Optical interferometry is a well established technique widely used in metrology. Sensitivities of digital interferometers for precision measurements of surface profiles (at angstrom resolution) and displacements (at picometer resolution) can be considered to be at the frontier of present experimental possibilities.

The applications of interferometry in physical chemistry are based on a simple principle whereby the object of investigation (for example, a growing crystal surface, solution or gas flow, or particle assembly) is placed in an optical cell and included in one arm of a Michelson, Twyman-Green or Mach-Zender interferometer.

The distribution of the interferometric fringes (phase map) corresponds to the surface profile and/or concentration or temperature gradients. Movement of the interferometric fringes gives information about the process dynamics. Interferometry can be combined with the other optical methods such as laser scattering, particle imaging velocimetry and fluorescence spectroscopy, complementing all these techniques.

Recently developed phase-shift interferometric systems (PSI) allow high vertical resolution of N/1000 (typical wavelength X = 633 nm) to be achieved with excellent repeatability and precision. The major problem is, however, the relatively long acquisition time (typically, about a second) for a typical PSI algorithm to be performed. This does not allow for fast dynamic measurements. For example, crystallization kinetics in highly supersaturated solutions require at least millisecond resolution in order to distinguish small morphological details (growth steps of 5 Å height). The typical time-scale of supercritical fluid expansion is smaller than one

microsecond. The same time-scale applies for the jet flows, combustion and plasma processes.

Thus, the principal problem is that it is not possible to combine the high vertical resolution of current PSI (A/1000-A/2000) with high temporal resolution (nanoseconds). In addition, in situ investigation into the dynamics of physio-chemical processes requires, firstly, special optical cells to study phenomena under high pressure and in high velocity flows, and secondly, new approaches in signal processing in order to provide real-time observation of surface morphology or flow gradients. The horizontal resolution of the interferometry is limited by the wavelength of light and is, typically, in the order of microns. This resolution is inferior to that of atomic force microscopy, the technique recently applied for in situ surface measurements. In many cases, however, the fast and non-invasive nature, and flexibility of interferometry are more important.

Up to date, there have been a few "instantaneous"

(i.e. high temporal resolution) PSI designs, which allow collection of several independent interferograms simultaneously. One approach is based on moir interferometer (see paper by M. Kujawinska, L. Salbut and K. Patorski: 'Three-channel phase stepped system for moir interferometry', Appl. Opt. 30 (1991) 1633) in which three independent interferograms were recorded using a single camera. This gave a certain advantage in signal acquisition (and lower system cost), but it also introduced problems with resolution and data processing.

Disclosed in papers by: (1) K. Onuma, K. Tsukamoto and S. Nakadate: 'Application of real time phase shift interferometer to the measurement of concentration field; J. Crystal Growth 129 (1993) 706; and (2) K.

Onuma, T. Kameyama, K. Tsukamoto: 'In situ study of surface phenomena by real time phase shift interferometry'; J. Crystal Growth 137 (1994) 610, is a system which used a real-time processing and, most importantly, applied the PSI system to in situ measurements of solution concentration and surface

crystallization kinetics. The disadvantages of this system lie in its optical design, which is based on three-camera acquisition. The polarisation plates, used for the signal decoding, appreciably decreased the signal-noise ratio. In this "instantaneous" interferometer, the acquisition (integration) time is defined by the camera frame rate (about 20 ms for the interlaced CCD). All of those methods were based on a continuous laser source and therefore unable to perform fast successive measurements.

Interferometry has a significant potential as a versatile, non-invasive, in-line analytical tool which allows different physico-chemical phenomena to be investigated and technological processes to be optimised for the needs of the industry. Any system in which phase borders and certain inhomogeneity are present can be investigated. The list includes surface crystallization kinetics in solution, gas, melt or gel; dielectro- and electrophoresis; transitions in colloid and liquid-crystal systems; flow dynamics and ballistics, combustion and plasma behaviour; mass- and

heat-transfer.

The present applicant's interest in the development of the novel PSI system lies in the area of crystallization technology, particularly that related to the production of solid organic pharmaceuticals from supercritical fluids and high-pressure solutions.

These pharmaceuticals are required for new, more efficient drugs which are designed for specific targeting options and controlled delivery. New crystallization methods have to be developed to manufacture the drug particles. These are, for example, rapid expansion of supercritical solution (RESS) and solution enhanced dispersion by supercritical fluids (SEDS), in which crystallization is produced by a rapid, uniform increase of supersaturation in SCF solution flow at pressures of 100-400 bars and very high, often supersonic, flow velocities. The crystallization can also be initiated rapidly by changing pressure (a few kbar) in a hydrostatic high pressure cell. The high pressure and high velocity conditions introduce very fast changes in

the solution bulk and on the growing crystal interface.

The fundamental physical chemistry of such processes (phase transitions, surface kinetics, diffusion mechanism, particle formation, etc.) have yet to be understood. From the industrial viewpoint, new crystallization techniques are able to afford fine, uniform crystal particles with engineered solid state properties without any post-crystallization mechanical treatment. In addition, the new technologies can provide a selective extraction of raw materials combined with highly efficient and environmentally friendly separation of organic substances.

Preferably, the main aim of this invention is to create an interferometric system for in situ applications in physical organic chemistry with emphasis on the investigation of crystallization phenomena under high-pressure and in supercritical fluids (SCF). This system will preferably be based on instantaneous phase-shift interferometry (PSI) and double-pulsed laser illumination. Preferably, the

interferometer will be able to obtain two successive high-resolution interferograms (A/1000-A/2000) within 5 nanoseconds and a digital processor will enable on- line real-time mode of operation.

Accordingly, in a first aspect the present invention provides a laser interferometer including: laser means for provision of two light pulses separated by a time interval; phase modulator means including means for: splitting each pulse into a pair of sub- pulses, each sub-pulse of each pair being directed along a respective separate path, directing the sub-pulses in one of the paths through or onto a sample to be investigated, and subsequently recombining the respective pairs of sub-pulses; and signal decoder means for directing the recombined pulses via respectively separate paths onto respective light sensitive devices or respective locations of a light sensitive device.

In this way, an interferometer is provided which can produce two interferograms which are temporally separated without the need to provide more than one camera and complicated moving optical parts. The signal decoder means produces a spatial shift between the two pairs of sub-pulses, i.e. produces two parallel interferometer beams.

Preferably the phase modulator means includes a combination of two interferometers, geometrically similar to Mach-Zender interferometers although other devices may be used, for example, a Michelson or Twyman-Green interferometer.

Preferably, the pulse modulator means includes means for shifting the polarisation of one sub-pulse of each pair of sub-pulses. This has the advantage that the subpulses can be recombined and directed together along a light path without interfering. In interferometers including such means, means are also provided for aligning the planes of polarisation of the recombined pairs of sub-pulses. Preferably the means

for shifting the polarisation includes a half-wave plate.

Preferably, the light sensitive device is a charge coupled device and a plurality of such devices may be provided. Preferably, both of the pairs of sub-pulses are included in the same exposure of the light sensitive device, i.e. there is no need to recharge the light sensitive device. Preferably the sample to be investigated is contained within an optical crystallisation cell.

In a second aspect, the present invention provides a method of performing interferometry including the steps of: (1) producing two light pulses separated by a time interval; (2) splitting each pulse into a pair of sub- pulses, each sub-pulse of each pair being directed along a respective separate path; (3) directing the sub-pulses in one of the paths through or onto a sample to be investigated;

(4) recombining the respective pairs of sub- pulses; (5) directing both pairs of sub-pulses via respective separate paths onto respective locations of a light sensitive device.

Embodiments of the present invention will now be described, by way of example, with reference to the accompanying drawings in which Figure 1 is a schematic diagram of an interferometer according to the present invention; Figure 2 is a schematic diagram of a second embodiment of a phase modulator unit for use in the interferometer of Figure 1; Figure 3 is a schematic diagram of the image processor of Figure 1; Figure 4 shows a second embodiment of the laser system; Figures 5 and 6 show an example of results obtained using an embodiment of the present invention.

The principal optical scheme is shown in Fig. 1.

The interferometer consists of the following modules: 1. Laser system. A dual-head Nd: YAG laser system (1) provides two laser pulses with the duration of e.g. about 5 ns and a variable delay time (e.g. as small as 1 ns) between the pulses. Such system is commonly used in particle imaging velocimetry (PIV) technique. The laser wavelength is preferably the second harmonic (green, 532 nm) of the residual infra-red beam. The polarisation of the two pulses may be orthogonal (achieved eg by rotating one of the laser heads or inserting a half-wave plate on one of the beams); this property being further used for discrimination of the two successive pulses (in the phase modulator). The double-pulsed laser beam is collimated and improved using a spatial filter (not shown in the figure). The delay/pulse generator (11) is connected to both the laser (1) and computer (12), providing timing/synchronisation of the whole optical system.

Alternatively, a single continuous laser can be used together with a beam chopper, as shown in Figure 4. Fig. 4 illustrates use of an eg He-Ne (633 nm) continuos laser beam 40 modulated by a beam-chopper 42 which can be, for example, an electro-optical or mechanical switch. A polarisation device 44 can be an electro-optical unit such as Pockels cell or mechanical unit such as rotating half-wave plate. Both devices are synchronised with each other to provide generation of the two perpendicularly polarised laser pulses.

2. Phase modulator. The optical layout of the phase modulator (2) is a combination of two interferometers, similar to Mach-Zender interferometers, defined by beamsplitters NPB1-PB1 and PB1-NPB2 correspondingly (the term "interferometer" is not entirely correct here because the orthogonally polarised beams do not interfere before they enter the signal decoder (3)). The phase modulator works as follows: the laser pulses (of both polarisations) are equally

divided by the non-polarising beamsplitter NPB1 into the reference beam (NPB1-M1) and the informative beam (NPB1-crystallization cell-M2).

M1 = Mirror 1 and M2 = Mirror 2. The half-wave plate (21) in the reference beam may rotates the polarisation by 90". The first half wave plate 20 adjusts the polarisation angle of both pulses to be 45" in respect of the optical axis of the device. This matches the intensity of the information and reference beams when recombined.

Then, the informative and reference waves of each pulse acquire a phase difference (defined by the chemical process in the optical cell) and arrive orthogonally polarised at the point of polarising beam splitter PB1. Their further optical path depends on the initial polarisation of the laser pulse as illustrated in Fig. 1. objective lenses 22,23 help produce a sharp image. If the initial pulse has its polarisation parallel to plane M1- NPB1-M2 (dotted arrows), both reference and informative waves are recombined in PB1-M3-NPB2

direction due to polarisation properties of the beamsplitter PB1. Accordingly, if the initial pulse had its polarisation perpendicular to the plane M1-NPB1-M2 (solid arrows), the reference and informative waves are recombined in PB1-M4-NPB2 direction. The mirrors M3 and M4 are shifted to produce a lateral displacement of the two successive laser pulses of different polarisation (see Fig. 1).

The first part of the system (NPB1-PB1) can be substituted by another phase-shifting device, for example, a Michelson or Twyman-Green interferometer as required.

3. Signal decoder (3). The function of this module is, firstly, to introduce a fixed phase shift between the informative and reference waves and, secondly, to align the polarisation vectors of these waves (s- and p-components) in the same direction of interference. The s-components and p-components interfere respectively. A laser

beam, exiting from the phase modulator (non- polarising beamsplitter NPB2) passes through the half-wave plates (31,32) which rotate the beam polarisations by 45". This rotation orients the beam so that the polarisations are at 45" to the optical axes of the polarising beamsplitters PB2 and PB3. PB2 analyses the p-components of the beam in one arm and s-components in the other arm.

An additional quarter-wave plate (33), between NPB2 and PB3, delays the phase of one of the orthogonal polarisation by 90". PB3 analyses these polarizations. Two signals produced by PB3 are sinusoidal and 1800 out-of-phase to each other. These signals are also 90" out-of-phase with those signals from the first beam half (PB2), therefore four interferograms, with 900 phase shift, are produced for each of the two initial laser pulses. The two displaced interferograms on each CCD sensor correspond to the different laser pulses. These four CCD sensors could be replaced

by a single CCD sensor.

This decoder may be a known optical unit.

4. Image processor (4). The interferometric pictures obtained are analysed using a 3+3 phase- shift algorithm, is explained more clearly in Fig.

3, which shows that the algorithm is able to correct for linear phase shift error. Simple 3- and 4-step algorithms can also be performed using the same processor. The image processor (4) consists of digital interface and look-up tables (memory devices) to perform all mathematical operations in tandem. High-resolution, 12-bit format pictures from Images 1-4 (Ij to 14) (provided by digital, non-interlaced, synchronised CCD cameras) are read pixel by pixel (using a buffer memory or common pixel clock).

The calculated and averaged phase-shift map cp is transferred into the PCI frame grabber and computer (12), and further subjected to

mathematical treatment such as phase map unwrapping (production of a smooth picture of surface or refractive index profiles), then a graphical presentation (2-D profiles, 3-D profiles, colour picture of the gradients, topographic map), and determining the kinetic data (rate, slope and concentration gradients versus time) for a double laser pulse. The double-pulse acquisition procedure will be repeated (the repetition rate of the laser is about 20 Hz) and new statistics accumulated concerning the dynamics of the process over a longer time-scale (several seconds).

5. Optical crystallization cell. A special crystallization cells provides measurements of the high-pressure solutions and high-velocity flows.

The cell complies with the optical quality of the system in Fig. 1. a. A supercritical fluid (SCF) crystallization unit consists of a steel cylindrical tube supplied with two co-axial glass/silica high-quality optical windows,

designed for pressures up to 400 bars. The entire tube is constructed on a modular principle in which the tube section with windows can be mounted in any position along the SCF flow (about 600 mm long). As a result, the flow dynamics can be monitored along the whole flow length. The tube is connected, through a nozzle, to a SCF extraction system. When the crystal surface is to be observed, the surface under investigation will be mounted parallel to the cell window and the whole tube filled with a supersaturated SCF solution.

Figure 2 shows a second embodiment of a phase modulator (6) for use in place of the phase modulator 2 of Figure 1. This phase modulator is for use where the sample to be investigated is not transparent and therefore the optical crystallisation cell 5 of Figure 1 is not appropriate. In the phase modulator 6 a reflective sample cell 61 is used instead.

The difference of the optical layout in Fig. 2 from that of the transparent mode of operation (Fig. 1) is that the informative laser beam, after being

produced by beamsplitter NPB10, is reflected back (from the surface under investigation) and then recombined with the reference beam (mirrors M10-M20) at the polarising beamsplitter PB10. The half-wave plate in the reference beam rotates the polarisation by 90".

Thus, the informative and reference waves of each laser pulse acquire a phase difference defined by the surface morphology of the sample. The further optical path (PB10-NPB20 phase decoder) is the same as for the transparent mode of the device (Fig. 1).

Figs 5 and 6 show an example of results obtained using the present invention. Figure 5 shows interferometric pictures of carbon dioxide flow obtained by depressurising liquid CO2 (at pressure about 55 bar) into the air. The polarising interferometer was arranged in the Mach-Zender configuration. The four pictures differ by the phase shift which is 0, n/2, n and 3n/2. The left and right part of each picture are time-resolved (time interval 5 ms) and show a fluctuation of the jet caused by precipitation of dry carbon dioxide in the nozzle. The

right part of the figures also shows a development of density:fluctuations within the jet (seen as dark and light fringes along the axis).

Figure 6 shows the result of phase-wrapping calculations using a four-step phase shifting algorithm: #(x, y) = tan-1[(I4-I2)/(I,-I3)] (1) where Il, 12, 13, 14 are the intensities of interferograms with the shift 0, n/2, n and 3n/2 correspondingly (see Figure 5) and (x, y) is the phase difference between the informative and reference beams.

Figure 6 shows a phase-wrapped picture obtained using interferograms in Figure 5 applying a four-step phase shifting algorithm.

The above embodiments of the present invention have been described by way of example only and various alternative features or modifications from what has been specifically described and illustrated can be made within the scope of the invention, as will be readily apparent to persons skilled in the art.