CA2294185A1 - Ablation profiler - Google Patents

Abstract

A calibration system uses error data from an ablation profiler or from a topographer to adjust the operating parameters of a laser ablation system. An ablation profiler, useful in calibrating systems that perform vision correction by ablating corneal tissue, is based on a scanning white light interferometer (SWLI) system. The profiler measures a sample (14) of material that has been ablated and maps the contours of the sample (14) to compare it with a database of desired contours. The SWLI system includes a novel method of controlling the movement of the SWLI's reference surface (22) using an inexpensive electromagnet assembly with a capacitive feedback control system, a pin-hole light source to improve interference patterns, and off-axis positioning of the light source and camera to avoid specular reflections. The system also includes a method of preparing the transparent sample for measurement by applying a rear-surface opaque coating and using front-surface illumination. Alternately, samples can make use of roughened surfaces to provide the required optical properties.

Description

ABLATION PROFILER
BACKGROUND OF THE INVENTION
1. Field of the Invention The invention relates to a profiler and more particularly to a surface profiling system.

2. Description of the Related Technolo,~y In recent years it has become possible to perform vision correction through various surgical procedures on the cornea of the eye. One of these procedures involves the use of an excimer laser to change the shape of the cornea by ablating tissue from an exposed surface. The process allows a surgeon to reshape a patient's cornea and minimize scars, without the need for invasive mechanical incisions. The effectiveness of this procedure is enhanced if the accuracy of the ablation equipment is caiibrated and maintained within very small tolerances. Unfortunately, present equipment for maintaining this calibration is expensive, making it prohibitively costly for many doctors to have in-house calibration. This expense might also create economic pressure to calibrate less often than otherwise optimal and thereby decrease the accuracy of the vision correction procedure.
One method used to test laser system operation is the ablation of polymethylmethacrylate ("PMMA") samples. The laser system ablates a lens like shape into one side of the PMMA test sample. The sample is then measured with a lensometer to determine the corrective power. (Note: The terms "sample" and "test sample" are used interchangeably throughout this . disclosure to denote the object whose surface contours are being determined.) This ablated shape is compared with a theoretical power to determine the quality of the ablation. This type of test is not sufficient to be used to enhance the accuracy of the procedure. To gain wide acceptance, the WO 99/0.1716 PCTNS98/13539 instrument for performing these test functions should be relatively small, portable, inexpensive and accurate, but prior to the current invention there was no known instrument on the market that would meet all these criteria.
SUMMARY OF THE INVENTION
It is an object of the invention to provide an ablation profiler to analyze the shape of ablated regions on a test sample.
It is a further object of the invention to provide an ablation profiler to generate data representing a three-dimensional map of the surface of an ablated test sample.
It is a further object of the invention to provide an ablation profiler to generate data which can be compared with theoretical data to determine the accuracy of the ablation.
It is a further object of the invention to provide an ablation profiler that can resolve surface features smaller than one micron.
It is a further object of the invention to provide an ablation profiler that is relatively small, portable and inexpensive.
It is a further object of the invention to provide a method of preparing a test sample to be ablated and subsequently measured by an ablation profiler.
It is a further object of the invention to decrease the smoothness of the test sample ablation.
It is a further object of the invention to provide a method for making a test ablation in a test sample by deactivating the exhaust suction of a surgical laser system in order to decrease the smoothness of the test sample ablation.
It is a further object of the invention to utilize the output of the ablation profiler directly, indirectly, or as feedback to adjust the operating and optical parameters of the excimer laser system to enhance the accuracy of the surgical procedures.
It is a further object of the invention to utilize the output of a corneal topographic map directly, indirectly, or automatically to adjust the parameters of a surgical laser system to enhance the results obtained.
It is a further object of the invention to adjust the iris in the beam path of the profiler to an optimum so as to increase the size of optical speckle and to optimize the speckle spot size for a single sensor pixel.
The invention uses a Scanning White Light Interferometer ("SWLI"), a device known in the art. Although the principles of a SWLI are well known, a brief description of its operation is given here to aid in understanding the invention.
As with other SWLI devices, the interferometer in this ablation profiler operates by using a beam splitter to split the light from a single light source into two parts, one part being directed onto an optically flat reference surface and the other onto the test sample being measured. The light scattered back from these two subjects is then combined by the beam splitter and directed onto an array of sensors which measure light intensity. The optics of the system are configured to assure that each sensor receives the light from a specific, unique location on the sample.
When considering the operation of a single sensor, it is obvious that the light it receives is a combination of light that has traveled along two separate optical paths. One path includes the reference surface while the other path includes a selected spot on the sample being measured. For most observations, the corresponding components of the two returning beams are randomly out of phase and will combine to produce rapidly fluctuating random noise which averages to a near constant single sensor output. But if the two path lengths are identical to within approximately one micron, each beam is essentially a duplicate of the other, so that their corresponding ' 30 components will constructively or destructively interfere with each other to produce a noticeable intensity change. This intensity change signifies that the observed spot on the sample is the same distance from the beam splitter as is the reference surface. When many such observations are made, each corresponding to a different location on the sample, the result is an image of those points on the sample that are at a particular elevation, similar to an elevation line on a terrain topographic map. This image is stored in the system as a data frame. A typical data frame is a 640 x 480 array of pixels that represent a corresponding rectangular area on the sample.
After storing the data frame, the reference surface is moved a very small distance closer to (or farther from) the beam splitter so that a new elevation on the sample is identified by a new image pattern. This new image pattern is saved as a new data frame. By moving the reference surface through a series of such incremental steps, multiple layers of elevation data are obtained. These layers are then processed to produce a complete elevation map of the selected surface area on the sample. By limiting each incremental step to a fraction of the average wavelength of the light being used, the resulting elevation map can be accurate to within a fraction of a micron. The significant movement is relative movement of the sample and reference. Alternatively, the sample can be moved and the reference held stationary, or both may be moved.
This is a brief and somewhat oversimplified description of the operation of a SWLI. A fuller description, along with some of its theoretical underpinnings, is described by Dresel, Hausler, and Venzke in the article "Three-Dimensional Sensing of Rough Surfaces by Coherence Radar", Applied Optics, Vol. 31, No. 7, March l, 1992. That article is incorporated herein by reference in its entirety.
Earlier versions of interferometers used a small number of sensors, multiplying their area of coverage by using mechanical motion to scan the area of interest. This complicated the mechanism and created additional sources of inaccuracy. Modern CCD cameras have provided an economical solution to this and other problems by providing a large rectangular array of sensors for simultaneous measurements of the entire region of interest, the optics for focusing each sensor on a specific portion of the sample, and the electronics to feed this information to a computer. The term 'scan' is still used, but it no longer refers to a mechanical surface scan. It now describes . a mechanical stepping of the reference surface along the optical axis. This so-y called z-axis scan should not be confused with the xy-axis scans of the earlier technology.
The invention contains a z-axis scanning mechanism, or translator, having several hundred microns of travel with movement resolution of approximately 1/10 of a micron. This translator is used to incrementally move the reference surface along the optical axis. The very small increments of travel permitted by this mechanism allow the profiler to reliably and accurately measure the shape of ablated regions on the PMMA sample. It is also possible to perform z-axis scanning by incrementally moving the sample instead of the reference surface, but the preferred embodiment moves the reference surface. To obtain accurate results, the device being stepped must be rigidly mounted to avoid unexpected mechanical shifting between each step. This requirement for rigidity is somewhat incompatible with the requirement for quick and easy insertion/removal of the target samples. By contrast, the reference surface does not have to be inserted/removed. It can be rigidly and permanently mounted on a z-translator without compromising other goals.
To satisfy the requirement of long travel with high resolution, the translator motion is controlled with a magnet forcer assembly made up of a ring-shaped magnet surrounding a cylindrical magnet, one or both of which is an electromagnet. Advantageously, these components may be of the type used in a conventional audio speaker. This results in a very low-cost alternative to the expensive piezoelectric crystals commonly used to control motion in interferometers. Such crystals also have a very limited range of travel, a problem not encountered with the magnet forcer. The forcer's - 30 extended travel range usually eliminates the need for precise vertical adjustments of the sample after it has been mounted.
_5_ One of the magnets of the forcer assembly is attached to the translator while the other is attached to the profiler frame. In a preferred embodiment, the electromagnet is attached to the translator so that its lighter mass won't create undue momentum when the translator is moved. By accurately controlling the current through the electromagnet, the translator position can be changed in very small amounts. A feedback mechanism is used to accurately measure actual movement of the translator by using a capacitive displacement sensor. This feedback causes an error-correcting adjustment in the current to the electromagnet, allowing very precise positioning performance from these inexpensive components. The sensor has a displacement resolution of approximately 0.02 microns.
In addition to the measurement device itself, a unique method of preparing the test samples is used so that the light scattered from the ablated surface accurately models the contours of the sample and is sufficient to achieve interference signals exceeding the noise Level. To satisfy the requirement of obtaining consistent uniform scatter from the ablated surface the following technique may be used: The test ablation may be performed on one side of a PMMA sample. Next the sample is inserted into the SWLI in such a manner that light from the light source is incident on the surface of the PMMA sample. If a clear PMMA sample is used, the light can enter the sample through the unablated side. Measuring the ablated surface through the back side of the sample is useful for measuring very smooth ablated surfaces. For this technique a highly opaque coating may be applied to the surface to be measured. The incident light then enters the sample from the uncoated side, passes through the material until it strikes the interface between the sample and the opaque coating, and is returned back through the sample where it exits from the front surface. This technique allows the shape of the ablated PMMA surface to be measured without being affected by the thickness of the coating applied to the sample. An alternate technique roughens the surface of the sample so that it scatters light rather than transmitting it. The surface can be roughened after ablation or more advantageously prior to ablation. One method is to use fine grit sandpaper or polishing compound (600 or finer) to lightly buff the sample. This is similar to the effect achieved with a ground-glass surface, which transforms a normally transparent material into a semi-opaque one. Various techniques are available for creating the proper degree of roughness to achieve a usable scattering effect.
As shown in Figure 5, error profiles can be used to adjust the operating parameters of the ablation system (200), resulting in more accurate ablations.
Error profiles from an ablation profiler (201) are based on test samples, while error profiles from a topographer (202) are based on actual corneas.
Profile data from one or more test samples can be compared with a theoretical profile to determine if the ablation device is performing within acceptable tolerances and is consistent from sample to sample. Deviations from the theoretical profile produce an error profile, which is a set of data defining the amount and location of the errors. This error profile can be used to recalibrate the ablation system, thus reducing future errors as shown by the error profiles of subsequently produced samples. In a preferred embodiment, this error profile can be transmitted from the profiler to the ablation system over a communications link, whereupon the ablation system automatically recalibrates its settings to eliminate this error from subsequent samples.
Alternately, the ablation system can be adjusted according to the results of a topographer. After a cornea has been ablated and allowed to heal, its shape is measured by using a topographer, a machine known in the art.
By comparing the actual corneal shape with the intended shape, a corneal error profile can be produced containing data on the differences. This corneal error profile can be transmitted to the ablation system for automatic adjustment, resulting in more accurate subsequent corneal ablations.
Both of these corrective mechanisms - feedback from the profiler and feedback from the topographer - can be used to overcome the variability which is inherent in current laser systems. Automatically providing this feedback and recalibrating allows accurate and immediate correction, without the risk of human error which is inherent if the data is manually transferred.
Although the invention was developed to fill a need in the field of corneal surgery, the application of the invention is not limited to that field.
It can also be used to accurately measure the surface contours of any sample which has the proper size, shape, and reflectivity characteristics, regardless of the sample's purpose, material composition, or method of creation.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 shows a block diagram of the overall ablation profiler.
Figure 2 shows the mechanical components of the ablation profiler.
Figure 3 shows a perspective view of a translator in the ablation profiler, including a magnet forcer assembly.
Figure 4 shows a logic diagram of a translator motion control circuit.
Figure 5 shows a block diagram of the overall calibration system.
DETAILED DESCRIPTION OF THE INVENTION
Overall S sY- tem Figure 1 shows a block diagram of the overall system. Frame (33) provides a mount for most of the major components, including a light source (1) and a mirror (9) for directing a beam of light from light source (1) onto beam splitter (12). Beam splitter (12) reflects a portion of the light onto a test sample (14). The beam splitter passes an approximately equal portion of the light onto reference surface (22). Both sample (14) and reference surface (22) scatter part of the incident light back to beam splitter (12).
The distinction between "reflect" and "scatter" is significant in this application. The term "reflect" implies the incident light is returned in a _g_ predictable manner in which the angle of reflection is equal to the angle of incidence. The operation of a mirror exemplifies this process. "Scatter"
implies that incident light is returned in many random directions simultaneously, with an approximately even distribution across that range of directions. The light returned from a rough piece of wood is an example.
For a SWLI to work properly, the sample and reference surface must each scatter light to achieve uniform intensity across their respective surfaces.
If they reflect instead of scatter, specular reflections or 'hot spots' will develop in some areas, causing the associated sensors for those areas to saturate and degrade the interference process. By contrast, the beam splitter must reflect (and pass) light without appreciable scatter.
The portions of light scattered from test sample (14) and reference surface (22) back to beam sputter (12) are combined by beam splitter (12) and directed to camera (45). Each of these two optical paths has a path length.
For those points on the sample that result in the two path lengths being identical, this combination results in a pattern of wavelength interference, or "fringe pattern", that is detected by the receptors of camera (45). This image is read from the camera and stored in a memory. Translator (24) is then moved a short distance, causing reference surface (22) to be farther from or closer to the beam splitter. This changes the path length of one of the two light paths, which changes the fringe pattern seen by camera (45). By repeating this step-read-store sequence as the reference surface is incremented along the optical axis, a series of fringe patterns can be obtained which are then processed to determine the physical contours of the sample.
The electronics for this operation are distributed primarily in electronics section (50), microcontroller (100), and system computer (51), each with sufficient memory to perform its functions. In a preferred embodiment, the electronics section (50) provides direct control and sensor functions, microcontroller (100) provides decision-making functions for electronics section (50), and system computer (51) provides overall system direction, computational facilities, and user interface. However, other functional _g_ distributions are feasible, depending on the specific design criteria of the particular system. This design choice is well within the skill of a typical systems designer.
Device Components Figure 2 shows most of the major components of the device, excluding some of the electronics. A lamp (1) having filament (2) and lamp base (3) is held in place by lamp mount (4). Some of the light (2a) emitted from lamp (1) is collected by lens (5) which is held in place by lens holder (6). This lens images light (2a) onto aperture (7), which is a small hole in an opaque plate, the aperture being held in place by aperture holder (8). The light (2a) is then reflected by mirror (9) which is mounted on mirror mount (10). The lens mount (6), aperture holder (8), and mirror mount (10) are held in place by bracket (11). The light (2c) reflected from mirror (9) is directed into beam splitter (12) which is held in place by splitter mount (13). The beam splitter (12) reflects a portion of the light (2c) onto a sample (14) which is held in place by retainer (15). The samples are held against locating surfaces on the retainer (15) by one or more ball-and-spring combinations (16,17). The springs (17) are held in place by spring retainer (18). The sample (14), retainer (15), and spring loaded balls (16) are mounted on an adjustable gimbal (19). The vertical position and horizontal level of the sample can be adjusted by three 80 thread-per-inch screws (20) (one of the screws cannot be seen in Figure 2 because it is obscured by another screw). The screws (20) may have ball tips which rest in supports (20a) (only one is illustrated). The gimbal (19) is held against instrument base (21) by 4 springs (not shown).
The portion of light (2c) transmitted by beam splitter (12) strikes reference surface (22) which is mounted on the reference plate (23). The reference plate (23) is screwed to the translator (24) through holes in one of the flat springs (25), the screw also serving to clamp this flat spring to the translator (24). The upper ends of the flat springs (25) are clamped at locations (26) to the frame top (27). The reference surface is perpendicular to the optical axis of the light traveling between the beam splitter and reference surface. Movement of the reference surface is parallel to this axis.
Mounted to the left side of translator (24) is cylindrical electromagnet (28) and its mounting plate (29). Mounting plate (29) is screwed to translator (24) through holes in one of the flat springs (25), the screw also serving to clamp this flat spring to the translator (24). The cylindrical electromagnet extends into the center of ring-shaped magnet (30) which is held by clamp (31) to the left end of the frame (32). The left end of the frame (32) and frame top {27) are bolted to the frame sides (33) which are bolted to the instrument base (21). The electronics used to provide a regulated current to cylindrical electromagnet (28) may be located in region (42).
A capacitor plate (35) is physically attached to the frame. A second capacitor plate (34) is physically attached to the translator (24). Both capacitor plates are electrically isolated from the translator and frame.
Electrical wires from capacitor plates (34) and (35) are connected to electronics which are located in the region (41). The two capacitor plates are parallel and closely spaced so as to form an electrical capacitor. The electronics located in region (41) together with the capacitor plates (34, 35) form a capacitive displacement sensor which is used to measure the position of the translator.
Since vibration within the profiler can destroy the fringe patterns, a damping cup (37) is used to minimize vibration caused by the stepping motions. The damping cup (37) is attached to the frame. Inside this cup are damping fins (38) which are mounted to translator (24) at location (39). The cup is filled with a viscous fluid which is sealed into the cup by a diaphragm (40).
Some of the light which is scattered from sample (14) and reference surface (22) is directed by the beam splitter through iris (43), where it is focused by lens {44) onto the sensors of CCD camera (45). The camera (45) may be held between the frame sides (33) by an adjustable camera mount (46). The camera (45) may be connected by a ribbon cable (4?) to the camera electronics board (48) which is mounted between the frame sides (33) by card mount (49).
Figure 3 shows a simplified perspective view of the translator. This figure shows how the translator (24) is suspended from the frame top (27) by flat springs (25). The flat springs (25) are clamped at the top at attachment points (26). The flat springs (25) are clamped at the bottom by the reference plate (23) and magnet mounting plate (29). Cylindrical electromagnet (28) is mounted to the plate (29) and is positioned coaxially with the ring-shaped magnet. (30). The cylindrical electromagnet (28) extends into the central opening of ring-shaped magnet (30).
Optical Desien The optical design of the interferometer can be seen by referring back to Figure 2. A quartz halogen lamp may be used as a light source. Some of the light from this lamp is collected with a lens and imaged onto a small hole.
From this hole a cone of light expands and is directed into one port of a beam splitter by a mirror. The beam splitter divides the light into approximately equal parts and directs one part to the test sample and one part to the reference surface. The test sample and reference surface each scatter a portion of the incident light back to the beam splitter. The beam splitter combines the light from the sample and the light from the reference into a combined beam of light which is directed to a camera lens having an iris stop.
The camera lens images this combined beam onto the pixel sensors of the camera.
The spectral width of the light source is approximately 150 nm. The light is passed through the small hole to increase the spatial coherence properties of the light source. To determine the size of the hole, a CCD pixel size may be mapped to a corresponding region on the sample surface. This region on the sample is effectively an optical receiver. The hole size is advantageously made smaller than the resolving power of the effective receiver. This resolving power is determined by the size of the sample region and the distance from the sample to the hole. It is possible to obtain interference without the hole but the amplitude of the fringes would be reduced.
. The mirror is used to provide better access to the sample region and it also provides a convenient mechanism for adjusting the incidence angle of the white light onto the sample. As seen in Figure 2, both the white light source and the camera are tilted away from normal incidence on the beam splitter. Adding this tilt has the advantage of rejecting light that is spectrally reflected from the smooth surfaces of the beam sputter.
The iris in front of the camera is made sufficiently small so that the camera optics cannot resolve details on the sample which are smaller than the effective receiver size described earlier. Making the iris small reduces the amount of light collected by the camera lens but it also increases the speckle size on the CCD array. If the speckle size is smaller than a camera pixel the fringe contrast is reduced making the fringes harder to detect. To reduce the effects of oversaturation due to reflected light, the system may include elements to alter the angle of incidence of the light. This may be done with a second light source or by changing the location and/or angle of the optics transmitting light from the source to the beam splitter. By considering results obtained separately by using light at two or more angles of incidence (at any one time illumination from only one angle of incidence is used), it is possible to reject the vast majority of oversaturated pixels in favor of corresponding pixels obtained with light from the other angle. Where neither angle results in an oversaturated pixel, the data from both can be averaged or one angle can be used in favor of the other. Where light from both angles results in oversaturated pixels, that data point can be rejected and either ignored or data from adjacent data points can be extrapolated. Other error . correcting or compensating schemes may be used without departing from the spirit of the invention.

Target Sample and Reference Surface To be an effective tool for use in connection with surgical lasers, the sample should be made from a material which is consistently uniform in the ablation process for the laser system being tested. For an excimer laser operating at 193 nm, PMMA works well even though it will not in general have the same ablation rate as corneal tissue. The ablation rate must therefore be converted when calibrating the laser ablation instrument. This calibration factor may vary slightly depending on the details of the laser system used or the material of the test sample.
PMMA is relatively transparent at the frequencies used in a typical SWLI. Therefore the incident light on a smooth uncoated PMMA sample will not be scattered back to the beam sputter but will pass through the sample.
An opaque coating can be used on the ablated surface to provide the proper scattering qualities. Unfortunately, if front-surface scattering is used the variable thickness of the coating can distort the apparent contours of the sample. Front-surface scattering means the coating is on the illuminated side of the sample, so that light is scattered from the surface of the coating without ever reaching the sample itself. The problem caused by a variable-thickness coating is overcome by using rear-surface scattering, in which the coating is on the opposite side from the illumination, so that light must pass through the sample before reaching the coating. In this method, the incident light strikes the non-ablated side of the sample, passes through the PMMA
material and strikes the coating at its interface with the ablated surface before being returned through the PMMA material and exiting from the non-ablated side. Since this interface conforms closely to the ablated surface, accuracy is maintained and the thickness of the coating becomes irrelevant.
There are other criteria, however. This coating is more effective if it satisfies several requirements:
1) The coating should scatter sufficient light to be measured by the camera. For example a silver colored coating will typically scatter sufficient light but black paint will not.

WO 99/01716 PC'f/US98/13539 2) The coating should conform closely to the ablated surface. If the coating does not conform to the ablated sample due to surface tension, large flake size or other reason then errors will result when profiling the surface shape.
3) The coating should be highly opaque. If the incident light penetrates into the coating to a distance greater than or equal to one half of the coherence length of the light, the fringe visibility at the camera will be reduced to near zero. For best performance light should not penetrate the coating by more than a small fraction of the light's coherence length.
4) The coated surface should not generate a large specular reflection.
A large specular reflection will cause saturation regions in the image.
Therefore the coating surface should be rough enough to prevent specular reflection yet smooth enough to follow the ablated surface.

5) The coherence properties of the light scattered from the coating should not be significantly altered from the incident Light. For example, light which is absorbed by a dye and reradiated will lose all coherence with the incident light and therefore generate no interference fringes.
One acceptable coating method is to paint the surface with a substance containing metal flakes such as silver paint. This method is simple to use but does not give the best results. With this method a vertical resolution of approximately +/- 4 microns can be obtained which is sufficient for some purposes. The painting method will give a reasonable value for the depth and volume of an ablation and is sufficient for determining the symmetry of the ablation.
A second coating method is to use a non-electrical plating technique.
For this technique chemicals containing metal (silver or copper for example) are mixed on the ablated surface. This mixing can be done using a two-nozzle sprayer. When the proper chemicals are mixed, metal precipitates out of the solution and plates the surface. By controlling the metal concentration in the solution the scattering properties of the ablated surface can be optimized.
In an alternate approach, PMMA samples which have been lightly roughened can be measured without using a coating. A root-mean-squared (RMS) surface roughness of approximately 1/4 micron causes this normally transparent surface to scatter enough light to be useable in the profiler, but won't significantly degrade the surface contours. The samples can be treated with the roughening process before ablation, since enough of the surface roughness is preserved during the ablation process to give valid measurements. If only the ablated side is roughened, the sample can be oriented in the profiler for either front- or rear-surface scattering. A
uniformly-roughened surface can be achieved through a controlled chemical etching process. But usable results can also be obtained by manually roughening the surface with fine sandpaper or polishing compound.
The operation of the ablation system itself can also be modified to achieve roughness on the sample in addition to the described advanced preparation or without requiring advanced preparation of the sample. The ablation process creates a quantity of airborne particles, which are normally removed by a vacuum smoke clearance device. This removal process helps create a very smooth ablated surface, which may be desirable for ablating the cornea but undesirable for ablating a test sample. By disabling the vacuum smoke clearance device, the airborne particles randomly block tiny portions of the laser beam, resulting in a rough ablated surface that is suitable for scattering light in the SWLI.
Finally, the sample may have markings scribed or printed on it or some other mechanism for positioning the sample relative to the ablating laser beam. Markings allow the laser system making the ablation to consistently target the ablation region on the sample, and allow the profiler to check the accuracy of this targeting. These markings may also provide a mechanism for angular alignment of the sample with the reference axis of the ablating laser system. By accurately measuring the angular dependence of the ablating laser system, the orientation and symmetry of astigmatic treatments can be tested.
The reference surface may be an optically flat, highly scattering material that scatters the incident light back to the beam splitter without significantly changing the coherence properties of the incident light, and without causing excessive specular reflection. This reference surface can be fabricated by chemically etching an optically flat glass surface. The etched surface can then be coated with a silver spray or by vacuum metalization.
The roughness can be controlled by varying the duration of the chemical etching process. The result is a surface that looks white and has little specular reflection at near-normal incidence.
Mechanical Considerations Because test ablations can be up to 200 microns deep and because there is some tolerance in the positioning and thickness of PMMA samples, the translator must have several hundred microns of travel. In addition, the translator must be sufficiently insensitive to vibrations to prevent the interference fringes from being destroyed. It is also desired that the translator be simple and inexpensive.
In this design the above requirements are satisfied by using a magnet/
electromagnet combination of the type used in the voice coil of a standard audio speaker. By accurately controlling the current in the electromagnet the horizontal force on the translator can be incremented in sufficiently small steps to cause a movement of less than 0.1 micron. At the same time the drive current can be sufficient to cause translations of approximately 1 mm.
This large translation range allows sufficient tolerance in the positioning of the sample that the end user of the profiler does not need to adjust the sample height. The sample can simply be placed in the sample retainer and a computer controlled routine can determine the limits of the ablated region and then generate a profile of the ablation.

A damping pot may be included to reduce sensitivity to vibrations. This damping pot contains a viscous fluid to prevent oscillations from building up.
The pot itself is attached to the frame while internal vanes are attached to the translator. When the translator moves, the motion of the vanes through the viscous fluid provides a smooth, non-sticking resistance that effectively damps out vibration. It is important that the damping pot does not have significant non-viscous friction. This type of friction would cause the translator to stick and jump as force is applied.
Finally, the position of the translator is measured using a capacitive displacement sensor. A capacitor is formed by two plates, one of which is mounted to the translator and one is mounted to the profiler frame. As the translator moves, the spacing between the two plates changes, thus changing the capacitance of the capacitor. The capacitor may be connected in a loop with an inductor and a resistor to form a resistor-inductor-capacitor (RLC) oscillator. The resonant frequency of this oscillator changes as the capacitance is varied. The electronics of a translator motion circuit may measure the resonant frequency of the RLC oscillator as described below.
This sensor has a resolution of approximately 0.02 microns. Other methods may also be used to measure the capacitance and convert the measurement into distance.
In spite of efforts to provide all affected surfaces with optimal scattering qualities, ablated PMMA samples can have regions which generate more specular reflection than scatter into a given camera light sensor. This is because the surface of the PMMA sample after ablation can have a large range of angles and some of these regions will directly reflect light into the camera. The specularly reflected light can saturate a light sensor so that fringe oscillations cannot be detected. To effectively reduce the number of saturated or underexposed pixels, multiple profiles of the same sample can be taken and the results merged through the use of appropriate algorithms. The profiles can be varied by either adjusting the gain of the camera or by changing the incident angle of the light on the sample. Changing camera sensitivity can be effectively implemented through computer control. To accommodate multiple light angles, the light source can be moved, or multiple light sources in different positions can be used, or multiple paths for directing the light from a single light source can be employed. The latter choice is the preferred embodiment, using a movable mirror to change the incident angle of light on the sample. Changing camera gain and changing the angle of incidence can each effect all pixels simultaneously, so both methods can have the side effect of losing some previously readable pixels. The algorithms used to merge these multiple profiles must check every pixel of interest in every profile and flag those that are out of range for special processing.
Electronics The system electronics may include a mechanism to control the current in the translator electromagnet to provide repeatable and stable displacements of approximately 0.1 micron. It is desirable for the electronics to also provide feedback of the translator position. The electronics can also be used to control the electronic gain of the signal from the CCD camera. In some cases only a portion of the camera field of view is of interest. By optimizing the CCD gain based only on the region of interest instead of the entire field of view, a more reliable profile is produced.
The electronics incorporated into the CCD camera and a frame grabber card in the system computer for acquiring and holding a frame of pixel data can also be used.
Figure 4 shows a block diagram of the electronics on the microcontroller card. A microcontroller (100) such as a Motorola HC11E9 may communicate with a system computer (51) via a communication path (101) such as an RS232 interface. The microcontroller receives a clock signal from an oscillator (102). Upon instruction from the computer via the communication path (101) the microcontroller (100) causes the current in the voice coil (28) to step by sending a step signal (103) and a direction signal (104) to an up/down counter (106). A third digital line (105) from the microcontroller can reset the up/down counter (lOfi) to a known state.
Several digital lines (107) connect the output of the up/down counter to the input of a digital to analog converter (108) (DAC). The analog output (109) from the DAC (108) controls the current in the electromagnet (28).
An oscillating signal (110) is received from the capacitive displacement sensor electronics. The frequency of this signal varies as the capacitance in the RLC circuit changes due to mechanical translation. This signal acts as an input clock to a frequency counter (I11). A timer counter (112) accurately counts cycles from the crystal oscillator clock (102) via line (113). A
control line (114) from the microcontroller can reset both counters (111) and (112).
After being reset, frequency counter ( 111) counts a predetermined number of oscillations coming in on line (110). When the preset number of counts is reached a signal (115) stops timer counter (112) and it holds its value. When counting stops the value stored in timer counter (112) is a measure of the time needed for the RLC oscillator to make the predetermined number of oscillations. By measuring this time the resonance frequency can be calculated and the capacitor plate spacing determined. The value stored on timer counter (112) is available on lines (116). These lines are input to digital input pins on the microcontroller (100) using a multiplexing circuit (117) controlled by address lines (119).
The combination of the capacitive displacement circuit, the counters, the microcontroller, and the digital-to-analog converter form a feedback control loop that permits accurate control of the translator without the need for the expensive standalone translation actuator that would be required in an open loop system, and without the need for a more expensive piezoelectric actuator.
Software Operation The ablation profiler system may include a microcontroller and a system computer. The microcontroller program can be provided either from the microcontroller ROM or by downloading a program from the system WO 99/01716 PCT/US9$/13539 computer over the communication path (101). The software on the system computer advantageously is capable of sending instructions and receiving data from the microcontroller, and of instructing the frame grabber to capture a video frame. The software analyzes data from captured frames to locate fringe oscillations as the position of the translator is stepped through a scan.
Once the location of the fringes for the pixels of interest are determined the software displays the data in a useful form or stores the data for later display.
During operation the microcontroller receives commands from the system computer over the communication path. The microcontroller performs tasks such as stepping and resetting the position of the translator, reading the position of the translator, and controlling the gain of the camera. The microcontroller sends data back to the system computer over the communication path.
The basic flow of the software which runs on the system computer is as follows:
1) Set up the frame grabber buffers and image buffers.
2) Set up I/O communications with the microcontroller.
3) Instruct the microcontroller to adjust the gain of the camera to provide the best dynamic range on the frame grabber. Because of the complex shape of the sample, some pixels may saturate the sensors when the slope of the sample is such that it gives a near specular reflection of the light source into the camera lens. If all pixels cannot be simultaneously brought within a useable range of brightness, the saturated pixels will be later rescanned.
4) Display the camera field of view to insure that the sample is properly positioned.
5) Go into rapid scan mode to locate the scanning limits. In this mode the profiler scans a predetermined number of frames, such as 21, to determine if fringes are present. Then the translator is moved a predetermined distance, such as 25 microns, and again scans multiple frames to look for fringes. This process is repeated until the highest and lowest points on the sample are located to within the aforementioned predetermined distance.
6) Once the scanning limits are determined, scan continuously between the lowest and highest points on the sample. This can be done in either increasing or decreasing direction. During the scan, circular buffers may be used to store the most recent frames, for example 21 frames. A discriminator algorithm may be used to check for oscillations in the intensity recorded for each pixel. The discriminator routine may calculate a discriminator for each pixel. The amplitude of this discriminator indicates the strength of oscillations having a period near the fringe. Each time a frame is captured the discriminator routine subtracts the contribution from the oldest stored frame and adds the contribution from the newest frame. The value of the discriminator for each pixel is then checked to see if it is larger than for any previous value for that pixel and if it is, the discriminator value and translator position are saved.
At the end of the scan, two data arrays have been filled. One array contains the position in the scan for each pixel that gave the largest value of the discriminator. This array is referred to as the elevation data because it contains the elevation for each pixel. The second array contains the largest value of the discriminator for each pixel.
The number of frames to be saved during a scan may depend on the translator step size and the coherence length of the light source. In general, the number of frames saved may be equal to the number of frames captured during a scan distance equal to the coherence length. The number of frames to save can easily be determined by storing the intensity for a single pixel during an entire scan and then plotting the portion of this data containing the fringe oscillations. Next count the number of data points in the oscillating region of the data. The single pixel scan data is also used to determine the oscillation period which is used to calculate the discriminator.

7) Reduce the camera gain for previously saturated pixels and rescan.
Alternately, change the angle of incidence of the light and rescan. For either method, the rescanning operation can cover the entire field of interest or it can be limited to the approximate area of the previously saturated pixels.

8) If some pixels were rescanned then merge the added data into the elevation and discriminator data. This merger may take different forms, depending on the reasons for it. If multiple scans give valid but differing values for the same pixel, the values may be averaged. If some scans show saturation or inadequate intensity for a given pixel, those values may be discarded from the calculations for that pixel.

9) If required, reduce the noise level through averaging. The elevations for a square grid of pixels are averaged together after being weighted by the value of the discriminator for each point. If a 3 by 3 grid of points are averaged then the size of the resulting elevation array is reduced by a factor of 9.

10) Analyze the perimeter of the elevation array to remove any overall tilt in the sample. This can be accomplished by a linear elevation shift in each pixel by an amount which makes the unablated portion of the sample lay in the same elevation plane. Due to noise there will be variations in the elevation of the unablated portion of the sample. However, a curve fitting routine can be used to minimize this variation.

11) Analyze the shape and symmetry of the elevation data and compare the cross sections of the profiles to theoretical curves. Then save the differences between the observed profiles and theoretical profiles and use these differences to calibrate the ablation device itself, or to calibrate the use of the ablation device in performing corneal ablation.

12) Display or print the data and cross sections if needed.
The aforementioned software steps have been described as being executed in the system computer. In alternate embodiments, various portions of these functions might be executed in the microcontroller or partially executed in the electronic circuitry. A single computer might combine the functions of both the system computer and microcontroller. The operation of the invention is not dependent on the exact distribution of these functions, and such design decisions are within the skill of a typical system designer.
Although the invention has been described in terms of preferred embodiments, these are not to be taken as limitations. The invention includes all variations and embodiments that fall within the scope of the claims.

Claims (34)

WHAT IS CLAIMED IS:

1. A scanning white light interferometer profiler system for mapping the three dimensional profile of a portion of a surface of a test sample, said system comprising:
a frame;
a reference surface;
a translator for controllably moving said reference surface with respect to said frame;
a test sample comprising at least one surface to be mapped; and wherein said scanning white light interferometer is adapted for determining the contours of said at least one surface to be mapped.

2. The profiler system of claim 1, wherein said translator further comprises:
a magnet forcer assembly comprising a stationary magnet and a movable magnet, at least one of said stationary magnet and said movable magnet comprising an electromagnet; and wherein said magnet forcer assembly is operatively attached to said reference surface and said frame, said magnet forcer assembly operating to control the physical position of said reference surface with respect to said frame.

3. The profiler system of claim 2, wherein said translator further comprises:
a capacitive displacement sensor comprising a first conductive plate and a second conductive plate, said first conductive plate and said second conductive plate being approximately parallel to each other and spaced apart from each other to form an electrical capacitor, said capacitor configured such that the capacitance of said capacitor varies with a change in the physical position of said first conductive plate with respect to said second conductive plate.

4. The profiler system of claim 3, wherein said translator further comprises:
a translator motion circuit, electrically coupled to said first conductive plate and to said second conductive plate, said translator motion circuit configured to operate cooperatively with said capacitive displacement sensor to determine the physical position of said first conductive plate with respect to said second conductive plate.

5. The profiler system of claim 4, wherein said translator motion circuit is configured to control the physical position of said reference surface with respect to said frame by sending control signals to said magnet forcer assembly.

6. The profiler system of claim 2, wherein said translator further comprises a viscous damper adapted to reduce vibration between said translator and said frame.

7. The profiler system of claim 1, said profiler system further comprising:
a light source;
a camera comprising an iris and a focusing lens; and a beam splitter configured to receive light from said light source and to direct light to said camera.

8. The profiler system of claim 7, wherein said light source and said camera are each tilted away from a forty five degree angle of incidence of said beam splitter.

9. The profiler system of claim 7 wherein:
said beam splitter is further configured to direct a portion of said light from said light source to said test sample; and said profiler system further comprises means for varying the angle of incidence with which said portion of said light from said light source strikes said test sample.

10. The profiler system of claim 7, wherein said light source generates light with a spectral width of approximately 150 nanometers.

11. The profiler system of claim 1, wherein said reference surface has been etched and coated with a metallic substance.

12. The profiler system of claim 1, wherein:
said test sample has a generally planar shape with first and second surfaces on opposing sides, said second surface containing contours to be mapped by said profiler system;
said test sample is composed of material which is relatively transparent to the wavelengths of light produced by said light source; and said test sample further comprises an opaque substance adapted for scattering incident light, said second surface being coated with said opaque substance.

13. The profiler system of claim 12, wherein said opaque substance has optical characteristics such that the coherence properties of the light scattered from said opaque substance will be essentially unchanged from the coherence properties of the incident light.

14. The profiler system of claim 13, wherein said opaque substance further comprises characteristics such that the incident light penetrates said opaque substance to a depth of less than one half the coherence length of the incident light.

15. The profiler system of claim 1, wherein:
said test sample has a generally planar shape with first and second surfaces on opposing sides, said second surface containing contours to be mapped by said profiler system; and wherein said second surface comprises a roughened texture, said roughened texture adapted to scatter light incident on said second surface.

16. The profiler system of claim 1, wherein said test sample comprises alignment markings, said alignment markings adapted to provide lateral and angular position information on said test sample.

17. A method of mapping the contours of a test sample in a profiler system, said method comprising the step of providing control signals to a magnet forcer assembly to control the physical position of a reference surface in a scanning white light interferometer.

18. The method of claim 17, further comprising the step of using a capacitive displacement sensor to provide positional information about said reference surface.

19. The method of claim 18, further comprising the step of using said positional information to adjust said control signals to accurately position said reference surface.

20. A method of mapping the contours of a test sample in a profiler system, said method comprising the steps of:
applying an opaque coating to the contours to be mapped; and using rear-surface scattering to provide scattered light from said contours.

21. A method of mapping the contours of a test sample in a profiler system, said method comprising the steps of:
providing a roughened surface on the test sample for the contours to be mapped; and using front-surface scattering to provide scattered light from said contours.

22. The method of claim 21, wherein said step of providing further comprises chemically etching at least one surface of said test sample.

23. The method of claim 21, wherein said step of providing further comprises abrasively roughening at least one surface of said test sample.

24. The method of claim 21, wherein said step of providing further comprises disabling a vacuum smoke clearing function on an ablation device during an ablation of said test sample.

25. A method of mapping the contours of a test sample in a profiler system, said method comprising the steps of:
providing a roughened surface for the contours to be mapped; and using rear-surface scattering to provide scattered light from said contours.

26. The method of claim 25, wherein said step of providing further comprises chemically etching at least one surface of said sample.

27. The method of claim 25, wherein said step of providing further comprises abrasively roughening at least one surface of said sample.

28. The method of claim 25, wherein said step of providing further comprises disabling a vacuum smoke clearing function on an ablation device during an ablation of said sample.

29. A method of mapping the contours of a test sample in a profiler system, comprising the steps of:
moving a reference surface through a series of positions;
creating a data set by producing and saving a data frame of pixel intensity values for each position of said reference surface;
changing the angle of incidence with which light strikes a test sample;
and creating a plurality of data sets by repeating the steps of moving, creating, and changing, wherein each data set corresponds to a different angle of incidence.

30. The method of claim 29, further comprising the steps of:
creating a single composite data set from said plurality of data sets by determining a composite value for corresponding pixel intensity values from each of said plurality of data sets; and processing said composite data set to create a three dimensional profile of the contours of said test sample.

31. A calibration system for adjusting the parameters of an ablation system, said calibration system comprising:
an ablation device, comprising a computer-controlled laser system adapted for ablation of biological tissue;
a profiler system; and a communications link, adapted for communicating data from said profiler system to said ablation device.

32. The calibration system of claim 31, wherein said ablation device is adapted for adjusting its operating parameters based on differences between said data from said profiler system and a reference profile.

33. A calibration system for adjusting the parameters of an ablation system, said calibration system comprising:
an ablation device, comprising a computer-controlled laser system adapted for ablation of biological tissue;
a topography system, adapted for measuring the shape of a cornea; and a communications link, adapted for communicating data from said topography system and said ablation device.

34. The calibration system of claim 33, wherein said ablation device is adapted for adjusting its operating parameters based on differences between data from said topography system and a reference profile.