca, Ba and Pb because the signals were similar to the background and therefore showed more fluctuation. However, Mn and Sb were the major additives in the production of the tape (III) according to the manufacturing information. In the case of Ca which showed s higher intensity than the background, a relative standard deviation of 6% was obtained, which indicated that the elemental distribution in the tape was homogeneous. Therefore, the Ca contamination could be corrected by subtracting the background. Only serious contamination for this tape was Mn and Sb (not shown on the figure because the intensity was out of range on the figure) which were major additives during the manufacture. Therefore, only when Mn or Sb are the elements of analytical interest in the powdered sample, tape (III) can not be used as a sample holder. However, care needs to be taken to








check the elemental contamination for the isotope of interest before using it as a sample holder. Comparison of Figures 3 through Figure 8 shows that tape (III) has cleaner and smoother surface feature compared with the other two tapes.

Tape (III) has been used as samples for soil and leaf samples. Several elements were chosen for the measurements. Since only a monolayer of powdered samples are left on the tape, it is important that the amount of mass ablated produce sufficient sensitivity and precision. The rsds obtained were close to that from measuring pellet samples when an internal standard was used. Therefore, tape (III) has been shown to be a feasible alternative sampling means for pressing pellets from powdered samples. Tape(I) and (II) were not feasible for this purpose because of high elemental contamination.

4.4. Conclusions

The use of tape as sample holder for powdered materials was studied. Three tapes obtained from different sources were studied. The special made tape (III) for research use purpose was demonstrated to be feasible for using as sample except when Mn or Sb were elements of analytical interest. However, the tapes obtained from super market were not feasible for this purpose because of significant elemental contamination. The use of a tape as a sample holder can save time and reduce the risk of contamination during sample handling in sample preparation compare with conventional method of pressing pallet. Also, the surface conditions can be kept constant when a monolayer of powdered materials are spread out on the tape.














Table 4-1. Typical LA-ICP-MS operating conditions.


Kt power
Coolant gas flow rate Auxiliary gas flow rate Nebulizer gas flow rate Ablation chamber flow rate Solution uptake rate

Scan conditions
ICP-MS-


Laser


Detector Scan range per isotop Number of passes Number of channels p Dwell time per chann

Energy per shot Repetition rate Rastering speed


IUU W 15 L/min 0.9 L/min
0.65 L/min 0.35 L/min 1.0 mL/min



electron multiplier e 0.25 amu
128
er amu 8 ms el 4 ms

-2 mJ
5 Hz
15 4zrns
































Figure 4-1. Tape (III) with leaf SRM 1547 unablated.


Figure 4-2. Tape (III) with leaf SRM 1547 ablated.
































Figure 4-3. Tape (I) unablated.


Figure 4-4. Tape (I) ablated.































Figure 4-5. Tape (II) unablated.


Figure 4-6. Tape (II) ablated.

































Figure 4-7. Tape (III) unablated.


Figure 4-8. Tape (III) ablated.







62






















mi ii mi





Mg Ca V Ni Cu Zn As Cd Ba Pb element


Figure 4-9. Elemental signals normalized to background for tape (I).


Mg Al Ca Co Cu Zn Rb Sr Ba Pb element

Figure 4-10. Elemental signals normalized to background for tape (II).








63


















































V Ni Cu Zn As Cd Ba Pb element


Figure 4-11. Elemental signals normalized to the background for tape (III).


� 1 1.5

0


Mg Ca














CHAPTER 5
STUDY OF SOLUTION CALIBRATION OF NIST SOIL AND GLASS SAMPLES BY
LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY

5.1. Introduction

Laser ablation provides significant benefit especially for the analysis of complex solid samples, such as ceramics, glasses, soils or acid resistant alloys, which are difficult or impossible to prepare for solution sampling. However, there are a number of problems that have prevented LA-ICP-MS from becoming a routine technique for the analysis of solid materials. One major concern has been selection of matrix-matched standards for calibration. Extensive studies have been made to calibrate the analysis of solid materials introduced by laser ablation without using reference samples. A number of approaches using desolvated aerosols generated by pneumatic nebulization, have been tested as alternative sample introduction systems for calibration.79-85 Hager85 developed a model that used response factors determined from solution nebulization and modified them based on element-dependent volatilization efficiencies. This work reported accuracy of �50% for aluminum, steel and copper standards. The use of dual-sample introduction appears to be a more promising alternative for analyzing materials without using matrix-matched standards. It was first used for quantitative LA-ICP- atomic emission spectrometry.42 However, some good results have been obtained by applying the method to LA-ICP-MS. Cronwel138 et al. applied the method using element relative








sensitivity factors (RSFs) to analyze steel Standard Reference Materials. An accuracy of �20% in general was obtained. We have applied the RSF method in this laboratory to determine the concentrations of trace elements in soil samples.39 The results obtained were mostly within �20% of the certified values using Ni or Ag as an internal standard. The internal standard proved to be critical in the measurements because of the potential problem of fractionation during the laser ablation. It was observed that the results for some elements were significantly different when different internal standards were used.

This work reports the application of a solution method by matching the signal intensity of a matrix element with a solution of a known concentration. A series of solutions with the same concentration of the matrix element but different concentrations of trace elements were made as standards. The method was reported by Falk86 et al. when they applied it to standard metal samples. An accuracy of � 10% was obtained. Metals were considered to be less of a problem because of their uniform chemical environment and their efficient coupling with laser energy. In this chapter, the feasibility of applying solution based calibration of more complex matrices such as soil and glass have been explored.

Three normalization factors were introduced in the current work: (1) correction for the variation of ICP-MS operating conditions; (2) correction for the response of each element measured in the solid sample for the calibration by solution standard, and correction for variation of laser power, and (3) correlation of signals from solution and solid for the calibration of trace element in solids. The homogeneity of the matrix








elements as well as some trace elements were studied by LA-ICP-MS as well as by laser induced breakdown spectroscopy (LIBS).

5.2. Experimental

5.2.1. Instrumentation

The experiments were performed with a Finnigan MAT (San Jose, CA, USA) SOLA ICP-MS and Finnigan MAT System 266 laser ablation accessory. Typical operating conditions for the ICP-MS are listed in Table 1. A combined flow of nebulized solution and carrier gas from the ablation chamber was introduced to the ICP-MS for all sample measurements. This provided dual channel introduction to obtain matched plasma conditions for the ablated material and a nebulized aqueous solution standard.

The laser was a Nd:YAG with a 266 nm output as described in Chapter 3. It operated at 5 Hz with a typical pulse energy of -7 mJ and pulse width of 8 ns. In this laser system a monitor and UV laser share a common lens having a focal length of -7.3 cm. A spot size of-50 gm in diameter is produced when the laser beam is focused on the surface of the sample. The irradiance produced at the surface of the sample is - 4 x 100 W/cm2. The system 266 laser ablation accessory was modified by using a separate computer to control the x-y-z translation stage. This allowed for the translation of the sample at approximately 15 [tm/s while the laser was repetitively fired. Typical analysis times were 80-100 s (signals measured over 400-500 laser shots). In general, sample translation during ablation provided more representative sampling of the surface; however, it also resulted in a larger mass removal rate and subsequently higher








sensitivities.87 The efficiency of detection for all analytes is a function of the ablation process, transport process, ionization in the ICP, and mass spectrometer response.

A microscopic Laser Induced Breakdown Spectroscopy (LIBS) instrument was used to measure the homogeneity of three matrix elements which were candidates for the internal standard study. A detailed description of the instrument has been reported elsewhere.88 It was used with a lOx working objective, the laser was focused to a -20 ptm diameter spot on the sample surface, corresponding to an irradiance of _1012 W/cm2. The conic crater left by the laser on a solid surface was -30 ptm deep and mass removal was on the order of 0.1 jtg per laser shot. The microscope stage (and hence, the sample) could be translated in a horizontal direction with a minimum speed of 4 jim/s. The laser repetition rate was 0.5 Hz. The spectrometer used was a compact dual-channel UV-VIS fiber optic CCD spectrometer (SD2000, Ocean Optics, Inc., Dunedin, FL, USA). The spectrometer was driven from a laptop computer (Travel Pro supplied with a Pentium II processor at 266 MHz, AMS Tech, USA) via a DAQCard-700 interface (National Instruments, USA).

A secondary electron microscopic spectrometry/energy dispersion spectrometry (SEM-EDS) instrument was used in this research. The JSM-6400 Scanning Microscope is a high-resolution scanning electron microscope with a modem digital image processing system. The image processing system enables image enhancement by averaging and integration and the storing of the image data on frame memories. An energy dispersive spectrometer was added to the JSM-6400 so that it can be used as an electron probe X-ray








micro-analyzer, allowing accurate and efficient nondestructive elemental analysis and elemental distribution study of micro areas.


5.2.2. Samples

Several soil samples and glass samples were studied to evaluate the feasibility of the solution calibration of these materials. These reference materials included NIST SRM 2709 (San Joaquin Soil), NIST SRM 2710 (Montana Soil, mildly contaminated), NIST SRM 2711 (Montana soil, highly contaminated), NIST Glass 611 and 612. The soil samples were weighed out in 1.00 g portions and pressed into pellets without a binder at a pressure of 35 MPa ( -5000 psi). To observe the image of the samples for laser ablation analysis, Au-Pd coating was applied to the soil samples. When measuring the concentration of the major elements in soils by SEM-EDS, carbon coating was applied to the samples. The solution standards were prepared from dilution of 10 or 1000 ppm standards (High Purity Standards) with Milli-Q water and Optima HNO3 (Fisher Scientic, St Louis, MO, USA) to a final acid concentration of 5%.

5.2.3. Solution-based Calibration.

It was desired to produce the same plasma conditions while analyzing solid and solution samples. A matrix element, Mg, was chosen to be the internal standard for the soil samples. The mass 25 isotope peak of Mg was chosen because it gave an appropriate signal intensity measured by the electron multiplier detector and was free of interference. A calibration curve was made using five solution standards; it was found that the Mg ion signal at mass 25 with laser ablation of soil 2709 corresponded to a Mg concentration of 2 ppm in a nebulized solution. Therefore, solution standards of different concentrations








of trace elements were prepared to have a concentration of 2 ppm Mg. A 2 ppm Mg solution with no added elements was used as the blank.

Three normalization (Fl, F2, F3) factors were introduced in this work to allow solution standards to be used for calibration of trace elements in solid samples. The F1, F2 and F3 factors are defined as follows.

F1: Corrects for the differences of signal intensity from the mass spectrum for internal standard in each solution standard and in the blank resulting from changes in the ICP-MS operating conditions from one set of measurement to another. For example, when Mg was used as internal standard,

F1 = Signal of Mg for each solution / Signal of Mg in blank

The Fl factor was used to correct all ICP-MS responses for each element in the solution standards needed in preparing the solution calibration curve.

Corrected signal of each trace element in standard solution = Measured Signal / F 1

F2: the ratio of the signal of internal standard in solid over the signal of internal standard in the solution. It was used to correct the signal of each element in each measurement in the solid samples. It has two functions:

1) Corrects the signal of each element to be measured in each solid sample so as to make them readily calibrated by solution standards. 3) Corrects for changes in laser energy in each measurement of the solid samples. Because each solid samples was measured five times, there were five F2 factors applied to each measurement.

For example, when Mg was used as internal standard,

F2 = Signal of Mg in each solid sample /Signal of Mg in blank








Corrected signal of each trace element in soil sample = Measured Signal / F2

In essence, we prepared a solution producing the same response for Mg as measured by LA in solid samples. This could only be done by applying F2 to each measurement of the solid sample. Although the response of each element in each measurement in the same solid should be the same when averaging -500 laser shots for the same sample, there are always differences in response in each measurement because of variation in the laser output energy as well as drifting of the ICP-MS detector. This was partly corrected by F2. By applying F2 to each measurement, the response of Mg in each measurement was forced to be the same; thus the response of each of the other elements in the solid sample were practically corrected for changes in the laser energy. F3: Correlates the signals of each element in the solid and solution to allow calculation of the concentration of trace elements in solid. The F3 factor correlates the signal of the solid and solution standards as follows: F3 = Concentration of internal standard in solid (ppm) / concentration of internal standard in solution (ppm)

Therefore, the concentration of element Z in the solid is given by: Czsolid = F3 x CzS'Ol

For example, when Mg was used as internal standard, the concentration of Mg in soil sample 2709 was 1.51%, which gave the same response as a 2 ppm Mg solution standard after correction by F2. Then

F3 = 0.0151 x 106 / 2 ppm Mg solution = 7550 ppm Mg solid/ppm Mg solution The concentration of element Z, Cz, in soil 2709 solid is then: Csolid = 7550 x Csoln








where C,"l is the concentration (in ppb) of Z in the solution. Therefore, when there were specified concentrations for each element in the solutions, a calibration curve could be made for each element. From the calibration curve, the concentration of a solution corresponding to the response (corrected by F2) measured for each element from solid samples could be obtained, and, the concentration of each element in each solid could be calculated using the factor F3. At most, we spiked 5 elements in the same series of solution standards. However, the results from the overall measurements indicated that the number of elements in the same calibration solution did not influence the accuracy. For NIST glass 611, 612 samples, a certified value of a matrix element was not available. Therefore, a trace element, Sr, was chosen as the internal standard. The concentration of Sr in 611 and 612 NIST was 515.5 and 78.4 mg/kg, respectively. It was found that a solution of 0.02 ppm Sr gave an F2 between Glass 611 and the solution close to 1. The F3 that correlated the signal of solid and solution in each case was 25775 and 3920 ppm Sr in solid / ppm Sr in solution.

5.3. Results and Discussion

5.3.1. Selection of an Internal Standard

A requirement for the use of internal standardization to obtain accurate results is that the analyte and internal standard be distributed similarly over the area sampled, and that they exhibit similar behavior in transport to and ionization in the ICP. A few elements were candidates as internal standards because they were present at significant levels in all the samples. The consideration here was to choose a matrix element which would respond similarly for plasma conditions for both solid and solution samples. The








ideal internal standard is one that is spatially homogeneous in the solid samples and constant in concentration in all samples. Silicon, the major element, was not chosen because it had interference at all three isotope masses. Two other elements Al and Na are monoisotopic and so they could not be measured by the electron multiplier at the same time along with trace elements of interest. To determine the homogeneity of the matrix elements, a microscopic laser induced breakdown spectroscopy system (LIBS) was used to spatially scan several matrix elements (Fe, Ca and Mg), in the solids. The results showed that for 15 laser shots randomly firing at the surface of the soil samples, Mg produced a more stable signal than Fe and Ca (see Table 5-2). This was concluded from comparing the standard deviation of the response measured by LIBS of Mg, Ca and Fe with the standard deviation when they were measured in a reference sample. Glass was used as a reference sample for Mg and Ca and a stainless steel sample was used as a reference for Fe. In all cases, the reference samples produced the smallest standard deviations. Ca was found to have the smallest standard deviation in the soil samples among the three matrix elements. Also, Mg has a minor isotope at mass 25 which can be measured by the electron multiplier at the same time as trace elements. Therefore, Mg was used as the internal standard for all soil samples in our work. In the case of glass samples, a trace element was chosen to be the internal standard since the concentrations of the major matrix elements were unavailable. However, it was known that the distribution of elements was much more homogeneous compared with soil samples because of the amorphous nature of the glass. Therefore, a trace element was used as the internal standard. In our case, Sr worked well as an internal standard for glass samples. However, when Sr was tested as an internal standard for soil samples (as a trace element),








the results showed a poorer accuracy compared with Mg for soil samples; relative errors of over �100% were obtained for some elements. Therefore it was not feasible to use a trace element as an internal standard for the soil samples.

To obtain the concentration of the internal standard in the solid sample, x-ray electron-microprobe analysis can be used. Also, in an earlier work, we have shown that it is feasible to use a spiked element as the internal standard for soil samples.39 An energy dispersive EPMA instrument was used to measure the concentration of matrix elements in the soil samples. The detection limit of this EPMA instrument is in the range of -1% for most elements so only matrix elements can be detected. The results are shown in Table 5-3. Carbon was used as coating for the soil samples when measuring the concentration of the elements. This is because carbon has fewer emission lines compared with Au-PD that would interfere with the emission lines from other elements. Comparison with the certified values showed that EPMA is a reliable method to obtain the concentrations of major elements in the soil samples.

5.3.2. Study of Fractionation Effect

Accurate analysis requires that the detected mass composition must be the same as the sample composition. For laser ablation sampling, fractionation (preferential mass removal during laser ablation), which can cause inaccurate analysis, can occur under some conditions. Numerous studies have shown that fractionation can be minimized or eliminated, depending on the sample and laser properties.89-92, 81

It is critical in this study that for different samples a constant sensitivity is obtained for each element to be measured relative to the internal standard. The change of








irradiance on the surface of the sample when ablating different samples may cause a change of sensitivity of different element relative to the internal standard. The change of RSF (defined as intensity/concentration relative to that of internal standard) for different elements resulting from a change in laser irradiance was investigated. There are two ways to change the laser irradiance at the surface of the sample: change of laser output energy by changing the flash lamp voltage and change of laser focus area by changing the focusing distance of the laser beam. The RSFs were measured when the flash lamp voltage was adjusted to 9.5, 9.8 and 10.1 kV, which corresponded to laser output energies of -6, 7 and 8 m.J, respectively. The RSFs were measured when the laser was focused below, on the surface or above the surface of the sample, a total range of 1.6 mm, which produced laser spot sizes of -40, 50 and 60 gtm in diameter, respectively. The result shows that the RSFs changed somewhat when the laser irradiance was changed. (Figure 5-1 and 5-2), indicating that the samples had to be maintained at the same condition during ablation to guarantee the sensitivity of each element would remain the same. This was done by weighing out the same mass of each soil samples and keeping the same pressure for each sample when pressing the pellet. The focusing and laser output energy was maintained the same during the measurement of each solid sample.

5.3.3. Analysis of NIST Standard Materials

The results of analyzing NIST soils 2709, 2710, 2711 and NIST glasses 611, 612 are shown in Tables 5-4 and 5-5. For most of the elements in soil samples, including Pb, V, Ba, Cr, Zn, Cd and Sb, the agreement of the measured value and the certified value are within +15%. In glass samples, measurements of elements, including Ni, Cu, Rb, Ag








and Co, are within �10% of the certified values. These results suggest that the use of solutions for calibration is effective for the analysis of these elements. It was also found that for elements in solids under 10 ppm, larger errors resulted, indicating the method was not capable of accurately measuring concentrations below 10 ppm in soil samples. Therefore, the results for Cd and Sb in soil 2709 are not listed in Table 5-4.

Several elements gave different behaviors in soil and glass samples. Poor accuracy was obtained for Cu and Ni in soil samples, whereas good results were obtained for them in glass samples. Systematically low concentrations were obtained for Pb and Th in glass samples; however, good results were obtained for Pb in soil samples. Several factors were investigated to determine their influence on the results.

It was expected that when factor F2 is close to 1, there is greater probability of producing identical plasma conditions for both solution and the solid. A series of solution standards were prepared for measuring soil 2711 which has the lowest concentration of Mg causing F2 to be 0.25 for the above analysis compared with -0.7 and

-0.9 for soil 2710 and 2709, respectively. When a solution concentration of 0.5 ppm was used, F2 was nearly 1 for 2711. However, the results from measuring 5 elements in soils showed that even with F2 equal to 0.25, it was not a factor that affected the accuracy of the measurement. Similar results were found for Ni and Cu using the new solution standard for soil 2711.

The larger errors for Ni and Cu in soil samples were investigated. A different internal standard, Ca, was used to measure these two elements in soil samples. The results showed that good accuracy for Ni was only obtained for 2709 soil. This is consistent with the results obtained for the three soil samples when Mg was used as the








internal standard. One suggestion of the cause of the error was the heterogeneity of the Ni distribution in the samples, since 2710 and 2711 were contaminated soils from unknown sources. An investigation was made by tracking the mass spectra for Ni and all elements that were measured in the current work when translating the ablating laser across the surface of the sample. Time resolved mass spectra show that Ni gave many more signal intensity spikes as compared with all the rest of the elements (Figure 5-3). In fact, the signals of Ni measurements showed bigger differences from one measurement to the other, also indicating a heterogeneous distribution of it in soil samples. The same experiment was repeated with the glass samples. Ni and all other elements showed nearly constant signals, indicating a homogeneous distribution of these elements across the surface of the glass samples. Therefore, we believe that heterogeneity of Ni distribution in the SRMs was the cause for the poor results for Ni in soil samples.

SEM images of the ablated and the un-ablated soil 2709 were shown in Figure 5-4 and 5-5. It was apparent that melting occurred during the ablation. This was indicated by the droplet formed along the ablation track. Matrix effect and fractionation caused by preferential ablation may cause the poor results of Cu. However, a detailed explanation is unattainable so far.

5.4. Conclusions

A method of using solution standards to calibrate LA-ICP-MS of NIST soil and glass samples was evaluated. In most cases, relative errors of less than +10% were obtained. Compared with results obtained using RSFs,39 improvements were made in the case of Pb and Ba, which are toxic pollutants in the environment. The solution








standards do not appear to be as convenient as single solution standards. However, once good accuracy is consistently obtained for certain elements, it is possible to set up a routine to control/monitor the content of these elements in different samples of the similar matrices. The limitation of the method is that an internal standard of known concentration must be present in the solid sample. The homogeneity of the matrix element distribution in the bulk was studied by LIBS. The error of Ni was found to originate from the heterogeneity of its distribution in the bulk. This is not surprising since the two soil samples 2710 and 2711 are samples contaminated from an unknown source. However, the poor accuracy for Cu in soil samples and Pb and Th in glass samples remains unclear. It is suspected that a matrix effect, probably at least in part, was the cause of these results.













Table 5-1. Typical ICP-MS operating conditions


Rf power Coolant gas flow rate Auxiliary gas flow rate Nebulizer gas flow rate Ablation chamber flow rate Solution uptake rate

Scan conditions
ICP-MSDetector
Scan range per isotope
Number of passes
Number of channels p Dwell time per chann
LaserEnergy per shot Repetition rate Rastering speed


1300 W 15 L/min 0.9 L/min
0.65 L/min 0.35 L/min 1.0 mL/min



electron multiplier e 0.25 amu
128
er amu 8 ms
4 ms

-7 mJ
5 Hz
15 /tm/s












Table 5-2. Intensity and relative standard deviation of signals of Mg, Ca and Fe from NIST soil 2709, 2710 and 2711 withl5 laser shots measured by LIBS. Steel (for Fe) and glass (for Mg and Ca) were used as references in which Fe or Ca/Mg were considered to be homogeneous.

Samples Mg Ca Fe
Intensity RSD (%)* Intensity RSD (%)* Intensity RSD (%)* 2709 827.1 11.7 1314 9.5 800.7 15.5 2710 470.3 16.2 866.5 32.0 589.4 19.1 2711 678.5 13.8 2270 16.5 722.8 17.9
Glass 352.4 8.00 2373 5.8
Steel - 3067 3.2
*Average of 5 measurements.










Table 5-3. Results from SEM-EDS analysis for soil samples.


0 Na Mg Al Si K 2709 meas. conc. 50.85 � 1.32 0.71 � 0.03 1.72 � 0.14 9.11 � 0.59 27.16 � 0.23 2.55 � 0.3
certified cone. - 1.16 1.14 1.21 0.91 2.03
%recovery - 61.2 114.4 121.5 91.5 143.3
2710 meas. conc. 49.54 � 0.54 1.43 � 0.25 1.31 � 0.11 7.88 � 1.09 29.69 � 1.36 2.64 � 0.56
certified conc. - 1.14 0.853 6.44 28.97 2.11 %recovery - 125 131 122 102 125
2711 meas. conc. 51.33 � 0.84 0.902 � 0.09 1.47 � 0.22 7.7 �0.19 28.17 � 1.16 2.59 � 0.16
certified conc. - 1.14 1.05 6.53 30.44 2.45
%recovery 79.1 140 117 92.5 105.7



Ca Ti Mn Fe Cu Zn 2709 meas. cone. 2.13 � 0.3 0.435 � 0.13
certified cone. 1.89 0.342
%recovery 112 127 -
2710 meas. cone. 0.99 � 0.09 0.272 � 0.08 1.00 � 0.13 3.84 � 0.59 0.32 � 0.13 0.55 � 0.098
certified cone. 1.25 0.283 1.01 3.38 0.295 0.695
%recovery 79.6 96.2 99.7 113 108 79 2711 meas. conc. 3.29 � 0.41 0.285 � 0.07 - 3.5 � 0.84 -
certified conc. 2.88 0.306 - 2.89 %recovery 114 93.9 - 121 -











Table 5-4. Results from calibration of NIST soil sample 2709, 2710 and 2711 using solution standards.


Elements V Ba Cr Pb Zn Cd Meas. Conc. (ppm) 112.8 932.1 101.7 20.9 79.5 �5.2 �11.3 �7.9 �3.7 �8.3 Cert. Conc. (ppm) 112 968 130 18.9 106 0.4
-%Recovery 100.8 96.3 78.2 110.6 75.0 -


Cu Ni Sr 12.9 96.6 137.4 � 2.2 � 5.8 � 8.7 34.6 88 231 37.2 109.6 59.5


Meas. Conc. (ppm) 74.6 814.1 29.8 5844 6443. 23.5 41.12 1650. 20.1 225.7
� 10.2 �26.7 �3.8 � 156 � 122 �4.2 �3.6 �35 �9.6 � 10.2 Cert. Conc. (ppm) 76.6 707 39 5532 6952 21.8 38.4 2950 14.3 240
-%Recovery 97.5 115.1 76.4 105.6 89.93 107.8 107.1 55.9 140.5 94

Meas. Conc. (ppm) 79.12 605.8 45.9 1073 402.1 43.6 21.18 43.4 55.4 107.6
�3.8 � 10.3 �5.5 �35 � 18.7 � 10.5 �3.4 �4.8 � 18.7 �8.9 Cert. Conc. (ppm) 81.6 726 47 1162 350.4 41.7 19.4 114 20.6 245.3
-%Recovery 96.9 83.4 97.7 92.4 114.2 104.5 109.2 38 264.3 43.9


Samples
2709


2710


2711














Table 5-5. Results from calibration of NIST glass sample 611, 612 using solution standards.

Sample Elements Ni Cu Rb Ag Co Pb Th 611 Meas. Conc. (ppm) 381 463.3 419.3 217.2 337.7 230.4 255.6 �28 �42.5 �21.8 � 11.4 � 10.6 �35.7 �9.6 Cert. Conc. (ppm) 458.7 444 425.7 254 390 426 457.2
%Recovery 83.1 104.3 98.5 85.5 86.6 54.1 55.9

612 Meas. Conc. (ppm) 42.1 44 31.9 24.8 39.9 25.2 17.9 � 8.3 � 12 � 5.1 � 2.3 � 5.8 � 8.2 � 4.3 Cert. Conc. (ppm) 38.8 37.7 31.4 22 35.5 38.6 37.8 %Recovery 108.6 116.6 101.5 112.7 112.4 65.4 47.4


















1.8


1.6


1.4 1.2 a, 08


0.6 0.4 0.2


0.
Mg V


OV=9.5kv =V=9.8kv *V=10.lkv


Cr Fe Ni


Cu Zn Sr Sb Ba Pb


elements


Figure 5-1. RSFs of each element in soil 2709 when change laser output energy.

















* focus 1 mm below sample surface
1.6

4 focus on sample surface 14
0 focus 1mm above sample surface

1.2







0.8



0.6







0,2



Mg V Cr Fe Ni Cu Zn Sr Sb Ba Pb Elements


Figure 5-2. RSFs of each element in soil 2709 when changing laer spot size.



















90000 450000
Ni-soil

80000 Ni-glass 400000


70000 . Mg-soil (right axix)
rooo 350000 60000 300000 50000 0 250000 040000 0 200000 30000t (10

200000



10000 100






0 10 20 30 40 50 60 70 80 90 100 time (s)


Figure5-3. Time resolved spectra for Ni, Mg in soil 2710 and Ni in glass 611.
































Figure 5-4. SEM image of SRM 2709 unablated.


Figure 5-5. Image of SRM 2709 ablated.














CHAPTER 6
ELEMENTAL ANALYSIS OF SPANISH MOSS USING LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY

6.1. Introduction

Tillandsla usneoides L., family Bromeliaceae, commonly called Spanish moss, is an epiphyte which festoons the trees in swamps and hammocks south from Virginia to Florida and west to Texas, and further southward throughout northern South America. The plant possesses no functional internal conducting system or cuticle, and water absorption occurs over the whole surface of the plant.93 Therefore these mosses can be used as biomonitors for direct monitoring of wet and dry deposits from the atmosphere. By observing and measuring the changes of a bioindicator, a conclusion as to the kind of pollution (e.g., a heavy metal), its source and possibly its intensity can be drawn.94 Since biomonitoring using mosses was first introduced in 1968 by Riihling and Tyler95 the use of mosses, lichens and barks for monitoring of heavy metal deposition from the atmosphere has found wide application.96-98 A number of analytical tools have been used to determine the elemental concentrations in mosses.99-101 The most often used techniques include ICP-AES and electrothermal atomization AAS following the digestion of moss samples. Spanish moss is considered as a difficult-to-digest sample which contains a variety of matrix constituents including organic compounds.102 Strong oxidizing agents with various acid mixtures as well as high pressure and temperature









conditions are necessary for complete digestion. Microwave digestion (MD) is also widely used.103-110 A problem associated with the determination of trace elements using any bomb digestion approach includes the risk of losing elements because of the pressure relief mechanism of the vessel during the digestion. Blank interference, which is mainly due to the contamination of the membrane filter as well as containers, may become an important factor during the preparation of samples. For samples of varied compositions and low elemental concentrations, a memory effect from previously digested samples may cause unreliable analytical results. 1 11 The acid concentration of its residue also may affect the reliability of the instrumental analysis. Direct analysis of Spanish moss samples has not been reported previously.

The use of pulsed laser ablation (LA) sample introduction for ICP-MS enables analysis of solid materials to be performed directly, without sample dissolution. 112,113 Although the relative sensitivity of LA-ICP-MS is poorer compared to solution ICP-MS, because much less material is injected into the ICP, trace detection with submicrogram per gram (ppm) sensitivity in the bulk is routinely achieved. Because the ablation yield varies with material properties, such as reflectivity, thermal conductivity, and melting and boiling points, it is important to obtain matrix-matching standards that contain all the elements of interest.38 There have been a few reports of full quantitative analysis by LAICP-MS using matched standards and multi-point calibration plots for each element of interest,12,114-116 e.g., using glass Standard Reference Materials (SRMs) for the analysis of minerals. 12








This chapter reports on use of LA-ICP-MS for the direct analysis of Spanish moss samples using NIST leaf SRMs as standards for calibration. The standard addition method was also studied. The results were compared to those obtained from MD-ICPAES. The sampling strategy, instrumental parameters, fractionation effects were also studied to characterize the strengths and limitations of this approach.

6.2. Experimental

6.2.1. Instrumentation

The experiments were performed with a Finnigan MAT (San Jose, CA, USA) SOLA ICP-MS and Finnigan MAT System 266 laser ablation accessory, which has been described in chapter3. Typical operating conditions for the ICP-MS are listed in Table 61. A combined flow of nebulized solution and carrier gas from the ablation chamber was introduced to the ICP-MS for all sample measurements. Plasma conditions were maintained the same for all the standards and samples analysis. Typical analysis times were 80-100 s (signals measured over 400-500 laser shots).

A photodiode, connected to a chart recorder, was used to monitor the laser output energy. A quartz plate ( 2'x2' )was placed at a 45 degree angle to the laser beam between the source and the ablation chamber. The quartz window transferred -96% of the laser beam to the ablation chamber while -4% of the laser beam was deflected to an UV filter and photodiode. The laser power was manually adjusted to maintain a constant signal.

Also, a microscopic LIBS system was used in this study to measure the Ca concentration. The LIBS system was described in Chapter 5.








6.2.2. Samples

The NIST archival leaf SRMs were used as matrix-matched standards for the elemental analysis of Spanish moss. These reference materials included NIST SRM 1515 (Apple leaves), NIST SRM 1547 (Peach leaves), NIST SRM 1570 (Spinach leaves) and NIST SRM 1573 (Tomato leaves). Mixed standards were also prepared to produce appropriate concentrations for the calibration curve. They were prepared by weighing out different portions from two standards and mixing in a Spex (Metuchen, NJ, USA) Mixer/Mill Model 8000 for 30 min to ensure homogeneity. The NIST standards were dried in a dissector for 5 days before pressing into pellets for use.

All Spanish moss samples were collected with plastic gloves and were placed in plastic storage bags for transport to the laboratory. The samples were then rinsed to remove wind-blown particles of dust and soil. Milli-Q double deionized water was used in the rinsing, and all samples were handled wearing plastic gloves. After rinsing, the wet moss samples were dried in an oven at 11 0C for a 4 hour period. The samples were crushed in a ceramic mill for 25-30 minutes, reducing them to fine homogeneous powder well suited to scooping and weighing. The powder was sealed in clean 60 mL polyethylene bottles for storage prior to pressing pellets and making measurements. The NIST SRM standards and Spanish moss samples were pressed into pellets without a binder at a pressure of 35 MPa ( -5000 psi).

The samples for standard addition analysis were prepared by adding 200, 300, 500 and 1000 .tL of 1 ppm Mn or Pb standard solution into each -0.5 g portion of solid powders. The samples were then dried at 110C in the oven for several hours. The




Full Text

PAGE 1

APPLICATIONS OF LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY FOR THE ANALYSIS OF SOLIDS By MELODY (XIANGQING) BI 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 2000

PAGE 2

To My Family

PAGE 3

ACKNOWLEDGMENTS I would like to express my immense gratitude to my advisor, Dr. James D. Winefordner who has offered not only his intellectual guidance for the project but also his patience for my mistakes and stubbornness in pursuit of the project. Dr. Benjamin W. Smith is the next person to whom I would like to express my gratitude for invaluable suggestions and comments on this research. He has also given me a lot of help on other issues. I would like to thank Dr. Antonio M. Ruiz and Dr. Igor Gomushkin who have contributed directly to my experiments. I would also like to thank Dr. Chenglong Yang for many useful discussions concerning my research. I would like to thank Dr. Stephen J. Pearton and Dr. Rajiv K. Singh (from the Materials Science & Engineering department) who provided semiconductor samples and useful information for my research. I would also like to thank Dr. Luisa Amelia Dempere and Mr. Wayne Acree from the Major Analytical Instrumentation Center, who helped me with the SEM-EDS study. I would like to thank Mr. Gill Brubaker from Energy Research Center for letting me use the ICP-AES instrument. My gratitude also goes to Drs. Willard W. Harrison, Daniel R. Talham, Martin Vala and Paul H. Holloway for their kindness in serving on my committee. They were iii

PAGE 4

also supportive of my research effort. I especially thank Dr. Harrison who attended our regular research group meeting and gave me useful advice every now and then. I would like to thank Dr. Burtron H. Davis from the Center for Applied Energy Research, University of Kentucky for recommending me to the University of Florida and for his help with my oral project. I feel very lucky to know him. I want to thank my brother Sean for his encouragement. I want to thank my mom Yuante Bi and my dad Junke Bi who live on the other side of the globe, and who have given me the love and support that 1 needed the most to pursue my life goal and dreams. I also thank my friends and colleagues, the entire Winefordner group. Merry Zhang, Lisa Lang, Detong Sun, Caijun Sun, Junqiang Sun, Johnathan Bishop, Qianrong Ma, Dale Cheng, Bob and Gay Greninger etc. who have made my time at the University of Florida a pleasant and interesting experience. Finally I would like to acknowledge financial support from a Texaco Fellowship for my research. IV

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS iii LIST OF ACRONYMS viii ABSTRACT ix CHAPTERS 1 INTRODUCTION 1 2 BACKGROUND 7 2.1. Principles of Laser Ablation 7 2.1.1. Principles of Lasers 7 2.1.2. Q-switch Lasers 9 2.1.3. Laser Material Interaction 9 2.2. Principles of Inductively Coupled Plasma Mass Spectrometry 12 2.2.1. Inductively coupled plasmas 12 2.2.2. Sample Introduction 15 2.2.3. Sampling Interface 15 2.2.4. Ion Lens in ICP-MS 17 2.2.5. Space Charge Effect 17 2.2.6. Quadruple Mass Spectrometer 18 2.2.7. Mass Detectors 19 2.3. Principles of LA-ICP-MS 20 2.3.1. Calibration Strategies 21 3 INSTRUMENT DIAGNOSTICS AND OPTIMIZATION 36 3.1. Introduction 36 3.2. Experimental 36 3.2.1. Instrumentation 36 3.2.2. Samples 39 3.3. Results and Discussion 39 3.3.1. Laser Output Energy 39 3.3.2. Laser Focusing Positions 41 3.3.3. Analysis of SiC 41 3.3.4. Plasma Operating Conditions 42 V

PAGE 6

4 STUDY OF SAMPLING STRATEGY FOR LA-ICP-MS ANALYSIS 52 4.1. Introduction 52 4.2. Experimental 53 4.3. Results and Discussion 54 4.4. Conclusions 56 5 STUDY OF SOLUTION CALIBRATION OF NIST SOIL AND GLASS SAMPLES BY LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY 64 5.1. Introduction 64 5.2. Experimental 66 5.2.1. Instrumentation 66 5.2.2. Samples 68 5.2.3. Solution-based Calibration 68 5.3. Results and Discussion 71 5.3.1. Selection of an Internal Standard 71 5.3.2. Study of Fractionation Effect 73 5.3.3. Analysis of NIST Standard Materials 74 5.4. Conclusions 76 6 ELEMENTAL ANALYSIS OF SPANISH MOSS USING LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY 87 6.1. Introduction 87 6.2. Experimental 89 6.2. 1 . Instrumentation 89 6.2.2. Samples 90 6.3. Results and Discussion 91 6.3.1. Elemental analysis of Spanish moss using NIST leaf SRMs as matrix matched standards by LA-ICP-MS 91 6.3.2. Study of Fractionation 93 6.3.3. Analysis of Moss Samples by LIBS 94 6.3.4. Standard Addition Method 94 6.4. Conclusions 96 7 LOW LEVEL COPPER CONCENTRATION MEASUREMENTS ON SILICONWAFER SURFACE USING DIRECT LA-ICP-MS AND SOLUTION SAMPLING ICP-MS 110 7.1. Introduction 1 10 7.2. Experimental 112 7.2.1. Instrumentation: 112 VI

PAGE 7

7.2.2. Samples 113 7.3. Results and Discussion 114 7.3.1. Direct Analysis of Wafer Standards 114 7.3.2. Analysis of Solution Samples 115 7.4. Conclusions 116 8 PROFILING OF PATTERNED METAL LAYERS BY LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY 120 8.1. Introduction 120 8.2. Experimental 122 8.2.1. Instrumentation 122 8.2.2. Samples 123 8.3. Results and Discussion 124 8.3.1. Effect of Ablation Chamber 124 8.3.2. Optimization of Argon Flowrates 124 8.3.3. Optimization of Laser Focus 125 8.3.4. Comparison of LA-ICP-MS with LA-OES 126 8.3.5. Profiling of Patterned Metal Layers on Silicon Wafers 128 8.4. Conclusions 129 9 CONCLUSIONS AND FUTURE WORK 142 9.1. Conclusions 142 9.2. Future Work 145 REFERENCES 150 BIOGRAPHICAL SKETCH 159 vii

PAGE 8

I JS^ OF ACRONYMS LA-ICP-AES Laser Ablation Inductively Coupled Plasma Atomic Emission Spectrometry LA-ICP-MS Laser Ablation Inductively Coupled Plasma Mass Spectrometry LA-ICP-OES Laser Ablation Inductively Coupled Plasma Optical Emission Spectrometry LIBS Laser Induced Breakdown Spectrometry MD-ICP-AES/MS Microwave Digestion Inductively Coupled Plasma Atomic Emission Spectrometry/Mass Spectrometry RSF Relatively Sensitivity Factor RSD Relatively Standard Deviation SEM Secondary Electron Microscopy SIMS Secondary Ion Mass Spectrometry SRM Standard Reference Material

PAGE 9

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 APPLICATIONS OF LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY FOR THE ANALYSIS OF SOLIDS By Melody Bi August 2000 Chairman; Dr. James D. Winefordner Major Department: Chemistry Laser-sampling techniques, such as laser ablation (LA) inductively coupled plasma (ICP) mass spectrometry (MS), have gained popularity for analyzing solid materials because little or no sample preparation is required. Practical concerns such as the difficulty of obtaining or making matrix-matched standards prevent such techniques from becoming analytical tools. This work was largely concerned with studies of calibration strategies for trace analysis in solids and the feasibility of surface profiling by LA-ICP-MS. The capabilities of surface profiling by LA-ICP-MS were also explored. By analyzing NIST soil and glass samples, a method for the determination of trace element concentrations by LA-ICP-MS using solution calibration and an internal standard has been studied and evaluated. In most cases, the measured element concentrations were within ±15% of the certified values. The choice of an internal IX

PAGE 10

standard and the fractionation effect caused by laser irradiance on the surface of the sample were studied and discussed. NIST archival leaf standards were used for reliable quantitative elemental analysis of Spanish moss samples by LA-ICP-MS. The results were compared with that obtained from microwave digestion (MD) ICP atomic emission spectrometry (AES) analysis. For most of the elements studied, the results for the two techniques agreed. A standard addition method was also studied and the results showed that it is an effective method when matrix-matched standards are not available. LA-ICP-MS was applied to profiling of patterned thin metal layers on a polymer/silicon substrate. The parameters of the laser and ICP-MS operating conditions have been studied and optimized for this purpose. LA-ICP-MS has good sensitivity and was able to profile thin metal layers on the order of a few nm on the silicon surface. A lateral spatial resolution of 45 pm was achieved. Finally, LA-ICP-MS was used to measure the amount of copper contamination on a silicon wafer surface. A detection limit of 6.1 x 10'^ atoms/cm^ was obtained. X

PAGE 11

CHAPTER 1 INTRODUCTION Inductively coupled plasma mass spectrometry (ICP-MS) has gained wide acceptance over the last 15 years as a technique for quantitative analysis and trace element detection. ICP-MS is most commonly used for the analysis of solution phase samples because accurate, multi-element calibration standards are readily available and matrix matching of the sample and standard is generally straightforward. Although ICPMS is a well-established analytical tool for solution analysis, an active area of interest in ICP-MS is the development, characterization and application of alternative methods for sample introduction. Popular methods that have been studied extensively and evaluated to date include: Electrothermal Vaporization (ETV).^'^ Flow Injection (FI),^>^ Direct Injection Nebulization (DIN) ^ and Laser Ablation (LA).10‘^2 Precise and accurate measurement of multiple elements at low levels is required in a variety of applications. 9 xhe presence of trace elements in soils, for example, can be the identification of contaminants and levels of toxic elements present. In another case, the trace elements in Spanish moss can be an indication of pollution source and levels. Direct analysis of these types of solid samples could provide numerous advantages over solution sampling including,20

PAGE 12

2 • Allows the sample to be analyzed directly and therefore introduces little dilution of the sample. The dissolution of the sample normally dilutes the sample and reduces the actual detection limits available. • Reduced sample preparation time. The time required for some solution preparation schemes can he extensive, typically several hours to one day. This preparation time may be reduced considerably for solid sample introduction where samples are typically ground to a fine powder and suspended in a fluid (slurry) medium, mixed with a binder and pelletized or analyzed directly. • Reduced risk of sample contamination. In general, contamination can be minimized for solids analysis because sample handling is minimal, and the agents added as binders, for example, tend to have an organic matrix. • When using laser ablation as a solid sampling technique, both bulk sampling and surface profiling can be carried out. In this way, detailed elemental or isotopic distribution through a sample can be investigated. Among the most popular direct solid sampling methods are laser ablation, glow discharge (GD),21 secondary ion mass spectrometry (SIMS)22 and spark/arc ablation (SA).23 GD converts a solid sample directly into an atomic phase by sputtering processes. Because sputtering is a primary sampling mechanism, heating is minimal and preferential vaporization is minimized. GD is best suited for conductive samples although non-conductive samples can be analyzed with a radio-frequency discharge.24-26 ^ qj) source coupled to a high-resolution mass spectrometer is a powerful technology for achieving excellent depth profiling and detection characteristics with limits of detection

PAGE 13

3 (LODs) of 10-100 ng/g. For arc/spark ablation, mass is eroded from the sample in the form of atoms, molecules, vapor, droplets, solid flakes, and particles. Normally, only conductive samples can be used with spark discharges. However, non-conductive materials can be analyzed by mixing the sample with conductive matrix. SIMS has been acknowledged as one of the most important techniques for surface analysis and for elemental imaging. 27, 28 strong points of SIMS are its outstanding sensitivity, which is in the ppb range (in the solid); and its extremely fine lateral-profiling resolution, which is on the order of 0.1 to 1 pm. 29 The use of pulsed laser ablation sample introduction for ICP-MS has attracted particular interest as a direct solids analysis technique. 2 0-3 6 unique characteristics of LA-ICP-MS include the following: • No vacuum is required in the sample chamber; however, an airtight seal is necessary; • LA-ICP-MS, unlike LA-ICP-AES, separates the ionization step from the sampling step the laser is used to ablate the sample only and the material is transported to the secondary plasma source in the torch of the ICP. Therefore, both steps can be independently controlled and optimized. • LA-ICP-MS reduces the amount of H 2 O and other polyatomic species in the plasma. • The high sensitivity of the ICP-MS allows small samples to be quantified, which is ideal for LA-ICP-MS in that spatial resolution can be used to investigate compositional gradients across a sample. If the sample of interest is the microscopic inclusion in a bulk material, or particles on a filter paper, it is not necessary to analyze

PAGE 14

4 the entire sample. Most of the mass in such cases is not of interest and leads to large background signals in the analytical source. Although the relative sensitivity of LA-ICP-MS is poorer compared to solution ICP-MS, because much less material is injected into the ICP, trace detection with submicrogram per gram (ppm) sensitivity in the bulk is routinely achieved. Becker et al. 37 determined Zn, B, Si, Ge, Sn, Sb, P, S, Se and Te in a synthetic GaAs standard using secondary ion mass spectrometry (SIMS), SN-ICP-MS, ICP-OES, spark source mass spectrometry (SSMS), radiofrequency glow discharge mass spectrometry (GDMS) and LA-ICP-MS. For the LA-ICP-MS study, ablation at 20 Hz using 10 ns pulses (266 nm) was used with detection by a double-focusing spectrometer. Inspection of Table 1-1 reveals that LA-ICP-MS competes favorably with other state-of-the-art solid state techniques for these analyses. A number of limitations of solid sampling include the following: • Sample inhomogeneity, which can to some extent be overcome when a sample is brought into solution. • Particles of solid material can fractionate because the sampling process is remote from the ICP, • The most serious problem is calibration and quantitative measurement. This usually requires that the sample and standard be of similar chemical matrix and physical form. Even when this is achieved, a mismatch of elemental response between the sample and standard may occur.

PAGE 15

5 • Deposition on the sampling cone can also occur if plasma loading is unacceptable or if the particles of the solid are too large for the dwell time provided by the sampling extraction position and gas flow rate. • LA has poor precision, partly because of fluctuation in the arrival and size of individual ablated particles transported to the ICP. Among these, the major obstacle that prevents the widespread applicability of LAICP-MS is the difficulty of obtaining matrix-matched standards. The use of relative sensitivity factors (RSF) obtained from solution to determine the elemental concentrations in solids has been studied for both LA-ICP-OES^^ and LA-ICP-MS.^^’^^ The method has obtained an accuracy of ± 20% of the certified values for metals, soils and glass materials. The present work applied LA-ICP-MS to the elemental analysis of solid samples including soils, glasses, Spanish moss, silicon wafers etc. A solution-based calibration methods for soils and glasses was studied and evaluated. We examined the utility of NIST leaf standards for calibration of Spanish moss samples. We also measured the surface contamination of Cu on silicon wafers by LA-ICP-MS. We studied sampling strategies for powdered materials for LA-ICP-MS analysis. In addition, we explored the surface profiling of thin metal layers on silicon wafer. Optimum experimental parameters were determined for these studies. Future research directions are suggested based on these studies.

PAGE 16

6 Table 1-1. Elemental concentrations in GaAs measured by different techniques. Dopant SIMS ICP-MS ICP-AES SSMS rfGDMS LA-ICP-MS Zn 1208 ±90 827 ± 22 910±50 870 ± 160^ 870 ± 100“ 870 ± 80“ B 17.8 ± 1.2 19.5 ±0.7 18±6 8.2 ± 1.5 8.2 ±2.9 8± 1 Si 11.7±0.7 11.5±0.8 <15 11 ±2 7.8 ±2.9 7.9 ±3.6 Ge 20.5 ±0.6 <40 11±4 36 ± 1 Sn 13.5 ±2.0'’ 6.0 ±0.2 <40 4.2 ± 1.2 <10 23 ± 1 Sb 49 ± 1 45± 12 14±4 132 ±4 P 328 ±30 (650 ± 1001) 1290 ±260 850 ± 100 S 450 ± 80 316±20 390 ± 100 720 ± 140 475 ± 62 74 ± 1 Se 400 ± 75 395 ± 12 420 ± 60 315±48 120 ±22 48 ± 1 Te 113±27 97 ±3 110±30 108 ± 15 43± 14 62 ± 1 “ Internal standard element. Possible inhomogeneity

PAGE 17

CHAPTER 2 BACKGROUND 2.1. Principles of Laser Ablation The use of a laser to extract material from a solid sample for analysis can be traced back to the early 1960s. Lasers also have been used an ion source, e.g. laser ionization mass spectrometry (LIMS) and laser microscope mass spectrometry (LMMS).40 Although LMMS is an established technique, its primary advantage of <25 pm spatial resolution is offset by the major disadvantage that it is a qualitative technique. Laser ablation for micro-sampling has been used as a sample introduction technique with ICP-AES for a number of years.'^lÂ’42 'ppe early work by Thompson et al.^^ showed good signal reproducibility for SRM steels. The first application of laser ablation for solid sample introduction into ICP-MS was published by Gray in 1985 43 using a ruby laser with the Surrey research ICP-MS system. 2.1.1. Principles of the Lasers Laser is an acronym for light amplification of stimulated emission of radiation. It was first achieved by Maiman^4 of the Hughes Research Laboratory. The general principles of the laser are illustrated by Figure 2-1. According to the resonance condition (AE = hy) and to the rules of quantum mechanics, an atom can change its energy level, which leads to the absorption or emission of a photon. Stimulated emission is the basis of laser behavior. Stimulated emission leads to the emission of radiation that is coherent 7

PAGE 18

8 with the incoming radiation. In order to have light amplification in a laser, the number of photons produced by stimulated emission must exceed the number of photons lost by absorption. Light amplification is only achieved when a population inversion from the normal distribution of energy states exists. Population inversion (activation of a laser material) is created by an external pumping source, so that a few photons of proper energy will trigger the formation of a cascade of photons of the same energy. As a consequence of its light-amplifying property, a laser produces spatially narrow and extremely intense beams of radiation with identical frequency, phase, direction and polarization properties. Figure 2-2 shows the four energy levels of the Nd:YAG (Y 3 AI 5 O 12 ) laser, which is used in the Finigan Mat SOLA LA-ICP-MS system. In this system, lasing occurs between the meta-stable levels. As the terminal level is essentially empty at room temperature, the population of El can be increased by a relatively small pump power above that of the E3 level. This is a significant advantage over other systems such as the ruby laser. Many of the characteristics that make the Nd:YAG laser attractive in terms of output power, pulse repetition rate, and pulse energy result from the properties of the yttrium aluminum garnet (YAG) host. Large YAG crystals of high optical quality are readily available and the relatively large thermal conductivity together with four-level operation make high average output powers feasible. Nd'.YAG systems are used widely because they are relatively simple and cheap, and they are incorporated into several commercial LA systems. Ablation using pulses, usually of 10 to 300 mJ at 1064 nm has become very common. Over the last 5 years, frequency doubling, tripling and quadrupling have been used to produce pulses of wavelengths of 532, 355 or 266 nm and

PAGE 19

9 employed for analysis, especially in geological applications, where laser-sample coupling is often very poor for matrices. 2. 1.2. 0-switch Lasers Stimulated emission depletes the upper lasing level much faster than the pumping rate of the flash lamp. The light output thus consists of many intense spikes. The normal or free-running mode is illustrated in Figure 2-3. Also illustrated is a Q-switched (Qmode) pulse, which is produced by rapidly changing the quality (Q) factor, a measure of the energy storage capacity of the device, by, for example rotating one of the end mirrors. During most of the rotation cycle, the mirror is tilted at an angle, so the two mirrors are not facing each other, and the laser is pumped very strongly with a pulsed flashlamp. Near the end of the pump pulse, a very large population inversion will have been built up in the laser material, but no lasing action can begin because the mirrors are not parallel. At this point, the rotation mirror is stopped parallel to the fixed mirror. Because of the large population inversion, the gain and the available stored energy in the laser rod are very high, and the round trip gain in the laser cavity is much greater than one. Consequently, the laser oscillation builds up much more rapidly than in nonnal operation and rises very rapidly to a peak power level much above the normal. Thus, a very short, very high power pulse is produced. Laser pulses of 10 to 100 ns and peak powers from 10Â’ to >10^ W are so obtained. 2.1.3. Laser Material Interaction Figure 2-4 shows a conceptual interpretation of laser ablation. Laser-material interactions involve coupling of optical energy into a solid, resulting in vaporization;

PAGE 20

10 ejection of atoms, ions, molecular species, and fragments, shock waves, plasma initiation and expansion; and a hybrid of these and other processes. The interaction between the incoming radiation and the solid sample depend on numerous variables related to the laser, the sample and the atmosphere above the sample. Among these are the wavelength, energy, spatial and temporal profile of the laser beam and the heat capacity, heat of vaporization and thermal conductivity of the sample.45,46 Many models have been developed to describe these processes, but no models completely describe the explosive laser ablation process. Two general descriptions are described on the basis of irradiance, namely, vaporization and ablation.47-52 When the laser pulse duration is microseconds or longer and the irradiance is less than approximately 10^ W/cm^ vaporization is likely a dominant process influencing material removal from a target. Phonon relaxation rates are on the order of 0.1 ps, and absorbed optical energy is rapidly converted into heat. Heat dissipation and vaporization are fast in comparison to the laser pulse duration. The thermal and optical properties of the sample influence the amount of material removed during the laser pulse. The interaction is predominantly thermal. Melting is common and elements of higher vapor pressure will be enriched in the vapor relative to the concentration in the solid. At higher irradiances, above 10Â’ W/cm^ with nanosecond and shorter laser pulses focused onto any material, an explosion occurs. Phenomenologically, the surface temperature is instantaneously heated past its vaporization temperature through linear one-photon absorption and multi-photon absorption of the laser pulse duration; energy dissipation through vaporization from the surface is slow relative to the laser pulse width.

PAGE 21

11 Before the surface layer can vaporize, the underlying material reaches its vaporization temperature. Temperature and pressure of the underlying material are raised beyond their critical values, causing the surface to explode. The pressure over the irradiated surface from the recoil of vaporized material can be as high as 10^ Mpa.^^ In this explosive interaction, melting often is not observed around the crater. Fractional vaporization is normally negligible. Power densities in the 10*10^ W/cm^ range cause vaporization, ablation, both of these processes simultaneously, or additional mechanisms that have not yet been identified. The incident beam is partially reflected by the sample surface to a degree that depends on the nature of the surface and that decreases as the temperature of the sample surface is increased. An advantage of using giant pulses is the very high rate at which energy is deposited at the surface, which decreases reflection to the extent it can be neglected. Absorption of incoming photons produces photoelectrons and ions. The conversion of incoming energy to heat is very rapid, leading first to melting, then boiling, over the area of laser impact. Evolution of this process depends on the parameters indicated above. Specifically, the time required for the sample to be raised to its vaporization temperature is given by: f = 7tKpC (T,-To)V4P\ where K is the thermal conductivity (W/cm K), p is the mass density of the sample (g/cm^), C is the heat capacity (J/g K), T^ is the initial temperature (K), T^ is the vaporization temperature of the sample (K) and P is the laser power density (W/cm^). As K, C and T^ depend on the composition of the sample, f, varies with composition. Hence different elements will vaporize at different rates. Moreover, at the periphery of the

PAGE 22

12 crater, high temperature gradients exist, allowing segregation of elements of high and low boiling point. Consequently, the ablated material, in the form of droplets and vapor, may not have a composition wholly representative of the original sample. Table 2-1 lists some of the mechanisms that have been studied. 2.2. Principles of Inductively Coupled Plasma Mass Spectrometry In 1980, Houk et al.^^ at Iowa State University first demonstrated the combination of an argon inductively coupled plasma (ICP) and a quadruple mass spectrometer for elemental analysis of aqueous sample solutions. The technique, now known as inductively coupled plasma mass spectrometry (ICP-MS), developed rapidly, especially after the launch of the commercial instruments in 1983-1984. It is now a standard method for multi-elemental and isotope ratio analysis of diverse biological and geological samples. ^ 5 Recognized advantages of ICP-MS include direct analysis of solutions, calibration against aqueous standards which are readily available, pg mU' detection limits for many elements, a wide elemental coverage and a linear dynamic range of up to 10 orders of magnitude. A typical ICP-MS is composed of an ICP, an interface system, ion lenses, a mass analyzer, and a detector. These are introduced individually. 2.2.1. Inductively coupled plasmas The inductively coupled plasma is an electrodeless discharge in a gas at atmospheric pressure, maintained by energy coupled to it from a radio frequency generator. The gas used is commonly argon. A schematic of the torch arrangement in ICP is shown in Figure 2-5. The plasma is generated inside and at the open end of an

PAGE 23

13 assembly of quartz tubes known as the torch. The torch has an outer tube within which there are two concentric tubes that terminate short of the torch mouth. Each annular region formed by the tubes is supplied with argon by a side tube entering tangentially so that it creates a vorticular flow. The center tube, through which the sample is introduced to the plasma, is brought out along the axis. The outer gas flow, i.e., coolant flow, protects the tube walls and acts as the main plasma support gas. The auxiliary flow is introduced to the inner annular space. It is used to ensure that the hot plasma is kept clear of the tip of the central capillary injector tube, to prevent its being melted. The central gas flow or nebulizer flow conveys the aerosol from the sample introduction system producing a high velocity jet of gas that punches a cooler hole through the center of the plasma. The coupling coil of 2-4 turns of fine copper tube, cooled by a water or gas flow, is located with its outer turn a few millimeters below the mouth of the torch. The RE current supplied from the generator produces a magnetic field that varies in time at the generator frequency so that within the torch, the field lies along the axis. The discharge is initiated in a cold torch by a spark from a Tesla coil, which provides free electrons to couple with the magnetic field. Electrons in the plasma travel around the magnetic field lines in circular orbits and the electrical energy supplied to the coil is converted into kinetic energy of the electrons. At atmospheric pressure, the free electron path before collision with an argon atom, to which its energy is transferred, is only about 1 to 3 mm, and thus the plasma is heated, forming a bright discharge. At the frequency used, the skin effect occurring in RE induction heating ensures that most of the energy is coupled into the outer region of the plasma. The cool injector gas flow punches a channel through

PAGE 24

14 the center of the plasma, carrying most of the sample aerosol, so that little appears in the outer annular part of the plasma. Gas in the center channel is heated mainly by radiation and conduction from the annulus and, while the temperature in the induction region of the plasma may be as high as 1 0,000 K, in the central channel, the gas kinetic temperature is between 5000 K and 7000 K. Power is coupled mainly into the outer region, which is physically distinct from the central charmel through which the sample aerosol travels. Thus the chemical composition of the sample solution can vary substantially without greatly affecting the electrical processes that sustain the plasma. Physical separation between the region where the electrical energy is added and the region containing the sample is one reason for the mildness of physical and chemical interference in the ICP compared to that seen in most other spectro-chemical sources. As an ion source, the ICP has several valuable properties. • Samples are introduced at atmospheric pressure, and may be interchanged readily. • The degree of ionization across the periodic table is relatively uniform and is mainly singly charged ions.^^ • Sample dissociation is very efficient at the gas temperatures experienced and few sample molecular fragments remain. • High ion populations of trace concentrations are produced and therefore potential sensitivity is high. The main disadvantages are the high gas temperature and the pressure at which the ions are produced which require an appropriate interface design to transfer the ions without significant distortion of their relative populations to a mass analyzer.

PAGE 25

15 2 .2 .2. Sample Introduction The ICP requires any sample to be introduced into the central channel gas flow as a gas, vapor or aerosol of fine droplets or solid particles. A wide variety of methods may be used to produce these such as pneumatic or ultrasonic nebulisation of a solution, electrothermal volatilization of microsamples from a hot surface, laser or spark ablation from a solid, and generation of volatile hydrides or oxides from a reaction vessel among others. Most ICP-MS systems are equipped with a pneumatic nebulizer. In a pneumatic nebulizer, a high velocity gas stream produces a fine droplet dispersion of the analyte solution. The larger droplets are removed by a spray chamber which allows only those below about 8 pm diameter to pass on to the plasma. These small droplets carry only about 1% of the solution, which is usually metered to the nebulizer by a peristaltic pump. Although it is not an efficient sample introduction system, the pneumatic nebuliser is convenient, reasonably stable and allows for multiple sample changes. Other introduction methods are used to meet more specialized requirements.^^ 2.2.3. Sampling Interface After formation in the plasma, the analyte ions must be transported into the analyzing mass spectrometer through an appropriate interface system. Figure 2-6 shows the layout of the sampling interface, which is used in the Finnigan Mat SOLA ICP-MS system. It is a three-cone system through which the pressure drops from atmospheric to approximately 10'^ torr. The first cone, the sampling cone, is located 14 mm from the end of the load coil and has an aperture of 1.1 mm at its tip. The inter-space behind it is evacuated by means of a 3 x 10^ mL/s single-stage rotary pump, and with the plasma lit, a

PAGE 26

16 pressure of 2 to 3 torr is maintained in this region of the system. The dynamics of the gas flow through the aperture are such that all the gas within a cylinder of some 4 mm diameter in the plasma flow through the aperture. The effective diameter D of the aperture is given by the relationship: D = (Go/Gt)1/2Dt Gj is the torch gas flow across a diameter Dj and Gq is the gas flow through the sampling orifice. Located 8 mm behind the sampling cone is the skimmer cone, which is used to select the center portion of the jet of plasma passing through the first aperture. This cone is located within the zone of silence behind the first cone where the components in the jet are moving at supersonic speed before the position in which the so called Mach Disc is formed. This is a shock wave where the supersonic jet meets the residual gas in the expansion chamber. The aperture at the end of this cone is 0.8 mm, which is sufficient to allow most of the ions sampled from the plasma to pass through to the next stage of the vacuum system while removing the cooler edge of the jet which are subsequently pumped away by the expansion chamber rotary pump. Behind the skimmer cone is located the accelerator cone. The inter-space between the skimmer cone and the accelerator cone is pumped by a 330 L/s turbo pump which maintains this section of the vacuum system at a pressure of 10'^ torr while the plasma is being sampled. A voltage of -2 kV is maintained on this electrode, and this has the effect of focusing all ions passing through the skimmer cone to a fine cross-over in the 1 mm aperture at the end of this cone. This aperture at the end of the cone also acts as the differential pumping aperture between the intermediate vacuum section and the high

PAGE 27

17 vacuum section of the instrument in which is located the quadruple mass analyzer. The operating pressure in this high vacuum housing is typically 5 x 10 ® torr for the multiplier. 2.2.4. Ion Lens in ICP-MS The ion lenses are also shown in Fig 2-6. In each lens, several electrodes are stung together to confine the ions on their way to the mass analyzer. Each lens incorporates a central disc to prevent photons from the plasma from reaching the detector. The juxtaposition of these non-ideal conditions in ICP-MS means that different ion optical conditions are required to transmit ions of different mlz and that the sensitivity for different elements is not as even across the mass range as the high ionization efficiencies of the different elements would indicate. Furthermore, the extent, and possibly even the direction, of the mass discrimination effect depend on ion lens settings and ion energy, the latter of which can be influenced by plasma potential and plasma operating conditions. 2.2.5. Space Charge Effect Few ions are lost to recombination during the extraction process. Thus, the ion current through the sampler is quite high (~0.1 pA), and the ion current through the skimmer is normally 1 mA. In the plasma and in the supersonic jet, this ion current is balanced by an equal electron current, so the beam acts more or less as if it were neutral. However, as the beam leaves the skimmer, the electric field established by the lens collects ions and repels electrons. The electrons are no longer present to keep the ions confined in a narrow beam, the beam suddenly is not quasi-neutral, and the ion density is still very high. The mutual repulsion of ions of like charge limits the total number of ions

PAGE 28

18 that can be compressed into a beam of a given size. This is called the space charge effect. The space charge effect should be substantial in the ICP-MS at total beam currents on the order of 1 pA. The ion current causes space-charge effects that are reasons for the nonideal behavior of ion optics in ICP-MS. Also, even a small change in the total ion current caused by the addition of just a modest amount of matrix element can change the fraction of analyte ions that get through the lens. Heavy matrix ions are themselves deflected less and stay closer to the center of the ion beam where they can do the most damage. Therefore, the space charge effects are a major cause of matrix interference in ICP-MS. 2.2.6. Ouadrupole Mass Spectrometer A diagram of quadruple mass filter is shown in Figure 2-7. Four straight metal rods or metallised surfaces are suspended parallel to and equidistant from the axis. Opposite pairs are connected together. DC and RF voltages of amplitude U and V, respectively, are applied to each pair. The DC voltage is positive for one pair and negative for the other pair. The RF voltages on each pair have the same amplitude but are opposite in sign. The ions to be separated are introduced along the axis into one end of the quadruple structure at the velocities determined by their energy and mass. The applied RF voltages deflect all the ions into oscillatory paths through the rods. If the RF and DC voltages are selected properly, only ions of a given m/z ratio have stable paths through the rods and will emerge from the other end. Other ions will be deflected too much and will strike the rods and be neutralized and lost there.

PAGE 29

19 12 . 1 . Mass Detectors The magnitude of the ion beam emerging from the quadruple mass analyzer is measured using either a DC Faraday amplifier or a pulse-counting electron multiplier. The arrangement is shown in Figure 2-8. Choice of collector depends on the size of the beam to be measured, and the appropriate collector can be selected electronically by deflecting the beam onto either the Faraday plate or the entrance aperture of the channeltron. The Faraday detector is based on electron counting. A 1-count-per-second (cps) signal on the Faraday cup corresponds to 5 x 10 '^ A, which is equivalent to 6 x 10'“ ions per second arriving the collector. Figure 2-9 shows a diagram of a channeltron electron multiplier. An open glass tube with a cone at one end is used. The interior of the tube and cone are coated with a lead oxide semiconducting material. The cone is biased at a high negative potential and the back of the tube near the collector is held near ground. Relative to either end, the resistance of the interior coating varies continuously with position. When a voltage is applied across the tube, a continuous gradient of potential exists with position inside the tube. Suppose a positive ion leaves the mass analyzer and is attracted to the high negative voltage potential at the cone. When the ion hits the surface, one or more secondary electrons are ejected. These secondary electrons hit another section of the coating and more secondary electrons are emitted. Thus, a discrete pulse containing as many as 10* electrons at the collector after an ion strikes the mouth of the detector. The largest current that can be measured on the system is normally 10'* A, which is equivalent to 6 x 10'“ ions per second arriving at the collector. For signals below 10 '^

PAGE 30

20 A, the multiplier is the preferred detector device. Each ion arriving at the front aperture of the multiplying element creates an electron cascade down the curved multiplying tube, resulting in a pulse of 10’ to 10* electrons being released at its output. The large flux (10^ ions per second) should not be allowed to reach the multiplier as these large flux beams can rapidly destroy the multiplier 2.3. Principles of LA-ICP-MS The roots of LA-ICP-MS lie in part with the logical extension of previous experiences with LA-ICP-OES^^’^^ and in part with the anticipated analytical advantages mentioned in the introduction section. Figure 2-10 shows a typical schematic of the LA-ICP-MS system. When the laser beam is focused on a sample, interaction between the laser beam and the sample allows the conversion of photon energy into thermal energy, which is responsible for the vaporization of most of the exposed solid surface. The material ablated is swept away with an argon stream to an inductively coupled plasma mass spectrometer (ICP-MS) and is analyzed. The laser may be mounted vertically or horizontally. In the latter, the beam is deflected onto the sample cell by a 45° mirror, typically incorporated into a viewing microscope. Optical and visual focusing may or may not be adjusted to coincide. Often the sample may be viewed remotely via a camera. The sample sits within a glass ablation cell, the upper face of which may be slanted at 45° to the vertical to reduce back reflection. For use with ultraviolet lasers, a fused silica window must be installed in the cell.^O A platform or turntable, which is usually under computer control, allows positioning of the sample in the X, Y, Z

PAGE 31

21 directions. Displacements as small as a few micrometers may be employed and the laser firing arranged to occur only between sample movements. According to the pattern of laser shots, individual features such as mineral grains in rocks may be analyzed, or depth profiling, line profiling, area or bulk analyses are possible. The cell is fed an inert gas to entrain the ablated material; this is the equivalent of the nebulizer flow in conventional solution nebulization ICP-MS. Argon or occasionally helium is employed; the latter is reported to improve scavenging. The sample remains at atmospheric pressure and this is an advantage over many other analytical techniques, such as X-ray photoelectron spectroscopy, where the sample must be under vacuum to permit analysis. As air must not enter the ICP, an arrangement for purging the cell to atmosphere must be included, usually by a three-way valve in the cell to ICP transport line. 2.3.1. Calibration Strategies Various strategies are explored to obtain elemental concentrations in solids by LA-ICP-MS. Quantification using a single-point calibration, based on the sensitivity obtained by analysis of a sample containing an element of known concentration, is simple but generally inaccurate. A more sophisticated approach was developed by Hagger et al. using elemental response factors determined by solution ICP-MS and modified based on element-dependent volatilization efficiencies. xhis method, derived from a thermal model of the ablation process, requires either knowledge of one analyte concentration and the ablation temperature T^ or two or more analyte concentrations from which T^ can be

PAGE 32

22 determined. Although external standards are not necessary, the obvious disadvantage is that the concentration data are not fully quantitative. From sensitivities obtained from analysis of a single multi-elemental standard, a response-mass curve may be constructed, allowing elements other than those of known concentration in the standard to be determined. The curve needs to be corrected using a Saha factor which accounts for differences in the degree of ionization in the ICP. The agreement between the determined and accepted concentrations of Mg, Mn, Fe, Cu, Zn, Sr, Ba and U in two limestone reference materials is within ± 10%.^1 Element-for-element calibration against sensitivities obtained by analysis of an external well-characterized matrix has been used for the multi-elemental analysis of diverse biological 63,64 geological samples. 65 Numerous difficulties, however, are associated with this approach: (i) the samples and standard need to be matrix-matched, (ii) the standard must be well characterized and must contain all the elements of interest at concentrations sufficient to give good sensitivity and (iii) the standards must be readily available and inexpensive. Much effort has been expended in finding ways to produce multi-element solid standards cheaply and rapidly. Approaches include the addition of individual compounds to a powdered matrix, mixing and pressing with (or without) a binder,66 the addition of liquid standard solutions to a powdered matrix, drying and pressing,67,68 the production of glass fusions69 and the production of sintered compacts. ^6 One area where calibration is easier is metal analysis, for which more standards exist. Moreover, complete matrix matching of metals is sometimes unnecessary. For

PAGE 33

23 example, Fe, Ni, Cu, Zn, Pd, Pt, Pt, Pb and Bi have been determined in gold, silver and gold-silver alloys using either gold or silver reference materials as standards.^ ^ The ease of ablation and the relatively homogeneous distribution of elements in metals simplify metals analysis. An alternative to solid standards is liquid-based calibration in which a dual flow system allows simultaneous introduction of laser ablated solids and nebulized aqueous solutions. As both forms of material are present in the ICP, calibration against the sensitivities obtained by analyzing aqueous standard solutions is possible. ^2 An obvious disadvantage, is the presence of oxide and hydroxide interference associated with the “wet” plasma. Owing to variations, for example in laser energy output between shots, sequential responses obtained under the same conditions are not identical. To compensate for such fluctuations, internal standardization (IS) is almost always necessary in LA-ICP-MS analysis. This is normally conveniently done using the responses of a minor isotope of a major element, which may be expected to be relatively uniformly distributed in the matrix. Improved precision results from IS on a set of replicate analyses; IS from standard to sample should also improve accuracy and is possible when the concentration of at least one suitable element is known. In addition to being homogeneously distributed, an IS element should behave in the ablation process and in the ICP in a manner identical with that of the analytes. An alternative to IS is to normalize elemental responses to the amplitude of the acoustic wave generated by laser ablation. This improves the precision of concentration

PAGE 34

24 data obtained by LA-ICP-AES.^^ Compensation for variation in transported mass may also be made using light-scattering measurements obtained using a photomultiplier tube (PMT) mounted above the cell to ICP transport line perpendicular to the beam of a He-Ne laser74 Although not as effective in improving precision as the use of IS, this method requires no previous knowledge of the sample homogeneity or elemental composition. A similar system denoted an in-line mass transport measurement (MTM) cell and employed a UV/VIS spectrophotometer.^^ Variations in the integrated total scattering effect of the ablated particles, obtained by summing the sub-peak areas of the intensities produced by this scattering, were used to normalize the analytical results. Accuracy was improved and precision (RSDs) lowered from 25 to 5% using this normalization in determination of multielements in arsenopyrite.

PAGE 35

25 Table 2-1. Laser ablation mechanisms. • Absorption (single, multiphoton, defect initiated, . . .) • Reflection (time-dependent) • Thermodynamics (melting, latency, phase change,. . .) • Plasma ignition • Shock waves (gas) • Stress waves (solid) • Laser-plasma interaction (inverse bremsstrahlung, . . .) • Plasma radiation/heating • Gas-dynamic expansion • Hydrodynamic expansion

PAGE 36

26

PAGE 37

Pumping level 27 Figure 2-2 Energy levels in Nd:YAG laser

PAGE 38

1 00% Mirror Partially reflecting mirrow 28 Figure 2-3. Q-switch laser

PAGE 39

29 Target Figure 2-4. Conceptual drawing of laser ablation

PAGE 40

30 aerosol gas flow (into axial channel) Figure 2-5. Torch layout in ICP.

PAGE 41

Accelerator cone 31 Figure 2-6.

PAGE 42

32 Detector Figure 2-7. Quadrupole mass analyzer.

PAGE 43

Faraday collector and amplifier 33

PAGE 44

34
PAGE 45

CCD camera 35
PAGE 46

CHAPTER 3 INSTRUMENT DIAGNOSTICS AND OPTIMIZATION 3.1. Introduction The purpose of this chapter is to provide a better understanding of the Finigan LA-ICP-MS system. Information that are additional to the knowledge from the manual and other descriptions are provided. The study was focused on the laser ablation system since a detailed study on the ICP-MS operating conditions were described in the dissertation by Baker.^^ the present work, the influence of laser output energy, flashlamp voltage, laser focusing conditions or mass removal etc. were studied and discussed. The effect of intermediate flow rate and forward RE power on the LA-ICP-MS analysis of solids and solutions were also studied and discussed. This information was provided for general purpose; however, more specific information or the effect of modification of the system on particular measurements can be found in the subsequent chapters. The parameters were studied based on measurements performed on NIST glass and High Purity Solution (Charleston, SC) standards. The goal is to provide information that is applicable to all matrices. 3.2. Experimental 3.2.1. Instrumentation An illustration of the LA-ICP-MS setup is show in Figure 2-10. It consists of a Firmigan MAT SOLA ICP-MS (Hemel Hempstead, UK) and System 266 laser ablation 36

PAGE 47

37 accessory. The ablation module consists of a Spectron SL 401 Nd: YAG laser which has been frequency quadrupled to produce an output beam of 266 nm. A CCD camera is used for remote viewing of the sample. The CCD camera is mounted in parallel with the laser; however, the focusing of the CCD camera is different from the focusing of the laser beam which will be discussed in this chapter. The system also includes an x-y-z translation stage for adjusting the focus of the laser beam and for selecting the area of the sample for analysis. This system was modified to allow for x-y-z translation of the sample while the laser is fired repetitively. Figure 3-1 shows the schematic of Finnigan 266 laser ablation system. The general components and their functions in the laser are described as follows: A single flashlamp is used in the cavity and amplifier. It is located in the lamp housing above the rod together with its ceramic reflector. The lamp has leads which terminate at fixing posts mounted on the outside of the housing together with the lamp supply leads. The pockels cell compromises a cube polariser, a wedged KD*P crystal and a glass compensating block. The telescopic oscillator is used to obtain a low beam divergence by discriminating against the higher order modes within the cavity together with efficient energy extraction. It also has a dual purpose in compensating for thermal lensing of the laser rod at different repetition rates. The fundamental wavelength output from a Nd:YAG laser is 1064 nm (Infrared). It is quadrupled by propagation through suitable nonlinear harmonic generating crystals to produce wavelength of 266 nm (Ultra Violet). The harmonic wavelength generated propagates collinearly with the fundamental output. Therefore, as the conversion efficiency is not 1 00%, a beam separator is used to separate the harmonic output from the residual fundamental output. In order to maximize

PAGE 48

38 the conversion efficiency of harmonic generation, the input beam must propagate through the crystal along a unique axis with respect to the crystalline axis. This phase matching angle is also dependent upon the crystal temperature. Consequently, all crystals are mounted in temperature stabilized ovens, held at elevated temperature. The 266 laser system has repetition rates of up to 5 Hz with pulse energies of 1~10 mJ and pulse widths around 8 ns in duration. When focused to a 50 pm spot, laser irradiances of ~8xl0* ~ 8x 10 ® W/cm^ are achieved. The efficiency of detection for all analytes is a function of the ablation process, transport process, ionization in the ICP, and mass spectrometer response. In general, sample translation during ablation provided more representative sampling of the surface; however, it also resulted in a larger mass removal rate and subsequently higher sensitivities.^^ For analyses, the sample is placed in an ablation cell consisting of a 6 cm glass tube with a quartz window for transmission of the UV laser beam. The cell has a total volume of ~80 cm^ The ablated material is transported into the ICP through ~I.O m of 3/16’ i.d. plastic tygon tubing with a stream of argon flow. In addition, the aerosol output from a concentric nebulizer and cooled Scott-type spray chamber can also be continuously introduced to the ICP-MS during ablation with a glass Y-coimector. The ICP is typically operated between 950 and 1300 W with coolant and intermediate flows of 15 L/min and 0.9 L/min, respectively. Sample is carried to the plasma, whether from the ablation chamber, nebulizer, or a combination of both, with a total argon flow of ~1 L/min.

PAGE 49

39 3.2.2. Samples The samples for this study include National Institute of Standards and Technology (NIST) (Gaithersburg, MD) glass 61 1 and 612. Solution samples were diluted from High Purity standard solutions. 3.3. Results and Discussion 3.3.1. Laser Power A single laser shot produces a transient response;^^ repetitive laser shots of a few Hz produce quasi-stable responses, which are more useful for quantification.39 xhe current work was done with multiple laser shots only. Typically, greater laser power produce greater ablated masses, greater transport to the ICP and hence greater responses for a constant analyte concentration. Sufficient response for good quantification (accuracy and precision) is required without saturation of mass peaks at m/z values of interest. Saturation of individual peaks may be tolerable if other elements present at trace concentrations are the only analytical interest. Excess ablated material, however, will have several deleterious results, including memory effects, cone blockage and plasma disequlibrium. Therefore, good control of laser power is necessary for obtaining the necessary sensitivity. The laser power can be adjusted by changing the flashlamp voltage. However, the change of laser power is not linearly corresponding to the change of flashlamp voltage. The laser power was measured using a power meter (Scientech 361). The correlation plot is shown in Figure 3-2. The mass removal of Zn corresponding to the change of flashlamp voltage is shown in Figure 3-3. The plot shows an increase in mass removal with increasing flashlamp voltage.

PAGE 50

40 The laser power at wavelength 1064 and 266 is measured by an oscilloscope (Tektronix, TDS 620A). It was found the voltage reading from the oscilloscope is directly proportional to the laser power, however, no effort was made to calibrate the reading for the actual laser power. It was found at the primary wavelength, 1064 run, the laser power was stable when averaging 4 laser shots over a 2 hour period of time (Figure 3-4). However, when the laser frequency quadrupled to 266 nm, the laser power showed significant fluctuation when averaging 32 laser shots (Figure 3-5). A more stable laser power versus time plot was obtained when averaging 128 laser shots (Figure 3-6). This has indicated that it is necessary to apply over 128 laser shots for any analysis to obtain a reasonably stable signal. An apparent flicker noise is seen from Figure 3-6. It was found that the laser power dropped steadily over a period of time, e.g., ~8% in half an hour time period. This corresponds with the signal decrease when measuring pure copper materials on ICP-MS. The depth and width of the laser track was measured using profilimetery. The mass removal was therefore estimated. The ICP-MS signal of copper was measured at the same time. Figure 3-7 shows the spectral intensity of Cu signal vs. mass removal over a 30 min period. The results show that the signal directly corresponds to the mass removal by laser ablation for pure copper samples. This phenomenon of power decreasing along with the time is most likely due to the temperature increase of the thermostat for the flashlamp, however, no definite reason is known. Because of the decay of the laser power, an internal standard is necessary to correct for the mass ablated for any quantitative analysis. In case an internal standard is not available, the laser power needs to be monitored and corrected.

PAGE 51

41 2 ) 3 . 1 . Laser Focusing Positions The focusing of the laser is crucial to the sampling spatial resolution, and affects sampling sensitivity. Defocusing the laser in order to sample larger areas of a target was critical for either enhancing the signal or obtaining more representative sampling of a heterogeneous sample. The Finnigan Mat LA-ICP-MS system has an optical system to focus the laser beam (UV light) on the sample surface and an optical system to focus the visible light on the sample so that the movement of the sample can be monitored. A digital displacement of the laser focusing position is designated as “d”. The two optical systems share the same lens which can be adjusted; however, the focusing point of the UV light and visible light did not coincide. The difference was measured by firing the laser beam at a piece of carbon paper. The focusing lens was adjusted to produce different focusing distance according to the digital display. The size of the laser beam was measured by measuring the spot size in the carbon paper left by laser beam. At the smallest beam size, the laser is focused on the surface of the object. A plot of focusing distance vs. spot size was obtained (Figure 3-8). When the digital display is 0.0, a sharp image of the sample is seen on the monitor. The plot has indicated that there is a 2 mm difference of the focusing distance of the visual image and the UV light. 3.3.3. Analysis of SiC A series SiC particles of different sizes were used to study the smallest individual particles feasible for the LA-ICP-MS analysis. The SiC particles were spread out on a foam tape so that single particle could be seen under the microscope on the monitor. The laser was repetitively fired at individual SiC particle of different mesh. The response of

PAGE 52

42 the Si signals on mass spectra were plotted vs. the particle size (Figure 3-9). This work demonstrates that particles larger than ~20 um (approximately focus laser beam diameter) should be used to maintain a constant mass signal. For smaller particles (<16 pm) mass signal intensities gradually decreased indicating that only a fraction of the laser radiation effectively vaporizes the particle and the rest of the laser radiation interacts only with the substrate material. Also, the transport of the ablated materials might not be efficient when the particle size is below 16 pm. 3.3.4. Plasma Operating Conditions A detailed study on the plasma operating was described by Scott Baker.^^ The present study is meant to provide additional information about the operating conditions of the plasma for LA-ICP-MS analysis. The effects of intermediate flow rate and forward RF power on solution and solid analysis were studied. It was found that in multi-element analysis a compromise is inevitable since maximum responses do not necessarily occur at the same flow rate and RF power. This is illustrated in Figure 3-10 and 3-11, which shows the sensitivity of different elements at different intermediate flow rate and forward RF power for glass 611,612 and 20 ppb multi-element solutions. In general, RF power at 1300 V produces the best sensitivity for a glass sample, and RF power at 1200 produces the best sensitivity for solutions standards. The sensitivity of each element reaches its maximum at different intermediate flow rates. These results indicate that elements in solids and in solution do not necessarily have their best sensitivity under the same plasma conditions. Tuning with a standard solution or solid materials is necessary if maximum sensitivity is required for a certain element in solid or solutions samples.

PAGE 53

output coupling mirrow and shutter 43 Figure 3-1 Finigan SOLA laser ablation system layout.

PAGE 54

44 Figure 3-2. Laser power vs. flash lamp voltage. Figure 3-3. Weight loss of Zn metal vs. flashlamp voltage.

PAGE 55

laser power (arbitrary unit) 45 150 -| 140 130 120 110 100 90 80 70 0 10 20 ~l ' I ' T" 30 40 50 time (min) 1 — ' — I — ' — I 60 70 80 Figure 3-4. Laser power measured by oscilloscope vs. time at 1064 nm.

PAGE 56

laser power (arbitrary unit) lagg, (arbitrary unit) 46 4 2 0 -2 “I ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 0 2 4 6 8 10 12 14 time(min) Figure 3-5. Laser power measured by oscilloscope vs. time at 266 nm averaging 32 laser shots. 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 "I ' 1 ' 1 ' 1 ' 1 ' 1 ' r 0 2 4 6 8 10 12 time (min) Figure 3-6. Laser power measured by oscilloscope vs. time at 266 nm averaging 128 laser shots.

PAGE 57

2.00E+07 T T 2000 47 A)|sua)U|

PAGE 58

diameter of laser spot (mm) 48 0.30 0.250 . 20 0.150 . 10 0.05T" T ' r -3-2-10 1 2 d (toward sample), (+ away from sample) Figure 3-8. Laser focusing position vs. diameter of laser beam size. At d = 0, a sharp image is seen on the monitor. At d = -2, the laser (266 nm) is focused on the surface of the sample.

PAGE 59

49 Figure 3-9. Intensity of Si in LA-ICP-MS measurement vs. particle size.

PAGE 60

Normalized intensity/Co 50 Figure 3-10. Normalized intensity of 20 ppb multi-element solution vs. forward RF power. Figure 3-11. Normalized signal intensity vs. forward RF power for glass 611.

PAGE 61

Normalized intensity/Ca 51 Figure 3-12. Normalized intensity vs. intermediate flow rate for 61 1 . Figure 3-13. Normalized intensity vs. intermediate flow rate for 20 ppb multielement solution.

PAGE 62

CHAPTER 4 STUDY OF SAMPLING STRATEGY FOR LA-ICP-MS ANALYSIS 4.1. Introduction The preparation of solid samples ean be time consuming, but much less so than many commonly used dissolution procedures, which yield good quantification in the LAICP-MS analysis. Contamination during the production of solid samples is not usually observed, although problems occasionally arise. For example, tungsten from the lining of a tungsten mill has been observed to contaminate sediments during ball-milling. 65 Sample preparations for laser ablation normally involves bringing the sample into powder form so as to obtain homogeneity. The powders are then pressed into pellets under a pressure of 35 Mpa (-5000 Psi) with or without binder. The pellets need to be pressed one by one in laboratory press and care needs to be taken when washing and cleaning the molds to reduce contamination. This can be troublesome when many samples are needed to be prepared. Also, pressing pellets involve the risk of contamination from the container. Pressure need to be carefully controlled when pressing each pallet in a set of samples. The differences in pressure may result in differences in the surface properties and thus can cause inaccuracy of the measurements. Adding binders for materials not easily pressed into pellets also involves the risk of contamination. An alternative method of sample preparation for laser ablation has been used in LIBS (laser induced breakdown spectroscopy) study. The powdered materials were 52

PAGE 63

53 placed on a double-sided tape stuck to a microscope slide. Good precision and accuracy were obtained. No significant background signals were found for the elements of interests. The IR laser used for LIBS study showed no sign of coupling with the tape or glass slides as substrates. Using tape as a sample holder for powdered samples is more convenient compare with pressing pellets provided no problems arise with the background from the tape. Less time for sample preparation is required and no binder is necessary for any powered materials or pressure control. In this study, the feasibility of using tapes as sample holder for LA-ICP-MS study was studied by analyzing soil and leaf standard materials. Three types of tapes were studied for this purpose. 4.2. Experimental The LA-ICP-MS system has been described previously (Chapter 3). Table 4-1 lists the typical ICP-MS operating conditions used in this work. A JSM-6400 Scanning Electron Microscope was also used in this study. It is a high-resolution scarming electron microscope with a modem digital image processing system. NIST leaf samples (1515, 1570, 1573 and 1547) and soils samples (2709, 2711 and 2710) were used in these studies. Three tapes were tested for their feasibility of being used as sample holders. They are double sided transparent tapes from supermarket (I) doubled sided foam tape (II), and specially-made double sided transparent tape (III) designed for research. They are all products of 3M company. A small ~2 cm x 2 cm tape sample holder was used to contain loose powders of leaf or soil samples. One side of the transparent tape was stuck to a microscope glass slide. The powdered samples were

PAGE 64

54 spread out on the tape using a glass slide so that only a mono-layer of the samples was left on the tape. Excess powders were tapped off the tapes. All the samples for the SEM study were coated with Au-Pd. 4.3. Results and Discussion The general considerations in this study were, • The laser output energy must be high enough to produce sufficient sensitivity for trace element detection; however, the laser beam can not penetrate the tape, in which case, the glass slides will be sampled and included in the background, • If the laser beam couples with the tape, it is necessary that the tape does not produce significant background signal for the isotopes of analytical interest. In other words, the tape should not contain that element, or the signal from the tape is constant and can be corrected as background, in which case, the element should homogeneously present in the tape. For a laser output energy of 7 mJ as in analyzing pellet samples, the laser track can be seen breaking the transparent tapes and coupling with the glass substrate beneath. High intensity signals of Si, Mg, Ca were observed in mass spectra indicating that the UV laser coupled efficiently with the polymer (tape) and glass materials. A series of laser energies were tested; if the laser output energy was adjusted to ~2 mJ, there were no sign of the laser beam breaking through the tapes into the glass substrates. Figure 4-1 and 4-2 are the SEM images of the powdered leaf SRM 1547 on tape (III) before and after the ablation. Figures 4-3 4-8 show tapes (I), (II) and (III) before and after laser ablation. A 2 mJ laser output energy was used. Comparison of the structures of the laser track in Figure 4-2 and Figure 4-8 reveals that at this output energy.

PAGE 65

55 the laser beam ablated the tape to some extent. However there were no signs from the ICP-MS measurement that the glass materials were ablated. The elemental content in the tapes was studied by measuring a few chosen elements; their intensities were normalized to the background. Figure 4-9, 4-10, 4-11 show the results obtained from the three tapes. It was found that for tape (I), all the elements chosen in the study showed higher intensity compare to the background. Also, the relative standard deviations (RSDs) were high (>40% in most cases), which indicated that the foam tape had serious elemental contamination in the matrix which could not be corrected since these elements were not evenly distributed in the matrix of the tape. The results for tape (II) indicate even more significant elemental contamination compare to tape (I). In both cases, elemental contaminants include Mg, Ca, Ni, Cu Zn and Pb. The results from tape (III) showed that the elemental contamination were very low and the RSDs were low for all the elements measured. The RSDs were relatively higher for Ca, Ba and Pb because the signals were similar to the background and therefore showed more fluctuation. However, Mn and Sb were the major additives in the production of the tape (III) according to the manufacturing information. In the case of Ca which showed s higher intensity than the background, a relative standard deviation of 6% was obtained, which indicated that the elemental distribution in the tape was homogeneous. Therefore, the Ca contamination could be corrected by subtracting the background. Only serious contamination for this tape was Mn and Sb (not shown on the figure because the intensity was out of range on the figure) which were major additives during the manufacture. Therefore, only when Mn or Sb are the elements of analytical interest in the powdered sample, tape (III) can not be used as a sample holder. However, care needs to be taken to

PAGE 66

56 check the elemental contamination for the isotope of interest before using it as a sample holder. Comparison of Figures 3 through Figure 8 shows that tape (III) has cleaner and smoother surface feature compared with the other two tapes. Tape (III) has been used as samples for soil and leaf samples. Several elements were chosen for the measurements. Since only a monolayer of powdered samples are left on the tape, it is important that the amount of mass ablated produce sufficient sensitivity and precision. The rsds obtained were close to that from measuring pellet samples when an internal standard was used. Therefore, tape (III) has been shown to be a feasible alternative sampling means for pressing pellets from powdered samples. Tape(I) and (II) were not feasible for this purpose because of high elemental contamination. 4.4. Conclusions The use of tape as sample holder for powdered materials was studied. Three tapes obtained from different sources were studied. The special made tape (III) for research use purpose was demonstrated to be feasible for using as sample except when Mn or Sb were elements of analytical interest. However, the tapes obtained from super market were not feasible for this purpose because of significant elemental contamination. The use of a tape as a sample holder can save time and reduce the risk of contamination during sample handling in sample preparation compare with conventional method of pressing pallet. Also, the surface conditions can be kept constant when a monolayer of powdered materials are spread out on the tape.

PAGE 67

57 Table 4-1. Typical LA-ICP-MS operating conditions. Rf power 1200 W Coolant gas flow rate 15 L/min Auxiliary gas flow rate 0.9 L/min Nebulizer gas flow rate 0.65 L/min Ablation chamber flow rate 0.35 L/min Solution uptake rate 1 .0 mL/min Scan conditions ICP-MSDetector electron multiplier Scan range per isotope 0.25 amu Number of passes 128 Number of channels per amu 8 ms Dwell time per channel 4 ms Laser~2 mJ 5 Hz 15 i^m/s Energy per shot Repetition rate Rastering speed

PAGE 68

58 Figure 4-2. Tape (III) with leaf SRM 1547 ablated.

PAGE 69

59 Figure 4-3. Tape (I) unablated. Figure 4-4. Tape (I) ablated.

PAGE 70

60 Figure 4-5. Tape (II) unablated. Figure 4-6. Tape (II) ablated.

PAGE 71

61 Figure 4-7. Tape (III) unablated. Figure 4-8. Tape (III) ablated

PAGE 72

normalized to bakg normalized Intensity 62 111 m Mg Ca V Ni Cu Zn As Cd Ba Pb Figure 4-9. Elemental signals normalized to background for tape (I). 50 ' 45 Â’ 40 1 35 30 Mg A1 Ca Co Cu Zn Rb Sr Ba Pb element Figure 4-10. Elemental signals normalized to background for tape (II).

PAGE 73

normalized intensity 63 3 n element Figure 4-11. Elemental signals normalized to the background for tape (III).

PAGE 74

CHAPTER 5 STUDY OF SOLUTION CALIBRATION OF NIST SOIL AND GLASS SAMPLES BY LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY 5.1. Introduction Laser ablation provides significant benefit especially for the analysis of complex solid samples, such as ceramics, glasses, soils or acid resistant alloys, which are difficult or impossible to prepare for solution sampling. However, there are a number of problems that have prevented LA-ICP-MS from becoming a routine technique for the analysis of solid materials. One major concern has been selection of matrix-matched standards for calibration. Extensive studies have been made to calibrate the analysis of solid materials introduced by laser ablation without using reference samples. A number of approaches using desolvated aerosols generated by pneumatic nebulization, have been tested as alternative sample introduction systems for calibration.^9-85 Hager^^ developed a model that used response factors determined from solution nebulization and modified them based on element-dependent volatilization efficiencies. This work reported accuracy of ±50% for aluminum, steel and copper standards. The use of dual-sample introduction appears to be a more promising alternative for analyzing materials without using matrix-matched standards. It was first used for quantitative LA-ICPatomic emission spectrometry.'^^ However, some good results have been obtained by applying the method to LA-ICP-MS. Cronwell^S et al. applied the method using element relative 64

PAGE 75

65 sensitivity factors (RSFs) to analyze steel Standard Reference Materials. An accuracy of ±20% in general was obtained. We have applied the RSF method in this laboratory to determine the concentrations of trace elements in soil samples.39 The results obtained were mostly within ±20% of the certified values using Ni or Ag as an internal standard. The internal standard proved to be critical in the measurements because of the potential problem of fractionation during the laser ablation. It was observed that the results for some elements were significantly different when different internal standards were used. This work reports the application of a solution method by matching the signal intensity of a matrix element with a solution of a known concentration. A series of solutions with the same concentration of the matrix element but different concentrations of trace elements were made as standards. The method was reported by Falk^^ ^ 1 . when they applied it to standard metal samples. An accuracy of ± 10% was obtained. Metals were considered to be less of a problem because of their uniform chemical environment and their efficient coupling with laser energy. In this chapter, the feasibility of applying solution based calibration of more complex matrices such as soil and glass have been explored. Three normalization factors were introduced in the current work: (1) correction for the variation of ICP-MS operating conditions; (2) correction for the response of each element measured in the solid sample for the calibration by solution standard, and correction for variation of laser power, and (3) correlation of signals from solution and solid for the calibration of trace element in solids. The homogeneity of the matrix

PAGE 76

66 elements as well as some trace elements were studied by LA-ICP-MS as well as by laser induced breakdown spectroscopy (LIBS). 5.2. Experimental 5.2. 1. Instrumentation The experiments were performed with a Finnigan MAT (San Jose, CA, USA) SOLA ICP-MS and Finnigan MAT System 266 laser ablation accessory. Typical operating conditions for the ICP-MS are listed in Table 1. A combined flow of nebulized solution and carrier gas from the ablation chamber was introduced to the ICP-MS for all sample measurements. This provided dual channel introduction to obtain matched plasma conditions for the ablated material and a nebulized aqueous solution standard. The laser was a Nd:YAG with a 266 nm output as described in Chapter 3. It operated at 5 Hz with a typical pulse energy of ~7 mJ and pulse width of 8 ns. In this laser system a monitor and UV laser share a common lens having a focal length of ~7.3 cm. A spot size of ~50 pm in diameter is produced when the laser beam is focused on the surface of the sample. The irradiance produced at the surface of the sample is ~ 4 x 10’° W/cm^. The system 266 laser ablation accessory was modified by using a separate computer to control the x-y-z translation stage. This allowed for the translation of the sample at approximately 15 pm/s while the laser was repetitively fired. Typical analysis times were 80-100 s (signals measured over 400-500 laser shots). In general, sample translation during ablation provided more representative sampling of the surface; however, it also resulted in a larger mass removal rate and subsequently higher

PAGE 77

67 sensitivities.^^ The efficiency of detection for all analytes is a function of the ablation process, transport process, ionization in the ICP, and mass spectrometer response. A microscopic Laser Induced Breakdown Spectroscopy (LIBS) instrument was used to measure the homogeneity of three matrix elements which were candidates for the internal standard study. A detailed description of the instrument has been reported elsewhere.^^ It was used with a lOx working objective, the laser was focused to a ~20 pm diameter spot on the sample surface, corresponding to an irradiance of ~10Â’^ W/cm^. The conic crater left by the laser on a solid surface was ~30 pm deep and mass removal was on the order of 0.1 pg per laser shot. The microscope stage (and hence, the sample) could be translated in a horizontal direction with a minimum speed of 4 pm/s. The laser repetition rate was 0.5 Hz. The spectrometer used was a compact dual-channel UV-VIS fiber optic CCD spectrometer (SD2000, Ocean Optics, Inc., Dunedin, FL, USA). The spectrometer was driven from a laptop computer (Travel Pro supplied with a Pentium II processor at 266 MHz, AMS Tech, USA) via a DAQCard-700 interface (National Instruments, USA). A secondary electron microscopic spectrometry/energy dispersion spectrometry (SEM-EDS) instrument was used in this research. The JSM-6400 Scanning Microscope is a high-resolution scanning electron microscope with a modem digital image processing system. The image processing system enables image enhancement by averaging and integration and the storing of the image data on frame memories. An energy dispersive spectrometer was added to the JSM-6400 so that it can be used as an electron probe X-ray

PAGE 78

68 micro-analyzer, allowing accurate and efficient nondestructive elemental analysis and elemental distribution study of miero areas. 5.2.2. Samples Several soil samples and glass samples were studied to evaluate the feasibility of the solution calibration of these materials. These reference materials included NIST SRM 2709 (San Joaquin Soil), NIST SRM 2710 (Montana Soil, mildly contaminated), NIST SRM 2711 (Montana soil, highly contaminated), NIST Glass 611 and 612. The soil samples were weighed out in 1.00 g portions and pressed into pellets without a binder at a pressure of 35 MPa ( ~5000 psi). To observe the image of the samples for laser ablation analysis, Au-Pd coating was applied to the soil samples. When measuring the concentration of the major elements in soils by SEM-EDS, carbon coating was applied to the samples. The solution standards were prepared from dilution of 10 or 1000 ppm standards (High Purity Standards) with Milli-Q water and Optima HNO 3 (Fisher Scientic, St Louis, MO, USA) to a final acid concentration of 5%. 5.2.3. Solution-based Calibration. It was desired to produce the same plasma conditions while analyzing solid and solution samples. A matrix element. Mg, was chosen to be the internal standard for the soil samples. The mass 25 isotope peak of Mg was chosen because it gave an appropriate signal intensity measured by the electron multiplier detector and was free of interference. A calibration curve was made using five solution standards; it was found that the Mg ion signal at mass 25 with laser ablation of soil 2709 corresponded to a Mg concentration of 2 ppm in a nebulized solution. Therefore, solution standards of different concentrations

PAGE 79

69 of trace elements were prepared to have a concentration of 2 ppm Mg. A 2 ppm Mg solution with no added elements was used as the blank. Three normalization (FI, F2, F3) factors were introduced in this work to allow solution standards to be used for calibration of trace elements in solid samples. The FI, F2 and F3 factors are defined as follows. FI: Corrects for the differences of signal intensity from the mass spectrum for internal standard in each solution standard and in the blank resulting from changes in the ICP-MS operating conditions from one set of measurement to another. For example, when Mg was used as internal standard, FI = Signal of Mg for each solution / Signal of Mg in blank The FI factor was used to correct all ICP-MS responses for each element in the solution standards needed in preparing the solution calibration curve. Corrected signal of each trace element in standard solution = Measured Signal / FI F2: the ratio of the signal of internal standard in solid over the signal of internal standard in the solution. It was used to correct the signal of each element in each measurement in the solid samples. It has two functions: I) Corrects the signal of each element to be measured in each solid sample so as to make them readily calibrated by solution standards. 3) Corrects for changes in laser energy in each measurement of the solid samples. Because each solid samples was measured five times, there were five F2 factors applied to each measurement. For example, when Mg was used as internal standard, F2 = Signal of Mg in each solid sample /Signal of Mg in blank

PAGE 80

70 Corrected signal of each trace element in soil sample = Measured Signal / F2 In essence, we prepared a solution producing the same response for Mg as measured by LA in solid samples. This could only be done by applying F2 to each measurement of the solid sample. Although the response of each element in each measurement in the same solid should be the same when averaging ~500 laser shots for the same sample, there are always differences in response in each measurement because of variation in the laser output energy as well as drifting of the ICP-MS detector. This was partly corrected by F2. By applying F2 to each measurement, the response of Mg in each measurement was forced to be the same; thus the response of each of the other elements in the solid sample were practically corrected for changes in the laser energy. F3: Correlates the signals of each element in the solid and solution to allow calculation of the concentration of trace elements in solid. The F3 factor correlates the signal of the solid and solution standards as follows: F3 = Concentration of internal standard in solid (ppm) / concentration of internal standard in solution (ppm) Therefore, the concentration of element Z in the solid is given by: C/”““ = F3 X C/“'" For example, when Mg was used as internal standard, the concentration of Mg in soil sample 2709 was 1.51%, which gave the same response as a 2 ppm Mg solution standard after correction by F2. Then F3 = 0.0151 X lO* / 2 ppm Mg solution = 7550 ppm Mg solid/ppm Mg solution The concentration of element Z, C^, in soil 2709 solid is then: C = 7550 X C/°'"

PAGE 81

71 where is the concentration (in ppb) of Z in the solution. Therefore, when there were specified concentrations for each element in the solutions, a calibration curve could be made for each element. From the calibration curve, the concentration of a solution corresponding to the response (corrected by F2) measured for each element from solid samples could be obtained, and, the concentration of each element in each solid could be calculated using the factor F3. At most, we spiked 5 elements in the same series of solution standards. Flowever, the results from the overall measurements indicated that the number of elements in the same calibration solution did not influence the accuracy. For NIST glass 611, 612 samples, a certified value of a matrix element was not available. Therefore, a trace element, Sr, was chosen as the internal standard. The concentration of Sr in 611 and 612 NIST was 515.5 and 78.4 mg/kg, respectively. It was found that a solution of 0.02 ppm Sr gave an F2 between Glass 61 1 and the solution close to 1. The F3 that correlated the signal of solid and solution in each case was 25775 and 3920 ppm Sr in solid / ppm Sr in solution. 5.3. Results and Discussion 5.3.1. Selection of an Internal Standard A requirement for the use of internal standardization to obtain accurate results is that the analyte and internal standard be distributed similarly over the area sampled, and that they exhibit similar behavior in transport to and ionization in the ICP. A few elements were candidates as internal standards because they were present at significant levels in all the samples. The consideration here was to choose a matrix element which would respond similarly for plasma conditions for both solid and solution samples. The

PAGE 82

72 ideal internal standard is one that is spatially homogeneous in the solid samples and constant in concentration in all samples. Silicon, the major element, was not chosen because it had interference at all three isotope masses. Two other elements A1 and Na are monoisotopic and so they could not be measured by the electron multiplier at the same time along with trace elements of interest. To determine the homogeneity of the matrix elements, a microscopic laser induced breakdown spectroscopy system (LIBS) was used to spatially scan several matrix elements (Fe, Ca and Mg), in the solids. The results showed that for 15 laser shots randomly firing at the surface of the soil samples, Mg produced a more stable signal than Fe and Ca (see Table 5-2). This was concluded from comparing the standard deviation of the response measured by LIBS of Mg, Ca and Fe with the standard deviation when they were measured in a reference sample. Glass was used as a reference sample for Mg and Ca and a stainless steel sample was used as a reference for Fe. In all cases, the reference samples produced the smallest standard deviations. Ca was found to have the smallest standard deviation in the soil samples among the three matrix elements. Also, Mg has a minor isotope at mass 25 which can be measured by the electron multiplier at the same time as trace elements. Therefore, Mg was used as the internal standard for all soil samples in our work. In the case of glass samples, a trace element was chosen to be the internal standard since the concentrations of the major matrix elements were unavailable. However, it was known that the distribution of elements was much more homogeneous compared with soil samples because of the amorphous nature of the glass. Therefore, a trace element was used as the internal standard. In our case, Sr worked well as an internal standard for glass samples. However, when Sr was tested as an internal standard for soil samples (as a trace element).

PAGE 83

73 the results showed a poorer accuracy compared with Mg for soil samples; relative errors of over ±100% were obtained for some elements. Therefore it was not feasible to use a trace element as an internal standard for the soil samples. To obtain the concentration of the internal standard in the solid sample, x-ray electron-microprobe analysis can be used. Also, in an earlier work, we have shown that it is feasible to use a spiked element as the internal standard for soil samples.39 An energy dispersive EPMA instrument was used to measure the concentration of matrix elements in the soil samples. The detection limit of this EPMA instrument is in the range of ~1% for most elements so only matrix elements can be detected. The results are shown in Table 5-3. Carbon was used as coating for the soil samples when measuring the concentration of the elements. This is because carbon has fewer emission lines compared with Au-PD that would interfere with the emission lines from other elements. Comparison with the certified values showed that EPMA is a reliable method to obtain the concentrations of major elements in the soil samples. 5.3.2. Study of Fractionation Effect Accurate analysis requires that the detected mass composition must be the same as the sample composition. For laser ablation sampling, fractionation (preferential mass removal during laser ablation), which can cause inaccurate analysis, can occur under some conditions. Numerous studies have shown that fractionation can be minimized or eliminated, depending on the sample and laser properties. ^9-92, 81 It is critical in this study that for different samples a constant sensitivity is obtained for each element to be measured relative to the internal standard. The change of

PAGE 84

74 irradiance on the surface of the sample when ablating different samples may cause a change of sensitivity of different element relative to the internal standard. The change of RSF (defined as intensity/concentration relative to that of internal standard) for different elements resulting from a change in laser irradiance was investigated. There are two ways to change the laser irradiance at the surface of the sample: change of laser output energy by changing the flash lamp voltage and change of laser focus area by changing the focusing distance of the laser beam. The RSFs were measured when the flash lamp voltage was adjusted to 9.5, 9.8 and 10.1 kV, which corresponded to laser output energies of ~6, 7 and 8 mJ, respectively. The RSFs were measured when the laser was focused below, on the surface or above the surface of the sample, a total range of 1 .6 mm, which produced laser spot sizes of ~40, 50 and 60 pm in diameter, respectively. The result shows that the RSFs changed somewhat when the laser irradiance was changed. (Figure 5-1 and 5-2), indicating that the samples had to be maintained at the same condition during ablation to guarantee the sensitivity of each element would remain the same. This was done by weighing out the same mass of each soil samples and keeping the same pressure for each sample when pressing the pellet. The focusing and laser output energy was maintained the same during the measurement of each solid sample. 5.3.3. Analysis of NIST Standard Materials The results of analyzing NIST soils 2709, 2710, 271 1 and NIST glasses 611,612 are shown in Tables 5-4 and 5-5. For most of the elements in soil samples, including Pb, V, Ba, Cr, Zn, Cd and Sb, the agreement of the measured value and the certified value are within ±15%. In glass samples, measurements of elements, including Ni, Cu, Rb, Ag

PAGE 85

75 and Co, are within ±10% of the certified values. These results suggest that the use of solutions for calibration is effective for the analysis of these elements. It was also found that for elements in solids under 10 ppm, larger errors resulted, indicating the method was not capable of accurately measuring concentrations below 10 ppm in soil samples. Therefore, the results for Cd and Sb in soil 2709 are not listed in Table 5-4. Several elements gave different behaviors in soil and glass samples. Poor accuracy was obtained for Cu and Ni in soil samples, whereas good results were obtained for them in glass samples. Systematically low concentrations were obtained for Pb and Th in glass samples; however, good results were obtained for Pb in soil samples. Several factors were investigated to determine their influence on the results. It was expected that when factor F2 is close to 1, there is greater probability of producing identical plasma conditions for both solution and the solid. A series of solution standards were prepared for measuring soil 2711 which has the lowest concentration of Mg causing F2 to be 0.25 for the above analysis compared with -0.7 and -0.9 for soil 2710 and 2709, respectively. When a solution concentration of 0.5 ppm was used, F2 was nearly 1 for 271 1. However, the results from measuring 5 elements in soils showed that even with F2 equal to 0.25, it was not a factor that affected the accuracy of the measurement. Similar results were found for Ni and Cu using the new solution standard for soil 2711. The larger errors for Ni and Cu in soil samples were investigated. A different internal standard, Ca, was used to measure these two elements in soil samples. The results showed that good accuracy for Ni was only obtained for 2709 soil. This is consistent with the results obtained for the three soil samples when Mg was used as the

PAGE 86

76 internal standard. One suggestion of the cause of the error was the heterogeneity of the Ni distribution in the samples, since 2710 and 2711 were contaminated soils from unknown sources. An investigation was made by tracking the mass spectra for Ni and all elements that were measured in the current work when translating the ablating laser across the surface of the sample. Time resolved mass spectra show that Ni gave many more signal intensity spikes as compared with all the rest of the elements (Figure 5-3). In fact, the signals of Ni measurements showed bigger differences from one measurement to the other, also indicating a heterogeneous distribution of it in soil samples. The same experiment was repeated with the glass samples. Ni and all other elements showed nearly constant signals, indicating a homogeneous distribution of these elements across the surface of the glass samples. Therefore, we believe that heterogeneity of Ni distribution in the SRMs was the cause for the poor results for Ni in soil samples. SEM images of the ablated and the un-ablated soil 2709 were shown in Figure 5-4 and 5-5. It was apparent that melting occurred during the ablation. This was indicated by the droplet formed along the ablation track. Matrix effect and fractionation caused by preferential ablation may cause the poor results of Cu. However, a detailed explanation is unattainable so far. 5.4. Conclusions A method of using solution standards to calibrate LA-ICP-MS of NIST soil and glass samples was evaluated. In most cases, relative errors of less than ±10% were obtained. Compared with results obtained using RSFs,^^ improvements were made in the case of Pb and Ba, which are toxic pollutants in the environment. The solution

PAGE 87

77 standards do not appear to be as convenient as single solution standards. However, once good accuracy is consistently obtained for certain elements, it is possible to set up a routine to control/monitor the content of these elements in different samples of the similar matrices. The limitation of the method is that an internal standard of known concentration must be present in the solid sample. The homogeneity of the matrix element distribution in the bulk was studied by LIBS. The error of Ni was found to originate from the heterogeneity of its distribution in the bulk. This is not surprising since the two soil samples 2710 and 2711 are samples contaminated from an unknown source. However, the poor accuracy for Cu in soil samples and Pb and Th in glass samples remains unclear. It is suspected that a matrix effect, probably at least in part, was the cause of these results.

PAGE 88

78 Table 5-1. Typical ICP-MS operating conditions Rf power Coolant gas flow rate Auxiliary gas flow rate Nebulizer gas flow rate Ablation chamber flow rate Solution uptake rate 1300 W 15 L/min 0.9 L/min 0.65 L/min 0.35 L/min 1.0 mL/min Scan conditions ICP-MSDetector electron multiplier Scan range per isotope 0.25 amu Number of passes 128 Number of channels per amu 8 ms Dwell time per channel 4 ms LaserEnergy per shot ~7 mJ Repetition rate 5 Hz Rastering speed 15/.im/s

PAGE 89

Table 5-2. Intensity and relative standard deviation of signals of Mg, Ca and Fe from NIST soil 2709, 2710 and 271 1 withlS laser shots measured by LIBS. Steel (for Fe) and glass (for Mg and Ca) were used as references in which Fe or Ca/Mg were considered to 79 (U U-( u 00 c/3 03 o (U a < *

PAGE 90

80 \D VO C<-1 o d d +1 p +1 +1 p lo iri CO p CO p P CN Os IT) d o (N (N o o (U o o "T3 0) (D > c o 4=: meas. '-H o o ON N 1 d d r' ' 1 m ’ — 1 d +1 3 (N m o^ CN 00 o u 1 d d ' 1 Ov VO d 00 +1 d o •vt 00 00 CO fO +1 p o^ 00 CN d 1 1 rd CO CO CN fO ’ — ' d +1 o f— ( Me o p ON 1 1 ON 1 1 CO oo o o d d d +1 +1 +1 CN (N CO CO r-* CN 00 > o o T3 0) w > o o T3 (D c/3 o o in 14=! o o c/3 t;=l o o S 'S
PAGE 91

Table 5-4. Results from calibration of NIST soil sample 2709, 2710 and 2711 using solution standards. 81 r-; cx5 m r-; in oo CN CN »n (N 00 o o NO CN O NO NO in CN On oo cn NO ON NO +1 r+1 rON r-* +i 00 ON . — . ^ ^ g s s Oh s cx s GDh ex W Oh w cx w d "T ^ d "7 b d G CJ c O CD G O C > o G > O •4-^ g U o o CJ ^ W (L) U O O CJ ^ N-^ (D U s d c3 d ^ o o o ^ H ^ U I ON o r^ (N o rCM CN

PAGE 92

82 T3 C C O o W) o s <2 U—i c/5 3 c/3 (D H H XI CL, o U W) < X Di US U 'O 'O C^j d d O) G\ d d (N +1 iTi d d 'O (N (N uS (N +1 yr) (N _0J s ID o. i 00 rX o (N o o w cd ^ oo ^ +1 m (N 00 X OO IT) +1 ro X 00 ^ Csi 0^ ’-H +1 ^ o^ 00 (N cn +1 (N (N 'Tf O (N +! +1 00 ro CN 00 yr) d oi 00 +1 00 +1 ,v«“v fi e c oCL 3 w X o c o U in d o o
PAGE 93

V = 9.5kv V = 9.8kv V=10.1kv 83 CM on CO TT CM o o d d c> 0) O o> Bvu/dSd W) Uh (U c D -*-* :3 cx :3 o CD W 03 (D W) o c (D X Os O r(N o in
PAGE 94

focus 1 mm below sample surface 84 Bwdsa Figure 5-2. RSFs of each element in soil 2709 when changing laer spot size.

PAGE 95

450000 85 6n JO^ S)uno3 o o o o CD IN JOj S)uno3 Figure5-3. Time resolved spectra for Ni, Mg in soil 2710 and Ni in glass 61 1 .

PAGE 96

86 Figure 5-4. SEM image of SRM 2709 unablated. Figure 5-5. Image of SRM 2709 ablated.

PAGE 97

CHAPTER 6 ELEMENTAL ANALYSIS OF SPANISH MOSS USING LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY 6.1. Introduction Tillandsla usneoides L., family Bromeliaceae, commonly called Spanish moss, is an epiphyte which festoons the trees in swamps and hammoeks south from Virginia to Florida and west to Texas, and further southward throughout northern South America. The plant possesses no functional internal conducting system or cuticle, and water absorption occurs over the whole surface of the plant. 93 Therefore these mosses can be used as biomonitors for direct monitoring of wet and dry deposits from the atmosphere. By observing and measuring the changes of a bioindicator, a conclusion as to the kind of pollution (e.g., a heavy metal), its source and possibly its intensity can be drawn.94 Since biomonitoring using mosses was first introduced in 1968 by Riihling and Tyler^^ the use of mosses, lichens and barks for monitoring of heavy metal deposition from the atmosphere has found wide application.96-98 number of analytical tools have been used to determine the elemental concentrations in mosses.99-101 jhe most often used techniques include ICP-AES and eleetrothermal atomization AAS following the digestion of moss samples. Spanish moss is considered as a difficult-to-digest sample which contains a variety of matrix constituents ineluding organic compounds. 1^2 Strong oxidizing agents with various acid mixtures as well as high pressure and temperature 87

PAGE 98

88 conditions are necessary for complete digestion. Microwave digestion (MD) is also widely used. 103-1 10 ^ problem associated with the determination of trace elements using any bomb digestion approach includes the risk of losing elements because of the pressure relief mechanism of the vessel during the digestion. Blank interference, which is mainly due to the contamination of the membrane filter as well as containers, may become an important factor during the preparation of samples. For samples of varied compositions and low elemental concentrations, a memory effect from previously digested samples may cause unreliable analytical results. ^ ^ ^ The acid concentration of its residue also may affect the reliability of the instrumental analysis. Direct analysis of Spanish moss samples has not been reported previously. The use of pulsed laser ablation (LA) sample introduction for ICP-MS enables analysis of solid materials to be performed directly, without sample dissolution. ^ 12,113 Although the relative sensitivity of LA-ICP-MS is poorer compared to solution ICP-MS, because much less material is injected into the ICP, trace detection with submicrogram per gram (ppm) sensitivity in the bulk is routinely achieved. Because the ablation yield varies with material properties, such as reflectivity, thermal conductivity, and melting and boiling points, it is important to obtain matrix-matching standards that contain all the elements of interest.28 There have been a few reports of full quantitative analysis by LAICP-MS using matched standards and multi-point calibration plots for each element of interest,^2,114-116 ^ using glass Standard Reference Materials (SRMs) for the analysis of minerals. 12

PAGE 99

89 This chapter reports on use of LA-ICP-MS for the direct analysis of Spanish moss samples using NIST leaf SRJVls as standards for calibration. The standard addition method was also studied. The results were compared to those obtained from MD-ICPAES. The sampling strategy, instrumental parameters, fractionation effects were also studied to characterize the strengths and limitations of this approach. 6.2. Experimental 6.2.1. Instrumentation The experiments were performed with a Finnigan MAT (San Jose, CA, USA) SOLA ICP-MS and Finnigan MAT System 266 laser ablation accessory, which has been described in chapters. Typical operating conditions for the ICP-MS are listed in Table 61 . A combined flow of nebulized solution and carrier gas from the ablation chamber was introduced to the ICP-MS for all sample measurements. Plasma conditions were maintained the same for all the standards and samples analysis. Typical analysis times were 80-100 s (signals measured over 400-500 laser shots). A photodiode, connected to a chart recorder, was used to monitor the laser output energy. A quartz plate ( 2'x2' )was placed at a 45 degree angle to the laser beam between the source and the ablation chamber. The quartz window transferred ~96% of the laser beam to the ablation chamber while ~4% of the laser beam was deflected to an UV filter and photodiode. The laser power was manually adjusted to maintain a constant signal. Also, a microscopic LIBS system was used in this study to measure the Ca concentration. The LIBS system was described in Chapter 5.

PAGE 100

90 6.2.2. Samples The NIST archival leaf SRMs were used as matrix-matched standards for the elemental analysis of Spanish moss. These reference materials included NIST SRM 1515 (Apple leaves), NIST SRM 1547 (Peach leaves), NIST SRM 1570 (Spinach leaves) and NIST SRM 1573 (Tomato leaves). Mixed standards were also prepared to produce appropriate concentrations for the calibration curve. They were prepared by weighing out different portions from two standards and mixing in a Spex (Metuchen, NJ, USA) Mixer/Mill Model 8000 for 30 min to ensure homogeneity. The NIST standards were dried in a dissector for 5 days before pressing into pellets for use. All Spanish moss samples were collected with plastic gloves and were placed in plastic storage bags for transport to the laboratory. The samples were then rinsed to remove wind-blown particles of dust and soil. Milli-Q double deionized water was used in the rinsing, and all samples were handled wearing plastic gloves. After rinsing, the wet moss samples were dried in an oven at 110°C for a 4 hour period. The samples were crushed in a ceramic mill for 25-30 minutes, reducing them to fine homogeneous powder well suited to scooping and weighing. The powder was sealed in clean 60 mL polyethylene bottles for storage prior to pressing pellets and making measurements. The NIST SRM standards and Spanish moss samples were pressed into pellets without a binder at a pressure of 35 MPa ( -5000 psi). The samples for standard addition analysis were prepared by adding 200, 300, 500 and 1000 pL of 1 ppm Mn or Pb standard solution into each -0.5 g portion of solid powders. The samples were then dried at 110°C in the oven for several hours. The

PAGE 101

91 samples were then transferred into plastic vials, mixed for 30 minutes, pressed into pellets without a binder at a pressure of 35 MPa (~5000 psi). 6.3. Results and Discussion 6.3. 1. Elemental analysis of Spanish moss using NIST leaf SRMs as matrix matched standards by LA-ICP-MS The elements chosen for this study were present and detectable in both leaf standard materials as well as in the moss samples. These elements were: Na, Ca, K, Fe, Mn and Zn. The calibration curves prepared using standards should cover the range of sample concentrations. This was true for most of the elements since the leaf standards and Spanish moss are essentially similar materials that have similar concentrations of these elements. However, for some elements, such as Cd, Ba, Pb etc., calibration curves were not available since the concentration of these elements were not sufficient to give a good sensitivity in ICP-MS measurement. Standard addition method need to be used for these elements which will be discussed later. Mixed standards were used to produce at least 3 data points for each calibration curve. The calibration curves for the 5 elements measured were shown in Figure 6-1 and 6-2. The calibration curves for most elements except K show reasonable linearity. The results obtained by LA-ICP-MS for the 6 moss samples were compared to those obtained from microwave digestion ICP-AES (Table 62). The results showed good agreement for Na, Ca, Zn, and Mn obtained by the two methods. However, discrepancies of >5 fold occurred for Fe. The MD-ICP-MS results were compared with that obtained from MD-ICP-AES. (Table 6-3). Good agreement for all elements measured including Fe was obtained for the two techniques. Fe has three isotopes, mass 54, 56 and 57. Mass 54 and 56 overlap

PAGE 102

92 with the major interference from ArN^ and ArO\ which leaves mass 57 the only isotope to be measured. The above results indicate that the interference from ArOH^ at mass 57 is more prevalent when using laser ablation for sampling because moss samples have a high content of organic compounds which would produce O during the ablation process. The increase of O would result in ArOH^ in the plasma. Since ArOH^ can not be corrected by background subtraction during laser ablation process, poor results for Fe were obtained using LA-ICP-MS. However, when solution samples were measured by ICP-MS, because of the stability of ICP-MS, the ArOH" formed from Ar and O from Hp in solution could be easily corrected by background subtraction. The calibration curve for K showed curving (Figure 6-2). Because the concentration of K is high (~2%), the linear dynamic range has been exceeded. Both Faraday and electron multiplier detectors were used to measure the Ca signal. There was good agreement between the results although the Faraday detector gave higher results when measuring Ca (44) (Table 6-2). This was believed to be caused by an interference of CO 2 at mass 44. It was also noticed that when the Faraday detector was used, the precision was poorer. In general, the results obtained indicate that leaf standards can be used for elemental analysis of Spanish moss. A serious limitation of the LA-ICP-MS analysis is the need of appropriate concentrations of each element of interest in the standard. In the case of Pb, the concentrations in the 3 leaf standards are below 1 ppm so they are not detectable by LA-ICP-MS. The standard addition method was evaluated to solve this problem and will be discussed later. The difference in results obtained by LA-ICP-MS with MD-ICP-AES/MS is a result of several causes. The SEM images of the surface of an leaf standard and moss

PAGE 103

93 sample prior to the laser ablation are shown in Figure 6-3 and 6-4. The surface appearance in the SEM are somewhat different for the leaf and moss samples. This difference can be caused by the different conditions for sample preparation, including the process of collection, treatment and pressing pellets. Fractionation and matrix effects can also affect the result of LA-ICP-MS analysis to some extent. A speciation effect was briefly studied using CaO, CaCOj, CaH4(P04)2«H20 and Ca3(P04)2 mixed with cellulose to prepare sample pellets with a Ca concentration of ~1%. Figure 6-5 shows the Ca response in the 4 samples. It shows that Ca3(P04)2 gave the highest Ca response even though it had lowest Ca concentration. A speciation effect may affect the consistency of LA-ICP-MS analysis and MD-ICP-MS analysis. However, when leaf standards were used for the calibration, this speciation effect can be neglected since a similar composition is assumed in the leaf and moss samples. 6.3.2. Study of Fractionation As in soil studies, it is critical in this study that constant sensitivity be obtained for each element to be measured in all samples. A change of laser fluence on the surface of the sample when ablating different samples would cause a change of sensitivity of different elements which would result in poor linearity of the calibration curve. The change of RSFs (relative sensitivity factor) for different elements resulting from a change in laser fluence was investigated. The laser fluence at the surface of the sample could be changed either by changing the flash lamp voltage or by changing laser focal area by changing the focusing distance of the laser beam. The RSFs were measured when the flash lamp voltage was adjusted to produce laser output energies of ~6, 7 and 8 mJ. The

PAGE 104

94 RSFs were measured when the laser was focused below, on the surface or above the surface of the sample, a total range of 1.6 mm, which produced laser spot sizes of ~40, 50 and 60 pm in diameter. The RSFs changed somewhat when the laser fluence was changed for most elements while a dramatic change is seen for Fe due to the interference from ArOH"^. (Figures 6-6 and 6-7). Maintaining constant laser fluence on all of the samples is important for minimizing fractionation. In order to minimize fractionation it was necessary to maintain the laser output energy constant for all measurements and to keep the same pellet thickness so that the laser focal area was maintained the same for all measurements. 6.3.3. Analysis of Moss Samples by LIBS A LIBS instrument was also used to measure the elemental concentration in Spanish moss. LIBS results agreed with LA-ICP-MS when 3 Ca emission line was used for the study (Figure 6-8). The result for Ca was ~2.0% for moss#6 from LIBS measurements for 3 Ca emission lines. This result agreed with the other two techniques. However, no other elements were measurable by LIBS due to the low sensitivity of our LIBS system. 6.3.4. Standard Addition Method The standard addition curves for the determination of Mn in moss10 and II are shown in Figure 6-9. The measured LA-ICP-MS intensities for the analyte, Mn, and for an internal standard, Ca in this case, were used to generate the calibration curve. Ca43 was used to normalize the measured Mn intensities to correct for variations in the amount

PAGE 105

95 of laser-ablated particles reaching the ICP torch during the standard addition procedure, using the expression. ^ UX„) = Ii (X) + Io (X)[l-Ii(IS)/Io(IS)] where Ij(X„) = normalized analyte intensity, plotted on the standard addition curve for i = 1-4 standard additions, I| (X) = intensity for the analyte (e.g., Mn) in each measurement, Ij (IS) = intensity for the internal standard (e.g., Ca) in each measurement for 1 = 1-4 standard additions, and Iq (X) and Io(IS)= intensity for the sample prior to any standard additions. Table 6-4 shows the normalized data for the calibration curve using the standard addition method for Mn in moss10. When Ca was used as internal standard to correct for the mass removal during ablation, the relative standard deviation for the analyte elements were below 6% which indicated a better precision and accuracy. Using this normalization approach, no prior knowledge of the concentration of the matrix component in the standard or samples was required. The precision of LA-ICP-MS intensities was within 10% rsd after the normalization. The concentrations determined by standard addition for moss 10 and 11 were 41.4, (38.8 from MD-ICP-AES) and 34.0 (34.6 from MD-ICP-AES) ppm, respectively. Although sample preparation time was increased, standard additions possessed several advantages. Strict matrix-matching was achieved and the concentration of the internal standard was not required. The latter aspect is a limitation associated with the solution-calibration technique. In general, the standard addition method is well-suited for the analysis of moss samples and other particulate samples, since an element can be homogeneously spiked into the sample of interest. Also, using the standard addition method can bypass the problem of obtaining

PAGE 106

96 the proper concentration of certain elements in the standards. This can be important for monitoring certain elements in Spanish moss samples. For example, in the case of lead, the concentration in the 3 SRM standards are below 0. 1 ppm which can not be detected by LA-ICP-MS. Also, the detection limit of Pb in solution ICP-AES is 10 ppb. The microwave digestion could not produce a Pb concentration high enough for accurate detection by ICP-AES. However, the standard addition method has combined the advantages of high sensitivity of LA-ICP-MS and the use of internal standardization to correct for the fluctuation of laser output energy, good results can be obtained. The concentration of Pb in Spanish moss 10 and 11 measured by LA-ICP-MS were 35.6 and 46.1 ppm respectively. The MD-ICP-MS of Pb in these two samples were 25.5 and 20.1 ppm respectively. The calibration curves for Pb in moss10 and moss11 using the standard addition method are shown in Figure 6-10. 6.4. Conclusions The use of NIST leaf standards for elemental analysis of Spanish samples using LA-ICP-MS is proved to be feasible for most of the elements studied. The results are compared with those obtained from MD-ICP-AES. Good agreement is obtained for Mn, Zn Na and Ca. Poor agreement for Fe and K is discussed. The standard addition method is demonstrated to be reliable for elemental analysis when matrix-matched standards are not available.

PAGE 107

97 Table 6-1. Typical ICP-MS operating conditions Rf power Coolant gas flow rate Auxiliary gas flow rate Nebulizer gas flow rate Ablation chamber flow rate Solution uptake rate Scan conditions ICP-MS1300 W 15 L/min 0.9 L/min 0.65 L/min 0.35 L/min 1 .0 mL/min electron multiplier 0.25 amu Detector Scan range per isotope Number of passes 128 Number of channels per amu 8 ms Dwell time per channel 4 ms LaserEnergy per shot Repetition rate Rastering speed ~4 mJ 5 Hz 15 iu.mls

PAGE 108

Table 6-2. Comparison of elemental analysis by microwave digestion (MD) ICP-AES and LA-ICP-MS 98 'O "rf l/^ rsi m o o o o o m fN| o NO o o 0 X) s <: s < s < s <; s < s c C/3 c/3 cL y y Ph y pL y 1 Oh O di y 1 Oh y oi. y 1 CL, y d y pi, y pi y 3 c/3
PAGE 109

99 c/3 CL, u § c/o U < u G o CO o o s Xi lA ' ifl 15 c: c3 15 4—* g s (U o c o w •C a Dh s o U s m (N A 00 o o o O d d +1 +1 o (N -H -H -H (N 00 y-^ CnJ (N oo X) rn
PAGE 110

Table 6-4. Matrix mass normalization of measurement Mn intensities for moss 11. 100 T3 "O ? D, D, cj o o ro Oc C S C O o o 00 rb 00 VO 'O Oc c S CO o C (U 4—* w •S oo 00 ro % ^ <4-r Q .ti ^ o O o o o W ^ § o o o o o r-o r-H 00 4-^ 4— ' V G O rri O') c: r-oo OV CTs /-*V o o .ti w o o o o o ^ G o o o o o o o o o o 2 o »— 1 oo (N o C ^ o o
PAGE 111

Ca calibration curve 101 sjunoo Figure 6-1. Calibration curves of Mn, Ca, Zn and Fe using electron multiplier detector

PAGE 112

Intensity 102 Na, r= 0.9998 K 6.2x10 6.0x10 5.8x10 5.6x10 5.4x10 5.2x10 5.0x10 4.8x10 4.6x10 4.4x10 4.2x10 4.0x10 3.8x10 8 cone. (wt% ) Figure 6-2. Calibration curves of Na and K using Faraday detector.

PAGE 113

103 Figure 6-3. Leaf 1515 unablated Figure 6-4. moss 12 unablated.

PAGE 114

counts 104 cone of Ca (wt% ) Figure 6-5. Response of Ca43 for different compounds having ~1% Ca in composition.

PAGE 115

6mJ B/mJ DStriJ 105 flashlamp voltage of laser system

PAGE 116

106 ->S/dSd Figure 6-7. Fractionation study of NIST leaf standard 1515, RSFs change vs. change of laser beam spot size (M^m in diameter).

PAGE 117

counts 107 Ca (I) 393.37 nm Ca (111) 422.67 nm cone. (%) Figure 6-8. Calibration curves for Ca using LIBS.

PAGE 118

counts 108 190000 1800001 70000 : 160000150000140000 : 130000120000: 110000: w 1 00000 : c 90000 o 80000 o 70000 60000 50000 40000 30000 20000 10000-1 I ' I ' I ' I ' I — '“n — ' — r -40 -35 -30 -25 -20 -15 -10 -5 1 — ' — r 5 10 15 20 cone, (ppm) a 1300000 -| 1200000 1100000 -^ 1000000 -^ 900000 800000 700000 600000 500000 400000 300000 200000 100000 -^ 0 -4Qy<35 -30 -25 -20 -15 -10 -5 — ' — I — ' — I — ' — I — 5 10 15 20 cone, (ppm) b Figure 6-9. Calibration curve of Mn in standard addition analysis, a) moss #10, b) moss #11.

PAGE 119

109 Figure 6-10. Calibration curves of Pb in Standard addition analysis, a). Moss #10. b) Moss #11.

PAGE 120

CHAPTER 7 LOW LEVEL COPPER CONCENTRATION MEASUREMENTS ON SILICON WAFERS SURFACE USING DIRECT LA-ICP-MS AND SOLUTION SAMPLING ICP-MS 7.1. Introduction The production of extremely clean silicon wafers is important because the electrical characteristics of very large-scale integrated circuits (VLSI) are degraded by the presence of elements such as copper, iron and zinc at surface concentrations as low as 10" atoms/cm^. Since copper has a high electronegativity compared to silicon, it is believed that Cu is reduced by electrons in the silicon wafer surface to reduce the copper and adhere directly onto the silicon surface in diluted HF solution. Several studies have been made seeking to reduce copper deposition on silicon wafers. jo produce a silicon wafer on which the surface concentration of copper is at its lowest level, it is necessary to develop precise and convenient analytical methods to measure the amount of copper contamination. The most often used analytical method for determining trace elements with high sensitivity is Total reflection X-ray fluorescence (TXRF)122,123 which metal concentrations as low as 10'° atoms/cm^ have been quantified. ^23 other methods have also been reported. Shigematsu et al measured copper on silicon wafers at 10'’10" atoms/cm^ using neutron activation analysis. ^^4 McDaniel et al. reported measuring copper on silicon wafers on the order of a few ppb level using trace element accelerator mass spectrometry (TEAMS). ^25 lowest 110

PAGE 121

Ill detection limit was obtained using vapor phase decomposition sampling combined with ICP-MS analysis (VPD-ICP-MS).^^^ and droplet-surface-etching-TXRF(VPD-DSETXRF).^27 Both reported detection limits of the order of 10* atoms/cm^ However, these two techniques require that the wafer surface be sampled in a reproducible manner and the liquid sample be collected for subsequent analysis without introducing trace contaminants. A potential alternative to the use of liquid reagents may be to use LA-ICP-MS. In LA-ICP-MS very little material is ablated and transported into the ICP-MS for analysis. Typical detection limits are in the low pg/g and even ng/g range for ICP-MS detection with laser ablation. However, lower detection limit can be achieved by increasing the amount of ablated mass per laser pulse, which can be achieved by increasing the laserenergy coupling efficiency to the sample. The biggest advantage of using LA-ICP-MS is that it allows samples to be analyzed directly without tedious sample preparation, which might cause contamination or loss of material. Surface analysis for ultra-trace elements has been one of the most important potential uses of LA-ICP-MS. LA-ICP-MS allows both conducting and non-conducting samples to be analyzed. This feature can be important for analyzing contamination on silicon wafers. Another popular direct sampling method, glow discharge (GD) spectrometry is suited for the analysis of conducting materials. This chapter explores the potential of LA-ICP-MS to measure surface copper concentration on silicon wafers. The best operating conditions for measuring copper

PAGE 122

112 signals are discussed. A comparison is made between using direct LA-ICP-MS analysis of standard wafer samples and using solution sampling. 7.2. Experimental 7.2.1. Instrumentation: LA-ICP-MS: The experiments were performed with a Finnigan MAT (San Jose, CA, USA) SOLA ICP-MS and Finnigan MAT System 266 laser ablation accessory as described in Chapter 3. Typical operating conditions for the ICP-MS are given in Table 7-1. The use of dry plasma conditions for sample introduction is necessary when silicon was used as an internal standard to correct for fluctuation of the mass removal. However, there were interference for all three silicon isotopic peaks at mass 28 (CO^, N 2 ^), mass 29 (NjH^) and mass 30 (NO^). These interference can be reduced by 2 orders of magnitude when a dry plasma is used, because some of the molecular species are derived from the aqueous solution. Also, when the electron multiplier is used to measure the copper signal as well as the silicon signal, the background signal must be controlled to guarantee it does not overload the detector. Therefore it is necessary to find an optimized laser output energy for the analysis. The RF power was reduced to 1150 W to reduce the reflected power when a dry plasma condition was used. The juxtaposition of the non-ideal conditions in ICP-MS indicates that different ion optical conditions are required to transmit ions of different m/z. So the sensitivity for different elements is not as even across the mass range as the high ionization efficiencies of the different elements would indicate. Further more, the extent, and possibly even the direction, of the mass discrimination effect depend on ion lens settings and ion energy.

PAGE 123

113 the latter of which can be influenced by plasma potential and plasma operating conditions. Therefore tuning of the instrument, i.e., to find the optimal ICP-MS operating conditions was carried out using the copper signal at mass 63 obtained from ablation of a homogeneous material, NIST 61 1 glass. Tuning of solution signals was carried out using a 1 ppm copper standard solution in 2% HNO3. Compared with the indium solution which is normally used for tuning the instrument, a higher RF power (1400 W) is needed for the optimized copper signal (1100 W for In). When measuring copper solutions prepared from dissolution from the surface of a silicon wafer using HF, a Teflon central tube was used in the torch. To measure solution samples with trace levels of copper, an ultrasonic nebulizer/membrane desolvator ( U-6000AT", Cetec, Inc.) was connected to the ICP-MS system. The ultrasonic nebulizer improved detection limits by enhancing analyte transport efficiency and reducing solvent loading to the plasma. Compared to pneumatic nebulization, signals levels were typically improved by an order of magnitude with the ultrasonic nebulizer. The membrane desolvator resulted in a high sample transport efficiency of the ultrasonic nebulizer and greatly reduces solvent loading in the ICP, minimizing solvent interference’s. 7.2.2. Samples The standards were obtained from IMEC (Interuniversity Micro Electronic Center, Belgium) and were silicon wafers with different amounts of copper contamination on the surface. The concentration of copper on the silicon surface of three standards were 3.4 ± 1.3 x 10'^ atom/cm^ 32 ± 13 x lO'^atom/cm^ and 287 ± 112 x

PAGE 124

114 lOÂ’^atom/cm^. Other solid samples obtained from the Material Science Department at the University of Florida were silicon wafers doped with Ippm CuClj solution (0.5% HF) for 5, 10, 20 and 40 minutes. Solutions standards were prepared by dilution of 10 ppm standards (High Purity standards, Charleston, S.C.) with double-deionized water and Optima HNO 3 (Fisher Scientific, St Louis, MO, USA) to give a final acid concentration of 2%. The standards prepared were 1, 5, 10 and 20 ppb copper solutions for the electron multiplier detector and 0.5, 1 and 2 ppm copper solutions for the Faraday detector. 7.3. Results and Discussion 7.3.1. Direct Analysis of Wafer Standards Since the copper layer on the silicon wafers is very thin, normally of the order of 0 . 01~1 nm, the mass spectrometer was set to measure an average of -600 laser shots to allow sufficient material to be ablated and averaged. The copper signal was taken as the Cu 63 and Cu 65 peak height/areas and used to obtain the isotopic ratio. The Cu63/Cu65 isotopic ratio was found to be 2.0. The same results were obtained when measuring pure copper metal and 1 ppm standard copper solutions. Stable copper signals were recorded for two standards with higher surface copper concentrations. When two wafer standards were used to prepare a calibration curve, the detection limit using LA-ICP-MS was 6.1 x 10'^ atm/cm^. The precision was improved when Si was used as an internal standard. This indicates the copper has diffuse into the silicon wafer matrix. It has been reported that using helium as the ablation environment could increase the sensitivity by 2-10 fold for nanosecond pulsed lasers. ^28, 129 have

PAGE 125

115 tested this by flowing 0.1~0.4 L/min He through the laser ablation chamber with Ar flowing 0.6-0. 9 L/min through nebulizer. No significant improvement in signal intensity was observed. 7.3.2. Analysis of Solution Samples When using the ultrasonic nebulizer, the detection limit for copper in solution was found to be 0.09 ppb, which corresponded to a concentration of 3.5 X 10’° atom/cm’^ of copper on the silicon wafer ( ~ 5 mL HF dissolving from a wafer of 15 cm^ area ). With the concentric nebulizer, a detection limit of 4.8 x 10" atom/cm^ was obtained. Compared with direct analysis by LA-ICP-MS, the detection limit was improved by two or three orders of magnitude when solution sampling was used. However, with solution sampling the copper must be chemically removed from the surface of the silicon wafer, which is lengthy and susceptible to contamination. Therefore, better detection limits result with solution samples, but direct measurement (LA-ICP-MS) is more convenient for samples over 10'^ atom/cm’^. Currently using LA-ICP-MS, typically it is possible to detect signals from surface concentrations on the order of -lO'"' atom/cm’^, whereas solution ICP-MS allows the detection of lower concentrations but only after dissolution of the surface material and sample solution preparation. However, one of the advantages of using dry ICP conditions for laser ablation sample introduction, is that plasma excitation/ionization temperatures and electron number densities are typically higher, compared to a wet plasma obtained with liquid nebulization. Dry plasmas favor ionization and excitation of the analyte. Therefore, improved detection characteristics result with laser ablation. Absolute

PAGE 126

116 detection limits (absolute amount of analyte mass) were shown to be two orders of magnitude lower for laser ablation ICP-MS compared to liquid nebulization sample introduction. In our case, we have shown that a stable signal can be obtained with a standard that has 10’^ atm/cm^. This is especially useful for research involving the measurement of relative amounts of copper on silicon wafers. Figure 7-1 shows a series of samples that were doped in copper solution for various times and measured by LAICP-MS. Fast and accurate results were obtained using LA-ICP-MS for the analysis. Figure 2 shows a SEM image of an ablated wafer sample. 7.4. Conclusions The application of LA-ICP-MS to measure copper contamination on silicon wafer surfaces was explored. Dry plasma measurement is needed when using silicon as the internal standard to correct for the laser energy as well as signal fluctuations. Tuning was carried out by ablating a homogeneous standard, NIST glass 611. The detection limit using LA-ICP-MS was found to be 6. x 10'^ atom/cm^ Direct analysis by LA-ICP-MS has the advantage of simple, rapid sample preparation but the disadvantage of a poorer detection limit. A stable copper signal can be obtained in ~4 min after the initiation of laser ablation because of diffusion of the analyte along the transport interface. Helium gave no significant improvement in the sensitivity when used as an environment gas for laser ablation. The detection limit for copper solution using ICP-MS with the ultrasonic/concentric nebulizer were found to be 4 x 10 '°atom/cm^ and 5 x 10''atom/cm^ respectively, if copper on the silicon surface was dissolved and nebulized into the ICPMS.

PAGE 127

117 Table 7-1. Typical ICP-MS operating conditions Rf power Coolant gas flow rate Auxiliary gas flow rate Nebulizer gas flow rate Ablation chamber flow rate 1150 W 15 L/min 0.9 L/min 0.65 L/min 0.35 L/min Scan conditions ICP-MSDetector electron multiplier Scan range per isotope 0.25 amu Number of passes 1 28 Number of channels per amu 1 6 ms Dwell time per channel 4 ms LaserEnergy per shot ~4 mJ Repetition rate 5 Hz Rastering speed 1 5 jum/s

PAGE 128

118 Figure 7-1 . Response of Cu over time using Si as internal standard.

PAGE 129

119 Figure 7-2. SEM image of ablated wafer standard

PAGE 130

CHAPTER 8 PROFILING OF PATTERNED METAL LAYERS BY LASER ABLATION INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY (LA-ICP-MS) 8.1. Introduction Laser ablation is particularly useful when spatial information is required. Laser ablation allows very small amount of materials to be sampled and transported into the ICP and analyzed by a mass spectrometer. ^30133 Common examples of spatially resolved analysis include grains and inclusions in minerals, depth analysis of coatings or thin films, small features in electronic devices and contaminants and particulates on surfaces. With laser sampling, it is possible to analyze these features on surfaces without the introduction of large amounts of background or substrate material. A lateral and depth resolution of 1 pm can be achieved with carefully designed optics. 38 However, laser ablation-ICP-MS (LA-ICP-MS) normally has fairly poor precision, mostly because of the fluctuation of shot to shot laser output energy. The fluctuations in the arrival time and size of the individual ablated particles transported to the ICP also contributes to the poor precision of the LA-ICP-MS measurement. In the semiconductor industry, it is desirable to develop techniques for controlling metal deposition on silicon wafers. Normally these metals are deposited as layers of different thickness on silicon wafers. The most used techniques for obtaining information about surface and depth profiles include SEM, TEM, SIMS and TRXF.135 However, 120

PAGE 131

121 TRXF does not give spatial information. SEM provides information only on the top layer. Therefore, laser ablation is currently being explored as a method of surface analysis and several applications in this area have been reported. 136138 distribution of metal impurities in silicon wafers was obtained using LA-OES with a spatial resolution of 750 p,m.l39 Wanner et a/.l^O applied a special autofocus system for the reproducible determination of the spatial resolution of trace elements by LA-ICPMS. This system allowed a point-to-point focusing precision of 10-50 pm. The use of laser ablation to characterize thin film or surface material has been reported in several papers. 141-143 However, the use of LA-ICP-MS for spatial profiling of metal layers has not been reported before. Because the laser beam can be focused to 20 pm in diameter, LA-ICP-MS has the potential to be applied to surface profiling. When the laser fluence can be finely controlled, it is possible to detect and identify each metal layer on any substrate, which is very desirable for semiconductor characterization. In the present study, the feasibility of using LA-ICP-MS as a tool to monitor metal patterns on different substrates was explored. The variables of laser and ICP-MS operating conditions were optimized. A new chamber was designed and built for the purpose of faster sample transport to the ICP and to reduce ablated material remaining in the chamber. The results obtained from LA-ICP-MS were compared to those obtained from LA-OES which monitored the emission of the plasma from the laser breakdown at the surface. LA-OES eliminates the time delay necessary for sample transport when using LA-ICP-MS. The results were also compared with that obtained from x-ray EDS (Energy dispersion spectrometry) measurements.

PAGE 132

122 8.2. Experimental 8.2.1. Instrumentation The experiments were performed with a Finnigan MAT (San Jose, CA, USA) SOLA ICP-MS and Finnigan MAT System 266 laser ablation accessory as described in Chapter 3. Typical operating conditions for the ICP-MS are given in Table 8-1. A Nd:YAG laser was used with an output at 266 nm and operated at 5 Hz with a typical pulse energy of ~7 mJ and pulse width of 8 ns. The focused spot diameter was of the order of 20 pm. A new laser ablation chamber was designed and built for the present study (Figure 8-1). The chamber was constructed from plexiglass with a quartz window glued to the top. The volume of the ablation chamber was reduced by approximately 2 fold compared to the old ablation chamber. The new chamber was constructed so that the argon entering and leaving the chamber did not encounter any comers minimizing the retention of particles. To monitor radiation from the laser spark when using laser ablation-optical emission spectrometry (LA-OES), an optical fiber was placed at a distance of 3 cm from the spark region (Figure 8-2). The fiber was terminated with a collection lens (5 mm diameter, 172) and mounted in a precision adjustable mount. The other end of the fiber was connected to an Ocean Optics mini-spectrometer (SD2000, Ocean Optics, Inc., Dunedin, FL, USA) with the following specifications: 230-310 nm spectral range, 3600 mm ' holographic grating, 25 pm slit, 0.16 nm spectral resolution, 2048 pixel linear CCD array. The spectrometer was driven from a laptop computer via a DAQCard-700 interface

PAGE 133

123 (National Instruments, USA). Software was developed to allow the collection of spatial information in both wavelength-integrated and spectrally resolved modes. The wavelength integrated mode was used when different patterns on the surface had to be distinguished without chemical characterization. A signal was a simple arithmetic mean over intensities on all pixels. In the spectrally resolved mode, separate spectral lines were monitored. Signals were peak intensities or peak areas. Therefore, in this mode, a change in chemical composition could be easily detected when the surface was scanned under the laser beam. 8.2.2. Samples The samples used for optimizing the LA-ICP-MS variables were computer chips having copper stripes on the surface of polymer substrates. These stripes were approximately 0.15 mm in width and the distance between each stripe was about 0.25 mm. The semiconductor samples were obtained from the Material Science Department of the University of Florida; they were silicon wafers on which were deposited different thin metal layers, such as Ni, Fe, Ta, Co. Each layer was on the order of a few nm thick. Some were patterned in a stripe shape that had a width ranging from 0.07 mm to 0.3 mm. The distance between each stripe was —0.1 mm. Some were in the shape of squares of two different sizes, 0.25 mm X 0.25 mm and 0.35 mm X 0.35 mm. A few SEM images were shown for the semiconductor samples in stripe and square patters (Figure 8-3, 8-4).

PAGE 134

124 8.3. Results and Discussion 8.3.1. Effect of Ablation Chamber Good gas flow dynamics, which result in fast and efficient sample transport (i.e., small chamber with high gas velocities), is obviously needed for good spatial resolution detection. However, since a quadruple mass spectrometer was used, it was desirable to have a cell (and transfer tubing) volume sufficient to allow dilution and mixing of the sample pulses so that the ICP-MS received a stable input of sample. The new ablation chamber was designed to increase the transport rate of the ablated material from the chamber to the ICP torch. The reduced volume of the new chamber allowed particles to remain for a shorter time in the chamber. Compared with the old chamber, most of the dead volume was removed which aided in the transport of particles to the ICP. When the flow rate and focus were properly adjusted (as discussed later), the new chamber gave improved sensitivity by a factor of 2 fold compared with the old chamber for monitoring copper stripes on computer chips (see Figure 8-5). 8.3.2. Optimization of Argon Flow Rates The normal operating conditions for argon flow was 0.35 L/min through the laser ablation chamber and 0.65 L/min through the nebulizer. This provided the best sensitivity for analyzing solution samples. The total argon flow rate passing through the plasma was required to be approximately 1 L/min. An increase in argon flow rate from the ablation chamber increased the rate of transport of ablated materials from the chamber to the ICP torch which improved the spatial resolution. However, a decrease in flow rate through the nebulizer correspondingly resulted in a decrease in the efficiency of the

PAGE 135

125 nebulizer. Therefore, it was necessary to find an optimized ratio between the two flow rates. A series of measurements were carried out by monitoring the copper signal on the computer chips while changing the flow rates; the nebulizer flow rate to flow rate through the chamber were: 0.80/0.20, 0.60/0.40, 0.40/0.60, 0.20/0.80 and 0.00/1.00 L/min. When the flow rate from the nebulizer was 0.40 L/min and the flow rate from the laser ablation chamber was 0.60 L/min, the copper signals had the best shape (see Figure 8-6), namely a sharp rise, a sharp fall and a definite plateau. At these flow rates, a spatial resolution of ~ 30 pm was obtained. At the combinations of 0.80/0.20 and 0.60/0.40, the sensitivity is better, but the spatial resolution is ~45 pm or higher. Therefore, we used the flow rate combination of 0.60/0.40 throughout the experiment for optimal spatial resolution. However, when sensitivity is the major consideration, for example when metal concentrations are low, it is necessary to adjust the flow rate to compromise spatial resolution for better sensitivity. The actual transport time for copper was measured when using different argon flow rate combinations from the nebulizer and the ablation chamber. A piece of pure copper foil was used in this study. A stop watch was used to measure the time it took for the copper signal to appear on the detector from the start of laser firing. The results are shown in Figure 8-7. It shows that increase in the argon flow rate from the laser ablation chamber decreased the time needed for ablated materials to reach the detector. 8.3.3. Optimization of Laser Focus The focusing of the laser was crucial to the sampling spatial resolution. Defocusing the laser in order to sample larger areas of a target was critical for either

PAGE 136

126 enhancing the signal or obtaining more representative sampling of a heterogeneous sample. Our Finnigan Mat LA-ICP-MS system has an optical system to focus the laser beam (UV light) on the sample and an optical system to focus the visible light on the sample so that the movement of the sample can be monitored. There is a digital displacement of laser focusing position designated as “d”. The two optical systems share the same lens which can be adjusted. In the present study, we found that defocusing the laser improved the sensitivity by 1.5 ~ 2.0 fold. The flow rates were set as discussed above. The sample was first focused at a position such that one could see the image of the sample surface clearly on the monitor (d = 4.5 on the laser focusing scale). Measurements were then made when the laser was defocused in both directions, i.e., towards or away from the sample. The laser was focused on the surface of the sample when d = 2.5. When the laser beam was defocused —0.5 mm above the surface of the sample (d = 3), the best sensitivity was obtained (see Figure 8-8). This result indicated that although the laser beam spot size was the major parameter influencing the spatial resolution, it was necessary to compromise spatial resolution and sensitivity for the best profiling results. While rastering the laser ablation across the sample surface, the sensitivity was improved by defocusing the laser while the spatial resolution was unaffected. 8.3.4. Comparison of LA-ICP-MS with LA-OES Emission spectra were obtained by monitoring the copper on computer chips using LA-OES (Figure 8-9). Since LA-OES requires the ablation chamber to have a transparent window, the old ablation chamber was used in this study (Figure 8-2). The

PAGE 137

127 spectra were detected directly from the laser spark induced on the sample surface. In this case, no material transport from the ablation plume was involved and there was no transport delay, as in LA-ICP-MS analysis. Therefore, the emission spectra represented real time analysis of the surface chemical composition. As we expected, the emission spectra showed the copper profiling without significant tailing which was more representative of the distribution of the copper stripes on the substrate. A spatial resolution of ~10 pm was obtained for LA-OES. Two operational modes (spectrallyresolved and wavelength-integrated) were applied for the LA-OES analysis of the Cucoated substrate. In the spectrally-resolved mode, the Cu II 237 line was used (Figure 89a). In the wavelength integrated mode, all copper spectral lines different from the background were used (Figure 8-9b). The sensitivity of our LA-OES fiber optic minispectrometer was lower than that of the LA-ICP-MS. In order to measure emission spectra, the laser pulse energy had to be increased to ~30 mJ. However, at this laser energy, the mass spectral signals were saturated when using the Faraday cup detector. Therefore, one of the advantages of using LA-ICP-MS for measurements of metal films on substrates was its high sensitivity. This advantage was also demonstrated in measuring the patterned metals on silicon wafers. Therefore, it can be concluded that LA-OES should be used for profiling if the signal is sufficiently strong and better spatial resolution is needed. However, in the case of thin layers of the order of a few nm thick, LA-ICPMS must be used.

PAGE 138

128 8.3.5. Profiling of Patterned Metal Lavers on Silicon Wafers It is shown in figure 8-10 that when the experimental variables are set at the optimal values for profiling, a well-defined spectral profile of Ta was obtained using LAICP-MS (Figure 8-lOa). The sample was Ta deposited on a silicon wafer in square pattern. The tailing of the peak was nearly completely removed. The spikes on the figure were mostly due to the fluctuation of laser power. The sharply defined signal (Figure 810a) of Ta showed a good quality of metal deposition while the poorly defined shape (Figure 810b) indicated diffusion of metal elements or poor deposition quality. The stripe patterned Ni layers on the silicon wafer were also monitored (Figure 8-11). Spatial resolution, defined as the distance between the intersection of the asymptote of the sloping edge with the baseline and the initial peak in the scan, was ~45 pm in Figure 810a and ~100 pm in Figure 8-11. The laser spot depth was measured by a profilometer. The average depth of each laser shot at ~7 mJ pulse energy was about 6 pm. Although the layers were of the order of nm thickness, it was impossible to perform depth profiling with the laser we used. When LA-OES was used to monitor the metal signals of wafer samples, no signal was captured due to the lower sensitivity of our LA-OES instrument. The use of x-ray EDS for monitoring the deposition of metals on wafers was studied. In EDS, x-rays bombard the surface of the wafer and transmit signal into the energy dispersion spectrometer. Because this transmission is completed within very short time, normally in the range of ps, the transport time can be ignored. Therefore, x-ray spectroscopy reflects the real time chemical compositions on the wafer surface. However, x-ray EDS spectra have a low intensity and so require a long counting time.

PAGE 139

129 The longer the x-ray bombard the surface sample, the higher the intensity of the spectra. When the surface metal layers are very thin, or when an underlying layer is to be detected, it takes long time to obtain signals with sufficient intensity for analytical purpose. Figure 8-12 shows that the Ni stripes were monitored by x-ray EDS result in a spectrum with reasonable intensity when the surface is bombarded for 20 min. However, the same spectra obtained by LA-ICP-MS takes only 1 min or less to obtain. 8.4. Conclusions The use of LA-ICP-MS for surface profiling has been explored. It was found when the LA-ICP-MS operating parameters were optimized for profiling, a spatial resolution of 100 pm was obtained when metal layers on silicon wafers were monitored. This application can be useful in many fields, especially where high sensitivity is required for the measurement, such as deposited metal layers on silicon surfaces in the semiconductor industry. LA-ICP-MS was compared with LA-OES in terms of spatial resolution. It was found that LA-OES gave better spatial resolution but lower sensitivity. LA was destructive, completely removing all layers interacting with the laser beam.

PAGE 140

130 Table 8-1. Typical ICP-MS operating conditions Rf power Coolant gas flow rate Auxiliary gas flow rate Nebulizer gas flow rate Ablation chamber flow rate Solution uptake rate Scan conditions ICP-MS1200 W 15 L/min 0.9 L/min 0.65 L/min 0.35 L/min 1.0 mL/min electron multiplier 0.25 amu Detector Scan range per isotope Number of passes 1 28 Number of channels per amu 8 ms Dwell time per channel 4 ms LaserEnergy per shot Repetition rate Rastering speed ~4mJ 5 Hz 15 //m/s

PAGE 141

131 Argon out > Figure 8-1 . The side view and top view of the new ablation chamber.

PAGE 142

CCD camera 132 Figure 8-2. Schematic diagram of LA-OES system.

PAGE 143

133 Figure 8-3. Square-pattern deposition sample. Figure 8-4. Strip-pattern deposition sample.

PAGE 144

2.00E+08 134 s^unoo Figure 8-5. Effect of new laser ablation chamber for LA-ICP-MS when the copper signal was monitored, nc: new chamber with reduced volume, oc: old chamber.

PAGE 145

6 00E + 008 6.00E + 008 135 (s^unoo) X)|SU9)U| (s^unoo) A;isu8)U| (s}unoo) A)jsu8)U| Figure 8-6. Effect of argon flow rate combinations (ratio of argon flow rate from nebulizer/argon flow rate from laser ablation chamber in LA-ICP-MS).

PAGE 146

136 0.65/0.35 0.6/0.4 0.55/0.45 0.5/0.5 0.45/0.55 0.4/0.6 Argon flow rate from nebulizer/from ablation chamber Figure 8-7. Transportation time affected by argon flow rates.

PAGE 147

137 (s)unoo) A}|sua)U| (s^unoo) Afsu8}U| (s)unoa) A)isue)u| (s)uno3) X)!$u8)U| Figure 8-8. Effect of focusing of laser in LA-ICP-MS.

PAGE 148

Cu line at 237 138 (s^unoo) Ajjsueiui Figure 8-9. Spatial profiling of copper strips on a computer chip using LA-OES. a. spectrally-resolved, b. wavelength-integrated

PAGE 149

8x10 139 03 (/) 0 E 00 s 1 C c/3 ft 03 o H kH 4-1 O 0) T3 Cl t3B C • ~a 1 1 O o 1-1 Ph B ,
PAGE 150

6.4x10 r T ^ 1 ' 1 « 1 » "* 1 ' 1 ' 1 r CO o CD o CO o CD o CD o CD o CD o T — T — c — ^ — T — ^ — ^ — X X X X X X X CM o C30 CD CO o hCO CD IT) CO CD CD iri (sjunoo) Ai!SU0JU| (siunoo) Ai!SU0JU| Figure 8-11. Spatial profiling of Ta and Ni stripes on silicon wafers by LA-ICP-MS. Spatial re

PAGE 151

141 c3 X) (U o c cS in 5 Figure 8-12. a. Profile from SEM image, b. Profile from EDS analysis.

PAGE 152

CHAPTER 9 CONCLUSIONS AND FUTURE WORK 9. 1 . Conclusions Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) has become increasingly versatile. In theory, any solid material can be analyzed provided the laser couples with the material, external standards are available and/or internal standards are available. The present work has utilized the two unique characters of laser sampling for ICP-MS analysis to study various solid materials. They are: • Analysis of solid samples is direct and requires no lengthy dissolution processing which may be incomplete and can also potentially introduce contamination to the sample; • The high sensitivity of the ICP-MS allows small samples to be quantified, which is ideal for LA-ICP-MS in that spatial resolution can be used to investigate compositional gradients across a sample. Calibration strategies have been developed and studied when external or internal standards are used. The use of solution calibration of NIST glass and soil samples were studied. An accuracy of ±15% of the certified value for most the elements studied were obtained. The results have shown that the use of aqueous standards for the calibration in laser-ablation ICP-MS can be a very flexible method to determine trace element concentrations in a wide range of matrices. This study has also shown that sample 142

PAGE 153

143 homogeneity information can be obtained by LIBS for matrix elements and LA-ICP-MS for trace elements. The poor accuracy obtained for Ni in soil samples was most likely due to the inhomogeneity of its distribution in the matrix. The effect of laser fluence at the surface of samples on fractionation was also studied by changing the laser out-put energy and beam spot diameter. NIST archival leaf standards were used as external standards for the calibration of elemental elements in Spanish moss. When an internal standard was not available, the laser ablation system was modified to maintain a reasonable constant output energy. Good agreement was obtained by LA-ICP-MS analysis and MD (microwave digestion) ICP-MS and MD-ICP-AES for most elements studied. The standard addition method was demonstrated to be reliable for elemental analysis when matrix-matched standards were not available. The application of LA-ICP-MS to measure copper contamination on silicon wafer surfaces was explored. Dry plasma measurement was needed when using silicon as the internal standard to correct for the variation in laser energy. Tuning was carried out by ablating a homogeneous standard, NIST glass 611. The detection limit using LA-ICPMS was found to be 6.1 x 10'^ atom/cm^. Direct analysis by LA-ICP-MS has the advantage of simple, rapid sample preparation but the disadvantage of poorer detection limits. Helium gave no significant improvement in the sensitivity when used as an environment gas for laser ablation. The detection limit for a copper solution using ICPMS with the ultrasonic/concentric nebulizer was found to be 4 x 10 '“atom/cm^ and 5 x 10''atom/cm^ respectively, if copper on the silicon surface was dissolved and nebulized into the ICP-MS.

PAGE 154

144 Since the laser beam can be focused to 20 in diameter, LA-ICP-MS has the potential to be applied to surface profiling. This use has been explored in the current study. It was found when LA-ICP-MS operating parameters were optimized for profiling. A spatial resolution of 45 pm was obtained when metal layers on silicon wafers were monitored. This application can be useful in many fields, especially where high sensitivity is required for the measurement, such as deposited metal layers on silicon surfaces in the semiconductor industry. LA-ICP-MS was compared with LA-OES in terms of spatial resolution. It was found that LA-OES gave better spatial resolution but poorer sensitivity. LA was destructive, completely removing all layers interacting with the laser beam. Finally, the sampling strategies of powdered materials for LA-ICP-MS analysis was studied. The use of tape as sample holder for powdered materials has been explored. Three tapes obtained from different sources were studied. The special tape (III) for research purposes was demonstrated to be feasible as sample holders without significant elemental contamination in the tape unless Mn or Sb were elements of analytical interest. However, the two tapes obtained from super market were not feasible because of significant elemental contamination. The use of tape as a sample holder can save time and reduce the risk of contamination during sample handling in sample preparation compare with conventional method of pressing pellet. Also, the surface conditions can be kept constant when a monolayer of powdered materials is spread out on the tape.

PAGE 155

145 9.2. Future Work This research has left a few questions to answer with future work. Also, some attempt has been made to open new areas of investigation. The poor precision of Cu in soil samples and Pb, Th in glass samples needs further explanation. The current work has studied the effect of laser fluence on fractionation; however, fractionation during transportation, fractionation between the isotopes etc still needs to be studied. Some tentative results have been obtained for these studies. The ablated particles were collected at the entrance of the ablation chamber and at a distance when the particles get in the plasma. Membrane filters (0.3 pm, Millipore Corp., Bedford, MA) were used to collect ablated particles at the mouth of the ablation chamber and at the entrance of the ICP . The collected particles along with the membrane were then transported into 2 mL optima nitric acid and heated to 50°C so that the metal salt would dissolve in the acid. The RSFs of each elements were studied using Sr as an internal standard (Figure 9-1). Some preliminary results have shown that copper showed significant transportation fractionation. This may in part contribute to the poor results for Cu in the analysis. However, care needs to be taken that all the particles were dissolved into the solution. Another study of RSFs (Fig 9-2 and 9-3) of isotopes have shown that in some cases, the RSF of different isotopes of the same element are different. Therefore, it is necessary to determine the source of interference and choose the correct isotope for any quantitative analysis. The use of an ICP coupled with an Ocean Optic spectrometer for elemental analysis of standard solution samples can be economical because of the low cost of the

PAGE 156

146 spectrometer. An optic fiber was mounted close to the plasma and the emission was captured and transported to the spectrometer. Figure 9-4 shows the calibration curve of Mg and Si for two emission lines. The detection limits for Mg and Si are ~1.0 ppm. However, the emission of the plasma gave a broad background so that most of the emission lines from the analyte were unidentifiable. Care must be taken to block the plasma light so as to obtain emission lines for most elements. The products from laser ablation include molecules, atoms, ions, particles etc. which provides the possibility of laser enhanced dissolution. For any refractory materials that are difficult to dissolve, the laser can be used to enhance the dissolution. In a test experiment, a piece of glass was placed in a Teflon container with ~5 mL 10% nitric acid which barely covered the surface of the glass. The container was sealed with a quartz window. A Nd:YAG laser at 1064 ran was used to ablate at the glass for a period of 15 min generating -500 laser breakdown at the surface of the glass sample. The resulting solutions were collected and measured by ICP-MS. The results showed that the concentration of Si in the solution was -0.50 ppm which corresponded to a mass removal rate -8 ng per laser shot. However, when the laser was fired at the glass sample in solution, the acid splattered to the quartz window which blocked the laser light within a few minutes. A stream of nitrogen guided through the top part of the container, however, did not eliminate the splattering of the acid. This problem needs to be solved before further studies can be carried out. Depth profiling and surface profiling/mapping are other applications of LA-ICPMS. Laser ablation was applied to mapping the surface elemental compositions, 145

PAGE 157

147 The distribution of the elements in depth and on surface can be very useful for the study of various materials. In addition to the ability of LA-ICP-MS for the analysis of solids, other novel use of LA-ICP-MS can be expected. Examples include LA-ICP-MS for the analysis of solutions^^h fQjspeciation studies. ^^7

PAGE 158

RSF/Sr 148 Figure 9-1 . SRFs of elements during transportation of SRM 2709.

PAGE 159

signal/conc/Ca43 counts/conc/Ca43 149 isotopes Figure 9-2. RSFs of isotopes in 50 ppb multielement sulution. isotopes Figure 9-3. RSFs of isotopes in glass SRM 612.

PAGE 160

count 150 Figure 9-4. Calibration curves for Mg and Si using ocean optic spectrometer.

PAGE 161

REFERENCES 1) R. S. Houk, Anal. Chem. 1986, 58, 91 A. 2) G. Holland, A. N. Eaton, Applications of Plasma Source Mass Spectrometry, Thomas Graham House: Science Park, Cambridge, U.K., 1991. 3) S. F. Durrant, J. Anal. At. Spectrom 1999, 1385. 4) C. J. Park and G. E. M. Hall, J. Anal. At. Spectrom 1987, 2, 473. 5) C. J. Park and G. E. M. Hall, J. Anal. At. Spectrom. 1998, 3, 355. 6) J. Wang, J. M. Carey and J. A. Caruso, Spectrochim. Acta 1994, 49B, 193. 7) J. R. Dean, L. E., H.M. Crews and R.C. Massey, J. Anal. At. Spectrom. 1988, 3, 349. 8) C. W. McLeod, J. Anal. At. Spectrom 1987, 2, 549. 9) D. R. Wiederin, R. E. Smyczek and R. S. Houk, Anal. Chem. 1991, 12, 1626. 10) E. R. Denoyer, K. J. Fredeen and J.W. Hager, Anal. Chem 1991, 63, 445A. 1 1) B. Fryer, S. Jackson and H. Longerich, Chem. Geol 1993, 109, 1. 12) S. E. Jackson, H. Longerich, G. Dunning and B. Fryer, Can. Mineralog 1992 30 1049. 1 3) M. Franek, V. Krivan, B. Gerchen and J. Pavel, Mikrochim. Acta 1994, 1 13, 25 1 . 14) B. Gerchec, J. Pavel and O. Suter, Applications of Plasma Source Mass Spectrometry IP, Thomas Graham House: Cambridge, 1993. 15) V. Balaram, Current. Sci. 1995, 69, 640. 16) K. H. Wedepohl, Handbook of Geochemistry, SpringerVerlag: New York 1970' Vol. l-II. 17) T. Kantor and G. Zaray, Spectrochim. Acta 1995, 5 OB, 489. 151

PAGE 162

152 1 8) A. S. Eppler, D. A. Cremers., D.D. Hickmott, M. J. Ferris and A.C. Koskelo, Appl. Spectrosc. 1996, 50, 1175. 1 9) J. C. V. Loon, Selected Methods of Trace Metal Analysis: Biological and Environmental Samples', John Wiley and Sons: New York, 1985. 20) K. E. Javis, A. L. Gray and R. S. Houk, Handbook of Inductively Coupled Plasma Mass Spectrometry, Chapman and Hall: New York, 1992. 21) W. W. Harrison, Inorganic Mass Spectrometry, John Wiley and Sons: New York, 1988. 22) A. Lodding, Inorganic Mass Spectrometry, John Wiley and Sons: New York, 1988. 23) N. Jakubowski, I. Feldmann and D. W. Steuwer, Spectrochim. Acta, Part B 1995 , 50, 639. 24) C. Lazik and R. K. Marcus, Spectrochim. Acta 1993 , 48B, 1673. 25) C. Lazik, R. K. Marcus, Spectrochim. Acta 1992 , 47 B, 1039. 26) S. A. McLuckey, G. L. Glish, D. C. Duckworth and R. K. Marcus, Anal. Chem 1992 , 64, 1606. 27) A. Benninghoven, B. Hagenhoff and E. Niehuis, Anal. Chem. 1993 , 65, 630A. 28) J. C. Vickerman, Chem. In Britain 1987 , 969. 29) M. W. Blades, y4/y?/. Spectrosco. 1994, 48, 11. 30) P. Arrowsmith, Anal. Chem 1995, 59, 1437. 31) N. J. G. Pearce, W. T. Perkins and R. Fuge, J. Anal. Atom. Spectrom. 1992 , 7, 595. 32) P. Van de Weijen, W. L. M. Baiten, M. H. J. Bekkers and P. J. M. G. Vulligns, J. Anal. At. Spectrom. 1992, 7, 599. 33) J. S. Crain, D. L. Gallimore, J. Anal. Atom. Spectrom. 1992, 7, 605. 34) S. Chenery, A. Hunt and M. Thompson, J. Anal. At. Spectrom. 1992, 7, 647. 35) E. R. Denoyer, J. Anal. At. Spectrom. 1992, 7, 1 187.

PAGE 163

153 36) L. Moenke-Blankenburg, Spectrochim. Acta. Rev. 1993 , 15, 1. 37) J. S. Becker, R. S. Soman, T. Becker, V. K. Ponday and H. J. Dietze, J. Anal. At. Spectrom. 1998 , 13,983. 38) E. F. Cromwell, P. Arrowsmith, Anal. Chem 1995 , 67, 131. 39) S. A. Baker, M. Bi., R.Q. Aucelio, B. W. Smith and J. D. Winefordner, J. Anal. At. Spectrom. 1999 , 14, 19. 40) L. Moenke-Blankenburg, Laser Microanalysis', Wiley: New York, 1989. 41) M. Thompson, M. Coulter and F. Sieper, Analyst 1981 , 106, 32. 42) M. Thompson, S. Chenery and. L. Brett, J. Anal. Atom. Spectrom. 1989, 4,\\. 43) A. L. Gray, Analyst 1985 , 7 70, 551. 44) H. T. Maiman, Phys. Rev. Lett. 1960 , 4, 564. 45) L. Moenke-Blankenburg, Laser Microanalysis', Wiley-Interscience: New York 1989. 46) L. Moenke-Blankenburg, Prog. Anal. At. Spectrosc. 1986 , 9, 335. 47) M. V . Allmen, Laser-Beam Interaction -with Materials: Physical Principles and Applications', Springer-Verlag: New York, 1987. 48) N. Bloembergen, 5o//7/ Interaction and Laser Processing', American Institute of Physics: New York, 1979. 49) J. F. Ready, Effect of High-Po-wer Laser Radiation', Academic Press: New York, 1971. 50) Y. V. Afanasyew, O. N. Kroklin and G. V. Sklizov, IEEE J. Quant. Elec. JQE-2 1966, 483. 51) J. F. Ready, J. Appl. Pity. 1 965, 36, 462. 52) A. A. Rostami, R. Grief and R. E. Russo, Int. J. Heat Mass Transfer 1992, 35, 2161. 53) T. J. Goldsack, J. D. Kilkenny, B.J. MacGowan, S. A. Veats, P. F. Cunningham, C. L. S. Lewis, M. H. Key, P. T. Rumsby and W. T. Toner, Opt. Commun. 1982, 42, 55.

PAGE 164

154 54) R. S. Houk, V. A. Fassel, G. D. Flesch, H. J. Svec, A. L. Gray and C.D. Taylor, Anal. Chem. 1980, 52„ 2283. 55) G. M. Hieftje, L. A. Norman. Int. J. Mass. Spectrom. Ion Process 1992, 519, 1 18. 56) R. S. Houk, J. J. Tompson. Inductively Coupled Plasma Mass Spectrometry In Mass Spectrometry Rewiews?-, Wiley: New York, 1988. 57) A. Montaser, D. W. Golightly, Inductively Coupled Plasma in Analytical Atomic Spectrometry, VCH: New York, 1987. 58) J. W. Carr, G. Horlick, Spctrochim. Acta. PartB 1983, 38, 519. 59) T. Ishizuka and Y. Uwamino, Spectrochim. Acta, Part B 1983, 38, 5 1 9. 60) B. J. Fryer, S. E. Jackson and H. P. Longerich, Can. Mineral. 1995, 33, 303. 61) S. M. Eggins, L. P. J. Kinsley and J. M. G. Shelley Appl. Surf. Sci. 1998, 729,278. 62) J. W. HdLggQX, Anal. Chem. 1989, 61, 1243. 63) N. I. Ward, S. F. Durrant and A. L. Gray, J. Anal. At. Spectrom. 1992, 7, 1139. 64) S. F. Durrant, N. I. Ward, Fresenius'J. Anal. Chem 1993, 345, 512. 65) S. F. Dxmani, Analyst 1992, 117, 1585. 66) A. Raith, R. C. Hutton, Fresenius'J. Anal. Chem 1994, 350, 242. 67) R. D. Evans, P. M. Outridge and P. Richner, J. Anal. At. Spectrom. 1994, 9, 985. 68) H. Cousin, B. Magyar, Mikrochim. Acta 1994, 113, 313. 69) W. T. Perkins, R. Fuge and N. J. G. Perkins, J. Anal. At. spectrom. 1991, 6, 445. 70) L. P. Bedard, D. R. Baker and N. Machado, Chem. Geol. 1997, 138, 1. 71) W. Kogan, M. W. Hinds and G. I. Ramendik. Spectrochim. Acta, PartB 1994, 49, 333. 72) S. Chennery, J. M. Cook, J. Anal. At. Spectrom. 1993, 8, 299. 73) H. -M. Pang, D. R. Wiederin, R. S. Houk and E. S. Yeung, Anal. Chem. 1991, 63, 390. 74) S. A. Baker, B. W. Smith and J. D. Winefordner, Appl. Spectrosc. 1998, 52, 154.

PAGE 165

156 93) S. D. Feurt, L. E. Fox, J. Amer. Pharm. Asso. Scientific Edition 1952, XLI. 94) O. Hutzinger, Toxicological and Environmental Chemistry 1993, 40. 95) A. Ruhling, G. Tyler. Bot. Notiser 1968, 121, 321 . 96) D. Briggs, Nature 1972, 238, 1 16. 97) L. Folkeson, Water Air Soil Pollut. 1979a, 11, 253. 98) B. Markert, Indicators for Heavy Metals in the Terrestrial Environment, VCH: Weinheim, 1993. 99) B. Markert, U. Herpin, U. Siewers, J. Berlekamp and H. Lieth The Sci. if the Total Environment 1996, 182, 159. 100) H. Th. Wolterbeek and P. Bode, The Science of the Total Environment 1995, 1 76, 33. 101) C. Y. Zhou, M. K. Wong, L.L. Koh and Y.C. Wee, Talanta 1996, 43, 1061. 102) C. F. Wang, J. Y. Yang and C. H. Ke, Anal. Chim. Acta 1996, 320, 207. 103) H. M. Kingston, L. Bassie, J. Anal. Chem 1986, 58, 2534. 104) P. J. Lamothe, T. L. Fries and J. J Consul, Anal. Chem 1986, 58, 1881. 105) C. S. E. Papp and L. B. Fisher, Analyst 1987, 112, 337. 106) H. Matusiewiz, R. E. Sturgeon, Prog. Anal. Spectrosc. 1989, 72, 21. 107) A. M. Paudyn, R. G. Smith, J. Anal. At. Spectrom. 1990, 5, 523. 108) J. Nieuwenhuize and C. H. Poley-Vos, Analyst 1991, 116, 347. 109) H.-M. Kuss, Fresenius'J. Anal. Chem. 1992, 343, 788. 110) W. Frenzel, Frenius'J Anal. Chem 1991, 340, 525. 111) G. Knapp, Int. J. Envenron. Anal. Chem 1985, 22, 71 . 112) D. R. Wiederin, R. S. Hovk, Appl. Spectrosc. 1991, 45, 1408. 113) P. Arrowsmith, yl«a/. Chem. 1987, 59, 1437.

PAGE 166

155 75) R. J. Watling, J. Anal. At. Speciom. 1998 , 13, 927. 76) S. A. Baker, Analysis of Glass, Ceramic, and Soil Samples Using Laser Ablation Inductively Coupled Plasma Mass Spectrometry, Baker, S. A., Ed.; University of Florida, 1998. 77) N. Imai, Anal. Chim. Acta 1990 , 235, 381 . 78) I. B. Gomushkin, A. R.-Madina., J. M. Anzao, B. W. Smith, J. D. Winefordner, J. Anal. At. Spectrosco. 2000 , 0, 1 . 79) M. Thompson, S. Chenery and L. Brett J. Anal. Atom. Spectrom. 1991, 1 1. 80) L. Moenke-Blankerhurg, T. Schumann, D. Gunter, H.-M. Kuss, and M. Paul J. Anal. At. Spectrom 1992 , 7., 251. 81) P. ArroÂ’wsmith, E. F. Crowell, Appl. Spectrosc. 1995 , 49, 1652. 82) J. M. Cook and S. Chenery, J. Anal. At. Spetrom. 1993 , 8, 299. 83) T. E. Jefferies, N. J. G. Pearce, W. T. Perkins and A. Raith, Anal Commun 1996, 33, 35. 84) L. Monke, P. Schumann., D. Gunther, H. M. Kuss and M. Paul, J. Anal. At. Spectrom. 1992 , 7, 25. 85) J. W. Hager, Anal. Chem. 1989 , 61, 1243. 86) H. F. Falk, B. Hattendorf, K. Krengel-Rothensee, N. Wiehemeit and S. L. Dannen Fresenius'J. Anal. Chem 1998 , 362, 468. 87) S. A. Baker, M. J. Dallavecchia, B. W. Smither and J. D. Winefordner, Anal. Chim. Acta 1997 , 355, 113 . 88) I. B. Gomushkin, B. W. Smith, H. Nasajpour and J. D. Winefordner Hna/. Chem. 1999 , in press. 89) R. E. Russo, Appl. Spectrosc. 1995 , 49, A 14. 90) L. A. Allen, H. M. Pang, A.R. Warren, and R.S. Houk, J. Anal. At. Spectrom 1995 , 10, 25. 91) W. -T. Chan, X. L. Mao and R.E. Russo Appl. Spectrosc. 1992 , 46, 1025. X. L. Mao, W. T. Chan, M. Caetano, M.A. Shannon, and R.E. Russo, Appl. Surf. Sci. 1996 , 96-98, 126. 92 )

PAGE 167

157 114) N. Imai, Anal. Chim. Acta. 1991, 46B, 1803. 115) A. A. V. Heuzen, Spectrochim. Acta. PartB 1991, 46, 1803. 1 16) E. R. Denoyer, K. J. Fredeen and J.W. Hager, Anal. Chem. 1991, 63, 455A. 117) D. P. Daldwin, D. S. Zamzow and A. P. D'Silvia, Anal. Chem. 1994, (5(5, 191 1. 118) W. Kern, D. A. Puotinen, RCA Rev. 1970, 31, 187. 119) T. Shimono, M. Tsuji, Extended Abstract of Electrochem. Soc 1991, Spring Meeting, P7-7, 278. 120) T. Ohmi, T. loaoka, I. Sigiyama and T. Kezuka, J. Electrochem. Soc. 1992, 139, 3317-3335. 121) T. Ohmi, T. loaoka, T. Kezuka, J. Takano, M. Kogure J. Electrochem. Sco. 1993, 140, 811. 122) M. Funabashi, T. Utaka and T. Arai, Spectrochimca ActaPartB 1997, 52, 887. 123) S. Hedge, B. S. Barron and H. H. Lamb. Material Research Society;, 1993; Vol. 315. 124) T. Shigematsu, M. Polasek, H. Yonezawa and M. Katoh, J. Radioanal. and Nuclear Chem. 1997, 216, 231 . 125) F. D. McDaniel, S. A. Datar, B.N. Guo, S. N. Renfrow, Z. Y. Zhao, and J. M. Anthony, H/y?/. Phys. Lett. 1998, 72, 8. 1 26) R. Henry, R&D 1 999, 6LS-1 ILS. 127) A. L. P. Rotondaro, T. Q. Hurd, H. F. Schmidt, I. Teerlinck, M. M. Heyns and C. Claeys, Mat. Res. Soc. Symp. Proc. 1995, 386, 183. 128) S. M. Eggins, L. P. J. Kinsley and J. M. G. Shelley, 1998, 127-129, 913. 129) A. P. K. Leung, W. T. Chan, X.. L. Mao, and R. E. Russo, Anal. Chem. 1998, 70, 4709. 130) D. Gunter, S. E. Jackson and H. P. Longerich. Spectrochim. Acta. Part B 1999, 54,381. 131) E. Schroeder, M. Hamester and M. Kaiser, Appl. Surf Sci 1998, 127-129, 292.

PAGE 168

158 132) H. F. Falk, B. Hattendorf, K.K. -Rothensee, N. Wiebemeit, and S.L. Dannen, FreseniusJ. Anal. Chem. 1998 , 362, 468. 133) S. Shuttleworth, D. T. Krenser, J. anal. At. Spectrom. 1998 , 13, 367. 134) V. Kanicky, I. Novotny, Jan Musil and Jean-Michel Mermet, Appl. Spectrosc. 1997 , 51, 1042. 135) V. Penka, W. Hub, Spectrochim. Acta PartB 1989 , 44, 481. 136) K. Yamamoto, D. A. Cremers, M.J. Ferris and L.E. Foster Spectrosc. 1996 , 50, 111. 137) A. E. Pichahchy, D. A. Cremers., M.J. Ferris, Spectrochim. Acta Part B 1997 , 52, 25. 138) B. C. Castle, K. Talabardon, B. W. Smithe and J.D. Winefordner yip/?/. Spectrosc. 1998 , 52, 649. 139) D. Romero, J. M. F. Romero, and J. J. Lasema. J. Anal. At. Spectrom. 1999 , 14, 199. 140) B. Wanner, C. Moor, P. Richner, R. Bronnimann and B. Magyar Spectrochim. Acta Part B \999, 54,289. 141) T. Gilbert, B. Dubreuil, M.F. Barthe and J.L. Debrun J. Appl. Phys. 1993 , 74, 3506. 142) K. Tanaka, N. Shirai and T. M. Chen, J. Appl. Phys. 1998 , 83, 7595. 143) A. I. Gusev. A. Proctor, Y. 1. Rabinovich and D. M. Hercules, Anal. Chem. 1998 , 67, 1805. 144) P. Krause, C. Hall, Kerry Joyce, Cameron W. McLeod ; Petra Krause, C. H., Kerry Joyce, Cameron W. McLeod, Ed.; ICP Information Newsleter, Inc., 2000, pp 212. 145) H. Toland, B. Perkins, N. Pearce , H. Toland, B. P., N. Pearce, Ed.; ICP Information Newsletter, Inc., 2000, pp 213. 146) P. K. Prabhu, S. Vijayalakshmi, T. T. Mahalingam, K. S. Viswanathan and C. K. Mathews, J. Anal. At. Spectrom. 1993 , 8, 565. 147) J. L. Neilsen, A. Abildtrup, J. Christensen, P. Watson, A. Cox and C. W. McLeod Spectrochim. Acta, Part B 1998 , 53, 339.

PAGE 169

BIOGRAPHICAL SKETCH Melody Bi was bom in Harbin, Heilongjiang Province, China in 1968. She received her Bachelor of Chemical Engineering degree from Shanghai Jiao Tong University, China, in June 1990. After that, she worked in the Beijing Tianyuan company as an assistant chemical engineer from 1990 to 1992. She Joined LufthansaAir China Joint Venture in November 1994 as a chemical engineer. In September 1994, she went to Western Kentucky University and obtained her masterÂ’s degree in chemistry in May 1996. She worked at the Center for Applied Energy Research in Lexington, Kentucky for the summer and joined the University of Florida in the Fall of 1996. She expects to receive her Ph. D. of Science degree in Chemistry from University of Florida in August 2000 . 159

PAGE 170

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. les D. Wineforcmer jraduate Research Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Willard W. Harrison Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Daniel R. Talham Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Martin Vala Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Paul H. Holloway Distinguished Professor of Materials Science Engineering