Laser Beam Diagnostics

Maximizing Laser Accuracy with Laser Beam Diagnostics

AN-4 MAXIMIZING LASER ACCURACY WITH LASER BEAM ANALYSIS

Abstract

With increasingly sophisticated applications, the demands on the quality of the laser beam have become much greater. Traditional methods of measuring laser beam intensity profile; i.e., burn spots, mode burns, and viewing the reflected beam, are woefully inadequate for assuring the laser quality needed for today's applications. Indeed, lasers are becoming of increasingly high quality. To a large extent this is due to the availability of electronic beam profile instruments. These instruments provide a real time view of the laser beam profile that provides infinitely greater intuition to enable laser optimization. Also, electronic laser beam profilers produce much more accurate quantification of laser beam properties. The accuracy of these measurements enables scientists to fine tune the laser properties to a greater extent than previously possible. New algorithms for laser beam property quantification are discussed, along with the performance improvement of these calculations. In addition, examples are presented of actual situations in which viewing the laser beam has significantly improved its performance.

Introduction

Obsolete Beam Profiling Methods

The 1990's have seen a period of increasingly sophisticated uses of lasers. Because the applications are becoming more complicated, the quality of the laser beam must be closer to the desired ideal to achieve success. Traditional methods of measuring the laser beam intensity profile, such as burn spots, mode burns, and viewing the reflected beam, do not provide sufficient information to enable a scientist to achieve optimum laser performance. Many of the emerging laser technologies would not be possible if they depended on the accuracy of these traditional methods for beam optimization.

 Figure 1. Burn Spots On Thermal Paper Figure 2. Acrylic Mode Burns.
Figure 3. Reflected Beam View.
 Figure 1. Burn Spots On Thermal Paper.  Figure 2. Acrylic Mode Burns.  Figure 3. Reflected Beam View.

Electronic Beam Profiling Accomplishments

On the other hand, the availability of electronic laser beam diagnostics instrumentation has enabled scientists and engineers to tune lasers to much higher standards. Two properties of electronic beam diagnostics facilitate this increased laser performance. The first is the ability to provide illuminating beam profile displays. These displays, in both 2D and 3D beam views, often provide sufficient intuition to enable the laser operator to make significant improvements to the laser beam in a very short time.

The other property of electronic beam profile analysis, which has had a high impact on laser performance, is the higher accuracy with which laser beam properties are measured. Precise measurements, especially the precise measurement of the laser beam width, are major keys in enabling a laser designer or user to maintain optimum performance. From accurate measurement of the laser beam width are derived much more accurate measurements of beam divergence, the M2 or Times Diffraction Limit of the beam, and other important beam properties. In addition, electronic laser beam diagnostics provides other measurements of the laser properties, such as Gaussian fit of near Gaussian beams, Top Hat measurements of flat top beams, and statistical measurements to determine the stability of many beam properties.

Modern Applications Demand Better Laser Beams

Nearly every application of laser beams requires very high performance from the laser. The property of lasers that makes them useful in nearly every application is the high energy density of light that is impinged upon a surface. If the intensity distribution, or beam profile, of this high power is not uniform, or is not as expected, the performance of the laser will be substandard. For example, in research and development with lasers, the reaction to the laser light is often proportional to the intensity squared or cubed. Imagine a beam profile pattern that has deteriorated to the point that much of the energy is in the wings of the beam rather than in the peak, such that the peak is 20% lower than expected. In a process that depends on the energy intensity squared, the result would only be 64% of what is expected. If the process depends on the energy cubed, then only 50% of the conversion efficiency would be achieved. Most research and development applications require better than 5% or 10% accuracy to have meaningful results. Thus, the beam intensity must be known within 3-5% to achieve meaningful research. This accuracy can be achieved only with the latest electronic beam profile instrumentation with sophisticated profiling algorithms.

Another very demanding area of laser work is the design, development, and manufacturing of lasers. More and more laser users are demanding that the laser beam be uniform within very tight limits. Thus the laser manufacturers are being forced more often to perform accurate beam profiling measurements on their lasers to ensure that they meet the requirements of their customers. An increasingly larger percentage of Spiricon's customers are laser manufacturers who are using Spiricon instrumentation to assure that the design of the laser can meet these demands. In addition, manufacturers are using beam profiling instrumentation on the production floor to assure that lasers being shipped to customers meet the specifications of the contract.

Medical applications of lasers are often in the public limelight, and have significant potential of improving the quality of life for everyone. Often these medical applications require a very careful characterization of the laser beam in order for the me dical operation to be applied successfully.

Industrial applications have proliferated very rapidly over the past 10-15 years. Nd:YAG and CO2 lasers are used widely in metal processing and other industrial processes. For years many of the industrial job shops have used the method of burn spots and mode burns to characterize the laser beam. However, increasing numbers of these job shops are realizing that these methods are inadequate. For example, they do not provide instantaneous views of beam profiles on a shot-to-shot basis. Often the laser can be varying over time, which causes, for example, uneven kerf on cutting through met al. By seeing the beam on a frame by frame basis the operator can tell when the laser is erratic, and thus tune it to a higher performance. Higher performance lasers produce less scrap and higher profits for the industrial application.

Accurate Laser Beam Width Measurements

Importance of Beam Width Measurement

There are three laser beam properties that are probably more important than any other. These are the total power or energy in the laser beam, the temporal pulse width, and the spatial width of the laser beam. Laser Power Measurement and temporal width are relatively straightforward, and can be measured accurately with a variety of commercial instruments. On the other hand, the spatial width of a laser beam is extremely difficult to measure. As Tom Johnston said, "Measuring the width of a laser beam is like trying to measure the size of a cottonball with a caliper". The difficulty in measuring laser beams arises from the fact that energy extends far out into the wings of the beam, and makes it difficult for an instrument to calculate or define what is the actual beam width. The reason this measurement is important is that it affects many parameters of the laser beam. For example, the beam width, in conjunction with the laser power or energy, determines the power or energy density of the beam. In addition, the beam width is part of the measurement of the divergence of a laser beam, wherein the divergence determines how large the beam will be at some point along the propagation path. Finally, the beam width is an essential measurement, and must be made very accurately in arriving at the Times Diffraction Limit, M2 , of the laser beam. The Times Diffraction Limit is a definition of a laser beam property that describes how close a beam is to a perfect Gaussian single mode. The ability to measure the beam width very accurately enables a scientist or engineer to determine precisely the quality of the beam and how well it will focus. As mentioned in the introduction above, the beam width affects all of the applications of lasers, including scientific, industrial, and medical, as well as the quality of the laser being manufactured.

Beam Width Measurement Methods

Part of the difficulty of measuring the width of the beam lies in the very definition of beam width itself. One of the traditional methods of measuring beam width with electronic beam profilers is to measure the width as a percent of the peak energy. For example, full-width-half-max (FWHM) has been a common measure of the beam width. Another method is to measure the width at 13% of the peak energy, which it is assumed would come close to containing 1/e2 or the 2s beam width. The beam width measurements which were made as a percent of peak were traditionally used because they were relatively insensitive to noise in the baseline of the measuring instrument or energy in the wings of the beam, and thus were relatively consistent. However, depending on the shape of the beam, they may or may not have much relevance to the actual properties of the laser beam.

A second method of beam width measurement is the percent of energy method. Typically the measurement is made such that it includes 86% of the energy, or describes the 1/e2 energy beam width. There are two methods of making this measurement. One is to place a circular aperture in the center of the beam and increase the aperture size until 86% of the raw beam energy passes through the aperture. This then defines the 1/e2 beam diameter. A second method in a camera system is to simply count all of the pixels and their energy content until 86% of the energy is counted. Both of these methods are subject to misinterpretation of actual beam width, depending on the shape of the beam. The pixel count method does not take into account the fact that there may be a hole in the center of the beam, and thus will report a beam width smaller than the actual width of a doughnut mode beam. The aperture method is useful only for beams that are round, and give no information about the potential ellipticity of the laser beam.

A beam width measurement method that has recently become popular is the knife-edge measurement. This method has traditionally been used with mechanical scanning beam width measurements, but is also very effective when used with camera based electronic beam profilers. In camera based systems a software knife-edge is drawn across the beam, and at a designed clip level, such as when 10% of the total energy is covered, one edge of the beam is defined. When 90% of the energy has been included, the other edge of the beam is defined. A correction factor is used to convert from 10% to 90% to the 1/e2 width. The knife-edge method can be used in both the X and Y axis, or on the major and minor axes of elliptical beams. Until the introduction of present innovations, the knife-edge method has provided the most accurate estimate of laser beam widths. The biggest difficulty with knife-edge measurements is that depending on the mode content of the beam, the clip level definition and correction factor can introduce error into the measurement.

Laser theoreticians have proposed, and the ISO standards committee has accepted, that the true measure of a laser beam width is the Second Moment, or D4s method. Second Moment is a method that integrates energy vs. distance from the centroid of the beam to obtain a properly weighted beam width. Theoretically this method defines precisely the width of the beam, taking into account any energy in the wings, and is independent of structure in the wings. That is, any structure in the wings is properly accounted for to define the real width of the beam.

Two difficulties exist with the Second Moment method. The first is that if there is diffraction in the beam, due to an aperture fairly close to the CCD camera, then there can be considerable energy in the wings of the beam that is not really part of the laser beam. That is, the diffracted energy is diverging rapidly, and would soon disappear in any beam application. Thus it becomes important when using the Second Moment method, to know the beam profile and to be sure that diffraction is not distorting the measurement. The second problem with the Second Moment method has totally prevented its use with camera systems until the present time. This is that noise from the CCD camera far out into the wings is integrated, and provides a heavier weight in making measurements of beam width. This noise in the camera is so serious that with camera based systems it has been impossible to accurately measure the D4s laser beam width. Thus even though the Second Moment method is theoretically the most accurate, it has, until now, been impossible to use for accurate measurements.

Noise Induced Errors In D4s Measurements

Beam Width Accuracy Simulations

One problem that has existed in defining the accuracy of a beam width lies in knowing precisely what is the width of the beam being measured; i.e., it is nearly impossible to tell the accuracy if the beam width is not already known. In the past, one method of beam width measurement calibration accuracy has been performed by first measuring a beam under the most ideal conditions. That is, the beam is at nearly saturation of the camera, and nearly filling the aperture of the camera. After measuring the beam under the most ideal conditions, the beam intensity is reduced, and/or the beam pixel count in the camera is reduced for the same size beam, and measurements are made to see how much the beam width calculation changes from the original estimate. Thus beam width accuracy calculations are based on the assumption that the most ideal measurement is accurate, and then extrapolated from that point.

To overcome these problems Spiricon engineers and programmers have performed beam simulations that enable them to precisely define the beam characteristics. That is, a mathematically derived perfect Gaussian beam is generated in the computer, and a beam profile for that beam is provided in the same format that a camera would provide a beam profile. Then computer generated random noise with a Gaussian distribution, with ±3s peaks, similar to the noise in a camera, is added to the simulated beam. Under these conditions the beam width measurement accuracy can be calculated precisely because the original beam width is known, having been computer generated.

 Figure 4. Beam Width Measurement Error Without Ultracal.  
 Figure 4. Beam Width Measurement Error Without Ultracal.
  Figure 5. Beam Width Measurement Error Using Ultracal.
 
 Figure 5. Beam Width Measurement Error Using Ultracal.

 

Using the above beam profile and noise simulations, the accuracy of various algorithms for calculating beam widths was calculated. All of the calculations are based on a perfect camera with no shading in the baseline. The calculations are also performed for no DC baseline offset. Thus the only mechanism that can contribute to error is the random noise from the camera. Figure 4 illustrates the beam width measurement error under the conditions of the Second Moment D4s measurements and the traditional knife-edge beam width measurements. These calculations were performed for the case of a 512 X 512 camera matrix, with the computer generating beams varying in size inside this matrix. It is seen that under these conditions, as soon as the beam is as small as 64 X 64 pixels, the D4s calculation creates an error of an astonishing 60%. Even at a 100 X 100 pixel beam size, the error is as high as 20%. This is in contrast to the knife-edge method, which creates only about 3% error at a beam size of 64 X 64 pixels. (All of the measurements included a signal magnitude of 256 counts; i.e., equal to saturation of an 8 bit digitizer.) Thus Figure 4 shows why the traditional knife-edge method has given accurate beam width measurements, whereas other researchers have found the Second Moment method impossible to use.

It should be noted that these measurements were made using the ideal conditions incorporated in Spiricon's Autocalibrate wherein the baseline is set precisely at zero, and negative noise components are used to offset positive noise components. Even though these negative noise components are used, the D4s still integrates noise in the wings of the beam to create a very large positive error.

New Measurement Algorithms

Using the knowledge gained from the simulations of Figure 4, Spiricon engineers and programmers have been able to develop a new calculation algorithm that enables accurate Second Moment measurements in the presence of camera noise. These algorithms are incorporated into patented algorithms called "Ultracal". Under the same conditions of noise and signal of the simulations of Figure 4, the simulated measurement accuracy of the new algorithms is shown in Figure 5. As shown, beam widths can be measured in the presence of noise, using these new algorithms with essentially no error; i.e., for a 512 X 512 camera array and a beam size of 64 X 64 pixels, the measurement error is now less than 0.1%, whereas it was formerly 60%. Even with a beam width as small as 16 X 16 pixels, the measurement error is less than 0.3%. Using these new algorithms, the Second Moment measurement can even make more accurate beam width measurements than the knife-edge method, which has a measurement error of 0.45% for a 16 X 16 matrix. The simulated measurement accuracy is improved by more than a factor of 200.

These simulations are extremely important in predicting the potential accuracy of measuring beam widths using the Second Moment method. The algorithms are incorporated into Spiricon's latest beam analyzer product. This method can now be used for calculating important parameters, such as divergence and the beam widths necessary for accurate M2 evaluation of the laser beam. Similar measurements have been made wherein the beam magnitude is reduced instead of the beam size. In these simulations the new "Ultracal" measurement algorithms also perform extremely well in maintaining high accuracy of the laser beam width measurement.

Examples Of Beam Profiling Applications

Industrial CW Nd:YAG Laser

An Nd:YAG laser was used in an industrial application for cutting metal. The best beam profile for this particular metal cutting application was a near Gaussian. The machine shop noticed that one laser had slight problems, wherein the cut width in the X direction was slightly different than the width in the Y direction. However, using mode burns on pieces of metal it was not possible to tell any difference between the X and the Y axis width. Figure 6 shows a 3D beam profile of this laser which appears very Gaussian and uniform in nature. Figure 7 shows a 2D beam profile of the laser, which begins to illuminate the characteristics. Notice in Figure 7 that the beam appears slightly oblong, and measures 8.7mm in the X axis and 6.8mm in the Y axis. Obviously the beam is elongated in one axis as seen in this view. Notice also that Gaussian fit along each of the major axes is extremely good. In Figure 8 where Gaussian fit is performed on the whole beam, the fit has dropped from about .95 on the axes to about .85 for the whole beam. This illustrates the oblong nature of the beam.

Figure 6. CW Industrial Nd:YAG Laser, 3D Beam Diagnostics
Figure 6. CW Industrial Nd:YAG Laser, 3D Beam Diagnostics

As soon as the 2D beam profile pattern was viewed, the technician began searching for what was causing the oblong nature. He discovered that a spatial filter inside the cavity of the laser was slightly misaligned, clipping the Y axis of the beam. By realigning the spatial filter, the beam was made closer to round, and gave much better cuts. The spatial filter in this laser is the object which made the beam so Gaussian, whereas in many cases YAG lasers are running multimode.

 Figure 7. CW Industrial Nd:YAG Laser. (Gaussian Fit On Major, Minor Axes)  Figure 8. CW Industrial Nd:YAG Laser. (Gaussian Fit on Whole Beam)

Figure 7. CW Industrial Nd:YAG Laser. (Gaussian Fit On Major, Minor Axes)

Figure 8. CW Industrial Nd:YAG Laser. (Gaussian Fit on Whole Beam)

Fiber Coupling of Laser Diode

Figure 9 shows the beam profile of a laser diode being used as an optical input to a fiber. The beam was collimated with a lens to bring it to a relatively uniform round profile, rather than the usual rectangular beam profile emitted by laser diodes. The focused spot of the laser diode was then coupled into an optical fiber. This coupling is critical in all three axes, X, Y, and Z.
Figure 9. Laser Diode Beam Profile. (Input Beam to Fiber)
Figure 9. Laser Diode Beam Profile. (Input Beam to Fiber)

Figure 10 shows the diode when it is coupled properly in the X and Y direction, but not in Z. In this case the focus is behind the surface of the fiber, so much of the beam energy is being coupled into the cladding, rather than the center of the fiber. Thus the output of the fiber appears as in Figure 10 with a significant amount of energy not in the central beam. The energy that is not in the central beam is diverging at a much higher rate than the central lobe, and would be easily lost in the application of this fiber. Figure 11 shows much better Z axis alignment of the diode into the fiber. In this case much more of the energy is in the central lobe.

Figure 10. Fiber Output With Input Beam Poorly Aligned.
Figure 11. Fiber Output With Input Beam Well Aligned.

Figure 10. Fiber Output With Input Beam Poorly Aligned.

Figure 11. Fiber Output With Input Beam Well Aligned.

Figure 12 shows the good coupling in 2D, and an aperture is drawn to determine how much energy is in the central lobe. In the case of the good coupling roughly 66% of the energy is in the central 1mm spot. Figure 13 shows the poor coupling, again with a 1mm aperture. It is seen that only 31% of the energy is within the central lobe. This means that 69% of the energy is being wasted into the highly divergent part of the beam.

Figure 12. Fiber Output With Input Beam Well Aligned. (65% of Energy In Central 1 mm Spot)
Figure 13. Fiber Output On Poorly Aligned Input Beam. (31% of Energy In Central 1mm Spot)

Figure 12. Fiber Output With Input Beam Well Aligned.
(65% of Energy In Central 1 mm Spot)

Figure 13. Fiber Output On Poorly Aligned Input Beam.
(31% of Energy In Central 1mm Spot)

Industrial CO2 Pulsed Laser

Figure 14 shows the beam profile of an industrial CO2 laser that did not operate as well as most lasers in the job shop. In this laser a significant amount of the energy is far out in the wings of the beam. This particular laser was used to cut steel, and it was found that with time there were variations in the quality of the cut. Nevertheless, the manager had been making acrylic mode burns, which appeared to be as high quality as the other lasers. It was found when electronic diagnostics were used with this beam, that the mode was changing quite rapidly. During rapid changes, the acrylic mode burns would create an average of the beam which appeared quite Gaussian, whereas individual pulses were highly structured. It was discovered that these variations of time were the cause of the ragged edges cut by this laser.
Figure 14. CO2 Laser With Large Structure In Wings. (78% of Energy Is In A Large 7.6 mm Diameter)
Figure 14. CO2 Laser With Large Structure In Wings. (78% of Energy Is In A Large 7.6 mm Diameter)

Figure 15 shows the same laser with an aperture drawn around it. It shows that only 33% of the energy of this beam is in the central lobe of 3.7mm diameter. From the previous figure we see that 78% is a 7.5mm diameter. Thus only one third of the energy is in the high intensity part that actually works the metal.
Figure 15. CO2 Laser With Large Structure In Wings. (Only 33% of Energy Is In 3.7 mm Diameter)
Figure 15. CO2 Laser With Large Structure In Wings. (Only 33% of Energy Is In 3.7 mm Diameter)

CO2 Laser Waveguide Medical Delivery System

Figure 16 shows the output of a flexible waveguide delivery system for CO2 lasers in medical application. When the waveguide is held straight the beam is quite uniform. This shows the beam pattern with slight bending of the waveguide. Notice that the beam is broken into many lobes and hot spots. However, these hot spots move around very rapidly with only slight bending of the waveguide delivery system, and thus a fairly uniform average energy is applied to the patient. Figure 17, however, shows the same waveguide with heavy waveguide bending. Now the beam has become much less uniformly disbursed, and the hot spots are coming together and leaving cold spots with large areas of low energy. In this case the motion did not make up for the poor distribution of energy, and uneven processing was performed with the laser. The hot spots became too concentrated to obtain uniform delivery to the patient.

Figure 16. CO2 Waveguide Delivery System Output. (Slight Waveguide Bending)
Figure 17. CO2 Waveguide Delivery System Output. (Heavy Waveguide Bending)

Figure 16. CO2 Waveguide Delivery System Output.
(Slight Waveguide Bending)

Figure 17. CO2 Waveguide Delivery System Output.
(Heavy Waveguide Bending)

Pulsed Nd:YAG Flat Top Laser

Figure 18 is a pulsed Nd:YAG laser used for thick film trimming. A hybrid semiconductor manufacturer was using this to trim active circuits. Burn spots of this laser appeared uniform, though obviously the intensity distribution is far from uniform. Other YAG lasers in the same facility had very flat top beams. As soon as electronic diagnostics were used to observe this beam, the alignment problems in the lasers were rectified, and they were able to obtain a uniform flat top and achieve reliable operation from the laser.
Figure 18. Pulsed Nd:YAG Laser For Thick Film Trimming. (Desired Beam Was A Flat Top)
Figure 18. Pulsed Nd:YAG Laser For Thick Film Trimming. (Desired Beam Was A Flat Top)

Optical Parametric Oscillator OPO at 3mm

Optical Parametric Oscillators are extremely useful lasers in that the wavelength can be changed over a fairly wide range. The output of OPOs is typically a strong function of the quality of the input laser beam. Figure 20 shows the output of an OPO that is being pumped by a high quality input beam, and thus has an output with a Gaussian fit of about .85, which is quite good for an OPO.

Figure 20. Output of OPO With High Quality Beam.
Figure 20. Output of OPO With High Quality Beam.

X-Ray Tube Alignment

Lasers are not the only beams that generate a profile than enables a person to determine the alignment. Figure 21 is the profile of an X-Ray tube that has poor filament alignment. Poor filament alignment causes a widely disbursed beam and poor intensity in the center. An X-Ray tube with this type of alignment would produce very low quality film for dental X-Rays. Figure 22 shows another X-Ray tube with excellent alignment. Prior to electronic beam diagnostics it was necessary to expose film with these X-Rays, develop the film, look at it, and then readjust the filaments. Using an X-Ray sensitive tube camera and beam diagnostics, the manufacturer was able to align the X-Ray tubes real time.

Figure 21. X-Ray Beam With Poor Filament Alignment.
Figure 22. X-Ray Beam With Good Filament Alignment.

Figure 21. X-Ray Beam With Poor Filament Alignment.

Figure 22. X-Ray Beam With Good Filament Alignment.

A CW Erbium YAG laser in the 1.5mm range was used in a medical application for tissue welding. It was found that with exactly the right energy density at this wavelength, animal tissue could be welded together. Figure 23 shows the output of this beam with poor alignment of the delivery system. Using this beam the tissue welds were very poor, and failed under a very small amount of stress. Figure 24 shows the same laser with the delivery system properly aligned. In this case the beam is a nice uniform flat top, and it gave excellent performance.

In laboratory experiments this laser was used to weld together a severed vein in a mouse's leg. It was determined that using a well aligned beam, the vein could be welded such that the weld area was as strong as the rest of the vein. With poor alignment of the laser, the veins ruptured very quickly under very low pressure.

Figure 23. 1.5µm YAG Laser With Poor Top Hat Structure. (Used for Tissue Welding)
Figure 24. 1.5µm YAG Laser With Good Top Hat Structure. (Needed for Good Tissue Welding Results)

Figure 23. 1.5µm YAG Laser With Poor Top Hat Structure.
(Used for Tissue Welding)

Figure 24. 1.5µm YAG Laser With Good Top Hat Structure.
(Needed for Good Tissue Welding Results)

 

Free Electron Laser

Figure 25 shows the output of a free electron laser at 100mm. This laser was measured with a pyroelectric matrix array camera. This was the first time that the scientists using this laser had ever seen the beam profile. They were quite thrilled to see that the intensity profile was as close to Gaussian as is illustrated in this slide.
Figure 25. Output of Free Electron Laser Focused Beam At 100 um Wavelength.
Figure 25. Output of Free Electron Laser Focused Beam At 100 um Wavelength.

Conclusion

Laser properties have improved dramatically in the last 5-10 years. Gains are increasingly made in many aspects. Pulses are getting shorter, wider wavelength range is achieved, higher powers and intensities are gained, and many other performance charact eristics have improved. One of the major gains has been improvement in the beam profile so that the intensity pattern performs its job better than previously possible. Electronic beam profile analysis has contributed significantly to this improvement in lasers. Many customers of Spiricon say that they could not build their lasers without the beam diagnostics provided by electronic profiling instrumentation.



Back to All Laser Measurement
Back to All Laser Beam Profilers