Measuring thin film thickness

Abstract

A method of measuring the thickness of a thin film structure during deposition or etching thereof comprising: illuminating a thin film structure with light, the light including at least a first wavele

Description

FIELD OF THE INVENTION The present disclosure relates to techniques for measuring the thickness of thin film structures during deposition or etching. BACKGROUND In situ monitoring of thickness of thin film structures during plasma assisted etching or deposition can be achieved by monitoring the total light reflected from the thin film structure. In conventional techniques, this may be achieved by counting the interference turning points in the intensity of reflected light at a particular wavelength. However, some processing chambers have a relatively small gap (~2cm) between electrodes or between the substrate and other chamber structures meaning that, when detecting light from the side viewport of a plasma processing tool, the angle of view of the thin film structure from the side is very shallow, typically above the Brewster’s angle for the thin film structure. If viewing polarised light, each polarisation (s or p) has slow moving sinusoidal ripples in intensity. This means that for typical film thicknesses 100-500nm there will be very few, if any, turning points in the trace, leading to difficulty in obtaining accurate thickness measurements for thinner films. Furthermore, polarisers add cost to the measurement hardware, and complexity in setting up the angles of polarisation, so it is desirable to use a system which does not require polarisers. Alternatively, the thickness measurement may be performed using unpolarised light. When detecting unpolarised light, the two polarisation traces are combined producing a more complex trace with far more detail. However, the turning points do not occur at equal intervals. Another problem with counting the turning points as the thickness of the thin film structure changes is that any disturbances in the plasma can produce fluctuations in the intensity trace, which may be incorrectly interpreted as turning points, in turn giving an incorrect thickness measurement. An improved method is therefore required which overcomes the related problems of incorrectly identified turning points and sensitivity to signal noise, as well as the problem that signal strength may reduce over time, for example due to window clouding or deposition on the window. Conventional methods are typically also unsuitable when pulsed or switched multi-frequency plasmas are used during etching or deposition. In such cases, the emission intensity of the plasma alternates between two or more levels, causing turning points to be missed or misinterpreted, a problem requiring resolution. SUMMARY OF INVENTION According to a first aspect of the invention, a method of measuring the thickness of a thin film structure during deposition or etching thereof is provided, the thin film structure being disposed on a substrate mounted on a table inside a processing chamber of a thin film processing tool. The method comprises: illuminating the thin film structure with light from a light source , the light including at least a first wavelength; detecting light of the first wavelength reflected by the thin film structure at a detection angle and generating a first measured signal which represents the intensity of the detected first wavelength of light versus time as the deposition or etching continues; at each of one or more time instances during the deposition or etching: comparing at least a portion of the first measured signal so far generated against at least a part of a first reference signal stored in a memory, the first reference signal representing the expected intensity of the first wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the first measured signal or of the at least a part of the first reference signal so as to fit the at least a portion of the first measured signal to some or all of the at least a part of the first reference signal; and identifying the best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal, and identifying the thickness of the thin film structure at the respective time instance based on the best fit to thereby give a first thickness measurement. The present invention overcomes the problems that can be caused by plasma instabilities when counting the turning points in a measured trace by evaluating the overall shape of (at least a portion of) the measured trace rather than focussing solely on the turning points. For the same reasons, the present invention is also less sensitive to signal noise and any reduction in signal strength over time, for example due to window clouding or deposition on the viewport of a etching or deposition tool. While the present invention is suitable for use when detecting polarised light, it may also be used to identify the thickness of a thin film structure using unpolarised light. This is advantageous as it obviates the need for any polarisers to be used to detect the reflected light, reducing the cost of the measurement hardware as well as the complexity of setting up this equipment. Identifying the best fit may comprise identifying one or more scaling factors which achieve the best fit between the at least a portion of the first or second measured signal and the at least a part of the first reference signal, the identified scaling factors being indicative of the rate of deposition or etching, thereby giving a first rate measurement. In other words, rather than directly identifying the thickness of the thin film structure from the best fit, the best fit is used to identify a measurement of the rate of etching or deposition. Identifying the thickness of the thin film structure at the respective time instance based on the best fit preferably comprises identifying the thickness of the thin film structure at the respective time instance based on the first rate measurement and the respective time instance. More preferably, during deposition, identifying the thickness of the thin film structure at the respective time instance based on the best fit comprises multiplying the first rate measurement and the deposition time at the respective time instance. During etching, identifying the thickness of the thin film structure at the respective time instance based on the best fit preferably comprises multiplying the first rate measurement and the deposition time at the respective time instance and subtracting this value from the initial thickness of the thin film structure. More preferably, in order to account for variations in the initial thickness of the thin film structure, a thickness offset may also be fitted alongside the first rate measurement, such that the initial thickness of the thin film structure is identified by adding the fitted thickness offset, which may be negative, to an expected initial thickness. The method may further comprise estimating a thickness of the thin film structure at a time between consecutive time instances based on the most recent rate measurement and the corresponding thickness measurement. In this way, an estimate of the thickness of the thin film structure may be obtained between time instances. During deposition, this preferably involves multiplying the most recent rate measurement and the deposition time at the current time. It is advantageous to only fit the most recent portion of the measured signal to the reference signal. To this end, the portion of the first measured signal so far generated preferably comprises at least a portion of the signal generated between an earlier time instance of the one or more time instances and the respective time instance. This allows for the measured signal to be fitted to the reference signal even if the rate of deposition or etching changes throughout the process of etching or deposition. However, in some embodiments the portion of the first measured signal so far generated may comprise substantially all of the signal generated between said earlier time instance and the respective time instance. The deposition or etching may comprise alternating between at least a first mode of deposition or etching and a second mode of deposition or etching. For example, a multi-frequency plasma may be used during the deposition or etching. In order to accurately fit the measured signal to the reference signal in such cases, generating the first measured signal may comprise excluding the light detected during the second mode of deposition or etching, thereby enabling fitting of the remaining portions of the measured signal to be fitted to the reference signal. The present invention is particularly advantageous when the detection angle is greater than the Brewster’s angle of the thin film structure. At these detection angles, obtaining an accurate measurement of the thickness of a thin film structure using conventional methods in which turning points in a signal are counted can be difficult, a problem that is overcome by the present invention. Advantageously, the detected light is reflected from a region of the thin film structure which is of uniform thickness. As such, it is preferably to detect light from a narrow range of acceptance angles around the detection angle and the detector used to detect the light of the first wavelength reflected by the thin film structure is therefore configured to receive light lying within 3° of the detection angle, preferably within 2° of the detection angle, and more preferably within 1° of the detection angle. This leads to improvements in the accuracy of the thickness measurements, for example of with an accuracy of ±1% when measuring thin film structures up to 3microns thickness. The method is especially suited to measuring the thickness of a single layer, but may also be used to measure the thickness of a multilayer film structure. The invention may be used with a variety of different etching or deposition techniques, but is advantageously used when the deposition or etching comprises a plasma enhanced deposition process or plasma enhanced etching process. In such cases, the light source preferably comprises the plasma source used in the plasma enhanced deposition process or plasma enhanced etching process, thereby obviating the need for an external light source. The plasma enhanced deposition process or plasma enhanced etching process may comprise one or more of: plasma enhanced chemical vapour deposition, plasma enhanced atomic layer deposition, and plasma enhanced atomic layer etching. The method finds particular utility when the thin film structure comprises a layer which is at least semi-transparent to light of the first wavelength, as conventional methods may not allow for accurate measurement of the thin film structure in such cases. In particular, if said layer is substantially fully transparent to light of the first wavelength, or if said layer is substantially fully transparent to all visible light, then significant improvements in the accuracy of the measured thickness may be achieved. Different techniques may be used to model the first reference signal. For example, the first reference signal may be modelled based on one or more of: the known or expected properties of the thin film structure; the first wavelength; and the detection angle. The first reference signal may also be calculated based on signals measured from light reflected from one or more thin film samples during one or more prior etching or deposition processes. In some cases, the method may comprise stopping the deposition or etching if the measured thickness reaches a predetermined thickness. The light may further include a second wavelength. Light of the second wavelength reflected by the thin film structure may additionally be detected at the detection angle and a second measured signal generated which represents the intensity of the detected second wavelength of light versus time, as the deposition or etching continues; and the method may further comprise: at each of the one or more time instances during the deposition or etching, comparing at least a portion of the second measured signal so far generated against at least a part of a second reference signal stored in a memory, the second reference signal representing the expected intensity of the second wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the second measured signal or of the at least a part of the second reference signal so as to fit the at least a portion of the second measured signal to some or all of the at least a part of the second reference signal; identifying the best fit between the at least a portion of the second measured signal and the at least a part of the second reference signal, and identifying the thickness of the thin film structure which corresponds to the detected intensity at the respective time instance based on the best fit, to thereby give a second thickness measurement. This second thickness measurement may be used in different ways. In some embodiments, for each of the one or more time instances, an estimated thickness of the thin film structure may be calculated based on the first and second thickness measurements, wherein the estimated thickness is preferably an average of the first. By using two wavelengths to identify the thickness of the thin film structure, an accuracy may be achieved of ±2nm in the thickness of films below 200nm in thickness and ±1% in films greater than 200nm in thickness. If the first wavelength is longer than the second wavelength, the estimated thickness may be calculated using a weighted average of the first and second thickness measurements which is weighted towards the first wavelength for greater thicknesses of the thin film structure, and towards the second wavelength for lesser thicknesses of the thin film structure. In such cases, during an etch process, the weighted average may be weighted towards the first wavelength at earlier time instances and towards the second wavelength at later time instances, while, during a deposition process, the weighted average may be weighted towards the second wavelength at earlier time instances and towards the first wavelength at later time instances. Alternatively, one of the two thickness measurements may be used at different points of an etch or deposition process. For example, if the first wavelength is longer than the second wavelength, the first thickness measurement may be used at greater thicknesses of the thin film structure and the second thickness measurement for lesser thicknesses of the thin film structure. In such cases, during an etch process, the first wavelength may be used at earlier time instances and the second wavelength will be used at later time instances, while, during a deposition process, the second wavelength may be used at earlier time instances and the first wavelength will be used at later time instances. For example, wavelengths of 660nm or more may be more suitable for films having a thickness of 500nm or more as there is less light absorption and signal loss in the film, whereas wavelengths shorter than 660nm may be used for films having a thickness of less than 500nm to give more accurate thickness measurements. Whether a single wavelength or two or more wavelengths are used, the selection of the specific wavelengths used may be dependent on the properties of the detector, such as by choosing wavelengths or ranges of wavelengths with reflection intensities above the noise threshold of the detection instrument and / or by avoiding any wavelengths with sufficient emission to saturate the detector. In the case of a plasma enhanced etching or deposition process, the selection may also be dependent on the emission spectrum of the plasma. Preferably wavelengths in the visible spectrum are chosen to enable the use of common optical components, but wavelengths in the ultraviolet or infrared may also be used. Indeed, infrared is particularly suited to thicker layers, such as those having a thickness of 2 micrometres or more, and ultraviolet to very thin layers, such as those having a thickness of 100 nm or less. In the case of two or more wavelengths being used, it is preferably that these be separated by at least 100nm. Identifying the best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal preferably comprises using a non-linear least squares fitting method, and preferably using a damped least squares fitting method such as the Levenberg-Marquardt fitting method. According to a second aspect of the invention, an apparatus for identifying the thickness of a thin film structure during deposition or etching thereof is provided. The apparatus comprises a processor and a memory, the processor being configured to, at each of one or more time instances during the deposition or etching: compare at least a portion of a first measured signal so far generated against at least a part of a first reference signal stored in the memory, the first measured signal representing the intensity of the detected first wavelength of light versus time as the deposition or etching continues and the first reference signal representing the expected intensity of the first wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the first measured signal or of the at least a part of the first reference signal so as to fit the at least a portion of the first measured signal to some or all of the at least a part of the first reference signal; identify the best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal; and identify the thickness of the thin film structure at the respective time instance based on the best fit to thereby give a first thickness measurement. BRIEF DESCRIPTION OF DRAWINGS Embodiments of the invention will now be described with reference to the drawings, in which: Figure 1 shows a plasma processing apparatus; Figure 2 shows a graph of the variations in intensity of light received at a detector during a deposition process performed according to embodiments of the present invention; Figure 3 shows a graph of the variations in intensity of light received at a detector according to a model of a deposition process performed according to embodiments of the present invention; Figure 4 shows a graph in which a trace of the variations in intensity of light according to a model of a deposition process performed according to embodiments of the present invention is overlayed on a trace of the measured variations in the intensity of light received at a detector during said deposition process; Figure 5 shows a comparative example of a graph in which a different trace of the variations in intensity of light according to a model of a deposition process performed according to embodiments of the present invention is overlayed on the trace of the measured variations in the intensity of light received at a detector during said deposition process; Figure 6 shows a graph in which a trace of the variations in intensity of light according to a model of a deposition process performed according to embodiments of the present invention is overlayed on a trace of the measured variations in the intensity of light received at a detector during the deposition of a silicon dioxide layer; Figure 7 shows the graph of Figure 6 along with an overlayed fitting window used in embodiments of the present invention; Figure 8 shows a graph of the variations in intensity of light of two wavelengths received at a detector during a deposition process performed according to embodiments of the present invention; Figure 9 shows a graph of the spectrum of light received by a detector during a deposition process at a first time period; Figure 10 shows a graph of the spectrum of light received by a detector during a deposition process at a second time period; Figure 11 shows a graph in which a trace of the ratio of the measured spectra shown in Figures 9 and 10 is overlayed with a modelled ratio of two spectra, showing the best fit and the corresponding film thickness; Figure 12 shows a graph in which a trace of the ratio of the measured spectra shown in Figures 9 and 10 is overlayed with a different modelled ratio of two spectra, where the modelled thickness is reduced by 1% to illustrate the change in fitting quality; Figure 13 shows a graph of the spectrum of light received by a detector during a deposition process at a second time period; Figure 14 shows a graph in which a trace of the ratio of the measured spectra shown in Figures 9 and 13 is overlayed with a modelled ratio of two spectra; Figure 15 shows a graph in which a trace of the ratio of the measured spectra shown in Figures 10 and 13 is overlayed with a modelled ratio of two spectra; Figure 16 shows a graph of the thickness of a thin film structure at a series of time instances during a deposition process calculated according to an embodiment of the present invention; Figure 17 shows a graph of the thickness of a thin film structure at a series of time instances during another deposition process calculated according to an embodiment of the present invention; Figure 18 shows a graph of the variations in intensity of light of two wavelengths received at a detector positioned at a first detection angle during a deposition process performed according to embodiments of the present invention; and Figure 19 shows a graph of the variations in intensity of light of two wavelengths received at a detector positioned at a second detection angle during a deposition process performed according to embodiments of the present invention. DETAILED DESCRIPTION In order to determine the endpoint of a deposition or etching process, it may be advantageous to monitor the thickness of a thin film structure in situ. That is to say, measuring the thickness of the thin film structure during deposition or etching thereof may be preferable to other methods of determining the endpoint of the deposition or etching process. Said in situ monitoring may be achieved by monitoring the light reflected from the thin film structure. Figure 1 shows an example of thin film processing tool 100 used in an etching or deposition process. A substrate on which a thin film structure 101 is to be etched or deposited is mounted on a table 102 inside a processing chamber 103 of the thin film processing tool 100. A viewport 104 is mounted in a wall 105 of the processing chamber 103 such that light 106 reflected by the thin film structure 101 during the etching or deposition may be received by a collimator 107 or lens 107. During an etching or deposition process, the detected intensity is sent via fibre optic cable 109 to a spectrometer or detector which generates a measured signal representing the intensity of the detected light versus time as the deposition or etching continues. A sapphire window, or similar etch-resistant window relevant to the processes used, can be used in the viewport 104 to minimise erosion during thin film processing, and hence to reduce clouding of the window. In alternative embodiments, the spectrometer or detector may be positioned directly adjacent the viewport 104. In the example shown in Figure 1, the light 108 illuminating the thin film structure 101 is emitted by a plasma used during the etching or deposition process, thereby obviating the need for an external light source and, consequently, a second viewport in another wall of the processing chamber. The detector or spectrometer may be configured to receive polarised light. When detecting polarised light of a given wavelength, the change in the intensity of light at said wavelength with thickness at each polarisation (s or p) exhibits slow moving sinusoidal ripples. This means that for typical film thicknesses of around 100 to 500nm there will be very few, if any, turning points in the measured trace of the intensity of the light, leading to difficulty in obtaining accurate thickness measurements for thinner films. Furthermore, polarisers add cost to the measurement hardware, as well as to increased complexity when setting up thin film processing tool. For this reason, it can be desirable to use a system in which the thickness of a thin film structure is measured using unpolarised light. The collimator 107 or lens 107 selects light which has reflected from the thin film structure 101 or from one or more regions of the thin film structure. It may also collect light emitted by a plasma in a plasma enhanced etching or deposition process in the direction of the detector in the path between the substrate and the detector. Although the additional background this adds to the reflected intensity is minimal, the sensitivity of the technique may nonetheless be increased by making the acceptance angle for light at the detector as narrow as possible, such as through the use of said collimator or lens. When detecting unpolarised light, the two polarisation traces are combined producing a more complex trace with far more detail. However, the turning points in this trace do not, typically, occur at equal intervals. This means that, whereas the thickness of a thin film structure may be measured by counting the ripples in a trace of the intensity of polarised light, this is not usually possible for a trace in the intensity of unpolarised light. Another problem with the ripple counting method is that disturbances in the plasma during the etching or deposition process (and, therefore, during the change in the thickness of the thin film structure) may produce fluctuations in the intensity trace. These fluctuations may then be interpreted as turning points, giving an incorrect thickness measurement. In order to address this issue, embodiments of the present invention use a modelled trace of the reflectance of a thin film structure. This may be generated using a modelling program (for example, using Abele’s matrix method) to determine the expected trace of reflectance versus thickness for a given thin film structure, monitored wavelength, and detection angle. Examples of this process is shown in the graphs of Figures 2 to 7. These figures relate to a deposition process, but the approach described below with reference to these figures may also be applied to an etching process. Figure 2 shows the measured trace of the intensity of reflected light versus time during deposition of a layer of silicon dioxide (SiO2) onto a silicon (Si) substrate. As can be seen, the intensity of light reflected from the silicon substrate varies between peaks of approximately the same intensity and troughs which alternate between a higher and a lower intensity. In this example, light of 661 nm was used to generate the measured trace. The trace is measured by illuminating the thin film structure with light from a light source during deposition of the silicon dioxide layer. In the example of Figure 2, the deposition occurs as part of a plasma deposition process and the light used to illuminate the thin film structure is provided by the plasma itself. The light reflected by the thin film is then detected by a detector and used to generate a signal representing the measured intensity of the detected light versus time as the deposition continues. In order to determine the thickness of the silicon dioxide a model of the thin film structure is used to generate a graph of the reflectance of the thin film structure versus thickness, as shown in Figure 3. This modelled reflectance is then compared with the measured signal, with Figure 4 showing the modelled reflectance trace overlayed on the measured signal. This comparison involves changing one or both of the horizontal and vertical scales of the modelled trace so as to fit said modelled trace to the measured signal. The best fit between the modelled trace and the measured signal is then used to identify the thickness of the silicon dioxide layer at a time instance. In this way the modelled trace is used as a reference signal which allows the thickness of the silicon dioxide layer to be determined at various time instances during the deposition. (For the avoidance of doubt, references to a modelled trace throughout this disclosure are to be taken as equivalent to a reference signal.) By way of comparison, Figure 5 shows an example of a different modelled trace fitted to the measured signal. It can be seen that the fit in Figure 5 is poorer than that of the example of Figure 4, meaning that the fit of Figure 4 will be chosen over the fit of Figure 5 when identifying the best fit. The modelled traces in Figures 4 and 5 each show the reflectance of the thin film structure across a range of thicknesses of the silicon dioxide layer. Both modelled traces start at substantially zero thickness, with the modelled trace shown in Figure 5 running to a greater final thickness than that shown in Figure 4. One way in which the best fit may be used is by identifying the final thickness of the modelled trace identified by the best fit, which in this case is around 2100nm for the modelled trace shown in Figure 4. However, this approach to using the best fit to identify the thickness of a thin film structure may not be suitable in some cases. For example, Figure 6 shows another graph in which a modelled reflectance trace is overlayed on a measured signal, and it can be seen that, although the model is fitted well to the portion of the signal between around 45 seconds and 60 seconds, the fit is poorer for earlier time instances during the deposition. The reason for this is that, whereas the measured signal shown in Figures 4 and 5 is taken during a substantially constant rate of deposition, the measured signal shown in Figure 6 is taken during the deposition of a silicon dioxide layer in which the initial deposition rate is slower than the rate of deposition at the subsequent time instances shown in the measured signal. In order to account for changes in the rate of deposition or etching, embodiments of the present invention preferably employ a rolling fitting window in which the modelled trace is fitted to only a portion of the measured signal (or, conversely, only a portion of the measured signal is fitted to the modelled trace), with this portion usually taken as the portion of the measured signal between a current time instance and a previous time instance. This fitting window is illustrated in Figure 7 for the graph of Figure 6. One or more scaling factors are then identified for the best fit between the portion of the measured signal within the fitting window and a corresponding portion of the modelled trace. The one or more scaling factors are then used to identify an average etching or deposition rate across the etching or deposition process. This average etching or deposition rate may then be multiplied by the current time instance to identify a total etch depth or deposition thickness. For example, if an average etching rate of 8nm per second was identified after 100s of an etching process, the etch depth at the current time instance would be identified as being 800nm. The fitting process may be performed as frequently or infrequently as desired, with more frequent fitting allowing for more accurate measurements of the thickness of a thin film structure during etching or deposition. However, the fitting is advantageously performed at regular intervals, either with the intervals between fittings being substantially the same across the whole of the etching or deposition process or with said intervals being substantially the same across a portion of the etching or deposition process. For example, in an exemplary deposition process the first 2 to 10 seconds of a signal may be excluded from the rate calculation to avoid rate errors due to initial settling of the process (such as due to changes in the plasma in a plasma enhanced deposition process). The first fitting may also occur after the first turning point in a measured signal so as to improve the accuracy of the identified thickness. For example, for a measured signal at 405nm this would typically be at thicknesses >32nm for silicon nitride (SiN) and at thicknesses >52nm for silicon dioxide. The thickness may also be estimated between fittings (that is to say between consecutive time instances at which the modelled trace and the measured signal are fitted) by using the most recently identified deposition or etch rate and multiplying by a current time to estimate a current thickness. The fitting process itself may be performed in any of a number of ways, but advantageously comprises fitting the modelled trace to the portion of the measured signal within the fitting window (or, conversely, fitting said portion of the measured signal to the modelled trace) using a non-linear least squares fitting method, such as the Levenberg-Marquardt fitting method. In some exemplary embodiments, the detector is configured to detect two or more wavelengths of light reflected by the thin film structure. Figure 8 shows the measured traces during a deposition process for wavelengths 405nm and 661 nm. The ripples in the 405nm trace occur at a faster rate, so potentially providing better measurement accuracy for thin films, but the shape becomes slightly distorted for thicker films. This is likely due to increased absorption of the light by the film material, and increased sensitivity to thickness non-uniformity within the viewed region of the wafer (i.e. the light collection area). However, the ripples in the 661 nm trace hold their shape throughout. For example, the process described above with reference to Figures 2 to 7 may also be performed using light reflected by the thin film structure at a different wavelength. In some embodiments, a first thickness measurement is identified using light reflected a first wavelength and a second thickness measurement is identified contemporaneously using light reflected at a second wavelength. An estimated thickness may then be calculated as a (preferably weighted) average of the first and second thickness measurements. The weighting of the average may vary depending on the thickness of the etched or deposited layer. For example, if the first wavelength is longer than the second wavelength, the weighted average may be weighted towards the first wavelength for greater thicknesses of the thin film structure and towards the second wavelength for lesser thicknesses of the thin film structure. In such cases, during an etch process, the weighted average may be weighted towards the first wavelength at earlier time instances and towards the second wavelength at later time instances, while, during a deposition process, the weighted average may be weighted towards the second wavelength at earlier time instances and towards the first wavelength at later time instances. Alternatively, one of the two thickness measurements may be used at different points of an etch or deposition process. For example, if the first wavelength is longer than the second wavelength, the first thickness measurement may be used at greater thicknesses of the thin film structure and the second thickness measurement for lesser thicknesses of the thin film structure. In such cases, during an etch process, the first wavelength may be used at earlier time instances and the second wavelength may be used at later time instances, while, during a deposition process, the second wavelength may be used at earlier time instances and the first wavelength may be used at later time instances. For example, wavelengths of 660nm or more may be more suitable for films having a thickness of 500nm or more as there is less light absorption and signal loss in the film, whereas wavelengths shorter than 660nm may be used for films having a thickness of less than 500nm to give more accurate thickness measurements. Whether a single wavelength or two or more wavelengths are used, the selection of the specific wavelengths used may be dependent on the properties of the detector, such as by choosing wavelengths or ranges of wavelengths with reflection intensities above the noise threshold of the detection instrument and / or by avoiding any wavelengths with sufficient emission to saturate the detector. In the case of a plasma enhanced etching or deposition process, the selection may also be dependent on the emission spectrum of the plasma. Preferably wavelengths in the visible spectrum are chosen to enable the use of common optical components, but wavelengths in the ultraviolet or infrared may also be used. Indeed, infrared is particularly suited to thicker layers, such as those having a thickness of 2 micrometres or more, and ultraviolet to very thin layers, such as those having a thickness of 100 nm or less. In the case of two or more wavelengths being used, it is preferably that these be separated by at least 100nm. Another approach to identifying the thickness of a thin film structure during a deposition or etching process will now be described with reference to Figures 9 to 17. In the embodiments described above with reference to Figures 2 to 7, while two or more thickness measurements may be identified using two or more correspond wavelengths of reflected light, the process used to identify each of these thickness measurements is based on analysing a single wavelength, the process being carried out, either sequentially or contemporaneously, separately for each wavelength of reflected light. However, in embodiments in which the detector is configured to detect a spectrum of light, the thickness of a thin film structure may be determined using the spectra of light reflected by a thin film structure at various time instances during an etch or deposition process. These embodiments involve first detecting the light reflected by a thin film structure at a baseline time instance and generating a baseline spectrum which represents the intensity of the detected light versus wavelength at the baseline time instance. The baseline spectrum may be updated throughout the etching or deposition, as will be described below, but the initial baseline spectrum is preferably taken at the start of the etching or deposition process when the etch depth or deposition thickness is substantially zero. However, as can be seen from Figure 9, which shows a baseline spectrum taken at 5 seconds into a deposition process, this may not be taken exactly at the start of the etching or deposition process. In the example of Figure 9, this is because the deposition process is plasma enhanced. The plasma used in this deposition process (and, similarly, the plasmas used in etching processes in embodiments of the present invention) may require an initial stabilisation period during which the emission spectrum of the plasma changes. Therefore, waiting until the plasma has stabilised allows for more accurate comparisons with a model of a deposition process. The stabilisation is rapid enough, however, that substantially zero etching or deposition occurs. The light reflected by the thin film structure is then detected at one or more time instances during the deposition or etching and a measured spectrum is generated which represents the intensity of the detected light versus wavelength. An example of a measured spectrum taken at 590 seconds into a deposition process is illustrated in Figure 10. As can be seen, the differences between the two spectra are small (less than 15%) but taking a ratio of the two allows for the differences (arising, for example, by interference effects which may oscillate during an etching or deposition process) to be enhanced allowing for more accurate fitting to modelled data. This can be seen by comparing Figure 11, which shows the ratio of a measured spectrum taken at 590 seconds into a deposition to a baseline spectrum taken at 5 seconds into a deposition process, with Figure 14, which shows the ratio of a measured spectrum taken at 295 seconds into a deposition to a baseline spectrum taken at 5 seconds into a deposition process, and Figure 15, which shows the ratio of a measured spectrum taken at 590 seconds into a deposition to a baseline spectrum taken at 295 seconds into a deposition process. Even though the measured spectrashown in Figures 9,10, and 13, the latter taken at 295 seconds into the deposition process, are very similar, the measured ratios are easily distinguished from one another. Therefore, while fitting a model directly to the measured spectra may be sensitive to noise in the measured spectra, leading to less accurate fitting and consequently to less accurate thickness measurements, the fitting of a model to the measured ratios is less susceptible to such noise, allowing for improved fitting and more accurate thickness measurements to be obtained. The identification of a thickness of the thin film structure comprises comparing the measured ratio with a plurality of ratioed reference spectra stored in a memory. Each these ratioed reference spectra corresponds to the expected spectrum at a corresponding thickness of the thin film structure divided by the expected baseline spectrum at the baseline time instance (or vice versa). These are then compared with the measured ratio by identifying which of the reference spectra provides a best fit with the ratioed spectrum, the thickness corresponding to the identified reference spectrum then being taken a measurement of the thickness of the thin film structure at the time instance at which the measured spectrum was detected. Figures 11 and 12 show two examples of a ratioed reference spectrum (also referred to as a modelled ratio) overlayed on the measured ratio of the spectrum shown in Figure 10 to the baseline spectrum shown in Figure 9, with the modelled ratio of Figure 11 based on a modelled thickness of 442.41 nm and that of Figure 12 based on a modelled of thickness of 438nm. It can be seen that even with a difference in thickness of only around 1% there are clear differences between the fit in Figure 11 and the poorer fit shown in Figure 12. The improved fit of the modelled ratio of Figure 11 as compared with the modelled ratio shown in Figure 12 and modelled ratios at other modelled thickness is advantageously identified using a non-linear least squares fitting method, and preferably using a damped least squares fitting method such as the Levenberg-Marquardt fitting method. Figure 13 shows another measured spectrum generated using light detected at 295 seconds into the deposition process. This can be fitted a modelled ratio as shown in Figure 14 according to the method described above to identify a thickness of 222.76 nm at 295 seconds. Alternatively, this measured spectrum may be used as an updated baseline spectrum used when fitting subsequently generated spectra to modelled ratios. For example, Figure 15 shows the measured ratio of the spectra shown in Figures 10 and 13 which has been fitted with a modelled ratio indicating a thickness of 442.28nm. In this instance, the identified thickness measurement is substantially the same as that identified in the fit shown in Figure 11. However, in other instances it may be preferable to update the baseline spectrum. For example, if a first layer of material is deposited (or etched) followed by a second layer of material being deposited (or etched) then it would be advantageous to update the baseline at the start of the deposition (or etching) of the second layer of material. It will be noted that the intensity of detected light shown in the spectra of Figures 9, 10, and 13 include frequencies that exceed the maximum intensity reading possible using the detector, known as saturation, as well as frequencies of very low intensities which may be more subject to noise. For this reason, intensity measurements above a maximum threshold intensity and / or intensity measurements below a minimum threshold intensity may be excluded from the baseline spectrum and / or from the measured spectrum prior to the calculation from the ratio. This can be seen in the trace of the measured ratio in Figures 11, 12, 14 and 15 which is discontinuous, the wavelengths between approximately 330nm and 390nm having been excluded, for example. Figures 16 and 17 show the thicknesses of layers of silicon dioxide and silicon nitride, respectively, plotted as a function of time during a deposition process. The dotted line in each graph shows a straight line fit between the measured thicknesses plotted on the graphs. The variations between the measured thicknesses and this line in each graph may be indicative of variations in the deposition rate during the deposition process. However, in some cases the straight line fit may be used to provide an estimate of the thickness of the thin film structure, such as if the detected intensities are subject to a high degree of background noise. The straight line fit may also be used to identify an average deposition rate that can be used to estimate the thickness of the thin film structure during the deposition process between fittings. Similarly, a straight line fit of the measured thickness of a thin film structure during an etching process may be used to estimate the thickness of the thin film structure during the etching process between fittings. The approach described above with reference to Figures 9 to 17 can be used instead of or in addition to the approach described above with reference to Figures 2 to 7. For example, a first thickness can be identified using the approach described above with reference to Figures 2 to 7 and a second thickness identified using the approach described above with reference to Figures 9 to 17 and the two thicknesses averaged to identify the thickness of the thin film structure. In some cases, it may be found that one approach is advantageous during some stages of an etching or deposition process and another approach is advantageous during other stages of the process. For example, the approach described above with reference to Figures 2 to 7 may be used to identify the thickness of a thin film structure during the early stages of the deposition of a thin film structure and the approach described above with reference to Figures 9 to 17 used at later stages of the deposition. The above described methods of measuring the thickness of a thin film structure overcome the problems that can be caused by plasma instabilities when counting the turning points in a measured trace by evaluating the overall shape of (at least a portion of) the measured trace rather than focussing solely on the turning points. For the same reasons, these methods are also less sensitive to signal noise and any reduction in signal strength over time, for example due to window clouding or deposition on the viewport. These methods are also advantageous during pulsed or switched multi-frequency plasma deposition or etching. Whereas the emission intensity from the plasma alternates between two or more levels when pulsed or switched multi-frequency plasmas are used, which can lead to difficulties in identifying turning points, these methods can be adapted to account for this by separating a measured signal into multiple curves corresponding to the two or more levels. For example, the deposition or etching may alternate between a first mode of deposition or etching and a second mode of deposition or etching with the measured signal separated into a first measured signal comprising only those portions of the signal detected during the first mode and into a second measured signal comprising only those portions of the signal detected during the second mode. The methods are especially advantageous when the light reflected from a thin film structure is detected at angles greater than the Brewster angle of the thin film structure. Figure 18 shows the modelled reflectance of light of 660nm for various thicknesses of a silicon dioxide film deposited on a silicon substrate when light is detected at an angle of 80 degrees to the normal to the substrate, which is above the Brewster’s angle for this thin film structure. As can be seen, the reflectance of the s-polarised light (shown using the dashed line) is out of phase with the reflectance of the p-polarised light (shown using the dotted line) leading to an overall trace in which the peaks are at a uniform reflectance value but in which the troughs alternate between a higher and a lower reflectance value. This can be compared with the graph shown in Figure 19 in which the light is detected at an angle of 45 degrees to the normal to the substrate, which is below the Brewster’s angle for this thin film structure. In this case, the reflectance of s-polarised light is in phase with the reflectance of p-polarised light. The graphs shown in Figure 18 and 19 therefore illustrate that a method of identifying the thickness of a thin film structure using conventional methods, such as by counting equally spaced turning points, is no longer possible when light is detected above the Brewster’s angle. Consequently, the identification of the thickness of a thin film structure is made more difficult since the turning points are not equally spaced. Conversely, the methods described herein are suitable for use when light is detected both above and below the Brewster’s angle. Indeed, the greater variation in the measured signal of unpolarised light detected above the Brewster’s angle may lead to improved accuracy in the measurements in the thickness of a thin film structure using embodiments of the present invention, owing to these embodiments fitting modelled data to the entirety of a modelled signal. This leads to the present invention being particularly suited to measurements in conventional plasma processing tools, since the position of a viewport in such tools is typically such that the angle of the view of a thin film structure subject to an etching or deposition process is very shallow, i.e. above the Brewster’s angle of the thin film structure. CLAIMS - REFLECTOMETRY VIA TIME-FITTING 1. A method of measuring the thickness of a thin film structure during deposition or etching thereof, the thin film structure being disposed on a substrate mounted on a table inside a processing chamber of a thin film processing tool, the method comprising: illuminating the thin film structure with light from a light source, the light including at least a first wavelength; detecting light of the first wavelength reflected by the thin film structure at a detection angle and generating a first measured signal which represents the intensity of the detected first wavelength of light versus time as the deposition or etching continues; at each of one or more time instances during the deposition or etching: comparing at least a portion of the first measured signal so far generated against at least a part of a first reference signal stored in a memory, the first reference signal representing the expected intensity of the first wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the first measured signal or of the at least a part of the first reference signal so as to fit the at least a portion of the first measured signal to some or all of the at least a part of the first reference signal; and identifying the best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal, and identifying the thickness of the thin film structure at the respective time instance based on the best fit to thereby give a first thickness measurement. 2. A method according to claim 1, wherein identifying the best fit comprises identifying one or more scaling factors which achieve the best fit between the at least a portion of the first or second measured signal and the at least a part of the first reference signal, the identified scaling factors being indicative of the rate of deposition or etching, thereby giving a first rate measurement. 3. A method according to claim 2, wherein identifying the thickness of the thin film structure at the respective time instance based on the best fit comprises identifying the thickness of the thin film structure at the respective time instance based on the first rate measurement and the respective time instance. 4. A method according to claim 3, wherein, during deposition, identifying the thickness of the thin film structure at the respective time instance based on the best fit comprises multiplying the first rate measurement and the deposition time at the respective time instance. 5. A method according to any of claims 2 to 4, the method further comprising estimating a thickness of the thin film structure at a time between consecutive time instances based on the most recent rate measurement and the corresponding thickness measurement. 6. A method according to any of the preceding claims, wherein the portion of the first measured signal so far generated comprises at least a portion of the signal generated between an earlier time instance of the one or more time instances and the respective time instance. 7. A method according to any of the preceding claims, wherein the portion of the first measured signal so far generated comprises substantially all of the signal generated between said earlier time instance and the respective time instance. 8. A method according to any of the preceding claims, wherein the deposition or etching comprises alternating between at least a first mode of deposition or etching and a second mode of deposition or etching and generating the first measured signal comprises excluding the light detected during the second mode of deposition or etching. 9. A method according to any of the preceding claims, wherein the detection angle is greater than the Brewster’s angle of the thin film structure. 10. A method according to any of the preceding claims, wherein the detector used to detect the light of the first wavelength reflected by the thin film structure is configured to receive light lying within 3° of the detection angle, preferably within 2° of the detection angle, and more preferably within 1° of the detection angle. 11. A method according to any of the preceding claims, wherein the thin film structure is a single film or a multilayer film structure. 12. A method according to any of the preceding claims, wherein the deposition or etching comprises a plasma enhanced deposition process or plasma enhanced etching process. 13. A method according to claim 12, wherein the light source comprises the plasma source used in the plasma enhanced deposition process or plasma enhanced etching process. 14. A method according to claim 12 or claim 13, wherein the plasma enhanced deposition process or plasma enhanced etching process comprises one or more of: plasma enhanced chemical vapour deposition, plasma enhanced atomic layer deposition, and plasma enhanced atomic layer etching. 15. A method according to any of the preceding claims, wherein the thin film structure comprises a layer which is at least semi-transparent to light of the first wavelength, said layer preferably being substantially fully transparent to light of the first wavelength, and said layer more preferably being substantially fully transparent to all visible light. 16. A method according to any of the preceding claims, wherein the first reference signal is modelled based on one or more of: the known or expected properties of the thin film structure; the first wavelength; and the detection angle. 17. A method according to any of the preceding claims, wherein the first reference signal is calculated based on signals measured from light reflected from one or more thin film samples during one or more prior etching or deposition processes. 18. A method according to any of the preceding claims, the method further comprising stopping the deposition or etching if the measured thickness reaches a predetermined thickness. 19. A method according to any of the preceding claims, wherein the light further includes a second wavelength, light of the second wavelength reflected by the thin film structure is additionally detected at the detection angle a second measured signal is generated which represents the intensity of the detected second wavelength of light versus time, as the deposition or etching continues; and the method further comprises: at each of the one or more time instances during the deposition or etching, comparing at least a portion of the second measured signal so far generated against at least a part of a second reference signal stored in a memory, the second reference signal representing the expected intensity of the second wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the second measured signal or of the at least a part of the second reference signal so as to fit the at least a portion of the second measured signal to some or all of the at least a part of the second reference signal; identifying the best fit between the at least a portion of the second measured signal and the at least a part of the second reference signal, and identifying the thickness of the thin film structure which corresponds to the detected intensity at the respective time instance based on the best fit, to thereby give a second thickness measurement. 20. A method according to claim 19, wherein for each of the one or more time instances, an estimated thickness of the thin film structure is calculated based on the first and second thickness measurements, wherein the estimated thickness is preferably an average of the first and second thickness measurements, most preferably a weighted average. 21. A method according to claim 19 or claim 20, wherein the first wavelength is longer than the second wavelength and the estimated thickness is calculated using a weighted average of the first and second thickness measurements which is weighted towards the first wavelength for greater thicknesses of the thin film structure, and towards the second wavelength for lesser thicknesses of the thin film structure. 22. A method according to any of the preceding claims, wherein identifying the best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal comprises using a non-linear least squares fitting method. 23. A method according to claim 22, wherein the non-linear least squares fitting method comprises a damped least squares fitting method such as the Levenberg-Marquardt fitting method. 24. An apparatus for identifying the thickness of a thin film structure during deposition or etching thereof, the apparatus comprising a processor and a memory, the processor being configured to, at each of one or more time instances during the deposition or etching: compare at least a portion of a first measured signal so far generated against at least a part of a first reference signal stored in the memory, the first measured signal representing the intensity of a first wavelength of light versus time reflected by the thin film structure as the deposition or etching continues and the first reference signal representing the expected intensity of the first wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the first measured signal or of the at least a part of the first reference signal so as to fit the at least a portion of the first measured signal to some or all of the at least a part of the first reference signal; identify the best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal; and identify the thickness of the thin film structure at the respective time instance based on the best fit to thereby give a first thickness measurement. 10 09 25 Amendments to the claims have been filed as follows :

Claims

1. A method of measuring the thickness of a thin film structure during deposition or etching thereof, the thin film structure being disposed on a substrate mounted on a table inside a processing chamber of a thin film processing tool, the method comprising:illuminating the thin film structure with light from a light source, the light including at least a first wavelength;detecting light of the first wavelength reflected by the thin film structure at a detection angle and generating a first measured signal which represents the intensity of the detected first wavelength of light versus time as the deposition or etching continues;at each of one or more time instances during the deposition or etching:comparing at least a portion of the first measured signal so far generated against at least a part of a first reference signal stored in a memory, the first reference signal representing the expected intensity of the first wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the first measured signal or of the at least a part of the first reference signal so as to fit the at least a portion of the first measured signal to some or all of the at least a part of the first reference signal; andidentifying a first best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal, and identifying the thickness of the thin film structure at the respective time instance based on the first best fit to thereby give a first thickness measurement;wherein the light further includes a second wavelength, light of the second wavelength reflected by the thin film structure is additionally detected at the detection angle and a second measured signal is generated which represents the intensity of the detected second wavelength of light versus time, as the deposition or etching continues; and the method further comprises:at each of the one or more time instances during the deposition or etching, comparing at least a portion of the second measured signal so far generated against at least a part of a second reference signal stored in a memory, the second reference signal representing the expected intensity of the second wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the second measured signal or of the at least a part of the second reference signal so as to fit the at least a portion of the second10 09 25measured signal to some or all of the at least a part of the second reference signal; identifying a second best fit between the at least a portion of the second measured signal and the at least a part of the second reference signal; andidentifying the thickness of the thin film structure which corresponds to the detected intensity at the respective time instance based on the second best fit, to thereby give a second thickness measurement;wherein for each of the one or more time instances, an estimated thickness of the thin film structure is calculated based on the first and second thickness measurements.

2. A method according to claim 1, wherein identifying the first and / or second best fits respectively comprise identifying one or more scaling factors which achieve the best fit between the at least a portion of the respective first or second measured signal and the at least a part of the respective first or second reference signal, the identified scaling factors being indicative of the rate of deposition or etching, thereby giving a respective first or second rate measurement.

3. A method according to claim 2, wherein identifying the thickness of the thin film structure at the respective time instance based on the first and / or second best fits respectively comprise identifying the thickness of the thin film structure at the respective time instance based on the respective first or second rate measurement and the respective time instance.

4. A method according to claim 3, wherein, during deposition, identifying the thickness of the thin film structure at the respective time instance based on the first and / or second best fits respectively comprise multiplying the respective first or second rate measurement and the deposition time at the respective time instance.

5. A method according to any of claims 2 to 4, the method further comprising estimating a thickness of the thin film structure at a time between consecutive time instances based on the most recent rate measurement and the corresponding thickness measurement.

6. A method according to any of the preceding claims, wherein the portion of the first and / or second measured signals so far generated respectively comprise at least a portion of the respective first or second signal generated between an earlier time instance of the one or more time instances and the respective time instance.10 09 257. A method according to any of the preceding claims, wherein the portion of the first and / or second measured signals so far generated respectively comprise substantially all of the respective first or second signal generated between said earlier time instance and the respective time instance.

8. A method according to any of the preceding claims, wherein the deposition or etching comprises alternating between at least a first mode of deposition or etching and a second mode of deposition or etching and generating the first measured signal comprises excluding the light detected during the second mode of deposition or etching.

9. A method according to any of the preceding claims, wherein the detection angle is greater than the Brewster’s angle of the thin film structure.

10. A method according to any of the preceding claims, wherein the detector used to detect the light of the first and second wavelengths reflected by the thin film structure is configured to receive light lying within 3° of the detection angle, preferably within 2° of the detection angle, and more preferably within 1° of the detection angle.

11. A method according to any of the preceding claims, wherein the thin film structure is a single film ora multilayer film structure.

12. A method according to any of the preceding claims, wherein the deposition or etching comprises a plasma enhanced deposition process or plasma enhanced etching process.

13. A method according to claim 12, wherein the light source comprises the plasma source used in the plasma enhanced deposition process or plasma enhanced etching process.

14. A method according to claim 12 or claim 13, wherein the plasma enhanced deposition process or plasma enhanced etching process comprises one or more of: plasma enhanced chemical vapour deposition, plasma enhanced atomic layer deposition, and plasma enhanced atomic layer etching.

15. A method according to any of the preceding claims, wherein the thin film structure comprises a layer which is at least semi-transparent to light of the first and second wavelengths, said layer preferably being substantially fully transparent to light of the first10 09 25and second wavelengths, and said layer more preferably being substantially fully transparent to all visible light.

16. A method according to any of the preceding claims, wherein the first and / or second reference signals are respectively modelled based on one or more of: the known or expected properties of the thin film structure; the respective first or second wavelength; and the detection angle.

17. A method according to any of the preceding claims, wherein the first and / or second reference signals are respectively calculated based on signals measured from light reflected from one or more thin film samples during one or more prior etching or deposition processes.

18. A method according to any of the preceding claims, the method further comprising stopping the deposition or etching if the measured thickness reaches a predetermined thickness.

19. A method according to any of the preceding claims, wherein the estimated thickness is an average of the first and second thickness measurements, preferably a weighted average.

20. A method according to any of the preceding claims, wherein the first wavelength is longer than the second wavelength and the estimated thickness is calculated using a weighted average of the first and second thickness measurements which is weighted towards the first wavelength for greater thicknesses of the thin film structure, and towards the second wavelength for lesser thicknesses of the thin film structure.

21. A method according to any of the preceding claims, wherein identifying the first and / or second best fits between the at least a portion of the respective first or second measured signal and the at least a part of the respective first or second reference signal comprises using a non-linear least squares fitting method.

22. A method according to claim 21, wherein the non-linear least squares fitting method comprises a damped least squares fitting method such as the Levenberg-Marquardt fitting method.10 09 2523. An apparatus for identifying the thickness of a thin film structure during deposition or etching thereof, the apparatus comprising a processor and a memory, the processor being configured to, at each of one or more time instances during the deposition or etching: compare at least a portion of a first measured signal so far generated against at least a part of a first reference signal stored in the memory, the first measured signal representing the intensity of a first wavelength of light versus time reflected by the thin film structure as the deposition or etching continues and the first reference signal representing the expected intensity of the first wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the first measured signal or of the at least a part of the first reference signal so as to fit the at least a portion of the first measured signal to some or all of the at least a part of the first reference signal;identify a first best fit between the at least a portion of the first measured signal and the at least a part of the first reference signal; andidentify the thickness of the thin film structure at the respective time instance based on the first best fit to thereby give a first thickness measurement;wherein the processor is further configured to, at each of the one or more time instances during the deposition or etching:compare at least a portion of a second measured signal so far generated against at least a part of a second reference signal stored in a memory, the second reference signal representing the expected intensity of a second wavelength of light reflected by the thin film structure versus thickness of the thin film structure, wherein the comparison comprises changing the scale of the at least a portion of the second measured signal or of the at least a part of the second reference signal so as to fit the at least a portion of the second measured signal to some or all of the at least a part of the second reference signal;identify a second best fit between the at least a portion of the second measured signal and the at least a part of the second reference signal; andidentify the thickness of the thin film structure which corresponds to the detected intensity at the respective time instance based on the second best fit, to thereby give a second thickness measurement;wherein for each of the one or more time instances, an estimated thickness of the thin film structure is calculated based on the first and second thickness measurements.

Images