Continuously decreasing feature size over large area fields with nanometer precision is a key challenge of todays optical metrology systems.1 There are solutions for one dimensional large scale nanometrology problems.2 But a still challenging part of these macroscopic nanoscale measurements is the high precision positioning in 3D space. Solutions to these challenges for small measurement volumes are the micro-CMM3 or the Nanopositioning and Nanomeasuring machine NMM1.4 The NPMM-200 at ITO, built by the University of Ilmenau,5, 6 is made for large scale high precision measurements. The machine allows picometer scale resolution positioning in 3D space over a large measurement volume of 200 mm x 200 mm x 25 mm. This precise positioning capability in combination with the flexible sensor mount of the machine can be used for different tasks, like measurements of height, shape or lateral dimensions. The machine can also be used for calibration of optical systems and sensors. One focus of our work with the machine lies in the measurement of topography errors of freeform surfaces.
In this paper we show the progress in using the NPMM-200 for freeform surface measurements. Section 2 reviews the principle of the realization of the NPMM. We show that the flexible sensor mount is a key feature that allows to address different measurement challenges. In chapter three we demonstrate this performance with stability tests and measurement examples of aspheric and freeform surfaces. We close with a short conclusion and an outlook on future developments.
PRINCIPLE SETUP OF THE NPMM
The NPMM-200 at ITO is a large area nanomeasuring machine which is described in detail in the work of Balzer.7 The concept of the machine is shown in figure 1. The probe system and the measuring systems for the moving sample are attached to the metrological frame made of Zerodur®, a glass-ceramic with near-zero thermal expansion. The kinematic concept of the machine is that the sensor, i.e. a nanoprobe system, is steady while the sample is moved on a three axis stage beneath the sensor. This stage is driven by direct electronic voice coil drives. The sample is located on top of a so called mirror corner. The position and the angular deviations of the mirror corner with respect to the metrological frame are measured interferometrically. The NPMM is avoiding Abbe errors in 3D space which is the key to its high precision. This is achieved by controlling the angular deviations of the mirror corner over the whole measurement range. The axis of the main interferometers intersect at one point, the Abbe-Point. Small samples can be placed inside the mirror corner, which allows to reduce all Abbe offsets to zero. The position and the angular deviations are controlled in close loop with a real time system based on National-Instruments PXI FPGA-Modules. Both the orthogonality as well as the flatness of the faces of the mirror corner are calibrated and fed into the control loop.
The interferometer signals are corrected using data of calibrated sensors which monitor the environmental conditions: temperature, pressure and humidity inside the measurement chamber. For high accuracy, however, the uncertainty of the refractive index of the air is limiting. Therefore, the machine is located inside of a vacuum chamber which is set on vibration damping feet. Due to the vacuum operation, the influence of refractive index modulation on the interferometric measurements is essentially eliminated. The opened vacuum chamber is shown in figure 2 on the left. The metrological frame, the mirror corner and three interferometers for the x and y axis are shown in figure 2 on the right.
Specifications of the NPMM-200
|Positioning volume||200 mm x 200 mm x 25 mm|
|Positioning speed||30 mm/s|
|Interferometer resolution||0.02 nm|
|Vacuum operation||3 mbar|
|Sample size||max. 310 mm x 310 mm x 90 mm|
|Sensor weight||max. 15 kg|
The machine is controlled through the open source measurement environment ITOM,8 which is based on C++ and uses PYTHON as macro language. There is a large amount of different sensors, cameras, measurement systems and actuators that can be simply used through ITOM in combination with the machine. At the present time we have used three different sensors for measurement of shapes and heights with the NPMM-200. An autofocus probe system built by the University of Ilmenau9 and the company SIOS is a versatile measurement system for optical surfaces. It has a high resolution but a small measurement range of 1.5 μm. After approaching the test sample surface to the middle of the measurement range, the measurement signal of the autofocus probe is used as a control signal for height control of the machine. Therefore linearity errors in the measurement signal can be minimized. In the evaluation of the measured height, the sensor signal is combined with the measured z-position of the machine. The measurable z-range is limited by the working distance of the sensor. We use a long working distance objective (NIKON CFI TU PLAN EPI ELWD 50x) with a working distance of 11.0 mm. For samples that include large steps, its advantageous to use a sensor with a large measurement range. Sensors such as the chromatic confocal sensor “Precitec Chrocodile” can be used in the same closed loop system.
Apart from the surface shape measurements the machine’s positioning accuracy can be used to assess a variety of metrology topics. E.g. for optical inspection investigations, a microscope is attached to the sensor mount of the machine. New optics distortion calibration methods can be based on the high positioning accuracy of the NPMM-200. Recently, we have demonstrated a distortion calibration of a telecentric objective using the NPMM-200.10 In this case, the objective is fixed on the metrological frame. A precise light source as target is positioned in the field of view of the objective. With the different positions of the light source and the position on the image of the light source on the camera a calibration function for the distortions can be determined.
In this section we show our current results in the field of freeform shape measurements with the NPMM-200. For all measurements the auto focus probe was used. With the autofcus probe a 4.00 mm step height target calibrated by the PTB Braunschweig was measured at a comparable machine NPMM-200 at TU Ilmenau and with the machine at ITO. In both measurements a reproducibility better than 80 μm could be achieved. The height value given from the PTB was 4077903 ± 22nm (k=2), the mean value measured at the ITO was 4077905.09 nm.
The position control interferometers of the NPMM have a resolution of 20 μm. To show the high precision positioning capabilities of the machine we made a motion in 100 μm steps along the z-axis of the machine. The mean values of the position using a moving average filter over time are shown in figure 3. It is clearly visible, that movements down to only five times of the interferometer increment are possible with the machine. This allows to address 1e27 different positions in the complete measurement volume.
Measurement of aspheric and freeform surfaces
In the European project FreeFORM-15SIB01 thermo invariant measurement samples were developed, see fig. 4. With the NPMM-200 we made measurements of two samples: an aspheric and a freeform surface. The substrates both are made of Zerodur® within a magnetorheological finishing process. We measured these samples to investigate the measurement stability of the sensor-metrological frame arrangement. The autofocus sensor was used for these measurements.
The measured data is shown in figure 5. The aspheric surface has additional steps and valleys added with a peak to valley height of 7 μm. We have done a meander-type scan-path over a rectangular field on the specimen. The focus sensor was used in the closed loop positioning mode, i.e. the surface was kept in focus during the scan. The specimen was moved with a velocity of 6 mm/s, leading to a measurement time of 30 minutes. Along the scan line in x-direction data was acquired every 1 μm. The line spacing was 300 μm along the y axis. This results in three million data points per measurement. We have done three measurements of the substrate to do analysis on the reproducibility of the sensor-machine combination. The machine follows the path with a mean deviation of 4.8 nm. To compare the three measurements, we linearly interpolated the three measurements to one
common grid along the x and y axis. In the following, we define a so called virtual reference surface, following the methodology of Schachtschneider.11 The virtual reference surface (VRS) is calculated as the mean value of all three measurements at every position. The mean differences between the measurements of the freeform surface and the VRS are shown in table 2. The measurement time for the three measurements was one and a half hour. The mean values are in the Angstrom regime, which is showing the high stability of the metrological frame.
Mean differences to the virtual reference surface
|Measurement number||Mean difference to virtual reference [Å]|
To show the capabilities for measurement freeform surfaces with higher surface deviations we measured two samples generated for the UPOB12 round robin test. The freeform surfaces are polynomial surfaces up to 4th order. The surface peak to valley difference is 200 μm for the milder freeform and up to 800 μm for the so called wild freeform. A photograph of the milder freeform surface is given in figure 7 on the left. The measurement data is shown in figure 6. The measurements were done in a meander scan with step sizes of 1 μm along the scanning range and 250 lines over the diameter leading to 5943100 measurement points. We moved our sample with 2.5 mm/s, the measurement time was around 50 minutes. The length of the scan line is adapted to the clear aperture of the circle at the different position on the diameter-axis. The scan path is shown in figure 7 on the right. For evaluation we subtracted a best fit polynomial from the measured data. The difference of the mild freeform to its best fit polynomial surface is shown in figure 8. The RMS difference to the best fit surface is 33.5 nm. In the residual error we can see marks from the production process.
In this contribution we have shown the capabilities of the large area nanomeasuring machine NPMM-200 at ITO. After a description of the working principle, we gave a short summary of the plentiful possibilities to use the high precision of the machine. We underline this with the measurement examples of precision positioning and surface shape measurements. We made measurements of a freeform and an aspheric surface and analyzed the reproducibility using the concept of a virtual reference surface. The difference between the individual measurements and the virtual surface are in the range of Angstroms. To show the measurement possibilities in freeform surfaces we measured more steep freeform surfaces with PV deviation up to 800 μm with a high point density. In the next steps we plan to integrate areal interferometric sensors for measurements of freeform surfaces with the NPMM-200.
This research was funded by the German Research Association DFG, Grant No. Os 111/44-1 and the EMPIR Project 15SIB01: FreeFORM from the European Union.