Chapter II – Instrumental developments
61
___________________________________________________________________________
CHAP II – INSTRUMENTAL DEVELOPMENT
L'appareillage développé est décrit dans ce chapitre. Il permet l'étude combinée de microindentation et de microdiffraction des rayons X en temps réel. Un microindenteur de la société Anton PAAR Gmbh est fixé verticalement sur une tourelle motorisée de microscope optique classique Olympus. Il peut être contrôlé à partir d'un boîtier de commande ou interfacé avec un ordinateur. Divers objectifs sont aussi rattachés à cette tourelle, ce qui permet de combiner l'axe optique et l'axe d'indentation. Un illuminateur coaxial et une caméra CCD permettent de visualiser l'échantillon et de définir un point d'intérêt à analyser. Le porte-échantillon est constitué d'un diamant d'environ 4 mm de diamètre qui n'a qu'une faible absorption aux longueur d'ondes utilisées. L'échantillon généralement collé sur sa surface est déplacé dans les trois directions de l'espace par trois moteurs de la société Micos avec une précision d'environ 1 µm. Tous ces éléments sont fixés sur un même cadre rigide qui peut donc être utilisé indépendamment en laboratoire ou monté sur le système de balayage de la ligne de microfocus de l'ESRF. Dans le premier cas, un calibrage optique est nécessaire pour localiser la pointe de l'indenteur. Dans le deuxième, le porte-échantillon est remplacé par une pointe au bout de laquelle est fixé un microcristal de GaN qui fluoresce dans le visible lorsqu'il est exposé à un rayonnement X de 13 KeV. Cela permet une calibration optique additionnelle portant sur la position du faisceau. L'échantillon peut alors être simultanément indenté et balayé dans le faisceau afin d'obtenir une information topologique de l'évolution structurale durant la déformation. Il est important de noter que l'appareillage décrit ici est le fruit d'une série d'essais et de modifications parfois importantes à partir d'une configuration de base. Diverses améliorations sont en outre décrites à la fin de ce chapitre.
Chapter II – Instrumental developments
62
___________________________________________________________________________ The microindenter set-up which was developed in this work to perform in-situ (real time) studies of microindentation (see section I.1) is described in this chapter. For this purpose, it was designed to be mounted on the existing X-ray scanning set-up described in section I.4.4. The present state of the device is the result of a series of developments and improvements starting from an initial configuration which will be described in a first part. Details of operation and calibrations will be given in a second part and further possible developments will be mentioned in a final part.
II.1.
FIRST DESIGN OF THE INDENTER SET-UP
II.1.1.
CHARACTERISTICS OF THE ANTON PAAR INDENTER
Hardness test instruments operating at low loads are delicate instruments and require a high degree of accuracy, both in the application of the load and in the measuring of the resulting indentation produced. The formula given for Vickers Hardness (equation (2), p.16) involves two main parameters: the load and the size of the diagonals of the impression. The control of the load applied is thus of extreme importance for a correct measurement. Furthermore it is essential for precise characterisation that the indentation gradient and dwell time be well known at all times. Those issues obviously become even more critical on applying microtesting methods to polymers and a further step is taken when considering realtime studies. The indenter (fig.II.1) chosen for our purposes is commercialised by Anton Paar GmbH under the trade name MHT-10 (MicroHardness Tester with video measuring system). A Vickers diamond pyramid with angles between faces of 136o ±15' and less than 0.25µm offset of diamond tip (relative lateral position of diamond tip with respect to ideal geometry) was used in our studies. Alternatively, a Knoop diamond tip is also available. The sensor holding the diamond indenter tip is controlled via a separate unit that can be interfaced with a PC. Its dimensions and weight allow mounting on the rotating turret of a microscope. The standard microscopy set-up allows using a variety of objectives and provides proper illumination devices for viewing and measuring of the indentation impression.
Chapter II – Instrumental developments
63
___________________________________________________________________________
Fig.II.1 : Anton-Paar microindenter sensor mounted on a microscope turet [1] Forces applied on the diamond tip are in the range of 5-4000mN with a resolution of 1mN and the force gradient can be computed between 1-999N.s-1 with a dwell (holding) time of 0-9999s. A feed back loop allows continuous control of all parameters during the measurement. It should also be noted that the diagonals of the impression left by the indenter are usually measured by optical means, which requires very accurate illumination. Although measuring hardness was not our primary goal throughout our studies, this matter will be dealt with in the subsequent description of the indenter set-up, which was developed for this work.
II.1.2.
ANALYSIS OF REQUIREMENTS
II.1.2.1. MINIMUM OPERATIONS
In order to match sample position, indenter position and X-ray beam position the following operations are required : - selection of a region of interest (ROI) of the sample under the microscope (in-micro position) - position of the ROI under the diamond tip of the indenter (in-indenter position) - position of ROI in the X-ray beam (in-beam position)
Chapter II – Instrumental developments
64
___________________________________________________________________________ The precision of positioning depends on the specific set-up used and is in general not better than 1-2 µm as will be discussed in more detail below.
II.1.2.2. SAMPLE VIEWING
Fig.II.2 shows an image of a number of indentations in an amorphous polyethyleneterephtalate (PET; trade name: Kapton) film. From other independent measures, the hardness was estimated to Hv = 155 MPa using equation (2) (p.16). At this magnification only indentations at forces > 200mN can properly be visualized. This shows that considering the size range of the impression left by the indentations, one needs a relatively important tuneable magnification. One solution to achieve this in a consistent way with compactness requirement is to use an optical zoom, which allows different samples (and thus different hardness) to be used.
0.5 N
1N
2N
3N
100µm FigII.2 optical micrograph of indented Kapton under various forces ; force gradient is F/10 mN.s-1and dwell time is 10 s for each indentation
II.1.2.3. SAMPLE POSITIONING
Three sets of xyz translations are required to permit each of the above-mentioned movement as described schematically in fig.II.3-a. In order to reduce these, both indenter and
Chapter II – Instrumental developments
65
___________________________________________________________________________ optical zoom were first to be fixed on a rigid frame at a given height with respect to the sample (fig.II.3-b). In this way, two translations in x and y would be sufficient to displace the sample and position it under either zoom or indenter. This frame is in turn fixed on the XYZ translation stage of the scanning set-up and can therefore be positioned with respect to the Xray beam. Due to steric problems, the zoom could not be fixed on the frame but rather had to be mounted on the scanning set-up as shown in fig.II.3-c a-
bZ
Z
I
I X
z
x
S
x
y
y
S
F
X Z
c-
X Y
M z S
I
F
x
Z y
X Z
X Y
Fig.II.3 : Schematical design of a possible set-up; F : frame, S : sample, I : indenter, Z : optical zoom, M : mirror and X : x-ray; diamonds indicate fixed position; a – three translations are needed to displace the sample under the different elements, b – both zoom and indenter are integrated on the main frame, c – zoom is fixed independently of the frame and sample is visualized using a mirror
Chapter II – Instrumental developments
66
___________________________________________________________________________ II.1.3.
TEST DESIGN
II.1.3.1. EARLY STAGE CONCEPT
Fig.II.4 shows an early stage design, which corresponds to fig.II.3-b later abandoned due to the size of available optical zooms. However, some of its features were used in later developments.
(2)
(1)
(3) (4) (5) (6) X (6)
Z
X Y
Fig.II.4 principle design of system coupling optical zoom to indenter set-up ; translation range is 5mm in x and 40mm in y It consists of a rigid aluminium frame (1) on which are fixed both optical zoom (2) and indenter (3) at a given height and at a fixed distance from one another. The sample (4) is
Chapter II – Instrumental developments
67
___________________________________________________________________________ fixed on an x (5) - y (6) stage which allows positioning, focussing and is further used to displace the sample from optical to indenter position. Finally, the frame is fixed on the scanning stage (6) allowing movements about the beam position.
II.1.3.2. ADAPTING INDEPENDENT OPTICS
The first engineering design [3] shown in fig.II.5 corresponds to the scheme of fig.II.3-c.
(3) (8) (4)
(7)
Z Y X (1)
(5) (6)
Fig.II.5 : test set-up within dependant side optics For convenience, the zoom attached to the main body of the scanning set-up is not represented. The first main difference is the addition of an "internal" (i.e. with respect to
Chapter II – Instrumental developments
68
___________________________________________________________________________ indenter / optics) z-motor (7) to the previous x-y stage (5)(6), which was found to be necessary to accurately position the sample (4) under the indenter (3). Secondly a mirror (8) is fixed on the rigid aluminium frame (1), which must therefore be used in conjunction with the zoom when visualization of the sample is necessary. This is achieved by displacing the whole frame using the scanning set-up XYZ-stage to a position where the mirror matches the zoom X, Y positions. Fig.II.6 shows a picture of the indenter device mounted on the scanning set-up for the first WAXS test and fig.II.7 the mirror (8).
(3)
(7)
(8)
(5) (6)
(1)
Z Y X
Fig.II.6 : indenter design mounted on the scanning set-up viewed in a position close to the CCD recording device; white frame indicates fig.II.7-b
Chapter II – Instrumental developments
69
___________________________________________________________________________
b-
a-
(3) (8)
II.1.1.1. opti
(8)
(4)
Fig.II.7 : a – mirror as viewed close to the zoom position; 3 screws allow fine tuning of mirror angle α with respect to optical path ; b – from behind, note the collimator in the background pointing in sample (4) direction
II.1.4.
DRAWBACKS
II.1.4.1. CALIBRATION PROBLEMS
Three calibrations have to be made in order to determine the absolute position of a – the optical axis b – the diamond of the indenter c – the beam position. Prior to any of these operations, the force sensor of the indenter has to be calibrated as described in [2]. It is important to understand that in this design, the sample is referenced with respect to four objects : the indenter, the mirror, the zoom and the X-ray beam. The first two are fixed on the frame while the latter are fixed on the scanning set-up (fig.II.10). In all the following, translations due to internal movements (i.e. with respect to the frame) of (5), (6), (7) in fig.3, 4, 6 will be noted xyz while those done using the scanning set-up motors (external) will be noted XYZ as in above paragraphs.
Chapter II – Instrumental developments
70
___________________________________________________________________________ OPTICAL CALIBRATION : In first place, the whole set-up is displaced in Y-Z (i.e. perpendicular to the X-ray beam) so that the mirror comes into the optical path of the zoom. Positions of the motors are then recorded as (Xo,Yo,Zo). The material is then brought into focus using the xyz translations and a selected area of interest is defined roughly as the centre of the image. In the case of fibres of diameters smaller than 150µm, due to imperfect symmetry it is often preferable to defocus (i.e. focus on the edges) rather than try to focus on the top. The (xo,yo,zo) position is recorded and a reference mark is taken on the zoom output monitor. It is evident at once that if the mirror is not positioned at 45o from the sample and the zoom, there will be distortions in the image and thus in the visual analysis. A special device shown in fig.II.7-a (above) allowing fine-tuning was therefore created [3] to overcome this difficulty. Nevertheless, this proved to be difficult to use. INDENTER CALIBRATION : Due to possible displacement of the diamond tip with respect to its main body in a range of ± 50 µm [1], fine tuning has to be made. This is done by indenting a standard sample which hardness is known from literature under well defined conditions of force, force gradient and dwell time. As an example, all calibrations were done using a Kapton film of 200 µm thickness under 350 mN, 350 mN.s-1 for 10 s.
zoom / mirror / video
(xo,yo,zo)
(xi,yi,zi)
indenter
40mm 40mm
X Y
supposed indenter position real indenter position (xi,yi’,zi)
(Xo,Yo,Zo) Fig.II.8 principle of indenter calibration
Chapter II – Instrumental developments
71
___________________________________________________________________________ Further calibration was also occasionally achieved using a lead foil [4]. In this operation, the sample is displaced by 40mm in y which corresponds to the distance between the mirror and the indenter. It is then indented and brought back to origin (xo,yo,zo). The position of the centre of the indentation is then recorded as (xi,yi,zi) (fig.II.8). In all the following, time given is the dwell time, that is the period during which the indenter is being held under the same load into the sample. The total time of the indentation must also take into account approach of the tip, indentation gradient and retraction of the tip. Should this dwell time be "long", creep effects would have to be taken into account in the case of polymer materials.
X-RAY CALIBRATION USING ABSORPTION PROPERTIES :
b-
c-
a-
Fig.II.9 : X-ray absorption scans of sharp glass tip; a- rough horizontal scan is followed by b- finer one ; centre of the glass tip is therefore determined using symmetry considerations and c- vertical absorption scan allows to further determine the end of the tip; note that several such operations are needed for accuracy
Chapter II – Instrumental developments
72
___________________________________________________________________________ The sample is replaced by a glass capillary, which has previously been split under tensile stress in a small furnace at a temperature close to fusion leading to a very sharp tip. Focus is achieved on the top of the tip (xo,yo,zt) From (xo,yo,zt,Xo,Yo,Zo), this tip is brought roughly in beam using Y-Z translations exclusively. Lateral and vertical absorption scans of the glass tip (fig.II.9) allow to determine the absolute position of the beam (xo,yo,zt, Xo,Yb,Zb) with a high degree of accuracy (within ± 2 µm) From above considerations, the minimum operations to do in-situ measurements are the following (fig.II.10) : 1- (x,y,z,X,Y,Z)
(xo,yo,zo,Xo,Yo,Zo)
2- (xo,yo,zo,Xo,Yo,Zo)
(xi,yi',zi,Xo,Yo,Zo) where yi'=yi+40mm
3- (xi,yi',zi,Xo,Yo,Zo)
(xi,yi',zi,Xo,Yf,Zf) where Yf=Yb-yi', Zf=Zb-zi
where the following indices are used : o : optics; i : indenter; b : beam; f : final
SCANNING SETFRAME SET-UP mirror / video monitor
indenter
2
1
3 (xi,yi’,zi)
(Xo,Yo,Zo) Zoom
(Xb,Yb,Zb)
X Y
Fig.II.10 : minimum operations in order to perform in-situ measurements
Chapter II – Instrumental developments
73
___________________________________________________________________________ Step 1- allows to select the sample under the zoom, step 2- zone of interest is then brought to indenter position; step 3- the whole frame is brought in beam position.
II.1.4.2. LOSS OF SPATIAL RESOLUTION DUE TO BACKLASH
As seen in previous discussion, the sample is translated over a large position range throughout the calibration. Step 2 (fig.II.10) requires 40mm displacements, which is already relatively high for conventional motors considering the precision required in final positioning. Backlash problems can occur therefore decreasing the degree of spatial resolution, which must remain within 2µm for our purposes. As a general rule of thumb, the loss of precision is directly proportional to the number of movements of the sample. In order to insure as high a reproducibility as possible, one would therefore need to reduce the number of translations described in the above section.
II.2.
FINAL DESIGN
II.2.1.
INTRODUCTION OF ROTATION SYMMETRY
After the first tests of the initial design described in the above section, it was found necessary
to reduce the large translations along the y axis (fig.II.8,10) during indenter
calibration. Therefore, to reduce uncertainties, an OLYMPUS rotating turret (11) of the type used in optical microscopes was added to the set-up (fig.II.11). This considerably simplifies the optical and indenter calibration as different objectives (10) and indenter (3), which are all screwed on the turret, (11) are successively rotated onto the optical path as shown in fig.11-b.
Chapter II – Instrumental developments
74
___________________________________________________________________________
a-
(2)
(11)
(9)
(1)
(6) (7) (5)
b(10)
c(3)
(4)
Fig.II.11 : actual set-up, a – global view and b – c – details of sample environment
All image distortions due to the use of the mirror in previous set-up are thus removed and the sample is not moved throughout the optical and indenter calibration process (step1 and 2 in fig.II.10). Furthermore, the frame can be brought directly close to beam position, which strongly favours stability of the design.
Chapter II – Instrumental developments
75
___________________________________________________________________________ II.2.2.
REDUCING CALIBRATION UNCERTAINTIES
II.2.2.1. COMBINING OPTICAL AND INDENTER AXIS
Optical calibration is now done by direct visualizing of the sample through the CCD camera (2) fixed above the turret along the optical axis (fig.II.11-a). Focus is achieved using internal x (5), y (6), z (7) motors. Once a region of interest is selected, a reference mark is taken on the control screen. Position of the motors are recorded as (xo,yo,zo,Xo,Yo,Zo). The indenter (3) is then rotated through remote control of the turret in place of the objectives (10) as shown in fig.II.12 below.
indenter
supposed indenter position real indenter position
2 CCD / objectives
(xo,yo,zo) turret rotation
∆P (xi,yi,zi) (Xo,Yo,Zo)
Fig.II.12 : optical & indenter calibration
1
Chapter II – Instrumental developments
76
___________________________________________________________________________ The sample is indented and the position of the centre of indentation is recorded as (xi,yi,zi,Xo,Yo,Zo). Following this, all selected region of interests will have to be translated by ∆P = (xi-xo, yi-yo, zi-zo, 0, 0, 0). It is of importance to note that only internal motions are used within this step and only take place in the order of ± 2 mm as opposed to the large ∆y = 40 mm from previous design. This also justifies the use of more compact and precise motors, which greatly reduced backlash problems.
II.2.2.2. X-RAY CALIBRATION USING A GaN CRYSTAL
The X-ray calibration described previously (fig.II.9) is precise enough, provided a sufficient number of vertical and horizontal absorption scans are made. This is however time consuming and requires a lot of attention due to the difficulty to visualize a 2 µm end tip. In order to improve this procedure, a GaN crystal was glued to the end of a sharp glass tip (fig.II.13). Such a crystal fluorescing in the visible range is therefore a powerful tool for final calibration as the observation is now done directly by visualizing the fluorescing crystal instead of analysing the absorption scans.
15µmGaN
Xtal
Fig.II.13 : GaN crystal mounted on sharp glass tip This step requires replacing the sample holder with the glass tip, and then focusing on the GaN 15µm crystal. Rough horizontal and vertical scans are used in a first step to define an approximate position of the crystal and finer scans allow visualizing the fluorescing crystal.
Chapter II – Instrumental developments
77
___________________________________________________________________________ II.2.3.
OTHER IMPROVEMENTS
II.2.3.1. X-RAY TRANSPARENT SAMPLE HOLDER
The WAXD pattern shown in fig.II.14 below was taken in a first test using a polyethyleneterephthalate (PET) fibre. It can be seen at once that the indenter and the brass sample holder shadow most of the region of interest for the analysis. This showed that in order to monitor any variation of the fibre crystalline structure during indentation it was necessary to 1 – remove the cover of the indenter (fig.II.16) and 2 – replace the brass sample holder by a material transparent to X-ray.
Indenter
brass sample holder
Fig.II.14 : WAXD pattern of a PET fibre showing shadows of indenter and sample-holder It was therefore decided to use a diamond, which has a very low X-ray absorption coefficient at the wavelengths generally used for polymer analysis (0.9755 Å) glued on a brass pin (fig.II.15).
Chapter II – Instrumental developments
78
___________________________________________________________________________
4mm
2mm 30µm PA6 Fig.II.15 : Diamond sample holder viewed a – sideways b – from above Furthermore, diamond being one of the hardest known materials, it also proved useful during the indentation process of relatively hard material as the deformation attributed to the sample holder can be neglected in the case of diamond but not in the case of brass. Fig.II.16 shows the set-up as currently used prepared for X-ray calibration.
Fig.II.16 : Sample environment as defined for X-ray calibration
Chapter II – Instrumental developments
79
___________________________________________________________________________ II.2.3.2. OPTICAL IMPROVEMENTS
Illumination was found to be a critical point due to the necessity to clearly identify low contrast indentations of small sizes. This is of importance as all calibration steps rely at one point on sample or calibrant viewing. Also, selecting a precise zone of interest of a particular sample requires proper illumination. Therefore, instead of using an external light source, a tuneable system of coaxial illumination (Olympus U-KMAS) was incorporated to the set-up. Also, as will be described in following chapters, in order to perform high quality WAXS and SAXS experiments, one needs to bring the sample as close as possible to the collimating device or to a guard aperture, which allows to reduce the background. However, usual microscope objectives such as MPL 20x (Olympus) allow only a working distance of 1.3mm, which is therefore a limiting factor. It was therefore preferred to use long working distance objectives such as LMPLFL 20x (Olympus) leaving a 12 mm working distance (fig.II.16).
II.3.
II.3.1.
POSSIBLE FUTURE TECHNICAL DEVELOPMENTS
HEATING / COOLING STAGE
As mentioned in section I.2 and I.3, the precise control of the temperature of the experiment is particularly important in the case of polymeric materials due to differences in mechanical behaviour above and below glass transition temperature Tg. Up to now, all experiments have been done in experimental hutch at room temperature (25 ± 2 oC). It could, however, be thought of a number of experiments making use of a heating / cooling device which would allow to control precisely the sample temperature. One could for instance run an experiment in a temperature range including Tg while indenting the sample.
Chapter II – Instrumental developments
80
___________________________________________________________________________ Various possible ways to achieve this could be to heat the diamond sample support, by fixing a small resistance at the base, or cool the sample in a Nitrogen flow. In both cases, one must found a way to accurately measure the temperature of the indented zone of the sample.
II.3.2.
ROTATING STAGE
Texture X-ray analysis were done using the microdiffractometer described in section I.4. In such experiments, the sample is fixed on a microgoniometer and subsequently rotated in the X-ray microbeam. Diffraction patterns can therefore be taken at various angles. The situation would be much more complicated if one was to consider real-time experiments making use of a rotation axis linked to the sample using the present set-up. One would in fact have to rotate the whole set-up around a fixed rotation axis as shown in fig.II.17.
Fig.II.17 principle of real-time indentation experiment using a rotation axis The main difficulty in doing so aside this of the weight of the set-up that the rotation axis would have to withstand is that the sample would have to be aligned on the optical axis, which is a challenging problem. However difficult, this could be considered in the future.
Chapter II – Instrumental developments
81
___________________________________________________________________________ II.3.3.
OTHER INDENTER TIPS
It is known that the force field developed under an indenter when pressed on the surface of a sample is specific to its shape [5]. In order to study the influence of the geometry of the indenter on the deformation, one can therefore replace the Vickers diamond pyramid following the procedure described in [2]. A Knoop diamond pyramid (see in section I.1), can be adapted to the Anton Paar indenter. Other indenter geometries such as steel ball or cone could also be developed for specific experiments.
II.3.4.
NANO-INDENTER
From the different experiments that will be described in chap.IV, it is clear that a major step would be taken in implementing nanoindentation techniques (see section I.I). This implies indentation at low loads, typically bellow 5mN in the case of polymers and therefore specific instrumentation is needed. In this field it is customary to use the Berkovitch triangular pyramidal indenter, since it provides a sharp point and a well-defined geometry [6]. However, it is usually difficult to resolve the impression left by the indenter (typically less than 1 µm2) by conventional visible light microscopy techniques. A possible alternative would lie in the use of a compact AFM instrument allowing sample visualization with a builtin nanoindentation Berkovitch module. Several features would, therefore, change considerably in the actual set-up as the optical part would no longer be necessary. As a direct consequence, the same applies for the rotating turret and objectives. The nanoindenter set-up could therefore be much more compact than the actual design. Nonetheless, several problems still remain : – calibration : one still has to match the position of the indenter tip with this of the beam. – the absolute precision of the motors would have to be well below the sub-µm range. – this type of device is known to be extremely sensitive to vibrations which might require specific beamline developments.
Chapter II – Instrumental developments
82
___________________________________________________________________________
REFERENCES
[1]
Web page : http://www.anton-paar.com/en/_mh/products/mht10a.htm
[2]
MHT-10 Instruction Handbook, Anton Paar GmbH
[3]
Lange
J.,
Aufbau
und
Erprobung
eines
Synchrotronstrahlungs-Messplatz, Diplomarbeit,
Mikrohärteprüfgerätes
an
einem
Fachhochschule Ravensburg-Weingarten
2001 [4]
Gourrier, A., Garcia, MC., Riekel, C. Macromolecules, 35, 8072, 2002
[5]
Tabor, D., The hardness of metals, Clarendon Press, Oxford, 1951
[6]
Tabor, D., Phil. Mag. A, 74, 1207, 1996
