Introduction

The study of micro and nano surface metrology is becoming commonplace in industrial and research environments as structures and surface features become smaller and smaller. Scanning interferometry is becoming increasingly important in metrology analysis because of a variety of factors including;

·         Non-destructive measurement - no sample contact or preparation is typically required

·         Accurate and quantitative surface characterisation

·         Fast and convenient sample loading and set-up

·         Capability of measuring a wide range of materials

·         High resolution, highly repeatable measurements

·         Fully automated measurement – ideal for process control

·         Roughness and step height analysis within a single measurement

·         Possibility of surface coating measurement – film thickness and real surface roughness measurement

It can address many of the challenging measurement problems that exist when studying samples at the micro and nano level. These include the measurement of critical dimensions, heights, angles, surface roughness, solving etch rate/time problems, measuring stress gradients, etc.

 

(Figure 1: A typical interferometry measurement of a MEMS device)

The technique offers nano/micro height measurements and sub-micro/micro/millimetre lateral measurements with high spatial resolution. The combination of a large field of view and high lateral sampling (>1,000,000 measurement points) allows it to offer a broad range of measured bandwidths for a single measurement. (Figure 1)

  Optical interferometry - theory

The use of optical techniques to study surfaces has been very popular for a long time and there have been many developments over the years to improve the accuracy of the metrology.

Interferometry has now become a very important metrology tool due to the introduction of modern electronics, computers and software to the basic technique.

A schematic of a scanning interferometer system is shown in Figure 2 (below). The upper beam splitter directs light from the light source towards the objective lens. The lower beam splitter in the objective lens splits the light into two separate beams. One beam is directed towards the sample and one beam is directed towards an internal reference mirror. The two beams recombine and the recombined light is sent to the detector. When the path length to the sample and the reference are the same interference is observed. The detector measures the intensity as the interferometric objective is scanned in Z.

(Figure 2: A schematic diagram of a typical interferometer setup)

The detector takes a series of snapshots as the sample is measured. This gives as series of images of the intensity map of the light being reflected from the surface. The area of interest for the measurement is the interference area. These intensity images are used to create a 3D image of the surface being measured. Different techniques are used to control the movement of the interferometer and also to calculate the surface parameters. The accuracy and repeatability of the scanning white light measurement are dependent on the control of the scanning mechanism and the calculation of the surface properties from the interference data.

Some of the areas where there have been changes in the hardware and software capabilities in recent years are discussed below:  

Step Height Measurements

Scanning interferometry is a powerful technique for the measurement of step heights. The sub nanometre vertical resolution and high spatial resolution lead to a highly detailed 3D representation of the specimen under test. This 3D information can be used to obtain very accurate and very repeatable step height information. The added advantage of a full 3D representation of the surface is the ability to measure the whole of the step and look for any variation in step height along the step edge.

One of the recent advances in step height measurement is the ability to obtain accurate step height and roughness in a single measurement. Previously the measurement of large step heights involved long measurement times or high speed measurement at lower z resolution. The advent of long range closed loop scanners have meant that it is now possible to use z stitching to accurately measure the surface of the bottom step and then move to the top surface and accurately measure the surface at the top of the step. The two measurements can then be combined to give an accurate step height measurement and sub nanometre roughness measurement

Roughness Measurements

Roughness is an important parameter for the control of the sample properties. Various roughness ranges are normally studied in order to define the overall properties of the surface and one of the limitations to the analysis is the bandwidth of the measurement method. Varying the size of the field of view makes comparison of datasets difficult because it normally involves changing the pixel size and it is not trivial to compare roughness data with varying pixel sizes. The X, Y bandwidth of interferometry measurements is normally restricted by the field of view of the optics used and the number of pixels of the camera. Therefore moving from a camera with 76 800 (320 x 240) pixels to a camera with > 4 000 000 (2 048 x 2 048) pixels leads to a large improvement in the range of roughness information that can be obtained. Another consideration is the dimensions of the camera, some analysis require a set of data that is square so a 320 x 240 camera actually gives a dataset that has 240 x 240 usable pixels this will give much less information than a 2 048 x 2 048 pixel camera.

Coated surfaces

Accurate analysis of coated surfaces is important for many samples. As the effects of films coatings on interferometry measurements are now becoming more widely understood there is growing interest in the measurement of thin film surfaces and thin film thickness

Thick Film Analysis

The thickness and real surface roughness of films can be obtained by measuring a step or by measuring through the film to the substrate.  Because of the λ/4 limit of PSI it is often not suitable for film thickness measurements and some form of scanning interferometry is used. Because of the high z resolution of the Coherence Correlation Interferometry (CCI) technique it is an ideal technique for measuring film thickness.

Measurement of the top surface of semi-transparent surfaces can also be a problem. Normally measurement of the roughness of top surfaces of semi-transparent films is complicated by the signal coming from other surfaces.  If possible the problem can be solved by setting the scan length so that only the top set of fringes are covered by the scan and the signal from the bottom set of fringes is not seen.  This can be a problem if the surface is thinner than about 4 microns or if the measurement is part of an automation process. In order to overcome this problem it is possible to instruct the software to just use the signal from the top surface alone. It is also possible to measure other surfaces by specifying which of the fringe envelopes is to be studied.

  For thinner films, the fringe envelopes may overlap. In this case, an intensity-based threshold can be applied to separate the two sets of fringes. This means that thicknesses below 2 microns can be measured with the same accuracy as the thicker films shown above. However as the fringe envelopes approach each other they will start to deform the overall shape of the waveform and therefore the thin film measurement.  .

Thin Film Analysis

Conventionally, thin film structures are measured using either a spectrophotometer or an ellipsometer but there is now a growing requirement for measurement of thin film structures that only exist on a local scale.  Using interferometry it is difficult to easily study the signal from the top of the thin film and the signal from the substrate if the film is less than about 1.5 microns because of the fringe overlap.  Recent studies have shown that through the development of the HCF function, it is possible to extend coherence correlation interferometry to include the calculation of layer thicknesses from this type of sample.

 An additional benefit is that, through the knowledge of the thin film structure, the phase-change on reflection may be compensated so that ‘true’ step-heights may be obtained

Interfacial Surface Roughness

 

(Figure 3: Thin film data obtained from a single measurement)

The measures surface roughness of a thin film sample is often a composite roughness of the top surface and the bottom surface. This can lead to large errors in the true surface roughness of the top surface. Because the HCF can be used to obtain separate information from the different layers it can be used to analyse the surface parameters for the interfaces, including the air - top surface interface.  This means that it is possible to measure the true roughness of the top surface of a thin film sample where the film thickness is greater than about 20 nm. It is also possible to measure the interface roughness. (Figure 3)

Conclusions

Scanning interferometry is a very powerful tool for micro and nano surface metrology. It has the ability to obtain step height, roughness and lateral dimension information from a single non – contact measurement. Other important advantages include the high speed and high repeatability of the technique. Recent developments such as long range closed loop scanners; dynamic measurements and coating analysis are increasing the usefulness of the technique.  Article written by Dr. Mike Conroy (CEMMNT Partner)