XRD Instrument

FACULTY OF SCIENCES

Department of Analytical Chemistry X-Ray Microspectroscopy and Imaging Group (XMI)

Implementation, Analytical Characterization and Application of a Novel Portable XRF/XRD Instrument

Thesis submitted to obtain the degree of Master of Science in Chemistry by

Robin DE WOLF

Academic year 2011 - 2012

Promotor: prof. dr. Laszlo VINCZE Copromotor: dr. Bart VEKEMANS

FACULTY OF SCIENCES

Department of Analytical Chemistry X-Ray Microspectroscopy and Imaging Group (XMI)

Implementation, Analytical Characterization and Application of a Novel Portable XRF/XRD Instrument

Thesis submitted to obtain the degree of Master of Science in Chemistry by

Robin DE WOLF

Academic year 2011 - 2012

Promotor: prof. dr. Laszlo VINCZE Copromotor: dr. Bart VEKEMANS

Acknowledgments First of all I wish to thank my promoter Prof. dr. Laszlo Vincze. Thanks to him I had the opportunity to work with this brand new portable XRF/XRD instrument, and the chance to visit the Deutsches Elektronen Synchrotron (DESY) facility in Hamburg which was a great and instructive experience. I also want to thank him for the time he spend answering my questions and the interest he showed in this master thesis topic. Secondly, I want to thank my supervisor dr. Bart Vekemans for the help with the XRF data processing software (AXIL and MICROXRF2) and the guidance during the measurements in the Mayer van den Bergh museum. Thank you Bart, for the interesting conversations, for the useful tips to perform a thorough research of the Surface Monitor, and for answering my questions at all times. I would like to thank dr. Ettore Di Masso and Andrea Bianco from the Assing S.p.A. company. Thank you Andrea, for teaching me how to work with the Surface Monitor, for the nice chats during the training, and for the repairs of the instrument. I also want to express my gratitude to Prof. dr. Maximiliaan Martens, Prof. dr. Peter Vandenabeele and dr. Claire Baisier (curator of the Mayer van den Bergh museum in Antwerp, Belgium) for involving the Surface Monitor in the characterization of the “Dulle Griet” painting and giving me the opportunity to perform measurements on this famous work of art. Furthermore, I would like to thank the Raman Spectroscopy Research Group for providing the painting “Colorful flowers in a vase”. This painting was used as one of the first real applications of the Surface Monitor. Special thanks go to Prof. dr. Peter Vandenabeele, for the use of the Olympus Innov-X DELTA handheld XRF analyzer. I would like to thank my college master thesis student Pieter Tack for the fine moments, the relaxing breaks and for his minimum detection limit program. I also want to thank the members of the X-ray Microscopy and Imaging (XMI) group for their support. Especially, Jan Garrevoet who helped me with the correction of my first thesis drafts. I would like to express my gratitude to dr. Tom Schoonjans for his support concerning the use of our computer infrastructure. I also want to thank Lien Van de Voorde for assisting me during the measurements with the Olympus Innov-X DELTA handheld XRF analyzer. iii

iv Verder wil ik nog Jarne, Birgit, Kevin, Heleen, Bram en Jonathan bedanken, voor de gezellige lunchpauzes van het voorbije jaar. Het uurtje kaarten tijdens de middag was de ideale ontspanning tijdens dit thesisjaar. Vervolgens zou ik ook nog graag mijn ouders willen bedanken, omdat ze mij de kans gaven deze opleiding te volgen. Ik wil hen ook bedanken voor de ondersteuning tijdens de voorbije vijf jaar, en omdat ze altijd klaarstonden om mij te helpen. Tenslotte, wil ik graag mijn vriendin Stephanie bedanken, voor haar motivatie en steun gedurende de laatste vijf jaar. Bedankt Stephanie voor de gezellige weekends, waardoor ik het thesissen even kon vergeten.

vi

Contents 1 Objectives and Outline of the Thesis

1

2 Principles of X-Ray Fluorescence and X-Ray Diffraction

3

2.1

X-Rays: Definition and Discovery . . . . . . . . . . . . . . . . . . . . . . . . . .

3

2.2

Principles of X-Ray Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . .

4

2.2.1

Interaction of X-Rays with Matter . . . . . . . . . . . . . . . . . . . . .

4

Principles of X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . .

8

2.3.1

Scattering of X-Rays by Electrons and Atoms . . . . . . . . . . . . . . .

9

2.3.2

Bragg’s Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9

Principles of the Surface Monitor’s Components . . . . . . . . . . . . . . . . . .

10

2.4.1

X-Ray Tube Based X-Ray Production . . . . . . . . . . . . . . . . . . .

11

2.4.2

Semiconductor Based X-Ray Detection . . . . . . . . . . . . . . . . . . .

12

2.4.3

Bragg-Brentano θ : θ Set-Up

14

2.3

2.4

. . . . . . . . . . . . . . . . . . . . . . . .

3 The Portable XRF/XRD Spectrometer: Assing’s Surface Monitor

15

3.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15

3.2

Description of Assing’s Surface Monitor . . . . . . . . . . . . . . . . . . . . . .

19

3.3

Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

3.3.1

Set-Up and Positioning of the Surface Monitor . . . . . . . . . . . . . .

22

3.3.2

XRF Acquisition and Data Processing . . . . . . . . . . . . . . . . . . .

22

3.3.3

XRD Acquisition and Data Processing . . . . . . . . . . . . . . . . . . .

24

4 Results: Characterization and Applications of the Surface Monitor 4.1

4.2

29

Surface Monitor’s Performance . . . . . . . . . . . . . . . . . . . . . . . . . . .

29

4.1.1

The Laser Interferometer’s Performance . . . . . . . . . . . . . . . . . .

29

4.1.2

Optimal XRF/XRD Acquisition Distance . . . . . . . . . . . . . . . . .

32

4.1.3

Beam Size and Position of the Beam on the Sample . . . . . . . . . . .

34

4.1.4

XRF Minimum Detection Limits . . . . . . . . . . . . . . . . . . . . . .

37

4.1.5

Influence of Slits and Pinholes on XRD Spectra . . . . . . . . . . . . . .

41

Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

44

4.2.1

XRF/XRD Analysis of the Painting “Colorful Flowers in a Vase” . . . .

47

4.2.2

“Dulle Griet” from Pieter Bruegel the Elder . . . . . . . . . . . . . . . .

48

vii

viii

Contents

5 Summary and Conclusions

55

A Standard Reference Material

59

A.1 NIST SRM 1412 Multicomponent Glass Standard . . . . . . . . . . . . . . . . .

59

A.2 NIST SRM 660b Diffraction Standard . . . . . . . . . . . . . . . . . . . . . . .

59

B Flo2spe-Converter

63

Bibliography

67

Nederlandstalige Samenvatting

71

Chapter 1

Objectives and Outline of the Thesis The aim of this master thesis is the establishment of the experimental methodology and the analytical characterization of a novel portable X-ray fluorescence (XRF)/X-ray diffraction (XRD) instrument called Assing’s Surface Monitor. The Surface Monitor was recently acquired from the company Assing S.p.A. (Rome, Italy) by the Analytical Chemistry Department of Ghent University (Belgium). This instrument is described to be used for studying art objects, e.g. paintings and murals, by combining two powerful non-destructive analytical techniques: Xray fluorescence and X-ray diffraction. Before the instrument can be applied for the study of cultural heritage objects, some questions need to be answered concerning the analytical performance of the portable instrument. In this master thesis, the analytical figures of merit and recommended handling of the Surface Monitor are determined. More specifically, how this instrument must be used in order to obtain reliable and high quality XRF/XRD results. Following the description of the objectives and outline, Chapter 2 describes the principles of Xray fluorescence and X-ray diffraction. Next to describing the physical properties of X-rays, also the production of X-rays by X-ray tubes, the detection of X-rays by semiconductor detectors and the Bragg-Brentano θ : θ set-up, used for the acquisition of XRD spectra, are discussed. Chapter 3 gives a brief overview of already existing portable XRF/XRD instruments. Furthermore, a full description of the Surface Monitor is given, including a description of the installation/set-up procedure of the instrument and the XRF/XRD data processing. Chapter 4 consists of two main parts. The first part describes the results of the Surface Monitor’s characterization: characterization of the laser interferometer and the determination of the optimal distance using this device, characterization of the beam size and its position on the sample, calculation of minimum detection limits for XRF, and finally the influence of the collimators on the XRD spectra. The second part of this chapter deals with two applications of the Surface Monitor. Finally, a summary, together with conclusions, is formulated in Chapter 5.

1

2

Objectives and Outline of the Thesis

Chapter 2

Principles of X-Ray Fluorescence and X-Ray Diffraction This chapter introduces the basic concepts of X-ray fluorescence (XRF) and X-ray diffraction (XRD) that are applied when using the novel portable XRF/XRD spectrometer of the company Assing S.p.A. [1]. Next to the physical properties of X-rays, also the production of X-rays by an X-ray tube, the detection of X-rays by using a semiconductor detector and the Bragg-Brentano θ : θ set-up, used for the acquisition of XRD spectra, are described in this chapter.

2.1

X-Rays: Definition and Discovery

X-rays can be classified as short wavelength electromagnetic radiation, generated in nature by slowing electrons in the outer field of an atomic nucleus or by changing the bound states of electrons involving transitions between the inner electronic shells of an atom. Generally the wavelength of this type of electromagnetic radiation is lower than 10 nm (see Fig. 2.1) [2]. In case of XRF analysis, typically radiation is used with an energy between 1 keV and 100 keV which corresponds to a wavelength of approximately 1 nm down to 0.01 nm [3].

Figure 2.1: Spectrum of electromagnetic radiation from infrared radiation to hard X-rays (IR, infrared; VIS, visible light; UV, ultraviolet; VUV, vacuum ultraviolet; EUV, extreme ultraviolet). Figure adapted from Beckhoff et al. [2].

3

4

Principles of X-Ray Fluorescence and X-Ray Diffraction

Wilhelm Conrad R¨ ontgen discovered X-rays on November 8, 1895 at the Physics Institute of Julius-Maximilian’s University of W¨ urzburg, Germany. He discovered X-rays by accident while studying cathode rays using a low pressure gas-filled discharge tube, a so called Hittorf-Crookes tube. Near the tube he noticed a weak visible photoluminescence which was emitted by a Barium Platino-Cyanide fluorescent screen whenever the cathode-ray tube was turned on. Röntgen recognized that this visible fluorescence was caused by “eine neue Art von Strahlen” originating from the Hittorf-Crookes tube. Due to its unknown character, he called the radiation X-rays. Shortly after the discovery it became clear that X-rays could be used to look into the structure of a living human body and the interior of optically non-transparent materials [2]. In 1901 Wilhelm C. Röntgen received the first Nobel prize in physics for his work [4]. The discovery of X-rays led to some other important breakthroughs. Barkla studied the nature of X-rays with respect to the atomic structure by observing the secondary X-rays radiated from the target sample. During his experiments he discovered the polarization of X-rays (1906), the gaps in atomic absorption (1909), and the distinction between continuous and characteristic X-rays, consisting of several series of X-rays referred to as K, L, M,. . . series (1911). Building on Barkla’s work, von Laue investigated the wave properties of X-rays. He demonstrated X-ray diffraction from a single crystal, which was composed of a 3D-structure with a regularly repeating pattern (1912). Attracted by the experiments of von Laue, Bragg observed X-ray diffraction patterns from single crystals of NaCl and KCl to be the regular patterns of an isometric system showing differences in the X-ray intensity when comparing Sodium and Potassium. The experiments of von Laue and Bragg formed the starting point of crystal structure analysis using X-rays [2].

2.2

Principles of X-Ray Fluorescence

The Surface Monitor can be used to perform X-ray fluorescence (XRF) analysis. The success of the XRF methodology is resulting from the fact that it provides a highly sensitive multi-element compositional analysis capability, in a non-destructive manner for most elements in the periodic system. Moreover, the interaction of X-rays with matter is essentially fully understood, giving the researcher a powerful and practical tool to quantitavely study samples. In the following the relevant interaction phenomena are described.

2.2.1

Interaction of X-Rays with Matter

If an X-ray beam interacts with matter, some of the incident photons are absorbed by the sample, others are scattered away from the original direction of the incident beam and the remaining part passes through the sample. The relation between the incoming beam intensity (I0 ), passing through a sample with given thickness (x) and density (ρ), and the outgoing beam intensity (I) is given by the law of Lambert-Beer:


5

I(x) = I0 e−µρ ρx = I0 e−µL x

(2.1)

with I0 [photons/s] the incoming X-ray beam intensity, I(x) [photons/s] the outgoing X-ray beam intensity, x [cm] the sample thickness, ρ g/cm3 the sample density, µρ cm2 /g the mass attenuation coefficient and µL = µρ ρ cm−1 the linear attenuation coefficient. If the

interaction occurs in the relevant energy region of 1 − 100 keV, the mass attenuation coefficient for a single element can be written as follows:

µρ (E, Z) = τρ (E, Z) + σρ,R (E, Z) + σρ,C (E, Z)

(2.2)

with τρ (E, Z) the photoelectric cross-section, σρ,R (E, Z) the Rayleigh (elastic) scattering crosssection and σρ,C (E, Z) the Compton (inelastic) cross-section, all function of the incident beam energy (E) and the atomic number (Z). The relative contribution to the mass attenuation coefficient of the three absorption effects as function of the photon energy is shown in Fig. 2.2 for three elements: Ar, Si and Ge. Eq. (2.2) is only valid if there is one element present in the sample. For multiple elements present in the sample the mass attenuation coefficient can be written as the sum of the weight fractions (wi ) times the mass attenuation coefficient for each element (µρ (E, Zi )) [3]. µρ (E, Z1 , . . . , Zn ) =

n X

wi µρ (E, Zi )

(2.3)

i=1

The three most important interactions between X-rays and atoms in the relevant region for X-ray analysis (1 − 100 keV) are briefly discussed below. It is worthwhile to note that if higher

X-ray energies are used, pair production and photonuclear absorption will also occur. Since these last two interaction types cannot occur at lower energies, they will not be discussed further in this work.

Figure 2.2: Mass attenuation coefficients for the elements Argon, Silicon and Germanium. Next to the total mass attenuation coefficient, the photoelectric, Compton and Rayleigh components are also indicated in the graphs. Reproduced from Beckhoff et al. [2].

6


Photoelectric Effect The photoelectric effect occurs when an incident X-ray photon with energy E interacts with a bound atomic electron, resulting in the absorption of the photon. In this case, the energy E of the photon is completely transferred to an electron, which is ejected from the atom. The resulting photoelectron has a kinetic energy equal to the incident photon energy (E) minus the binding energy (E0 ) of the ejected electron (see Fig. 2.3). After the ejection of the photoelectron an inner shell vacancy is created which causes an electron cascade from the outer to the inner shells accompanied by the emission of X-ray fluorescence photons (see Fig. 2.4), or Auger-electrons. The energies of the emitted photons are characteristic for each element and this relationship makes X-ray fluorescence (XRF) a suitable tool for elemental composition analysis. It was around the beginning of the twentieth century that Moseley discovered the linear relationship between the square root of the frequency of the emitted photons and the atomic number (see Fig. 2.5) [2].

Figure 2.3: Photoelectric effect, an X-ray photon is absorbed by an atom and a photoelectron is ejected. Reproduced from Amptek [5].

Figure 2.4: Electron cascade from the outer to the inner shell, causing characteristic X-ray fluorescence radiation (or the emission of Auger electrons, not shown in this figure). Reproduced from Amptek [5].


7

Figure 2.5: The Moseley plot shows a linear relationship between the square root of the frequency of the emitted characteristic X-ray fluorescence photons and the atomic number. Reproduced from Vincze [3].

Scattering Two different scattering effects occur while using X-rays: Rayleigh scattering and Compton scattering. The difference between these two types of scattering is described in the next two paragraphs. During Rayleigh scattering (see Fig. 2.6) an incident photon interacts with a tightly bound atomic electron via an elastic scattering process which implies that throughout this scattering event the photon loses no energy. The wavelength λ1 of the incident beam is therefore the same as the wavelength λ2 of the scattered beam. Compton scattering is an inelastic scattering process between photons and outer shell electrons, during which a part of the incoming photon energy is transferred to an atomic electron, the scattered photon has a longer wavelength and lower energy than the incident photon and is deflected by an angle θ compared to the initial direction (see Fig. 2.7). The resulting energy of the scattered photon is given by the Compton equation: E=

1+

E0 m e c2

E0 (1 − cos θ)

(2.4)

with E [keV] the scattered photon energy, E0 [keV] the incident photon energy, me [kg] the electron rest mass, c [m/s] the speed of light and θ the angle between the incident photon and

8


the scattered photon. The term me c2 equals 511 keV.

Figure 2.6: Rayleigh scattering (λ1 , wavelength incident beam; λ2 , wavelength scattered beam where λ1 = λ2 ). Reproduced from Vincze [3].

Figure 2.7: Compton scattering (λ1 , wavelength incident beam; λ2 , wavelength scattered beam; θ, scattering angle). Reproduced from Vincze [3].

2.3

Principles of X-Ray Diffraction

Next to X-ray fluorescence (XRF), the Surface Monitor can be used to perform X-ray diffraction (XRD) measurements on the sample under investigation. Where XRF is capable to determine quantitatively the elements present in the sample, XRD gives information about the arrangement of the atoms in solid (crystalline) samples. It is important to mention that XRD can only be applied on crystalline structures. Thus, a combination of these two techniques, XRF and XRD, can provide an accurate compositional and structural characterization of the material being studied. In a crystal, atoms are arranged in a regular pattern, which makes that a small volume can be identified, that by repetition in three dimensions describes the whole crystal [6]. This small volume is called the unit cell and can be described by three axes: a, b and c, and the angles between them: α, β and γ (see Fig. 2.8). A brief description is given in the next sections.

Principles of X-Ray Diffraction

9

Figure 2.8: Unit cell parameters: α, angle between b and c; β, angle between a and c; γ, angle between a and b. Reproduced from Jeffrey [7].

2.3.1

Scattering of X-Rays by Electrons and Atoms

From the theory of classical electromagnetism it is known that electrons, accelerated by an alternating electric field, are oscillating with the same frequency of the field and are emitting radiation of the same frequency of the alternating electric field (see Fig. 2.9). If atoms, from a crystal, are exposed to an X-ray beam, all electrons around the nuclei are oscillating and emitting radiation. Since atoms are arranged in a regular three dimensional way, constructive interference of the scattered waves in a few directions will take place [8]. This can be described mathematically by the von Laue equations for diffraction:

a (cos Ψa − cos Ψa,0 ) = hλ b (cos Ψb − cos Ψb,0 ) = kλ

(2.5)

c (cos Ψc − cos Ψc,0 ) = lλ where Ψx,0 is the incident X-ray beam, Ψx is the diffracted X-ray beam, a, b and c are lattice spacings, h, k and l are natural numbers and λ is the wavelength of the incident X-ray beam [6]. The von Laue equations (Eq. (2.5)) are very difficult to implement because in total six angles, three lattice spacings and three natural numbers need to be defined. A mathematically more accessible function is given by W. L. Bragg and is describe in the next section.

2.3.2

Bragg’s Law

Instead of studying the diffraction of a single electron, Bragg looked at diffraction in terms of reflection from crystal planes (see Fig. 2.10). Two identical beams in terms of wavelength and phase approach two layers of a crystalline solid. After interaction with two different atomic layers, they are scattered. The beam that is scattered by the second crystal plane travels an

10


Figure 2.9: Scattering by a free electron. Reproduced from Wikipedia [9].

Figure 2.10: Bragg diffraction (θ, angle between incident beam and the crystal plane; d, distance between two crystal planes; d sin θ, path length difference between upper and lower beam). Reproduced from Wikipedia [11].

extra length equal to 2d sin θ compared to the beam that is scattered by the upper crystal plane. In order to obtain constructive interference this length must be equal to a positive multiple of the wavelength of the incident radiation. This relation is mathematically described by the famous Bragg law: 2dhkl sin θ = nλ

(2.6)

where dhkl [nm] is the distance between two crystal planes, θ is the angle between the incoming beam and the crystal plane and is also equal to the angle between the diffracted beam and the crystal plane. λ [nm] is the wavelength of the incident beam and n is the diffraction order [10].

2.4

Principles of the Surface Monitor’s Components

The previous sections provided a general outline on the principles of XRF and XRD. In what follows the generation and detection of X-rays as being done with the Surface Monitor is explained. X-rays can be generated through a variety of ways by using: an X-ray tube, synchrotron radiation or the radioactive decay of specific isotopes (e.g.

55 Fe).

Since the Surface


11

Monitor only uses an X-ray tube the last two possibilities of generating X-rays are not discussed in this thesis. The same remark can be made for the detection of X-rays. X-rays can be detected by using photographic plates, ionization chambers, semiconductor detectors, etc. Since the Surface Monitor uses a semiconductor detector, only this type of detector is explained in the following section. Next to the general working principle of the X-ray source and the semiconductor detector, the Bragg-Brentano θ : θ set-up is briefly mentioned because the Surface Monitor uses this method to acquire XRD spectra. It should be noted that there are other ways to acquire XRD spectra, but they are out of scope of this thesis.

2.4.1

X-Ray Tube Based X-Ray Production

The first X-ray tube ever used for experimental applications is the Hittorf-Crookes tube. This X-ray tube was used by Wilhelm C. Röntgen during his experiments on cathode rays. But the Hittorf-Crookes tube has several disadvantages: short lifetime, instability and difficult to control. Therefore W.D. Coolidge made an improved version of the Hittorf-Crookes tube, the so called Coolidge X-ray tube (see Fig. 2.11). Coolidge introduced a so called hot filament electron emitter in a high vacuum X-ray tube. The improvements of Coolidge are still used nowadays in modern X-ray tubes [2]. In general a modern X-ray tube consists of a negatively charged metal filament (cathode) and a positively charged target material (anode) e.g. Cu, Mo, Rh, Ag, etc. The negatively charged filament is heated by an electric current which produces electrons by thermionic effect. Due to the application of a high voltage current between the anode and cathode, electrons are accelerated towards the anode. Once they reach the anode, the electron beam interacts with the target material and X-rays are produced. During the interaction with the target material the accelerated electrons lose their energy via a number of processes. A part of the incident electrons decelerate continuously which results in a gradual energy loss, this process gives rise to a continuous spectrum or Bremsstrahlung. Incident electrons can also interact with atomic electrons of the target material. If an incident electron transfers its energy to an inner electron of an atom, a vacancy is created through impact ionization. After the interaction, the vacancy is filled up by another orbital electron. This process produces the characteristic lines in the X-ray emission spectrum. E.g., Fig. 2.12 shows a typical spectrum of an X-ray tube with a Copper anode at different voltages. Because 99 % of the incident energy is converted to heat, water-cooling is essential. Due to this inefficient production of X-rays, the dissipation of heat at the focal spot forms a main limitation on the power which can be applied on the Coolidge X-ray tube. In order to apply a higher power, a rotating anode X-ray tube was developped as an improvement of the Coolidge X-ray tube. This rotating anode tube consists of a disk-like anode fixed on a rotor with a bearing system supporting the rotation. So, by rotating the anode past the focal spot the heat load can be spread over a larger area, which implies that higher powers can be used. Nowadays there are also low-power X-ray tubes available. The applied power is only a few watts and therefore air-cooling is sufficient.

12


Figure 2.11: Conventional X-ray source (C, cathode; A, anode; Win , water inlet; Wout , water outlet). Reproduced from Wikipedia [12].

Figure 2.12: Typical X-ray spectrum from a Copper target at different voltages. Reproduced from Cockcroft [13].

2.4.2

Semiconductor Based X-Ray Detection

The sensitive material of the semiconductor detector typically consists of Silicon or Germanium crystals. In such a single crystal of semiconductor material, the sharply defined atomic electron states are broadened into bands of energy states. The outer electrons of the semiconductor are kept in the valence band that on its turn is separated from the conduction band by an energy gap or band gap. This band gap contains forbidden states and an electron can only be promoted if it receives energy at least equal to that of the band gap. If such an electron receives enough energy it is promoted to the conduction band (see Fig. 2.13). At the conduction band, the electron can move under the influence of an externally applied electric field and can be collected at an electrode. The promoted electron also creates a vacancy or hole in the valence band. The hole on his turn can also move under influence of the applied electric field, but in the opposite direction of the electron. The excitation of an electron to the conduction band can be achieved by interaction with an X-ray photon. The number of electron-hole pairs created by the incident X-ray beam is proportional with the energy of the beam.


13

Figure 2.13: Band structure of a semiconductor material. Reproduced from Beckhoff et al. [2].

Next to impurity-free semiconductors or so called intrinsic semiconductors, dopants are added to influence the conduction properties of the material. Adding for example Phosphorous to a semiconductor increases the conductivity. The increase in conductivity is due to the fact that only four of the five valence electrons of Phosphorous are involved in binding it with Silicon or Germanium atoms. The fifth electron is free and can therefore be promoted to the conduction band with only a small amount of energy. Dopants like Phosphorous are called donors and the doped semiconductor is a so called n-type semiconductor. On the other hand, adding an acceptor, for example Boron, creates free holes. Boron has only three electrons available to bind with the semiconductor material, therefore free holes are created in the semiconductor material. This is also increasing the conductivity of the material and the resulting semiconductor is a so called p-type material [2]. A so called Si-PIN detector makes use of the n- and p-type semiconductor. The PIN junction is a combination of a p-doped semiconductor layer, a ndoped semiconductor layer and an intrinsic silicon layer in between. The working principle of a Si-PIN detector is represented in Fig. 2.14. Here, the PIN junction is reverse biased by an external source. This external voltage results in the formation of a strong electric field in the intrinsic layer. If an X-ray photon reaches the intrinsic layer, free electrons and holes are created. Due to the applied electric field, electrons are moving to the n-layer and the holes are moving in the opposite direction (toward the p-layer). The magnitude of the resulting current or pulse is proportional to the energy of the incident X-ray. In the next step, the pulse is amplified and passed to a computer which is acting as a multichannel analyzer (MCA). The MCA determines to which of e.g. 2048 channels, each representing a different X-ray energy, the pulse should be registered in. Finally, when plotting the number of counts as a function of the channel number a spectrum is obtained [14].

14


Figure 2.14: Scheme of a reverse bias PIN photodetector. Reproduced from Azadeh [15].

2.4.3

Bragg-Brentano θ : θ Set-Up

The Surface Monitor is based on the Bragg-Brentano θ : θ acquisition mode which is a popular set-up for recording an XRD spectrum. The powder that needs to be analyzed is placed in a static sample holder. Both, X-ray tube and detector axes, make an angle θ with the horizontal plane of the sample. Also the distance between sample and tube, and sample and detector are identical. During the measurement both X-ray tube and detector move simultaneously over an angular range, defined at the beginning of the experiment, see Fig. 2.15 [6].

Figure 2.15: Bragg-Brentano θ : θ set-up, with on the left the X-ray tube and on the right the X-ray detector. The sample is placed in between. Detector and X-ray tube move with the same angle θ with respect to the sample plane in order to detect the diffracted X-ray beams. Reproduced from the Surface Monitor User’s Manual [16].

Chapter 3

The Portable XRF/XRD Spectrometer: Assing’s Surface Monitor This chapter gives an overview of already existing portable XRF/XRD instrumentation, followed by a full description of the Surface Monitor’s main parts, going from the available X-ray tubes and the X-ray detector, to the XRD protractor. Also, the description is given how to set up the Surface Monitor for an experiment, including the actual XRF/XRD acquisition and data processing.

3.1

Introduction

A first prototype X-ray fluorescence (XRF)/X-ray diffraction (XRD) instrument was developed in 1992 at NASA (National Aeronautics and Space Administration) Ames Research Center [17, 18]. The intended applications for this new instrument, that can do both XRF and XRD measurements simultaneously, are for planetary exploration and as a portable instrument for terrestrial use. During Mars missions in the past, only methods (i.e. XRF analysis) were used that give the elemental composition of the sample and suggesting certain minerals, but the traditional method of mineral identification is by XRD. Therefore, a new instrument was developed which property makes it possible to obtain correct mineral identification. This prototype makes use of XRF to determine the elemental composition, suggesting certain minerals, while XRD is used for definitive mineral identification. In this way, a small portable XRF/XRD instrument makes it possible to gather this information simultaneously and provides accurate mineral information on Mars as well as on Earth. In 2000, at MOXTEK, Inc. a breadboard setup was constructed in the attempt to construct a portable XRF/XRD instrument (see Fig. 3.1) [19]. The main parts of this instrument are the rotating Copper anode X-ray source and a CCD (charged-coupled device) camera. The CCD detector can be seen as the key component of the instrument because it is able to record the spatial position and the energy of an X-ray event si15

16

The Portable XRF/XRD Spectrometer: Assing’s Surface Monitor

Figure 3.1: Picture of the breadboard set-up, constructed at MOXTEK, Inc. and used by Cornaby et al. [19], with the different components.

multaneously. In order to extract the fluorescent and diffraction information from the recorded data dedicated software algorithms were applied. The X-ray energy detection range is between 1.7 keV and 12 keV, the angular range of the detected diffraction peaks is between 2° 2θ and 50° 2θ. The main disadvantage of this instrument is the large X-ray source, which will later on be replaced by a smaller low power X-ray tube, and the CCD camera which has less resolution in the diffraction peaks. On the other hand, the CCD makes it possible to gather XRF and XRD information at the same time. Important in the improvement of this portable XRF/XRD instrument is the development of a low-power X-ray tube, enabling field operations. These low-power tubes are characterized by their small size and low power consumption (circa 5 W) compared to conventional X-ray tubes. Fig. 3.2 shows an example of a low-power X-ray tube. This tube requires a maximum voltage of 20 kV with an emission current of 100 µA, which results in a maximum input power of 2 W. According to Cornaby et al. [20] they were able to obtain XRF and XRD data with the low power X-ray tube at a comparable rate with respect to the rotating Copper anode, used for the breadboard setup. Due to the low power requirements, a battery can supply the required power for the acquisition of XRF and XRD spectra. In 2004, Uda [21] combined the low-power X-ray tube with a Si-PIN photodiode in a portable energy dispersive X-ray diffraction and fluorescence (ED-XRDF) spectrometer. This combination provides several advantages: short measuring time, non-destructive and contact-free measurement, no restriction on sample size and shape, small dimensions, light weight, no need for special coolant and the operations are assisted by a personal computer. The spectrometer was successfully tested by the determination of crystal structures and chemical composition of ancient plasters and pigments on the field. Later on the ED-XRDF spectrometer was improved to acquire an XRD pattern and an XRF spectrum simultaneously and it was made possible to collect the XRD pattern under angle dispersive mode next to the energy dispersive mode. The

Introduction

17

Figure 3.2: Scheme of a low-power Cu anode X-ray tube, the tube is 42 mm long and has a diameter of 15 mm. The dimensions of the high-voltage power supply are 3 cm x 7 cm x 17 cm. Reproduced from Cornaby et al. [20].

portable XRF/XRD spectrometer was specially designed to be used for archaeological studies and this for the following reasons: some objects are not allowed to be transfered from open air to a vacuum, since the spectrometer works in open air a vacuum environment is not necessary; some objects can not be moved from their original sites; some parts are extremely fragile and therefore impossible to measure them under vacuum; and chemical composition and structural information can be obtained from the same area of an object which offers a better understanding of the constitutive materials of the object [22]. Despite of the several advantages offered by this new instrument, it is still difficult to analyze complex mixtures of compounds due to a high background signal and weak peaks in the XRD spectra [23]. Another portable system that simultaneously performs XRF and XRD analysis was developed by Gianoncelli et al. [23] in 2007. The equipment is represented in Fig. 3.3. The main components (Copper anode X-ray tube, Silicon drift XRF detector and an imaging plate) are mounted on a frame which makes it possible to move along the surface of the object to be analyzed. Two laser pointers intersect at the position where the X-ray beam impinges the sample and at this point the analysis takes place. Their instrument gave satisfactory and reasonable performances but one important drawback is the reproducibility of the XRD measurement results. Moving the measurement head 1 mm away from the reference point, induced a 3° shift in the 2θ-scale. Next to this problem the fluorescence lines of the chemical elements present in the sample were collected on the imaging plate causing a substantial background signal in the XRD spectrum. In 2010, this portable device was also used by Duran et al. [24] for the analytical study of Roman and Arabic wall paintings. They concluded that the portable XRF/XRD system was able to successfully characterize the paintings but for a thorough identification of the components of the pigments SEM-EDX (Scanning Electron Microscopy - Energy Dispersive analysis of X-radiation) was needed.

18


Figure 3.3: Picture of the XRF/XRD system used by Gianoncelli et al. [23] in 2007. The Copper anode X-ray source is on the right, the imaging plate is on the left and the Silicon drift XRF detector is in between (dimensions: 75 cm x 45 cm x 45 cm) [25].

Figure 3.4: Picture of the modified XRF/XRD instrument used by Pifferi et al. [26] in 2008 (dimensions: 60 cm x 30 cm x 45 cm; mass: 25 kg).

In 2008, some efforts were made by the company Assing S.p.A. [1] in order to provide a small instrument for simultaneous XRF/XRD measurements. According to Pifferi et al. [26] the results are still unreliable due to peak broadening and mechanic instability. In order to overcome these problems Pifferi et al. modified the prototype developed by Assing S.p.A. The hardware of the modified instrument consists of a Copper anode X-ray tube and a Si solid state detector (Si-PIN) both moving on a horizontal Theta-Theta goniometer (see Fig. 3.4). A laser interferometer ensures a reliable positioning of the instrument with respect to the sample. The X-ray tube and detector can move in the range of 10° < 2θ < 140° symmetrically around the sample. The software and hardware makes it possible to operate the instrument in three different modes: XRF mode, XRD mode and XRF/XRD mode. The later implies that the acquisition of an XRF and XRD spectrum is performed simultaneously. The energy selection for the XRD measurements is performed via software, without the use of a monochromator. Last year (2011), Assing’s Surface Monitor was acquired by the UGent Analytical Chemistry department. The description of this novel instrument is given in the next section.

Description of Assing’s Surface Monitor

3.2

19


The Surface Monitor is a new portable device developed by the Italian company Assing S.p.A. [1]. Its design allows to perform elementary analysis applying XRF and simultaneous analysis of mineral phases performing XRD. As can be seen in Fig. 3.5 the main parts of the Surface Monitor are: the X-ray tube, the detector, the laser interferometer and the XRD protractor (for further details see the following paragraphs). All these components are assembled into the instrument head that can be fixed on a tripod. This tripod allows easy positioning of the probe head; this is described more in detail in Section 3.3.1 on Page 22. Furthermore, the Surface Monitor is equipped with a control box, a control portable PC and an extra arm with webcam. Fig. 3.6a shows a general overview of the Surface Monitor’s set-up, the webcam is shown in Fig. 3.6b. This webcam is not part of the original Surface Monitor as provided by the company Assing S.p.A, but was added by the UGent-XMI Group to improve the safety. The webcam allows the researcher to watch the status of the device head, on a distance, using the control portable PC. The arm with the webcam is fixed on the horizontal arm of the tripod, watching the probe head.

Figure 3.5: Top view of the Surface Monitor’s instrument head, acquired by the UGent Analytical Chemistry department in 2011, with on the left side the Cu anode X-ray source, at the right side is the X-123 Si-PIN detector (both fixed on the XRD protractor) and the instrument head with the laser interferometer is situated in the center of the picture.

For the acquisition of XRF and XRD spectra two X-ray tubes are available, one with a Copper (Cu) anode and one with a Molybdenum (Mo) anode. Both X-ray tubes have a maximum input voltage of 30 kV and a maximum current of respectively 500 µA and 300 µA. The generation of X-rays through these low power tubes is the same as in conventional X-ray sources. More information about the generation of X-rays through X-ray tubes can be found in Section 2.4.1 on Page 11. The X-ray sources of the Surface Monitor are not equipped with a monochromator to select the characteristic X-rays originating from the X-ray tube, so the X-ray beam originating from both X-ray tubes is polychromatic.

20


(a) Overview Surface Monitor’s set-up

(b) Webcam with arm

Figure 3.6: (a) Overview of the Surface Monitor’s set-up, reproduced from the Safety Manual (1, control portable PC; 2, control box; 3, instrument head; 4, tripod). (b) Webcam with arm. This arm is fixed on the tripod’s horizontal arm.

In order to detect the X-rays, the Surface Monitor uses a X-123 Si-PIN complete X-ray detector of the Amptek Inc. company [27]. The working principle of a Si-PIN detector has been explained in Section 2.4.2 on Page 12. Both, X-ray tube and detector, can be equipped with pairs of vertical slits or pinholes with respectively a different width or internal diameter. The available pairs of collimators are shown in Fig. 3.7. The X-ray tube and detector are each provided with two notches and two screws to insert and fix the collimators. During measurements these screws are removed in order to replace the collimators easily and to prevent damage on the object that is analyzed. The choice of the pinholes and slits has important consequences on the quality of the acquired XRD spectra. If a narrow slit is chosen, the obtained peaks in the spectra will also be narrower and the narrower the peaks the higher the resolution will be. On the other hand, choosing a narrow slit, the positioning of the sample becomes critical. The positioning of the analytical head to the correct distance of the sample is done using the laser interferometer that is built into the analytical head. Next to the indication of the distance, the use of the laser ensures the correct aiming of the X-ray beam on sample area of interest. For a correct XRD measurement the device should be positioned with its main axis perpendicular to the sample and the distance between the instrument head and the sample should be between 94.5 mm and 95 mm. When the instrument was delivered this distance was set between 91.8 mm and 92.2 mm by the company Assing S.p.A., but unreliable data were obtained at this distance. Further details concerning this problem will be described in Section 4.1.2 on Page 32. Changes in the state of perfect perpendicularity and distance will result in an offset and different peak widths. Fig. 3.8 shows


21

Figure 3.7: Available pinholes (left side) and slits (right side). The pinhole diameters are, starting from the top, approximately 0.5 mm, 1 mm, 1.5 mm and 2 mm. The width of the slits are, starting from the top, approximately 0.5 mm, 1 mm, 1.5 mm and 2 mm, the length of each slit is approximately 8 mm.

Figure 3.8: Positioning of the analytical head with respect to the position of the sample for XRD measurements. The distance between the analytical head and the sample surface is a subject of investigation of this work; in the Surface Monitor User’s Manual it is mentioned to keep the distance to 92 mm in order to obtain a reliable XRD spectrum. Reproduced from the Surface Monitor User’s Manual [16].

the positioning of the instrument head with respect to the sample. The last part of the Surface Monitor is the XRD protractor. The protractor connects the instrument head with both the X-ray tube and the energy dispersive detector. This allows movement of the X-ray tube and the detector over a specified angular range 2θ according to the Bragg-Brentano arrangement that is described in Section 2.4.3 on Page 14. In practice, the range of angles is limited at one side due to the physical constraint of possibly hitting the sample surface [16]. The next section describes how the Surface Monitor is correctly positioned in front of an object before starting the actual XRF/XRD measurement and the XRF/XRD data analysis.

22

3.3 3.3.1


Methodology Set-Up and Positioning of the Surface Monitor

The first step during the installation of the Surface Monitor in front of the object to be investigated, is the positioning of the tripod at the approximate height for the analysis. When the tripod is placed in a stable position, the instrument head is fixed on the tripod and the cables, connecting the instrument head with the control box, are plugged in. Once the connection with the computer is made and the initialization of the instrument is completed, the laser interferometer is switched on. The laser is pointing at the sample area where the X-ray beam is going to hit the sample. In this way, while being used as a distance monitor, it is also providing a very useful tool to position the X-ray beam in the two other directions on the sample area of interest. Before starting the actual analysis of the object, the instrument needs to be calibrated. This calibration, with the LaB6 diffraction standard, is necessary to provide an indication of the diffraction peak offset. Once the calibration is completed, the instrument is placed perpendicular to the object’s surface. When the laser is pointing at the surface that needs to be analyzed, the tripod is moved step by step towards or away from the object until the ideal distance is reached. To obtain reliable XRD spectra the distance must be between 94.5 mm and 95.0 mm. The distance between the object and the instrument can be monitored with the laptop using the laser interferometer. In the next step, the smallest angles for the tube and detector are determined by visually checking (while moving to devices to smaller and smaller angles) at which angles these devices are still not hitting the sample’s surface. An example of a set-up is shown in Fig. 3.9. Once the positioning is performed, and the range of angles is known, the actual measurement of XRF and XRD spectra can be done. The next two sections will describe how the acquisition parameters can be adjusted and how the obtained data is processed.

3.3.2

XRF Acquisition and Data Processing

The instrumental parameters are set by using the Surface Monitor’s control-software. From the interface in Fig. 3.11a one can see that the following parameters can be changed: the angle of the X-ray tube and detector, the voltage and current applied on the X-ray tube, the acquisition and preheating time (setting a preheating time is only necessary after a long inactive period of the instrument), and the filename. Some optional parameters such as operator and sample type can also be specified, but these settings will not affect the result of the acquisition. The evaluation of the X-ray fluorescence data, obtained with the Surface Monitor, is executed

Methodology

23

(a)

(b)

Figure 3.9: Set-up of the Surface Monitor in front of the Dulle Griet painting at the Mayer van den Bergh Museum in Antwerp (Belgium). (a) General view with in the front the control box and in the back the instrument head mounted on the tripod, (b) detailed picture of the instrument head in front of the painting.

using the AXIL (Analysis of X-rays by Iterative Least Squares) software package [28]. The Surface Monitor generates output files with a flo-extension. In order to convert the output files into the spe-extension, used by AXIL, the flo2spe program was made. The source code and further details can be found in Appendix B on Page 63. The AXIL software uses the non-linear least squares fitting strategy to fit a mathematical function to experimental spectra. The resulting mathematical function describes the fluorescence peaks and the spectral background. The applied fit-routine tries to minimize the weighted sum of differences χ2 between the experimental data y and a mathematical fitting function yf it : 1 X [yi − yf it (i)]2 χ = n−m yi 2

(3.1)

i

with yi the content of channel i, yf it (i) the calculated content of the fitting function in channel i, n the number of channels in the fitting window and m the number of parameters of the fitting function that are estimated during the fitting process. The fitting function is the sum of the spectral background and the fluorescence peaks:

yf it (i) = yback (i) + ypeak (i) X = yback (i) + yj (i) j

(3.2)

24


Figure 3.10: Spectrum evaluation of the NIST SRM 1412 multicomponent glass standard using AXIL.

(a)

(b)

Figure 3.11: Interface for (a) the XRF acquisition and (b) the XRD acquisition.

where the first term represents the spectral background and the second term corresponds to the fluorescence peaks included in the model. The index j runs over all characteristic line groups. During the iterative process, the parameters are optimized to obtain the best match between the model and the spectral data [29]. In this work, the spectrum evaluation module of the AXIL software is used to determine the fluorescence lines, escape peaks, sum peaks, etc. in a rapid and reliable way. Once such a spectrum evaluation is completed, the obtained net peak areas can be used to calculate e.g. minimum detection limits.

3.3.3

XRD Acquisition and Data Processing

The interface for the XRD acquisition set-up (see Fig. 3.11b) is somewhat similar to the interface for XRF acquisition set-up but some additional instrumental parameters need to be defined. These extra parameters are the start and stop angle, the angular step size and the XRD energy (it is recommended to set this value to 8.05 keV when the Cu X-ray source is used). Note that the values for the start (2θi ), stop (2θf ) and step angles (∆) are in the 2θ-scale. This implies that if one wants to move the X-ray tube and detector during an XRD acquisition from

Methodology

25

an angle of 15° to 45° with a step size of 0.1° one should enter the following: 30° as start angle, 90° as stop angle and 0.2° as step angle. The Surface Monitor can be equipped with either a Cu anode or a Mo anode X-ray tube, see Section 3.2 on Page 19. Both X-ray tubes can be used to perform XRD measurements, but depending on the tube that is selected for the experiment, there will be a different representation at the 2θ-scale. E.g., the influence of the anode selection can be illustrated by looking at the theoretical XRD spectrum of the Calcite (CaCO3 ) mineral. Fig. 3.12 and Fig. 3.13 show the XRD spectra for this mineral, obtained with the Cu X-ray tube at 8.047 keV and the Mo X-ray tube at 17.479 keV. The XRD spectrum in the 2θ representation thus depends on the used energy in agreement to Bragg’s Law (see Eq. (2.6)). As can be seen in Figs. 3.12 and 3.13 the XRD spectrum obtained with the Mo X-ray tube, will be more compressed at smaller angles compared to the XRD spectrum obtained with the Cu X-ray tube. It should be noted that shallow angles are practically limited because the detector and the X-ray source are approaching the sample when rotating. Therefore it is recommended to use the Cu X-ray tube during the acquisition of XRD patterns. It is worth mentioning that the Surface Monitor does not use a monochromator to select one specific wavelength, e.g. the Cu Kα -line. Instead of using a monochromator to select one wavelength for the acquisition of the XRD spectrum, a software filter is used to extract the XRD spectrum at one specific wavelength. This method makes it possible to obtain diffraction patterns at energies different from the characteristic radiation of the X-ray tube, but the peak intensity at these energies will be significantly lower. The identification of minerals in an acquired XRD spectrum is based on the comparison with XRD spectra from a mineral database in order to find the correct fingerprint. The control portable computer of the Surface Monitor is equipped with the program “XRD Match!” that is linked to the library of the International Center for Diffraction Data (ICDD) (PDF-4/minerals 2010 relational database; version 4.1011) [30]. The analysis of an XRD spectrum can be helped with the selection of the chemical elements present in the sample and the phase, e.g. mineral, pigment, metal & alloys, etc. (see Fig. 3.14). The extra information concerning the chemical elements can be obtained by first analyzing the XRF spectrum. The next step is to select the peaks in the obtained XRD spectrum, to subtract the noise and to apply a Gaussian fit on the selected peaks (see Fig. 3.15). After completing these steps, a search can be performed in the ICDD database to see which phases match with the profile that consists of the peaks selected in the previous step. The matched phases are ranked in a way that the phases with the best match are placed first.

26


Figure 3.12: XRD spectrum for the mineral Calcite (PDF No. 00-047-1743), measured with a Cu X-ray tube (E = 8.047 keV). Reproduced from the International Center for Diffraction Data database [30].

Figure 3.13: XRD spectrum for the mineral Calcite (PDF No. 00-047-1743), measured with a Mo X-ray tube (E = 17.479 keV). Reproduced from the International Center for Diffraction Data database [30].

Methodology

27

Figure 3.14: “XRD match!” interface; the user can select the phase (left part) and the elements present in the sample (right part). The spectrum is showed in the upper part of the window.

Figure 3.15: The “XRD match!” software is able to determine the profile of the XRD pattern, by extracting the areas of the selected peaks and applying a Gaussian fit after the subtraction of the background. A tolerance of deviation in intensity and d-space values between the acquired experimental values and the ICDD values can also be chosen.

28


Chapter 4

Results: Characterization and Applications of the Surface Monitor Before using the portable XRF/XRD instrument in real applications, several aspects concerning methodology and performance were investigated. This includes the determination of the laser interferometer’s reliability, the ideal XRF/XRD acquisition distance, the beam size and the position of the X-ray beam with respect to the laser interferometer, minimum detection limits for XRF, and finally the influence of using different slit and pinhole sizes during an XRD acquisition. Also shown in this chapter are the results from the very first applications done with the Surface Monitor. I.e. the identification of pigments in paintings: a painting with colorful flowers made available to the UGent-XMI lab by the Raman Spectroscopy Research Group, and an on-the-field session on the “Dulle Griet” in the museum Mayer van den Bergh (Antwerp, Belgium).

4.1 4.1.1

Surface Monitor’s Performance The Laser Interferometer’s Performance

When performing XRF and especially XRD experiments with the Surface Monitor, the positioning of the instrument (i.e. distance and angle) relative to the sample surface is of major importance. The laser interferometer plays a crucial role during this process. The laser not only enables the user to place the Surface Monitor at the correct distance in front of e.g. a painting, but it also gives an indication about the position of the X-ray beam at the surface of the sample. Because this laser is such an important tool, it is obvious to study how precise this device is performing. To test the laser interferometer the following experiment was performed: the laser was pointing perpendicular on a flat surface, i.e. a paper taped on a holder that is mounted on a manual translational stage (Standa; type 7T184-13) with a reading accuracy of 5 µm. The set-up is shown in Fig. 4.1. The stage was moved with steps of 50 µm over a range of 1 cm, starting at 29

30

Results: Characterization and Applications of the Surface Monitor

Figure 4.1: Set-up used to test the Surface Monitor’s laser interferometer. The set-up consists of a piece of paper taped on a sample holder, which on its turn is mounted on a manual translational stage (Standa; type 7T184-13) with a reading accuracy of 5 µm.

a distance of approximately 87 mm up to approximately 97 mm, this is the distance between the object and the Surface Monitor’s instrument head. This range was selected because the optimal distance to perform an XRF/XRD acquisition was assumed to be in this interval. Fig. 4.2 shows the interferometer distance read-out versus the manual stage movement; a trend line is fitted. The histogram in Fig. 4.3 shows the frequency as a function of the difference between the measured distances by the laser interferometer and the trend line. Fig. 4.3 shows a distribution which is similar to a Gaussian distribution, the cumulative distribution function is also plotted in this figure. In addition the skewness and the kurtosis were calculated. The skewness gives an indication about the symmetry of the distribution, symmetric data should have a skewness near zero [31]. Negative skewness indicates that data are skewed left, a positive value indicates that data are skewed right. Kurtosis is a measure of whether data are peaked or flat relative to a normal distribution. A standard normal distribution has a kurtosis of zero. Skewness, kurtosis, mean and standard deviation are shown in Table 4.1. This table indicates that 95 % of the values are within the the following interval: ±0.11 mm or two times the standard deviation.

From this experiment with the interferometer we can conclude that the distances measured with the Surface Monitor’s laser interferometer are somewhat shifted to lower values (negative skewness) and that the distances, measured by the laser interferometer, are within the following interval: ±0.11 mm. Now that the precision of the laser interferometer is known, it is important to determine the optimal acquisition distance as shown in the next paragraph.

Surface Monitor’s Performance

31

Figure 4.2: Distance read-out measured by the laser interferometer as a function of the stage movement. The manual translational stage moved over a range of 1 cm in steps of 50 µm.

Figure 4.3: Measured distribution of the difference between the trend line of Fig. 4.2 and the measured distances by the laser interferometer. Next to the histogram, the cumulative distribution is plotted.

Table 4.1: Table with mean, standard deviation, skewness and kurtosis. These values were calculated from the histogram shown in Fig. 4.3.

Mean (mm)

0.00036

Standard deviation (mm)

0.054

Skewness

-0.49

Kurtosis

0.14

32

4.1.2


Optimal XRF/XRD Acquisition Distance

When the Surface Monitor was delivered, the distance to acquire XRF and XRD spectra was set in the control software by the company Assing S.p.A. between 91.80 mm and 92.20 mm. However, it was clear that from the very first experiments performing XRF/XRD acquisitions, using slits and pinholes on respectively the detector and the X-ray tube, the proposed distance of 92.0 mm was not the optimal distance for the instrument. An example of these spectra with the acquisition parameters is shown in Fig. 4.4. As can be seen from Fig. 4.4, the number of counts per peak (maximum 1 count per peak) are rather low, so probably only a background signal was measured.

Figure 4.4: Diffraction spectrum of the NIST SRM 660b. Acquisition parameters: Cu X-ray tube (28 kV/250 µA), 2θi = 60°, 2θf = 90°, step size = 0.1°, life time = 20 s LT/step, distance = 92 mm, 2 x 1 mm pinhole at the X-ray tube and 2 x 0.5 mm slit at the detector side.

In order to determine the optimal distance, the following experiments were performed using the NIST SRM 660b diffraction standard. Details about the sample preparation and XRD peak positions of the NIST SRM 660b can be found in Appendix A on Page 59. The diffraction standard was placed at different distances from the instrument head to see where the most intense peak of the standard reaches a maximum and deviation of the certified 2θ-angle reaches a minimum. The XRD acquisition parameters were set as follows: the measurements started at a 2θi of 15° to a 2θf of 90° with steps of 0.2° and each step was measured for 5 s LT (live time). Two 1 mm pinholes were inserted at the X-ray tube and two 1 mm slits were used at the detector. All spectra were acquired using the Cu anode X-ray source (28 kV/250 µA). At each distance the 2θ-value and the number of counts of the diffraction standard’s most intense peak, with a certified 2θ-value of 30.385°, were recorded and summarized in Table 4.2. Fig. 4.5 shows the diffraction spectrum obtained at a distance of 95 mm with an indication of the peak that was used to compose Table 4.2. Next to the diffraction spectrum a simultaneously acquired XRF spectrum was available at each distance; these spectra are shown in Fig. 4.6.


33

Figure 4.5: Diffraction spectrum of the NIST SRM 660b obtained at a distance of 95 mm. The most intense peak of the diffraction standard is indicated with a red line.

Table 4.2: Table with the number of counts, the measured 2θ-angle and the deviation from the certified value (2θ = 30.385°), acquired at different distances of the NIST SRM 660b’s most intense diffraction peak. At 93 mm and 98 mm it was impossible to distinguish the diffraction peak.

Distance (mm)

Counts

2θ (°)

Deviation (°)

93.0

-

-

-

94.0

185

32.0

1.615

94.5

234

30.8

0.410

95.0

214

30.0

-0.385

96.0

205

31.0

0.610

97.0

43

29.2

-1.190

98.0

-

-

-

From Table 4.2 one can see that if the distance between the diffraction standard and the instrument head is between 94.5 mm and 95.0 mm, a relatively higher number of counts are obtained and that the deviation with respect to the certified value is smaller compared to deviations at other distances. The XRF spectrum, shown in Fig. 4.6, also confirms that the Surface Monitor must be placed at a distance between 94.5 mm and 95.0 mm from the object. The number of counts is lower when measuring at a distance of 97.0 mm. However at a distance of 94.0, 94.5, 95.0 and 96.0 mm the number of counts is approximately equal. Knowing the optimal distance, in the next step the beam size(s) can be measured at this distance.

34


Figure 4.6: Simultaneously acquired XRF spectra of the NIST SRM 660b. Only the XRF spectra obtained at the following distances are shown: 94.0, 94.5, 95.0, 96.0 and 97.0 mm.

4.1.3

Beam Size and Position of the Beam on the Sample

From the practical point of view, it is important to know the size of impact of the incident X-ray beam on the sample. The size of the beam can be changed by using pinholes of different sizes, while the view of the detector can be restricted by using slits of different widths. The following experiments were performed to determine the beam size of the X-ray beam. The size of the beam and its position was determined by applying the same conditions as used during a general XRF/XRD measurement: Cu anode X-ray tube (28 kV/250 µA), the tube moved over an angle of 12° to 45° with steps of 0.1° and the measuring time was set to 5 s LT/step. Two measurements were performed using this set-up, one by using two pinholes with a diameter of 1 mm and one by using two pinholes with a diameter of 1.5 mm. In all cases a self-developing radiochromic film (GAFChromic

® RQTA2), placed at a distance of 95 mm, was used to de-

termine the beam size. When X-rays are interacting with the film, a black spot is appearing. This spot gives an indication about the size of the beam. The results are shown in Table 4.3 and Fig. 4.7. Table 4.3: Beam sizes measured at a distance of 95 mm by using two different pinhole diameters.

Pinholes

Width (mm)

Height (mm)

2 x 1 mm

16

3

2 x 1.5 mm

21

4


35

As shown in Fig. 4.7, the footprint of the laser interferometer beam is approximately in the middle of the black spot corresponding to the X-ray beam footprint, i.e. in the middle of the irradiated surface. This information is very important when performing measurements on e.g. a painting. When the painting consists of a large area of the same pigment it is better to use a large pinhole size because this improves the count rate which leads to diffraction peaks with a higher intensity. On the other hand, when a small area needs to be analyzed it is preferable to use a smaller pinhole size, to minimize the irradiation of adjacent areas colored with other pigments.

(a) 2 x 1 mm pinholes

(b) 2 x 1.5 mm pinholes

Figure 4.7: Beam sizes measured, with different pinhole sizes, at a distance of 95 mm by using a GAFChromic

® X-ray film.

With on the left side the detector and on the right side the

Cu anode X-ray tube. The X-ray tube moved over an angle of 12° to 45°. Next to the dimensions of the X-ray beam, the laser interferometer point was indicated.

Another way to minimize the impact area during an XRD measurement is to avoid shallow angles. However, the most distinct and significant information can be retrieved from the lower 2θ-angles (i.e. large d-spacing values), because this area only contains a few peaks that on their turn are very characteristic for a crystalline material. In some cases it is impossible to measure at these small angles because the X-ray source and/or detector will touch the object. Therefore, just to complete the information, the beam size was also determined starting at a

36


larger X-ray source angle. The following conditions were applied during the measurement: Cu anode X-ray tube (28 kV/250 µA), the tube moved over an angle of 20° to 45° with steps of 0.1° and a measuring time of 5 s LT/step. A total of four measurements were performed by using

® film

two equal pinholes of respectively 0.5 mm, 1 mm, 1.5 mm and 2 mm. The GAFChromic

was placed at a distance of 94.5 mm. The results of this experiment are shown in Table 4.4. When using the smallest pinhole no beam size was measured. Table 4.4: Beam sizes measured at a distance of 94.5 mm by using four different pinhole diameters. The Cu anode X-ray tube moved over an angle of 20° to 45°. No beam size was measured using the smallest pinhole.

Pinholes

Width (mm)

Height (mm)

2 x 0.5 mm

-

-

2 x 1 mm

9

2.5

2 x 1.5 mm

13

3.5

2 x 2 mm

15

4.5

The beam size was also measured at a distance of 92 mm to see how the position of the beam changes with respect to the laser point when measuring at a closer distance. The following conditions were applied during this measurement: Cu anode X-ray tube (28 kV/250 µA), the tube moved over an angle of 30° to 45° with steps of 2.5° and a measuring time of 100 s LT/step. During the measurement only two 1 mm slits were used to collimate the beam. The result is shown in Fig. 4.8. This figure shows that the laser’s point is not at the center of the X-ray beam anymore and that the X-ray beam size is rather large when using slits as collimators. These results at the distance of 92.0 mm confirm that this distance, as proposed by the manufacturer, is not the optimal distance for the XRF/XRD measurements.

Figure 4.8: Beam size measured at a distance of 92 mm by using a GAFChromic

® X-ray film. With

on the left side the detector and on the right side the Cu anode X-ray tube. The X-ray tube moved over an angle of 30° to 45°. Next to the dimensions of the X-ray beam, the laser interferometer point was indicated.


37

From the obtained results it is clear that the smallest X-ray beam is obtained by using two pinholes at the X-ray source and to only use the slits if a large homogeneous area needs to be analyzed with the Surface Monitor. It should be noted that the set of 0.5 mm pinholes did not give any satisfying results, probably because of misalignment of the two holes when put in their holders in front of the X-ray source. Next to the size of the beam, the position of the X-ray beam with respect to the laser interferometer’s point was investigated. When measuring at a distance of approximately 95 mm the laser’s point indicates the center of the X-ray beam. This result gives an extra indication that the object must be placed at a distance of 95 mm, this value was already obtained in Section 4.1.2. Knowing the influence of the selected collimators at the optimal distance, the performance of the instrument can be investigated as shown in the following.

4.1.4

XRF Minimum Detection Limits

The performance of an analytical instrument can be measured by determining minimum detection limits (MDLs). With regard to XRF analysis, the minimum detection limit for a given element is the concentration of that element related to the peak that can be still distinguished from its background. Therefore, the MDLs have direct impact on the quality of the acquired XRF spectra and the recorded XRD spectra. In order to be able to compare the performance of the Surface Monitor with other contemporary instrumentation, MDLs are also determined with other state-of-the-art instrumentation, such as the EDAX Eagle III micro-XRF spectrometer, and the Olympus Innov-X DELTA handheld XRF analyzer. The micro-XRF spectrometer is a laboratory instrument that is based on a Rh anode X-ray tube coupled with polycapillary optics, while the latter is a dedicated tool for on-field measurement using a Rh anode X-ray tube providing typically a 5 mm X-ray beam. To determine the MDLs for these three instruments, the NIST SRM 1412 was measured in optimal circumstances with each instrument. Information concerning the certified concentrations of this standard can be found in Appendix A on Page 59. The obtained XRF spectra were evaluated by using the AXIL spectrum evaluation software and the MDLs were calculated using Eq. (4.1). This equation gives the mathematical definition of the MDL for a given element i. The equation consists of the following parameters: CM DL,i , NB,i , NN,i and Ci , which respectively represent the minimum detection limit for element i, the number of background counts, the net number of counts and the concentration of element i in the standard reference material. The results for the three instruments are shown in the following paragraphs. CM DL,i

p NB,i × Ci = NN,i 3

(4.1)

In case of the Surface Monitor, the MDLs were determined by using three different combinations of slit and pinhole sizes at respectively the detector and X-ray source. Two slits of 2 mm, 1.5 mm

38


and 1 mm were combined with respectively two pinholes of 2 mm, 1.5 mm and 1 mm. For each combination, the standard was measured for 1000 s live time (LT) at a distance of 95.0 mm using the Cu X-ray source (28 kV/250 µA). The spectrum is shown in Fig. 4.11 and the minimum detection limits are shown in Fig. 4.9. For the Olympus Innov-X DELTA handheld XRF analyzer, the multicomponent glass standard was measured for 300 s LT. The resulting spectrum is shown in Fig. 4.12 and the obtained detection limits are normalized to 1000 s LT in order to compare with the Surface Monitor and the EDAX Eagle III. For the EDAX Eagle III four point measurements were performed of 250 s LT at different places on the standard, using a 300 µm beam. These measurements were performed in air and under vacuum conditions. After the measurement the four separate spectra were combined in a sum spectrum of 1000 s LT. The sum spectra (air and vacuum measurement) are shown in Fig. 4.13. The MDLs for the EDAX Eagle III and the handheld XRF analyzer are shown in Fig. 4.10. The MDLs for the Surface Monitor are between 100 ppm and 3000 ppm, except for Fe for which a lower MDL was obtained. If these values are compared with the other two instruments, the EDAX Eagle III and the handheld XRF analyzer, the MDLs for the Surface Monitor are at least one order of magnitude higher. This difference is mainly due to the large distance between the Surface Monitor and the object that was analyzed, in this case the NIST SRM 1412. By measuring at a larger distance there is a lot of air between the X-ray source and the sample, and between the sample and the detector. The large detector distance also reduces considerably the detection solid-angle, which has a negative influence on the MDL. In case of the Olympus Innov-X DELTA handheld XRF analyzer, there are less geometrical constraints so that the probe head can be placed at a much smaller distance from the sample’s surface. However, this instrument will not allow long measurement times when operating it in hand. Executing experiments with the EDAX Eagle III micro-XRF spectrometer, samples need to be brought to the hosting laboratory. The sample surface is brought in view of the polycapillary that confines the X-ray beam. The major advantage of this instrument is that, since all components are fixed, all instrumental parameters are optimized in order to acquire data in optimal conditions. E.g., measurements in vacuum allows detection down to Na.


39

Figure 4.9: Minimum detection limits (MDLs) as a function of the elements for the Surface Monitor, determined with the NIST SRM 1412. (1000 s LT; Cu X-ray tube; 28 kV; 250 µA; θtube = θdetector = 45°)

Figure 4.10: Minimum detection limits (MDLs) as a function of the elements determined with the NIST SRM 1412 for the EDAX Eagle III micro-XRF spectrometer in air and under vacuum conditions, and for the Olympus Innov-X DELTA handheld XRF analyzer. (Settings EDAX Eagle III micro-XRF spectrometer: 1000 s LT; Rh X-ray tube; 40 kV; air: 350 µA; vacuum: 220 µA; 300 µm beam/Settings Olympus Innov-X DELTA handheld XRF analyzer: 1000 s LT; Rh X-ray tube; 40 kV; 79 µA; 5 mm beam diameter)

40


Figure 4.11: XRF spectrum of the NIST SRM 1412. The glass standard was measured with the Surface Monitor by using the different combinations of slits and pinhole sizes at respectively the X-ray source and the detector. (1000 s LT; Cu X-ray tube; 28 kV; 250 µA; θtube = θdetector = 45°; distance = 95.0 mm)

Figure 4.12: XRF spectrum of the NIST SRM 1412. The glass standard was measured by using the Olympus Innov-X DELTA handheld XRF analyzer. (300 s LT; Rh X-ray tube; 40 kV; 79 µA; 5 mm beam diameter)


41

Figure 4.13: XRF spectrum of the NIST SRM 1412. The glass standard was measured with the EDAX Eagle III micro-XRF spectrometer in air and under vacuum conditions. (1000 s LT; Rh X-ray tube; 40 kV; air: 350 µA; vacuum: 220 µA; 300 µm beam)

4.1.5

Influence of Slits and Pinholes on XRD Spectra

As already mentioned in Section 3.2 the Surface Monitor can be equipped with different collimators (slits and pinholes). These collimators will certainly have an influence on the intensity and the quality of the diffraction peaks obtained from the XRD measurement. It is good to remember that during a field mission, e.g. investigation of art objects in museums, there is a certain time pressure to retrieve information in a restricted time. Therefore it is important to know the performance of the different combinations of pinholes (X-ray tube) and slits (detector) and in particular how these choices affect the quality of XRF/XRD spectra. Very often the scientist is balancing between performing long spectral measurements, acquiring only few spectra of good quality, and executing short experiments resulting in many more spectral results though of less quality. In the following, the effect of the choice of collimators is investigated, so that it may help the operator of the Surface Monitor when time-dependent decisions have to made. The effect of a few, but different combinations of collimators was studied on the basis of measurements of the NIST SRM 660b diffraction standard. Only pairs of vertical slits of the same size were used at the detector side, and also only pairs of pinholes of equal size were used at the X-ray tube. The 0.5 mm pinholes were not used at the tube because there is no beam coming through these collimators, see Section 4.1.3. So three different pinhole sizes were combined with four different slit sizes which makes a total of twelve combinations that were tested.

42


The acquisition parameters for all experiments for comparative studies: the distance was set at 94.5 mm, the Cu X-ray tube (28 kV/250 µA) was used, the acquisition started at a 2θ-angle of 15° to 90° with steps of 0.2° and a measuring time of 5 s LT/step. The obtained XRD spectra were processed by the “XRD match!” software. The noise of the spectrum was removed by setting the smooth factor to 10. In the next step, the peaks of the diffraction standard were selected and a Gaussian fit was applied. From this Gaussian fit the intensity of each peak was recorded. Table 4.5 shows how the intensity I changes relative to the intensity obtained with a pair of 1 mm X-ray tube pinholes coupled with a pair of 0.5 mm sized detector slits when operating at 94.5 mm and using the Cu based X-ray source. Three XRD spectra are shown in Fig. 4.14. Each spectrum is representing four measurements by varying the slit size and keeping the pinhole size constant.

(a) 2 x 1 mm pinhole at X-ray source

(b) 2 x 1.5 mm pinhole at X-ray source

(c) 2 x 2 mm pinhole at X-ray source

Figure 4.14: XRD spectra acquired by varying pinholes and slits sizes at respectively the X-ray tube and the detector. Each spectrum is representing four measurements with different slit sizes and by keeping the pinhole size constant.

Next to the XRD peak intensities, the quality of the spectra also depends on the width of the peaks observed in the XRD spectra. Changes in the full width at half maximum (FWHM) of a diffraction peak will give an indication how the peak width can be improved by varying the size of the collimators but it also provides useful information for the data processing. Table 4.6 shows the FWHM of the most intense diffraction peak of the LaB6 standard for the


43

different pinhole and slit combinations. This table shows that the FWHM only changes when the slit sizes are changed, changing the pinhole size at the X-ray tube has only a minimal or no influence on the FWHM of these peak. The width of the XRD peaks gives information on the uncertainty of the d-spacing values of the observed peaks, and therefore very useful when performing a mineral identification search in the ICDD database as mentioned in Fig. 3.15 in Section 3.3.3 on Page 27. Table 4.5: This table shows how the intensity (I) of the XRD peaks changes relatively to the smallest slit and pinhole size when replacing them by larger collimators. This table is only valid for the Cu X-ray tube, when measuring at a distance of 94.5 mm and when inserting two slits of the same size at the detector and two pinholes of the same size at the X-ray tube.

Detector slit size (mm)

X-ray tube pinhole size (mm)

0.5

1

1.5

2

1

I

2xI

3xI

3.5 x I

1.5

2xI

5xI

7.6 x I

9xI

2

2.5 x I

6.3 x I

9.5 x I

11.8 x I

Table 4.6: This table shows the change in FWHM, for the most intense diffraction peak (see Fig. 4.5) of the LaB6 diffraction standard, when different pinholes and slits are used on respectively the X-ray source and the detector. Each time two pinholes of the same size were inserted at the X-ray source and two slits of the same size were used at the detector.

Pinhole size X-ray tube

Slit size detector

(mm)

(mm)

1

1.5

2

FWHM ˚) (A

0.5

0.08

1

0.08

1.5

0.13

2

0.15

0.5

0.09

1

0.11

1.5

0.12

2

0.13

0.5

0.09

1

0.14

1.5

0.18

2

0.14

Combining the results shown in Tables 4.5 and 4.6, together with the results obtained in previous paragraphs, the following conclusions can be made. First of all, concerning the selection of the beam size, when a large homogeneous area needs to be analyzed, a large pinhole and slit

44


size is recommended. By using the largest collimators the acquisition time can be shortened to only 1 s LT/step and the intensity of the diffraction peaks will still be twice more intense when using the smallest collimators with 5 s LT/step. However, one needs to take into account that the width of the XRD peaks increases when using larger collimator sizes. In another situation, when a small area needs to be analyzed it is better to use the smallest pinhole size since this results in the smallest beam size of 16 mm x 3 mm. This beam size is obtained by measuring at 95 mm and by varying the tube angle from 12° to 45°. Concerning the slit sizes there are two options. When a large slit is used, the intensity of peak increases though its width also broadens. The choice of small slit sizes results in narrower but less intense peaks. To increase the number of counts, the acquisition time can be increased, but this leads to a longer measuring time. Finally, combining the results from Table 4.5 and Fig. 4.9, due to increased intensity when choosing larger collimators, better MDLs are obtained. The performance of the Surface Monitor equipped with the Mo anode X-ray tube was not investigated due to the lack of time. However, as already explained, it is expected that the Surface Monitor will best perform when equipped with the Cu anode X-ray source. In the next section, the results of the very first applications of the Surface Monitor are presented.

4.2

Applications

After the characterization, the methodology of the Surface Monitor was applied to two different paintings. The primary aim was to identify the pigments used in these works of art. One painting was provided by the Raman Spectroscopy Research Group and is shown in Fig. 4.15; this work of art could be investigated in the laboratories of the UGent-XMI Group. The other painting is the so called “Dulle Griet” or Mad Meg, painted by the Flemish renaissance artist Pieter Bruegel the Elder in 1561-1562. The painting is currently exhibited at the Museum Mayer van den Bergh in Antwerp (Belgium). A photo of the “Dulle Griet” is shown in Fig. 4.16. A few zones were selected from each painting and on each zone a simultaneous XRF/XRD measurement was performed. The results of these measurements are shown in the next two sections starting with the painting provided by the Raman Spectroscopy Research Group. All measurements were performed by using the Cu anode X-ray source (28 kV/250 µA).

Applications

45

Figure 4.15: “Colorful flowers in a vase” painted by Perot. This painting that was provided by the Raman Spectroscopy Research Group and it was investigated in the laboratories of the UGent-XMI Group. The zone that was analyzed is indicated on the photo.

46


Figure 4.16: “Dulle Griet” or Mad Meg, by Pieter Bruegel the Elder (1561-1562). The painting is exhibited at the Museum Mayer van den Bergh in Antwerp (Belgium). On this photo the four zones that were analyzed with the Surface Monitor are indicated. Fig. 4.19 shows four detailed pictures of these zones.

Applications

4.2.1

47

XRF/XRD Analysis of the Painting “Colorful Flowers in a Vase”

One zone of this painting was analyzed with the Surface Monitor and is indicated on Fig. 4.15. Details and results of the measurement are summarized in the next paragraph. The white pigment was characterized using the following acquisition set-up: 2θi of 30° to 2θf of 90°, step size of 0.2°, 5 s LT/step, two slits of 1.5 mm were used at the detector and two pinholes of 1.5 mm were inserted at the X-ray tube. The instrument was placed at a distance of 94.6 mm from the painting. The resulting XRF and XRD spectrum are shown in respectively Figs. 4.17 and 4.18. The XRF spectrum gives more information about the elements that are present in the white pigment, these elements are: Ca, Fe, Ba, Zn, Pb and Sr. For the element Cu it is impossible to determine if it is present in the sample or not because the Cu anode X-ray source was used. With the information obtained from the XRF spectrum the following white pigments could be present: BaSO4 , ZnO and lead white. To see which of these mineral phases are actually present in the white pigment, the XRD spectrum is analyzed. From the XRD spectrum, shown in Fig. 4.18 it is clear that the white pigment consists of the mineral BaSO4 . The other two possible minerals are also added but it is impossible to conclude that they are also present.

Figure 4.17: XRF spectrum of the lab painting’s white pigment (see Fig. 4.15). The spectrum was obtained using the Cu anode X-ray source (28 kV/250 µA).

48


Figure 4.18: XRD spectrum of the lab painting’s white pigment (see Fig. 4.15). Three mineral phases are added in the spectrum, the code before the mineral phase can be used to retrieve the XRD diffraction patterns from the ICDD database.

4.2.2

“Dulle Griet” from Pieter Bruegel the Elder

During the analysis of the “Dulle Griet” with the Surface Monitor four different zones were measured. These four zones are indicated on Fig. 4.16. Before the start of each measurement a photo was taken with the webcam, these detailed photos are shown in Fig. 4.19. Practically, there was a time limit of a one day experiment time (approximately 6 hours in total). During the measurements the following conditions were applied: 2θi 40°, 2θf 90°, step size 0.2°, 5 s LT/step, two 1.5 mm pinholes were used at the X-ray source and two 0.5 mm slits were used at the detector. The results, an XRF and XRD spectrum, of the four measurements are shown below. Zone 1 The first zone was measured by placing the Surface Monitor at a distance of 94.8 mm. The XRF and XRD spectra are shown in respectively Fig. 4.20 and Fig. 4.21. From the XRF spectrum it can be seen that the following elements are present: K, Ca, Mn, Fe, Co and Pb. The copper K-lines are also visible in the spectrum but since the Cu anode X-ray source was used, it is impossible to determine if there is Cu present or not. From the XRD spectrum there was only one mineral phase (CaCO3 ) found with a good match. Zone 2 The second zone was measured at a distance of 94.0 mm and the resulting XRF and XRD spectra are shown in respectively Fig. 4.22 and Fig. 4.23. Seven elements are present in the

Applications

49

XRF spectrum: K, Ca, Mn, Fe, Co, Hg and Pb. The XRD spectrum shows the possible presence of CaCO3 , CaSO4 and the pigment HgO.

(a) Zone 1

(b) Zone 2

(c) Zone 3

(d) Zone 4

Figure 4.19: Detailed photos of the four zones of the “Dulle Griet” that were measured with the Surface Monitor.

Zone 3 The third zone was measured by placing the instrument at a distance of 94.0 mm. The acquired XRF and XRD spectra are show in respectively Fig. 4.24 and Fig. 4.25. The XRF spectrum shows the presence of K, Ca, Mn, Fe and Pb in this zone. The XRD spectrum shows the possible presence of CaCO3 and CaSO4 . Zone 4 The final zone was measured at a distance of 94.0 mm. The obtained XRF and XRD spectra are shown in respectively Fig. 4.26 and Fig. 4.27. The XRF spectrum shows the presence of the following elements: K, Ca, Mn, Fe, Co, Hg and Pb. The XRD spectrum shows the possible presence of CaCO3 and CaSO4 . From the obtained data it was only possible to determine the mineral phase CaCO3 , probably originating from the preparation layer of the painting. No pigments could be determined with certainty, due to the complexity of the painting. The painting is in fact a multilayer

50


consisting of different pigment layers and a varnish layer, which makes it hard to find the different components when the noise level in the XRD spectrum is to high. The relatively large size of the X-ray beam can also form a problem because adjacent areas colored with other pigments are irradiated next to the area of interest. For future measurements it should be considered to measure only one or two points for a longer time to obtain at least one high quality result instead of measuring as much points as possible, using a shorter measuring time, resulting in less useful data.

Figure 4.20: XRF spectrum of the first zone of the Dulle Griet painting.

Figure 4.21: XRD spectrum of the first zone of the Dulle Griet painting.

Applications

Figure 4.22: XRF spectrum of the second zone of the Dulle Griet painting.

Figure 4.23: XRD spectrum of the second zone of the Dulle Griet painting.

51

52


Figure 4.24: XRF spectrum of the third zone of the Dulle Griet painting.

Figure 4.25: XRD spectrum of the third zone of the Dulle Griet painting.

Applications

Figure 4.26: XRF spectrum of the fourth zone of the Dulle Griet painting.

Figure 4.27: XRD spectrum of the fourth zone of the Dulle Griet painting.

53

54


Chapter 5

Summary and Conclusions The aim of this master thesis was the establishment of the methodology, the characterization and the application of a novel portable X-ray fluorescence (XRF)/X-ray diffraction (XRD) instrument, called the Surface Monitor. This instrument was recently acquired by the Analytical Chemistry Department of the Ghent University (Belgium) from the company Assing S.p.A. (Rome, Italy). This instrument is described to be used for studying art objects, e.g. paintings and murals, by combining two powerful non-destructive analytical techniques: X-ray fluorescence and X-ray diffraction. The success of the XRF methodology is resulting from the fact that the interaction of X-rays with matter may be fully understood, giving the researcher a powerful and practical tool to study samples. Where XRF is capable to characterize the elements present in the sample, XRD gives information about the arrangement of these elements in the sample, the so called crystal structure. Thus, a combination of these two analysis techniques, XRF and XRD, can be used to obtain a detailed chemical/structural characterization of the investigated materials. The main parts of the Surface Monitor are: the X-ray tube (Cu or Mo anode), the X-123 Si-PIN detector, the laser interferometer and the XRD protractor. All these components are assembled on the instrument head that is fixed on a tripod. Furthermore, the Surface Monitor is equipped with a control box, a control portable PC and an extra arm with webcam. This webcam is not part of the original Surface Monitor as provided by the company Assing S.p.A, but was added by the UGent-XMI group to improve the safety. The webcam allows the operator to watch the status of the device head, on a distance, using the control portable PC. Finally, the manufacturer included collimators in the form of slits and pinholes. These collimators are used to adjust the beam size and the area analyzed by the detector. For the acquisition of XRD spectra, the Surface Monitor is equipped with the XRD protractor. This part ensures the simultaneous movement of the X-ray tube and detector according to the Bragg-Brentano θ : θ set-up. In order to obtain reliable and reproducible XRD spectra, the Surface Monitor should be positioned with its main axis perpendicular to the sample’s surface and at the optimal acquisition distance. The Surface Monitor is positioned at the 55

56

Summary and Conclusions

optimal acquisition distance by using the laser interferometer. The performance, i.e. the precision, of the laser interferometer was a first point of concern during the characterization of the Surface Monitor. The precision of the laser was tested by moving a flat surface with a manual translational stage over certain range. A precision of ±0.11 mm was obtained from the

results, which is sufficient for further applications. Beside the laser interferometer’s precision, the optimal distance for the acquisition of XRF/XRD spectra was determined. When the Surface Monitor was delivered, the distance to acquire XRF and XRD spectra was set by the manufacturer between 91.8 mm and 92.2 mm. During the first tests with the instrument it became clear that this was not the optimal distance for the acquisition of XRF and XRD spectra. Therefore a new range was defined by performing XRD measurements with the LaB6 diffraction standard at different distances. From the obtained data it is now recommended to place the instrument at a distance between 94.5 mm and 95.0 mm from the object to obtain reliable XRD spectra. The next step in the Surface Monitor’s characterization was the determination of the beam size and more importantly the position of the X-ray beam with respect to the footprint of the laser interferometer beam. The X-ray beam can be collimated by inserting different collimator sizes at the X-ray source. During the measurements only pairs of collimators of the same size were used to regulate the size of the X-ray beam. From the obtained results it is recommended to use the pinhole collimators, since the beam size obtained with the slits is to large for many further applications. Next to the size of the beam, the position of the beam with respect to the laser interferometer has also been determined. When the instrument is positioned at a distance of 95 mm, the laser interferometer is indicating the center of the X-ray beam, which gives an extra indication that the object must be placed at this distance. Once the position of the beam and its size on the sample were known with high accuracy, the minimum detection limits for certain detectable elements could be determined and compared to other state-of-the-art instrumentation: the EDAX Eagle III micro-XRF spectrometer and the Olympus Innov-X DELTA handheld XRF analyzer. Minimum detection limits were obtained by using the NIST SRM 1412 multicomponent glass standard. The resulting minimum detection limits for the Surface Monitor are between 100 ppm and 3000 ppm, depending on the atomic number of the detected element, which are approximately one magnitude higher than the minimum detection limits obtained with the EDAX Eagle III micro-XRF spectrometer and the Olympus Innov-X DELTA handheld XRF analyzer. This difference is mainly due to the large distance between the Surface Monitor and the object that was analyzed. By measuring at a larger distance there is a long air-path between the X-ray source and the sample, and between the sample and the detector. The large detector distance also reduces considerably the detection solid-angle, which has a negative influence on the minimum detection limit. In case of the Olympus Innov-X DELTA handheld XRF analyzer, there are less geometrical constraints so that the probe head can be placed at a much smaller distance from the sample’s surface. Executing experiments with the EDAX Eagle III micro-XRF spectrometer, all instrumental

57 parameters are optimized in order to acquire data in optimal conditions. The last step in characterizing the portable XRF/XRD instrument dealt with the influence of changing the collimator sizes, at both X-ray source and detector, on the quality of the XRD spectra. The results were obtained by acquiring XRD spectra of the LaB6 diffraction standard while using different combinations of slits and pinholes at respectively the detector and the X-ray source. From the results we can conclude that it is better to use the largest collimators when measuring a large homogeneous surface, this reduces the measuring time but also increases the width of the diffraction peaks in the XRD spectrum. For small areas, the collimators with the smallest opening are recommended but the measuring time must be sufficiently long to obtain good results. The measurements provided very useful information in selecting the correct collimators before starting an XRD measurement. This information will allow future operators of the Surface Monitor to make fast decisions when measuring on the field. The final part of this master thesis dealt with the analysis of two paintings: “Colorful flowers in a vase” by Perot and the famous “Dulle Griet” by Pieter Bruegel the Elder (1561-1562). The first painting was provided by the Raman Spectroscopy Research Group and it was available at the laboratory of UGent-XMI Group. One zone, consisting of a white pigment, was analyzed by the Surface Monitor and the presence of the pigment BaSO4 was demonstrated. The analysis of the “Dulle Griet” at the Mayer van den Bergh museum in Antwerp (Belgium), was the first application of the Surface Monitor on the field. Due to a limited amount of time (approximately 6 hours in total) only four measurements were performed on the painting. From the obtained data it was only possible to determine the mineral phase CaCO3 , probably originating from the preparation layer of the painting. No pigments could be determined with certainty, due to the complexity of the painting. The painting is in fact a multilayer consisting of different pigment layers and a varnish layer, which makes it hard to find the different components when the noise level in the XRD spectrum is to high. The relatively large size of the X-ray beam can also form a problem because adjacent areas colored with other pigments are irradiated next to the area of interest. For future measurements it should be considered to measure only one or two points for a longer time to obtain at least one high quality result instead of measuring as much points as possible, using a shorter measuring time, resulting in less useful data. Finally, it should be mentioned that the Surface Monitor is still under development and can still be improved in some ways. The detection of X-rays can be improved by using a Silicon Drift Detector (SDD) instead of the Si-PIN detector. The SDD has a lower dead time and higher resolution compared to the Si-PIN detector that is now used by the Surface Monitor. The Cu anode X-ray source can be changed by the Mo anode X-ray source, but this is only improving the quality of the XRF spectra. For XRD measurements it is recommended to use the Cu anode X-ray tube. The geometry of the XRD protractor can be changed to decrease the optimal distance between the probe head and the sample. If this distance could be decreased,

58

Summary and Conclusions

the detector and the X-ray source are positioned at a closer distance towards the sample’s surface which improves the count rate.

Appendix A

Standard Reference Material A.1

NIST SRM 1412 Multicomponent Glass Standard

The NIST SRM 1412 multicomponent glass standard was used to determine the minimum detection limits for the Surface Monitor and to compare them with two other instruments: the EDAX Eagle III micro-XRF spectrometer and the Olympus Innov-X DELTA handheld XRF analyzer. The certified mass fractions of this standard are shown in Table A.1. Table A.1: Certified concentrations for the NIST SRM 1412 multicomponent glass standard. Reproduced from NIST SRM 1412 certificate [32].

Element

A.2

Mass fraction (ppm)

Element

Mass fraction (ppm)

Si

198074

Li

20904

Al

39798

B

14068

Ca

32376

Ba

41827

Mg

28280

Zn

35992

Sr

38474

Pb

40846

Na

34797

Cd

38342

K

34368

Fe

217

NIST SRM 660b Diffraction Standard

The NIST SRM 660b diffraction standard was used to optimize the Surface Monitor’s XRD acquisition mode. The standard was used e.g. to determine the ideal range for XRD acquisition, to determine the offset of the diffraction peaks, and to study the consequences of using different slit and pinhole sizes on respectively the detector and the Cu anode X-ray tube. A unit of the SRM 660b consists of approximately 6 g of lanthanum hexaboride (LaB6 ) powder bottled under argon. Before one can start measuring the LaB6 -powder it is necessary to bring the powder in a stable formulation so that there is no loss of the material during transport or during the analysis and to prevent contamination with other crystalline materials. Therefore, a sufficient 59

60

Standard Reference Material

amount of the LaB6 -powder is placed between two layers of 4 µm Ultralene

® Window Film

(purchased from SPEX CertiPrep Group) and fixed on an 31 mm Double Open-Ended X-Cell (purchased from SPEX CertiPrep Group). The result is shown in Fig. A.1. Since the polymer

® Window Film also gives rise to diffraction peaks,

used for the manufacturing of the Ultralene

a second sample was made without the NIST SRM 660b. From both samples, one with the diffraction standard and one without diffraction standard, an XRD spectrum was record and shown in Fig. A.2. By comparing both spectra in Fig. A.2 one can see that the peaks at 17.4° and 25.6° are originating from the Ultralene

® film. The information values for peak positions

between 20° and 90° for the NIST SRM 660b are shown in Table A.2.

® and

Figure A.1: NIST SRM 660b (LaB6 ) diffraction standard placed between two layers of Ultralene

fixed on a Double Open-Ended X-Cell that can be stored in a box for easy transportation.

(a)

Figure A.2: Diffraction spectra of the Ultralene SRM 660b diffraction standard.

(b)

® Window Film (a) with and (b) without the NIST

NIST SRM 660b Diffraction Standard

61

Table A.2: Information values for peak positions computed for the NIST SRM 660b, using the Cu-Kα radiation. Only the values between 20° and 90° 2θ are shown. Reproduced from NIST SRM 660b certificate [32].

h

k

l

2θ (°)

h

k

l

2θ (°)

1

0

0

21.358

3

0

0

67.548

1

1

0

30.385

3

1

0

71.746

1

1

1

37.442

3

1

1

75.844

2

0

0

43.507

2

2

2

79.870

2

1

0

48.958

3

2

0

83.846

2

1

1

53.989

3

2

1

87.792

2

2

0

63.219

62

Standard Reference Material

Appendix B

Flo2spe-Converter The program flo2spe-converter can be used to convert the flo output files, generated by the Surface Monitor’s software package, into spe files. The flo output file contains all information concerning an XRF acquisition such as: tube voltage and current, detector and tube angles, measuring time, etc. Next to this information it also contains the number of counts per channel. The number of counts per channel will be extracted from the flo file and written in a new spe file by the program flo2spe-converter. This spe file can be used to process the obtained XRF spectrum by the AXIL software package. The program flo2spe-converter was written in the C language on a CentOS (release 6.1) operating system and compiled with gcc (version 4.4.5). The source code is given below. Source code flo2spe-converter 1

// Program to convert flo to spe // Written by Robin De Wolf during Master Thesis (2011-2012)

3

#include<stdio.h> 5

#include<stdlib.h> #include<stdbool.h>

7

#include<string.h>

9

#define CHANNELS 2048 /*Change channel number here if necessary*/

11

int main (int argc, char *argv[]) { int fileCount;

13

int buffer[CHANNELS];

15

bool ReadFlo (char *inputfile, int output[CHANNELS]); bool MakeSpe (char *inputfile, const int counts[], const int channels);

17

if (argc == 1) { 19

/*Check if there is at least 1 input file*/ printf ("Need one or more .flo files!\n");

return EXIT_FAILURE; } 21

63

64

Flo2spe-Converter fileCount = argc - 1; /*fileCount is equal to the number of input files*/

23

do { 25

if (ReadFlo (argv[fileCount], buffer) == false) { printf ("Couldn’t make spe file from %s!\n", argv[fileCount]);

27

return EXIT_FAILURE; } else if (MakeSpe (argv[fileCount], buffer, CHANNELS) == false) {

29

printf ("Couldn’t make spe file from %s!\n", argv[fileCount]); return EXIT_FAILURE; }

31

fileCount -= 1; 33

35

} while (fileCount != 0); return EXIT_SUCCESS; }

37

// Function to read .flo file

39

bool ReadFlo (char *inputfile, int output[CHANNELS]) { bool CheckFlo (char *inputfile);

41

if (CheckFlo (inputfile)) { 43

int charNumber, i = 0;

45

FILE *in = fopen (inputfile, "r");

47

if (in == NULL) { /*Check if it is possible to read the .flo file*/

/*Open the .flo file*/

printf ("Can’t read %s!\n", inputfile); 49

return false; }

51

// Skip the first 12 lines

53

do { charNumber = fgetc (in);

55

if (charNumber == 10) /*Check if char is equal to newline character (\n)*/ ++i; } while (i < 12 && charNumber != -1);

57

59

i = 0;

61

if (charNumber == -1) return false;

63

// Read .flo file starting from line 13 65

do { 67

fscanf (in, "%i %*c %*f %*c", &output[i]); ++i;

69

} while (fgetc (in) != EOF && i < CHANNELS);

65

71

fclose (in);

73

return true; }

75

/*Close flo file*/

return false; }

77

// Function to check flo file

79

bool CheckFlo (char *inputfile) { if (strstr (inputfile, ".flo") != NULL)

81

return true; else {

83

printf("%s isn’t a correct flo file!\n", inputfile); return false; } }

85

// Function to create the .spe file 87

89

bool MakeSpe (char *inputfile, const int counts[], const int channels) { char *outputfile; int i;

91

void NewName (char *old, char *new); 93

// Allocate memory to pointer 95

if ((outputfile = (char *) malloc ((strlen (inputfile) + 1) * sizeof (char))) == NULL) { /*strlen doesn’t count the terminating ’\0’, therefore the +1 is 97

inserted*/ printf("Memory allocation error!"); return false; }

99

// Create file name 101

NewName (inputfile, outputfile); 103

// Open the new .spe file 105

FILE *out = fopen (outputfile, "w"); 107

if (out == NULL) { 109

printf("Can’t write %s!\n", outputfile); return false; }

111

// Layout .spe file 113

fprintf(out, "$DATA:\n"); 115

fprintf(out, "1 %i\n", channels);

117

// Print data to .spe file

66

119

Flo2spe-Converter

for (i = 0; i < channels; ++i) { fprintf(out, "%i\n", counts[i]); }

121

// Close .spe file 123

fclose (out); 125

// Free the allocated memory 127

free (outputfile); 129

return true; } 131

// Function to create a name for the new .spe file 133

135

137

void NewName (char *old, char *new) { char *p; p = strchr (old, ’.’); strcpy(new, old);

139

strcpy (new + (p - old), ".spe"); }

flo2spe.c

Bibliography [1] Assing S.p.A. (2011). Adress: Via E. Amaldi 14 - Monterotondo (Italy). URL http: //www.assing-group.it/index_eng.html. [2] B. Beckhoff, N. Langhoff, B. Kanngiefer, R. Wedell & H. Wolff, eds. (2006). Handbook of Practical X-Ray Fluorescence Analysis. Springer Verlag, Berlin. [3] L. Vincze (2011). Advanced Spectroscopic Methods of Analysis. Course Master Chemistry, Ghent University 2011-2012. [4] Nobelprize.org (2011). Wilhelm Conrad Röntgen. URL http://www.nobelprize. org/nobel_prizes/physics/laureates/1901/rontgen-bio.html. [5] Amptek (2011). X-ray Fluorescence Spectroscopy. URL http://www.amptek.com. [6] D. Depla (2009). Kristalchemie, Een inleiding op X-straaldiffractie. Course Bachelor Chemistry, Ghent University 2009-2010. [7] P. Jeffrey (2012). X-ray Data Collection Course. URL http://xray0.princeton. edu/˜phil/Facility/Guides/XrayDataCollection.html. [8] C. Suryanarayana & M. Grant Norton, eds. (1998). X-Ray diffraction: a practical approach. Plenum Press, New York. [9] Wikipedia (2011). Thomson Scattering. URL http://en.wikipedia.org/wiki/ Thomson_scattering. [10] C. Hammond, ed. (2009). The Basics of Crystallography and Diffraction. Oxford University Press Inc., New York, third edit ed. [11] Wikipedia (2011). Bragg’s law. URL http://en.wikipedia.org/wiki/Bragg’s_ law. [12] Wikipedia (2011).

X-ray tube.

URL http://en.wikipedia.org/wiki/X-ray_

tube. [13] J. K. Cockcroft (2011). Generation of X-rays. URL http://pd.chem.ucl.ac.uk/ pdnn/inst1/xrays.htm.

67

68

Bibliography

[14] P. J. Goodhew, J. Humphreys & R. Beanland, eds. (2001). Electron Microscopy and Analysis. Taylor & Francis, London, third ed. [15] M. Azadeh, ed. (2009). Fiber Optics Engineering. Optical Networks. Springer US, Boston, MA, 1 ed. [16] Assing (2011). Assing Surface Monitor User’s Manual. [17] D. Blake & C. Bryson (1992). Design of an x-ray diffraction/x-ray fluorescence instrument for planetary applications. Lunar and Planetary Institute, 13:117–118. [18] D. Blake, C. Bryson & F. Freund (1993). X-ray diffraction apparatus, U.S. Pattent No. 5,491,738. [19] S. Cornaby, A. Reyes-Mena, H. K. Pew, P. W. Moody, T. Hughes, A. Stradling, D. C. Turner & L. V. Knight (2001). An XRD/XRF instrument for the microanalysis of rocks and minerals. Measurement Science and Technology, 12(6):676–683. [20] S. Cornaby, A. Reyes-Mena, P. Moody, A. Stradling, T. Grow & L. Knight (2002). Simultaneous XRD/XRF with Low-Power X-ray Tubes. Adv. in X-ray Analysis, 45(c):34. [21] M. Uda (2004). In situ characterization of ancient plaster and pigments on tomb walls in Egypt using energy dispersive X-ray diffraction and fluorescence. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 226(1-2):75–82. [22] M. Uda, A. Ishizaki, R. Satoh, K. Okada, Y. Nakajima, D. Yamashita, K. Ohashi, Y. Sakuraba, A. Shimono & D. Kojima (2005). Portable X-ray diffractometer equipped with XRF for archaeometry. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 239(1-2):77–84. [23] A. Gianoncelli, J. Castaing, L. Ortega, E. Dooryhee, J. Salomon, P. Walter, J. Hodeau & P. Bordet (2008). A portable instrument for in situ determination of the chemical and phase compositions of cultural heritage objects. X-Ray Spectrometry, 37(4):418–423. [24] A. Duran, J. Perez-Rodriguez, M. Jimenez de Haro, M. Franquelo & M. Robador (2011). Analytical study of Roman and Arabic wall paintings in the Patio De Banderas of Reales Alcazares’ Palace using non-destructive XRD/XRF and complementary techniques. Journal of Archaeological Science, 38(9):2366–2377. [25] S. Pagès-Camagna, E. Laval, D. Vigears & A. Duran (2010). Non-destructive and in situ analysis of Egyptian wall paintings by X-ray diffraction and X-ray fluorescence portable systems. Applied Physics A, 100(3):671–681. [26] A. Pifferi, G. Campi, C. Giacovazzo & E. Gobbi (2009). A new portable XRD/XRF instrument for non-destructive analysis. Croatica Chemica Acta, 82(2):449–454.

Bibliography

69

[27] Amptek Inc. (2012). Adress: De Angelo Drive 14 - Bedford (U.S.A.). URL http://www. amptek.com. [28] B. Vekemans, K. Janssens, L. Vincze, F. Adams & P. Van Espen (1994). Analysis of X-ray spectra by iterative least squares (AXIL): New developments. X-Ray Spectrometry, 23(6):278–285. [29] B. De Samber (2010). Spatially resolved X-ray micro/nano-spectroscopy and imaging on the model organism Daphnia magna using laboratory and synchrotron sources. Ph.D. thesis, Ghent University. [30] International Center for Diffraction Data (2012). Adress: Campus Boulevard 12 - Newtown Square (U.S.A.). URL http://www.icdd.com/. [31] NIST (2012). Measures of Skewness and Kurtosis. URL http://www.itl.nist.gov/ div898/handbook/eda/section3/eda35b.htm. [32] NIST (2012). index.cfm.

Standard Reference Materials.

URL http://www.nist.gov/srm/

70

Bibliography

Ingebruikname, Analytische Karakterisering en Toepassing van een Nieuw Draagbaar XRF/XRD Instrument Robin De Wolfa,∗ , Bart Vekemansa , en Laszlo Vinczea a

X-Ray Microspectroscopy and Imaging Group (XMI), Department of Analytical Chemistry, Ghent University, Krijgslaan 281, S12, B-9000 Ghent, Belgium ∗ E-mail: [email protected]

De Surface Monitor, i.e. een draagbaar X-straal fluorescentie (XRF)/Xstraal diffractie (XRD) instrument, werd in 2011 aangekocht bij de firma Assing S.p.A (Rome, Italië) door de Vakgroep Analytische Chemie van de Gentse Universiteit (België). Vooraleer dit toestel kan worden toegepast om metingen uit te voeren op kunstobjecten werden verschillende aspecten van de Surface Monitor getest. Dit houdt in: het bepalen van de laser interferometer’s betrouwbaarheid, het bepalen van de optimale XRF/XRD meetafstand, het bepalen van de bundelgrootte en de positie van de X-straal ten opzichte van de laser interferometer, bepalen van de minimum detectie limieten voor XRF, en tenslotte de invloed onderzoeken van collimatoren op XRD spectra. De Surface Monitor werd voor de eerste keer getest tijdens de karakterisatie van het schilderij de “Dulle Griet” in het museum Mayer van den Bergh in Antwerpen (België).

Trefwoorden: draagbaar, XRF, XRD, Surface Monitor Inleiding De analyse van kostbare kunstobjecten is een moeilijke onderneming omdat het meestal onmogelijk is om stalen te nemen van deze kostbare voorwerpen, en deze objecten kunnen niet altijd naar een laboratorium gebracht worden om te worden geanalyseerd met niet-destructieve analytische technieken. Daarom werd er de laatste jaren ge¨ınvesteerd in onderzoek naar draagbare XRF/XRD systemen welke het mogelijk maken de kunstobjecten ter plaatse, bv. in een museum, te gaan analyseren. Hieronder volgt een kort overzicht. In het NASA (National Aeronautics and Space Administration) Ames Research Center werd in 1992 het eerste prototype van een XRF/XRD instrument ontwikkeld [1, 2]. Dit instrument combineert twee krachtige niet-destructieve analytische technieken nl. XRF en XRD. XRF wordt ingezet om al een eerste indicatie te geven van de aanwezige mineralen, aan de hand van de elementaire samenstelling van het monster. De definitieve identificatie van de aanwezige mineralen is met behulp van XRD. Het uiteindelijke toestel zou worden gebruikt voor planetaire exploratie en als een draagbaar instrument voor het analyseren van kunst objecten. In 2000 werd er door MOXTEK, Inc. een poging ondernomen om een draagbaar XRF/XRD instrument te ontwikkelen [3]. Figuur 1a toont een foto van deze opstelling, met de bijhorende componenten. De CCD (charged-coupled device) detector kan gezien worden als één van de belangrijkste onderdelen van dit toestel. Deze detector maakt het immers mogelijk om gelijktijdig zowel een XRF als een XRD spectrum op te nemen. Met behulp van een software algoritme worden de XRF en de XRD data van elkaar gescheiden. Het grootste nadeel aan dit instrument is de grote X-stralen buis, maar deze zal voor later toepassingen worden vervangen door een laag vermogen X-stralen buis. Naast de te grote X-stralen buis is er ook nog de CCD detector waarvan de resolutie te laag is om bepaalde diffractie pieken van elkaar te kunnen onderscheiden. 71

In 2004 combineerde Uda [4] een laag vermogen X-stralen buis met een Si-PIN detector in een draagbare energie dispersieve X-straal diffractie en fluorescentie (ED-XRDF) spectrometer. Deze combinatie had vele voordelen: kleine dimensie, licht gewicht, korte meettijd, nietdestructieve en contact vrije meting, geen restricties met betrekking tot de vorm en grote van het monster, er is geen speciale koeling nodig, en het toestel wordt bestuurd met behulp van een computer. Het toestel werd gebruikt om op archeologische sites de samenstelling van pigmenten te bepalen. Desondanks de vele voordelen van dit instrument is het nog steeds moeilijk om de samenstelling van complexe pigmenten te bepalen door een te hoge achtergrond en minder intense pieken in het diffractie spectrum [5]. Een nieuw XRF/XRD instrument werd in 2007 ontwikkeld door Gianoncelli et. al [5]. Figuur 1b toont een foto van dit instrument. Dit toestel bestaat uit een koper anode X-stralen buis (rechts op de foto), een silicon drift XRF detector (centraal op de foto) en een imaging plate (links op de foto). Alle onderdelen zijn bevestigd op een frame wat het mogelijk maakt om te bewegen langsheen het oppervlak van het te analyseren object. Belangrijkste nadeel aan dit instrument is de reproduceerbaarheid van de XRD resultaten: 1 mm afwijken van het referentie punt zorgt al voor een afwijking van 3° in het XRD spectrum. In 2008 werd door de firma Assing S.p.A. (Rome, Italië) een compact instrument ontwikkeld waarmee het mogelijk was gelijktijdig XRF en XRD metingen uit te voeren [6]. Volgens Pifferi et. al [7] waren de resultaten, bekomen met dit nieuwe toestel nog steeds onbetrouwbaar door de mechanische instabiliteit. Om deze problemen te overkomen werd het prototype aangepast met een koper anode X-stralen buis, een silicon solid state detector en een horizontale Theta-Theta goniometer (zie Figuur 1c). Met behulp van een laser interferometer wordt het instrument op de juiste afstand van het te analyseren object geplaatst en het toestel kan in drie verschillende modi worden gebruikt: een XRF, een XRD en een XRF/XRD modus. De laatste modus houdt in dat tegelijkertijd een XRF en XRD spectrum kan worden opgenomen. Verder maakt de hardware het mogelijk dat de X-stralen buis en de detector symmetrisch over een interval van 10° < 2θ < 140° kunnen bewegen.

(a) XRF/XRD instrument ontwikkeld door MOXTEK, Inc. gebruikt door Cornaby et. al [3].

(b) XRF/XRD instrument gebruikt door Gianoncelli et al. [5].

(c) XRF/XRD instrument gebruikt door Pifferi et al. [7].

Figuur 1: Foto’s van XRF/XRD instrumenten beschreven in de literatuur.

72

In 2011 werd de Surface Monitor aangekocht door de Vakgroep Analytische Chemie van de Gentse Universiteit. De beschrijving van de componenten van dit toestel en de bijhorende dataverwerking worden beschreven in het volgende deel van dit werk. Beschrijving van de Surface Monitor De Surface Monitor van Assing S.p.A (model 2011-2012) bestaat uit de volgende onderdelen: een meetkop, een statief, een controle box, een laptop en een extra arm met een webcamera. De webcamera werd door de UGent-XMI groep aan de Surface Monitor toegevoegd om de veiligheid te bevorderen. De webcamera zorgt ervoor dat de operator vanop afstand de status van de Surface Monitor kan controleren met behulp van de laptop. Figuur 2a toont de volledige opstelling van de Surface Monitor: de bovenvermelde meetkop is gemonteerd op een statief en verbonden met de controle box met de nodige elektronica, de besturing gebeurt met behulp van de laptop. De webcamera is niet te zien op deze foto, maar wordt bevestigd aan de horizontale arm van het statief en wordt naar de meetkop gericht. Figuur 2b toont een bovenaanzicht van de meetkop met de belangrijkste onderdelen: X-stralen buis, X-123 Si-PIN detector en tussenin de laser interferometer.

(a) Opstelling Surface Monitor

(b) Meetkop

(c) Collimatoren

Figuur 2: Foto’s van (a) de opstelling van de Surface Monitor volgens het veiligheidshandboek (1, laptop; 2, controle box; 3, meetkop; 4, statief), (b) bovenaanzicht van de meetkop en (c) beschikbare collimatoren (links: pinholes gebruikt voor de X-stralen buis; rechts: verticale slits gebruikt voor de detector); de collimatoren worden steeds in paren gebruikt.

Twee X-stralen buizen zijn beschikbaar om metingen met de Surface Monitor uit te voeren: één koper anode en één molybdeen anode X-stralen buis. Enkel de koper anode werd gebruikt tijdens het uitvoeren van de experimenten omdat deze X-stralen buis de beste resultaten geeft op vlak van XRD in vergelijking met de molybdeen X-stralen buis. De maximale spanning en stroomsterkte op de koper anode bedragen respectievelijk 30 kV en 500 µA. Tijdens de metingen werd er telkens een spanning van 28 kV en een stroomsterkte van 250 µA aangebracht op de koper anode X-stralen buis. Verder dient nog te worden opgemerkt dat beide X-stralen buizen niet uitgerust zijn met een monochromator. Voor de detectie van X-stralen is de Surface Monitor uitgerust met een X-123 Si-PIN detector van de firma Amptek Inc. [8]. Beide onderdelen, X-stralen buis en detector, kunnen voorzien worden van collimatoren met verschillende diameter of met een variërende breedte. De beschikbare collimatoren zijn weergegeven in Figuur 2c. De pinhole diameters zijn, van bovenaan te beginnen, ongeveer 0, 5 mm, 1 mm, 1, 5 mm en 2 mm. De breedte van de verticale slits zijn, van bovenaan te beginnen, ongeveer 0, 5 mm, 1 mm, 1, 5 mm en 2 mm, de lengte van elke slit is ongeveer gelijk aan 8 mm. De collimatoren kunnen in de voorziene ruimtes van de X-stralen buis en detector worden geschoven en ze worden gefixeerd met behulp van schroeven. Deze schroeven worden tijdens de metingen verwijderd om te voorkomen dat deze het te analyseren object zouden beschadigen en om de collimatoren handig te kunnen verwisselen tussen twee metingen door. De keuze van het type 73

collimatoren heeft een belangrijke invloed op de XRD spectra. Indien een collimator met een kleine opening wordt gekozen, worden de pieken in het XRD spectrum smaller en de resolutie hoger. Daartegenover staat dat het instrument zeer precies moet worden gepositioneerd. Tijdens de metingen werden paren van pinholes enkel gebruikt voor de X-stralen buis en paren van verticale slits voor de detector. Het volgende onderdeel van de Surface Monitor, is de laser interferometer. Dit onderdeel wordt gebruikt om de Surface Monitor op de optimale afstand van het te analyseren object te plaatsen, wat belangrijk is om betrouwbare en reproduceerbare XRD spectra te bekomen. Tenslotte is er nog de XRD protractor. Dit onderdeel zorgt voor de simultane beweging van de X-stralen buis en de detector tijdens een XRD meting, volgens de Bragg-Brentano θ : θ methode. Deze meetmethode houdt ook in dat de afstand tussen de X-stralen buis en het staal, en tussen de detector en het staal identiek moeten zijn om betrouwbare resultaten te bekomen. In de praktijk is het bereik van de detector en de X-stralen buis beperkt omdat deze bij kleine hoeken het te analyseren object kunnen raken. Dataverwerking De evaluatie van de XRF spectra werd uitgevoerd met behulp van het AXIL software pakket [9]. Deze software werd onder andere gebruikt voor het bepalen van de karakteristieke fluorescentie lijnen, sompieken, escape pieken, enz. Eens de spectrum evaluatie is afgerond, kan met behulp van de netto piek oppervlakten bv. de minimum detectie limieten berekend worden. De interpretatie van de XRD spectra werd uitgevoerd via het programma “XRD match!”, dat behoort tot het software pakket van de Surface Monitor. De interpretatie van de XRD spectra is gebaseerd op het vergelijken van de bekomen diffractie pieken met gegevens uit een database. Voor dit werk werd een database van het International Center for Diffraction Data (ICDD) (PDF-4/minerals 2010 relational database; version 4.1011) gebruikt [10]. Vooraleer het bekomen spectrum kan vergeleken worden met de gegevens uit de database dient de gebruiker eerst een aanduiding te geven van de aanwezige elementen (deze informatie wordt bekomen uit het XRF spectrum) en moeten de verschillende diffractie pieken geselecteerd worden. Hieruit volgt een lijst met mineralen die mogelijk aanwezig zijn in het staal. De XRD spectra van deze mineralen worden vervolgens vergeleken met het experimenteel bekomen XRD spectrum tot er een match wordt gevonden. Experimenteel Zoals reeds vermeld speelt de laser interferometer een cruciale rol in het positioneren van de Surface Monitor voor het te analyseren object. Omdat dit zo een belangrijk onderdeel is, werd bepaald hoe precies de afstandsmeting gebeurt. Hiervoor werd de laser loodrecht op een vlak oppervlak gericht, i.e een stuk papier geplakt op een houder dat op zijn beurt op een “manual translational stage” (Standa; type 7T184-13) werd geplaatst. De “manual translational stage” had een afleesnauwkeurigheid van 0, 005 mm. Het oppervlak werd verschoven met stappen van 0, 050 mm over een afstand van 10 mm. Naast de prestaties van de laser interferometer werd de optimale XRF/XRD meetafstand bepaald. Toen het toestel werd afgeleverd, werd aangeraden door de firma Assing S.p.A. om de objecten op een afstand tussen de 91, 8 mm en 92, 2 mm te plaatsen. Uit de eerste metingen op deze afstand bleek echter dat dit niet de optimale afstand kon zijn. Om een nieuwe optimale afstand te definiëren werd de NIST SRM 660b diffractie standaard gebruikt. Deze standaard werd op verschillende afstanden van de Surface Monitor gemeten. Telkens werd het aantal tellen van de meest intense diffractie piek genoteerd en de afwijking van deze piek ten opzichte van de gecertificeerde 2θ-waarde, i.e. 30,385°. Volgende parameters werden gebruikt voor de XRD meting: de meting starte bij 2θi van 15° tot een 2θf van 90° met stappen van 0,2°, en elke stap werd gedurende 5 s LT (live time) gemeten. Twee pinholes van 1 mm werden gebruikt bij de X-stralen buis en twee verticale slits van 1 mm werden gebruikt bij de detector. Verder werd ook nog de bundelgrootte van de X-straal bepaald. Om de bundelgrootte te 74

bepalen, werden dezelfde omstandigheden gebruikt als tijdens een XRF/XRD meting. De X-stralen buis bewoog van 12° tot en met 45° met stappen van 0,1° en een meettijd van 5 s LT/stap. Via deze opstelling werden twee metingen uitgevoerd. Voor de eerste meting werden twee pinholes met een diameter van 1 mm gebruikt en voor de tweede meting werden twee pinholes van 1, 5 mm gebruikt. De bundelgrootte werd bepaald met behulp van een zelf ontwikkelende GAFChromic (type: RQTA2) film, welke op een afstand van 95 mm werd geplaatst. De volgende stap in de karakterisatie van de Surface Monitor was het bepalen van de minimum detectie limieten (MDL). De MDL is de concentratie van een bepaald element die nodig is om de piek van dit element te onderscheiden van de achtergrond. De MDL werden bepaald met de NIST SRM 1412 glas standaard en met behulp van volgende formule: p 3 NB,i CM DL,i = × Ci (1) NN,i

®

met CM DL,i de minimum detectie limiet voor element i, NB,i het aantal tellen van de achtergrond, NN,i het netto aantal tellen van element i en Ci de concentratie van element i in de standaard. De MDL voor de Surface Monitor werden bepaald door gebruik te maken van drie verschillende collimator combinaties. Twee verticale slits (detector) van 2 mm, 1, 5 mm en 1 mm werden gecombineerd met respectievelijk twee pinholes (X-stralen buis) van 2 mm, 1, 5 mm en 1 mm. Elke combinatie werd gedurende 1000 s LT gemeten op een afstand van 95, 0 mm. Ter vergelijking werden de MDL nog bepaald voor volgende state of the art toestellen: EDAX Eagle III micro-XRF spectrometer en de Olympus Innov-X DELTA handheld XRF analyzer. De EDAX Eagle III is een laboratorium instrument bestaande uit een rhodium anode X-stralen buis en polycapillaire optica. Het tweede toestel bestaat ook uit een rhodium anode X-stralen buis met een bundelgrootte van 5 mm, dit toestel is geoptimaliseerd om op verplaatsing te gebruiken. De NIST SRM 1412 werd met de EDAX Eagle III op vier verschillende punten gemeten voor 250 s/punt en dit zowel onder vacu¨ um als in lucht. De vier spectra werden telkens gecombineerd zodat een somspectrum van 1000 s werd bekomen om hieruit dan de MDL te berekenen. De NIST SRM 1412 werd met de handheld XRF analyzer in totaal 300 s gemeten en de resultaten werden herleid naar een meting van 1000 s om te kunnen vergelijken met de resultaten van de Surface Monitor en de EDAX Eagle III. Tenslotte werd nog de invloed van verschillend collimatoren op een XRD spectrum nagegaan. Deze informatie is van belang tijdens metingen op locatie, bv. in een museum. Tijdens deze metingen wordt de gebruiker geconfronteerd met een zeker tijdsdruk, daarom is het nuttig om de invloed van de verschillende pinholes en slits op het XRD spectrum na te gaan en hoe de kwaliteit kan verbeterd worden door andere combinaties te gebruiken. Het effect van de collimatoren werd getest met behulp van de NIST SRM 660b diffractie standaard. De standaard werd telkens op een afstand van 94, 5 mm geplaatst en de meting startte bij een hoek van 15° 2θ tot 90° 2θ met stappen van 0,2° en een meettijd van 5 s LT/stap. In totaal werden twaalf collimator combinaties getest. Naast de karakterisatie van de Surface Monitor werd nog een meting uitgevoerd op het schilderij de “Dulle Griet” in het museum Mayer van den Bergh in Antwerpen (België). Tijdens deze meting werden volgende parameters toegepast: 2θi 40°, 2θf 90°, stapgrootte 0,2°, 5 s LT/step, twee 1, 5 mm pinholes werden gebruikt voor de X-stralen buis en twee 0, 5 mm slits werden gebruikt voor de detector. De meting werd uitgevoerd door het toestel op een afstand van 94, 8 mm van het schilderij te plaatsen. Resultaten en Bespreking Vooraleer het draagbaar XRF/XRD instrument kan gebruikt worden voor de karakterisatie van kunstobjecten, dienen enkele vragen in verband met de prestaties van het instrument te worden opgelost. Dit houdt in: het bepalen van de precisie van de laser interferometer, bepalen van de optimale XRF/XRD meetafstand, bepalen van de bundelgrootte en de positie van de X-straal ten opzichte van de laser interferometer, bepalen van minimum detectie limieten voor 75

XRF, en tenslotte het bepalen van de invloed van collimatoren op XRD spectra. Verder wordt ook nog het resultaat van de Surface Monitor’s eerste applicatie gegeven. Karakterisatie van de Surface Monitor Voor het bepalen van de precisie van de laser interferometer, startte de meting bij een afstand van ongeveer 87 mm tot ongeveer 97 mm. Dit interval werd gekozen omdat verwacht werd dat de optimale afstand voor XRF en XRD metingen hiertussen zou liggen. De output van de laser interferometer werd geregistreerd in functie van de verplaatsing en er werd een trendlijn door deze waarden getrokken. Uit het verschil tussen de gemeten waarden en de trendlijn werd een histogram opgesteld om de verdeling van de waarden beter te visualiseren. Het resultaat wordt getoond in Figuur 3. Aan de hand van het histogram is te zien dat de verdeling Gaussiaans is, wat het mogelijk maakt om een standaardafwijking te berekenen. Naast de standaardafwijking werden ook nog het gemiddelde, de scheefheid en de kurtosis berekend. De resultaten zijn samengevat in Tabel 1. Uit Tabel 1 kan men besluiten dat de afstanden die met behulp van de laser interferometer gemeten worden wat verschoven zijn naar lagere waarden (negative scheefheid) en dat 95 % van de meetwaarden in een interval van ±0, 11 mm liggen.

Figuur 3: Histogram met cumulatieve frequentiecurve.

Tabel 1: Tabel met gemiddelde, standaardafwijking, scheefheid en kurtosis. Deze waarden werden berekend aan de hand van het histogram in Figuur 3. Gemiddelde (mm) Standaardafwijking (mm) Scheefheid Kurtosis

0,00036 0,054 -0,49 0,14

De resultaten met betrekking tot de optimale XRF/XRD meetafstand zijn samengevat in Tabel 2. Uit deze tabel blijkt dat als de afstand tussen de Surface Monitor en de diffractie standaard tussen de 94, 5 mm en 95, 0 mm ligt, de beste resultaten worden bekomen, i.e. een kleine afwijking ten opzichte van de gecertificeerde waarde en een hoog aantal tellen. Nu de optimale XRF/XRD meetafstand bekend is, werd de bundelgrootte op deze afstand bepaald. Tabel 3 en Figuur 4 tonen de resultaten van dit experiment. Uit Figuur 4 is duidelijk dat het punt van de laser interferometer zich in het midden van de bundel bevindt en dat een bundelgrootte van 16 mm bij 3 mm en een bundelgrootte van 21 mm bij 4 mm werden bekomen voor respectievelijk twee pinholes van 1 mm en twee pinholes van 1, 5 mm. 76

Tabel 2: Tabel met het aantal tellen, de gemeten 2θ-hoek en de afwijking ervan ten opzichte van de gecertificeerde waarde. Bij 93 mm en 98 mm was het onmogelijk om de diffractie piek te onderscheiden van de achtergrond. Afstand (mm) 93,0 94,0 94,5 95,0 96,0 97,0 98,0

Tellen 185 234 214 205 43 -

2θ (°) 32,0 30,8 30,0 31,0 29,2 -

Afwijking (°) 1,615 0,410 -0,385 0,610 -1,190 -

Tabel 3: Tabel met bundelgroottes, bepaald op een afstand van 95 mm door gebruik te maken van twee verschillende pinhole diameters. Pinholes 2 x 1 mm 2 x 1, 5 mm

Breedte (mm) 16 21

(a) 2 x 1 mm pinholes

Hoogte (mm) 3 4

(b) 2 x 1, 5 mm pinholes

®

Figuur 4: Bundelgroottes bepaald op een afstand van 95 mm door gebruik te maken van een GAFChromic film.

De resultaten voor het bepalen van de minimum detectie limieten (MDL) voor XRF worden in Figuur 5 getoond. De MDL voor de Surface Monitor liggen tussen de 100 ppm en 3000 ppm. Minimum detectie limieten werden ook bepaald voor twee state of the art toestellen: EDAX Eagle III micro-XRF spectrometer en de Olympus Innov-X DELTA handheld XRF analyzer. De MDL voor deze twee toestellen worden in Figuur 6 weergegeven. De MDL voor deze toestellen zijn minstens een grootteorde lager in vergelijking met de Surface Monitor. Het verschil tussen de Surface Monitor en de Olympus Innov-X DELTA handheld XRF analyzer ligt hem in het feit dat de handheld XRF analyzer tot tegen het oppervlak van het staal kan worden gebracht. Bij de EDAX Eagle III micro-XRF spectrometer werden de instrumentele parameters geoptimaliseerd om optimale resultaten te bekomen.

Figuur 5: Minimum detectie limieten (MDL) in functie van het element voor de Surface Monitor, bepaald met de NIST SRM 1412 glas standaard. (1000 s LT; Cu X-stralen buis; 28 kV; 250 µA; θtube = θdetector = 45°)

77

Figuur 6: Minimum detectie limieten (MDL) in functie van het element voor de EDAX Eagle III micro-XRF spectrometer (in lucht en onder vacu¨ um), en voor de Innov-X DELTA handheld XRF analyzer, bepaald met de NIST SRM 1412 glas standaard. (Set-up EDAX Eagle III: 1000 s LT; Rh X-stralen buis; 40 kV; lucht: 350 µA; vacu¨ um: 220 µA; 300 µm bundel/Set-up handheld XRF analyzer: 1000 s LT; Rh X-stralen buis; 40 kV; 79 µA; 5 mm bundel diameter)

Tenslotte werd nog de invloed van verschillende collimatoren op een XRD spectrum nagegaan, de resultaten werden samengevat in Tabellen 4 en 5. Tabel 4 toont hoe de intensiteit van de XRD pieken wijzigt als een andere collimator combinatie gekozen wordt. De wijziging in intensiteit is relatief ten opzichte van de kleinste collimator combinatie. Tabel 5 toont hoe de breedte of de “full width at half maximum” (FWHM) van de meest intense XRD piek van de diffractie standaard gaat veranderen als de collimatoren gewijzigd worden. Tabel 4: Deze tabel toont hoe de intensiteit (I) van de XRD pieken veranderd relatief ten opzichte van de kleinste collimator combinatie.

X-ray tube pinhole size (mm)

1 1,5 2

Detector slit size (mm) 0.5 1 1.5 2 I 2xI 3xI 3,5 x I 2xI 5 x I 7,6 x I 9xI 2,5 x I 6,3 x I 9,5 x I 11,8 x I

Tabel 5: Deze tabel toont hoe de breedte (FWHM) van de meest intense diffractie piek veranderd in functie van de gebruikte collimatoren. Pinhole size X-ray tube (mm) 1

1,5

2

Slit size detector (mm) 0,5 1 1,5 2 0,5 1 1,5 2 0,5 1 1,5 2

FWHM (˚ A) 0,08 0,08 0,13 0,15 0,09 0,11 0,12 0,13 0,09 0,14 0,18 0,14

De resultaten van Tabellen 4 en 5 kunnen nu worden gecombineerd met de resultaten uit de vorige paragrafen, wat leidt tot de volgende conclusies. Als een groot homogeen oppervlak moet worden geanalyseerd met de Surface Monitor is het aan te raden grote pinholes en slits te gebruiken. Hierdoor kan de meettijd verkort worden tot 1 s LT/stap en is de intensiteit van de XRD pieken nog steeds dubbel zo groot in vergelijking met de kleinste collimatoren met een meettijd van 5 s LT/stap. Nadeel hiervan is dat de breedte van de XRD pieken zal toenemen. In het andere geval, als een klein oppervlak moet worden geanalyseerd, is het aan te raden de kleine pinholes te gebruiken. Hierdoor kan een oppervlakte van 16 mm bij 3 mm worden geanalyseerd (toestel op een afstand van 95 mm plaatsen en X-stralen buis laten bewegen tussen 12° en 45°). Afhankelijk van de tijdsdruk kan men dan kiezen tussen brede of smalle slits. Voor snelle metingen zijn de brede slits te verkiezen, deze keuze resulteert in een hogere intensiteit van de XRD pieken maar ook voor een slechtere resolutie. Daarentegen zorgen smalle slits voor 78

een betere resolutie in het XRD spectrum maar de meettijd zal voldoende hoog moeten liggen om een duidelijk XRD spectrum te bekomen. In het volgende deel wordt de eerste toepassing van de Surface Monitor beschreven, nl. een meting op het schilderij de “Dulle Griet” van Pieter Bruegel de Oude (1561-1562) in het museum Mayer van den Bergh in Antwerpen, België. Toepassingen van de Surface Monitor De resultaten zijn weergegeven in Figuur 7. Uit het XRF spectrum is te zien dat volgende elementen aanwezig zijn: K, Ca, Mn, Fe, Co en Pb. Uit het XRD spectrum blijkt dat CaCO3 aanwezig is, dit is waarschijnlijk afkomstig van de preparatielaag van het schilderij. Uit de resultaten van de meting is het niet mogelijk om met voldoende zekerheid de aanwezige pigmenten te bepalen en dit door de complexiteit van het schilderij. Het schilderij bestaat uit verschillende lagen: een vernislaag met daaronder verschillende lagen van pigmenten. Door de aanwezigheid van deze lagen is het moeilijk om de verschillende componenten van elkaar te onderscheiden in het XRD spectrum. De bundelgrootte kan ook een probleem vormen omdat aanliggende gebieden, bestaande uit andere pigmenten, ook worden bestraald. Voor verdere metingen is het belangrijk om één of twee punten te meten gedurende een lange tijd, zodat minstens één resultaat van hoge kwaliteit kan worden bekomen.

(a) XRF spectrum

(b) XRD spectrum

Figuur 7: Resultaten van de karakterisatie van de “Dulle Griet”.

Samenvatting en Conclusies In 2011 werd door de Vakgroep Analytische Chemie van de Gentse Universiteit de Surface Monitor aangekocht van de Italiaanse firma Assing S.p.A. Na de aankoop rezen er vragen met betrekking tot de prestaties van het draagbaar XRF/XRD instrument. Om een idee te hebben over de mogelijkheden van het instrument, werden in dit werk de volgende aspecten onderzocht: de betrouwbaarheid van de laser interferometer, bepalen van de optimale XRF/XRD meetafstand, bepalen van de bundelgrootte en de positie van de bundel ten opzichte van de laser interferometer, minimum detectie limieten voor XRF, en de invloed van de collimatoren op de XRD spectra. Het eerste onderdeel van de Surface Monitor dat werd getest, was de laser interferometer omdat dit onderdeel van groot belang is tijdens het positioneren van de Surface Monitor voor het te analyseren object. Uit de test bleek dat de precisie van de laser interferometer gelijk is aan 0, 11 mm wat voldoende is voor verdere toepassingen. Een tweede aandachtspunt was de optimale XRF/XRD meetafstand. De firma Assing S.p.A. raadde aan om het toestel tussen de 91, 8 mm en 92, 2 mm van het object te plaatsen. Uit de eerste tests bleek dat deze afstand niet de optimale afstand kon zijn, dus werd er een nieuwe optimale afstand gedefinieerd aan de hand van metingen met de LaB6 diffractie standaard. De beste resultaten werden bekomen tussen de 94, 5 mm en 95, 0 mm. De volgende stap in de karakterisatie van de Surface Monitor was het bepalen van de bundelgrootte en de positie van de bundel ten opzichte van de laser interferometer. Uit dit onderzoek 79

bleek dat de laser interferometer het midden aanduidt van de X-stralen bundel. Nadat de positie van de bundel op het te onderzoeken object was gekend, werden de minimum detectie limieten voor XRF bepaald. Voor de Surface Monitor liggen deze tussen de 100 ppm en 3000 ppm. De laatste fase in de karakterisatie van het draagbare XRF/XRD instrument ging over de invloed van collimatoren op de XRD spectra en hoe de kwaliteit van de XRD spectra kan worden verbeterd. Tenslotte werd de Surface Monitor ingezet tijdens de analyse van de Dulle Griet. Pigmenten konden niet met zekerheid worden bepaald door de complexiteit van het schilderij (multilaag van vernis en pigmenten). Voor verder metingen wordt aangeraden om één punt voor lange tijd te meten zodat een duidelijk XRD spectrum wordt bekomen. Erkenning Graag wil ik Prof. dr. Laszlo Vincze bedanken om mij de kans te geven om met dit nieuwe toestel te werken. Verder wil ik nog dr. Bart Vekemans bedanken voor de begeleiding en de hulp met de XRF dataverwerkingssoftware (AXIL en MICROXRF2). Ook wil ik dr. Ettore Di Masso en Andrea Bianco van Assing S.p.A. bedanken voor de training met de Surface Monitor en de bijstand toen het toestel defect was. Verder wil ik Prof. dr. Maximiliaan Martens, Prof. dr. Peter Vandenabeele en dr. Claire Baisier bedanken om de Surface Monitor te betrekken in de karakterisatie van de Dulle Griet. Tenslotte wil ik Prof. dr. Peter Vandenabeele bedanken voor het gebruik van de Olympus Innov-X DELTA handheld XRF analyzer en Lien Van de Voorde voor de assistentie tijdens de metingen met dit toestel. Referenties [1] D. Blake & C. Bryson (1992). Design of an x-ray diffraction/x-ray fluorescence instrument for planetary applications. Lunar and Planetary Institute, 13:117–118. [2] D. Blake, C. Bryson & F. Freund (1993). X-ray diffraction apparatus, U.S. Pattent No. 5,491,738. [3] S. Cornaby, A. Reyes-Mena, H. K. Pew, P. W. Moody, T. Hughes, A. Stradling, D. C. Turner & L. V. Knight (2001). An XRD/XRF instrument for the microanalysis of rocks and minerals. Measurement Science and Technology, 12(6):676–683. [4] M. Uda (2004). In situ characterization of ancient plaster and pigments on tomb walls in Egypt using energy dispersive X-ray diffraction and fluorescence. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 226(1-2):75–82. [5] A. Gianoncelli, J. Castaing, L. Ortega, E. Dooryhee, J. Salomon, P. Walter, J. Hodeau & P. Bordet (2008). A portable instrument for in situ determination of the chemical and phase compositions of cultural heritage objects. X-Ray Spectrometry, 37(4):418–423. [6] Assing S.p.A. (2011). Adress: Via E. http://www.assing-group.it/index eng.html.

Amaldi

14

-

Monterotondo

(Italy).

URL

[7] A. Pifferi, G. Campi, C. Giacovazzo & E. Gobbi (2009). A new portable XRD/XRF instrument for nondestructive analysis. Croatica Chemica Acta, 82(2):449–454. [8] Amptek Inc. (2012). Adress: De Angelo Drive 14 - Bedford (U.S.A.). URL http://www.amptek.com. [9] B. Vekemans, K. Janssens, L. Vincze, F. Adams & P. Van Espen (1994). Analysis of X-ray spectra by iterative least squares (AXIL): New developments. X-Ray Spectrometry, 23(6):278–285. [10] International Center for Diffraction Data (2012). Adress: Campus Boulevard 12 - Newtown Square (U.S.A.). URL http://www.icdd.com/.

80

XRD Instrument

Recommend Documents