Open Access
Add to BibSonomyAbstract
Femtosecond lasers have a unique ability of processing bulk transparent materials for various applications such as micromachining, waveguide manufacturing, and photonic bandgap structures, just to name a few. These applications depend on the formation of micron or submicron size features are known to be index modifications to the bulk substrate [2, 11], which were thought to persist indefinitely. However, it has been observed that some of these bulk transparent materials recover or “heal” with time. This “healing”process is studied and quantified using Nomarski Differential Interference Contrast optical microscopy and diffraction efficiency measurements of micro-machined gratings. We find healing to be accelerated in dye doped polymers.
©2009 Optical Society of America
Introduction
The morphology and the physical mechanisms for index changes of femtosecond laser micromachined features processed in crystalline and amorphous transparent bulk materials; 6H-SiC semi-insulating, doped and undoped PMMA, soda lime glass (SLG), and other amorphous samples were studied for their recovery process after being irradiated with single and multi-femtosecond laser pulses. Previously studied femtosecond interaction with crystalline and amorphous transparent substrates describes the different processes used to irradiate and characterize various micromachined features [2,10,11]. It has been observed that the modified index of refraction recovers for amorphous substrates from femtosecond laser processing, while crystalline bulk transparent materials do not recover (“heal”). Quantitative measurements were performed using Nomarski DIC microscopy, diffraction efficiency measurements, amplified spontaneous emission (ASE) [12–14], and two-photon absorption spectroscopy (TPAS) [15] to quantify the “healing”process with time. This work contributes to the prior imaging measurement techniques performed on SiC crystalline materials [11], which assist in formulating theories presented here to the cause of the physical mechanism for the index of refraction modifications with femtosecond laser pulses in transparent bulk substrates.
Diffraction efficiency measurements
Grating structures were selected in order to measure the diffraction efficiency, η, initially, to approximate the change in index magnitude (Δn) [11], but η was also used to measure changes in the material over time. Grating structures, however, were only micromachined into the SiC and SLG samples. The PMMA samples were also femtosecond laser processed but with a 250mm lens yielding circular features. Our method of micromachining gratings uses an anamorphic lens designed to redistribute the ultrafast (UF) laser pulse from a 5.5mm round Gaussian profile to a 3µm by 190µm line at the focal plane [2]. Each individual grating line consists of three separate line pulses exposed in sequence along the horizontal direction to make a single 500µm line, and 25 parallel vertical lines, resulting in a ~500µm×500µm grating. These gratings are typically about 1µm to 10µm deep below the surface depending on the focus alignment of the anamorphic lens. Figure 1 shows a 500µm×500µm grating in semi-insulating SiC with a spacing of 20µm.
Fig. 1. (Left) SiC grating view with an optical microscope using Nomarski DIC for semi-insulating SiC with 10X magnification; (right) 50X magnification. Image processing was performed in order to better resolve the modified surface lines [11].
The following measurement of the diffraction efficiency (η) was measured for a 1st order diffracted HeNe beam through the grating structure and calculated using Eq. (1). An example of a diffraction pattern along with the theoretical mathematical model is given by DesAutels et al [11]. Measured were the 0th-order and diffracted beams resulting from a HeNe laser at 632.8nm, 1.5mm 1/e2 beam diameter, and 1.5mW output power. A long focal length lens was used to keep the HeNe beam diameter close to 500µm to illuminate the full grating structure. The HeNe beam power and diameter were measured with Spiricon software and a Cohu 4812 CCD camera in the absence of the gratings. It was verified that the + 1 and -1 orders were of the same power to within ±5%. However, only the + 1 order was subsequently monitored. The diffraction efficiency, η, is given by [11],
where the factor of two accounts for the fact that only the + 1 order diffracted beam is measured. In Eq. (1), P1 is the power of the 1st order measured beam, P0 is the power of the zero-order beam, ND0 is the neutral density filter’s optical density-it is placed in front of a Cohu 4812 camera when measuring the zero-order beam (no grating present only the unprocessed sample); and, ND1 is the optical density of the neutral density filter placed in front of the Cohu camera while measuring the 1st order diffraction beam. Neutral density filters were needed to keep the CCD camera from saturating. P1 and Po are measured using Spiricon laser beam analyzer (LBA) software with a Cohu 4812 CCD camera, which is calibrated using a Coherent FieldMax II TOP meter with an OP2-VIS detector (both traceable to NIST). The uncertainty in the diffraction efficiency measurement was determined using standard propagation of uncertainties and found to be ±1% for the meter accuracy and ±4% for the optical density values, thus leading to an overall uncertainty in the diffraction efficiency of just under ±5%.
Fig. 2. Diffraction efficiency, η, versus fluence for 2, 6, 12, and 36 femtosecond laser pulses in semi-insulating 6H-SiC; date: January 2006.
Figure 2 shows how the diffraction efficiency increases with fluence and the number of pulses. This data was taken in 2005 shortly after the semi-insulating 6H-SiC sample was laser processed. A similar analysis was prepared for amorphous soda lime glass (SLG) as shown in Fig. 3 only for a single laser pulse.
Figure 3 represents single femtosecond laser pulse diffraction efficiency versus fluence in SLG recorded shortly after being processed in December of 2005.
Index modification degradation over time
The grating structures micromachined in semi-insulating 6H-SiC, soda lime glass (SLG) and PMMA as mentioned, were re-measured via Nomarski DIC microscopy. The η was also remeasured for SiC and SLG samples to further investigate the “healing”process. It was during periodic inspections that it was observed that the subsurface structures prepared in amorphous substrates (SLG) have completely vanished. However, the femtosecond laser processed regions in all of the SiC crystalline substrates are completely intact. The features in SLG vanish, or “heal” in approximately 2 years after laser processing.
All the PMMA, SLG and SiC samples were kept under the same conditions. Each sample was tested then stored at room temperature between 20 and 22°C. Testing exposed samples to fluorescent lighting but were stored in the dark.
Soda lime glass (SLG)
Initially, at approximately 6–8 months after processing gratings into SLG, it was difficult to detect the structures under the optical microscope even with Nomarski DIC. This observation prompted the re-measuring of the diffraction efficiency and is shown that it is down by approximately half.
Fig. 4. η versus fluence for two times: time t=0 is defined by the time the diffraction efficiency is measured directly after writing the gratings and time t=1year shows an approximate drop in η by about 2 times. After two years the features have completely vanished. Note that this plot uses a linear scales (as opposed to log scale as in previous figures) to best depict the change in η.
Figure 4 shows that after about 1 year (i.e. approximately between 8 to 16 months), the efficiency drops by a factor of 2. After approximately 2 years even the high-fluence grating structures completely recovered. Efficiency versus time was not measured for the SiC samples because there was no detectable change in the crystalline subsurface grating structures. This effect is best described visually in Fig. 5 and Fig. 6.
Fig. 5. Subsurface SLG gratings: (Top) are the optical microscope images, (Bottom) are the lineout images; the lineout lengths are not calibrated, but the length of the grating structure (which is the length of the lineout) is approximately 500µm. The lineout images represent contrast resolution amplitude in arbitrary units.
The images in Fig. 5 are very difficult to show with clarity (contrast resolution) due to the fact that they are subsurface and without destroying the sample using FIB/TEM. This is why graphical charts are being used to demonstrate the “healing” effect.
Figure 5 shows how a subsurface grating micromachined in SLG has degraded over time, that is, the features made from femtosecond processing are nearly completely reversed. The amorphous substrates, shown in Fig. 5, are “healing” over time to the point where the subsurface features are barely visible even using Nomarski DIC microscopy. These images are difficult to observe due to the fact that the features are subsurface, which is why lineout charts are presented to quantify the “healing”processes.
We suggest that the healing in SLG is that a multi-photon absorption process leads to the trapping of the laser-ejected outer bound atomic electrons through fast femtosecond electron cooling. The cooled or trapped electrons then become part of electron-hole pairs, which after time or annealing, electrons re-combine with holes and essentially “heal” the modified femtosecond processed area [1,3,4]. This argument results from the experimental Nomarski DIC microscope image/lineout data, which represents recovery of the amorphous subsurface substrates. The super-cooled electron-hole pairs recover at different rates depending on the doping levels as shown in data. We propose that the recovery is due to electron-hole pairs since the refractive index depends on electronic dipole vibrations. When those dipoles are altered slightly that will modify the refractive index. When the sample recovers the modified index features disappear suggesting that the electronic dipoles are in their original states. We also conclude that polymer bond breaking due to laser irradiation is probably irreversible thus further supporting electron-hole pair recombination. It is suggested that the electron-hole pair lifetime is greatly extended because the charges are suspended/trapped in the polymer.
6H-SiC semi-insulating
Unlike the amorphous substrates, SiC crystalline materials were also processed and their subsurface features remain without any indication of “healing”, even after being annealed.
Fig. 6. Subsurface semi-insulating 6H-SiC gratings: (Top) optical microscope images, (Bottom) lineout images; the lineout lengths are not calibrated, but the length of the grating structure (which is the length of the lineout) is approximately 500µm. The lineout images represent contrast resolution amplitude in arbitrary units. (Top) the color of these images is false color that is dependent on the polarization setting of the polarizer and is difficult to match precisely between measurements (2006 and 2008).
Figure 6 consist of different colors that is due to the Nomarski DIC color microscopy gives slightly different colors depending on the polarizer position not exactly the same for each image. What is important is the consistent contrast resolution of both images.
As observed from Fig. 6 the SiC gratings have not “healed” to any significant amount. In fact, the 2008 semi-insulating SiC grating was also repeatedly annealed to 399°C. Thus, neither time nor heat will alter the SiC subsurface grating structures, which suggests their crystal lattice structure was altered by laser processing. This suggests that the lattice has been deformed, resulting in a displacement of lattice dipoles [7], inducing an index change.
In crystalline structures, we suggest that the femtosecond pulse may induce a multi-photon absorption process freeing electrons. Because the substrate is highly organized this would result in broken atomic bonds at the surface and the subsurface features are attributed to lattice deformation [7]. Lattice deformation distorts bonds that could possibly result in a new structure with the atomic dipoles in different locations from the original crystalline lattice structure. Any photon incident on these new dipoles will interact differently than the originally crystal dipole mapping. This is supported by the fact that annealing the SiC crystalline-processed samples below the melting point did not “heal” or change the resulting femtosecond micro-machined grating lines.
PMMA host and PMMA doped/AF455
PMMA amorphous samples were also investigated to compare with initial findings from SLG and SiC samples. If the theory that the amorphous-type samples “heal” due to electron-hole recombining [11] after time of ~1–2 years is correct then a doped amorphous sample should recover more rapidly if there are an excess of electrons supplied by the dopant. Therefore, our studies compare undoped PMMA with PMMA doped with AF455 [18]. Instead of gratings, circular structures were formed in PMMA to eliminate the possible spatial dependence of fluence. A 250mm lens was used to focus the single femtosecond pulse onto the samples resulting in an 80µm spot diameter.
Fig. 7. PMMA host (without dopant): (Top) are the Nomarski DIC microscopy images, (Middle) are the line outs of those images, and then (Bottom) are the same images from the Top view only image processed for better viewing.
Figure 7 illustrates that approximately 6 months after processing, no notable changes have been observed in the PMMA host, which is similar to the SLG samples. If the mechanisms responsible in PMMA are the same as in SLG, we would expect recovery in about 1 year, and complete recovery in ~2 years. The hypothesis that the underlying mechanisms are the same in SLG and PMMA is based on their common properties of being transparent amorphous materials. However, given the different mechanical and physical properties of these materials, the recovery rates may be somewhat different.
Fig. 8. PMMA with AF455 doped: (Top) are the Nomarski DIC microscopy images, (Middle) are the line outs of those images, and then (Bottom) are the same images from the Top view only image processed for better viewing.
Next we tested PMMA polymer doped with AF455 dye. Figure 8 demonstrates how the PMMA with a dopant completely recovers from single femtosecond laser pulse processing within approximately 3 months. The two large multi-pulse laser processed areas are fiduciary points that define the start of the damage matrix (i.e. the array of points). The fiduciary points, being formed by multiple shots, remain visible after the other features have healed.
Figure 9 shows the AF455 molecule and Fig. 10 shows the dye doped sample just after laser exposure and one month after exposure. The upper left spot in each image is the fiduciary point formed with multiple laser shots. Each point in a particular row corresponds to the same laser exposure. The energy per pulse and fluence is shown for each row on the right-hand portion of Fig. 10. All damage tracks were created using a 125mm focusing lens.
At high fluence, the polymer experiences laser oblation, leading to physical pits. The first four rows clearly show pits of smaller diameter for lower fluence. At fluence of 1.05J/cm2 and below, most of the damage is below the surface. The sub-surface modifications of the material appear as gray spots.
After only one month, a large portion of the subsurface damage has healed; however, interestingly the diameter of the pits appears to be shrinking. This effect is most pronounced in the third and fourth rows. Damage area shrinkage after oblation is not observed in the neat polymer. It thus appears that healing of both surface and subsurface features are enhanced by the presence of the dye.
The dye can lead to several potential decay and recovery mechanisms. As mentioned above, the dye can be a source of charge, which is ejected from the molecule by a laser pulse and trapped in the polymer, leading to a modification of the polymer refractive index due to the resulting local electric field. Additionally, the electronic structure of the dye can change after charge ejection, leading to a change in refractive index that originates in the dye. These two effects together can account for the observed refractive index change.
Alternatively, the absorbed energy from the laser may damage the polymer by breaking bonds. This effect, in contrast to charge trapping, might be irreversible. Similarly, the molecules may photo-degrade through bond breaking, also leading to a refractive index change.
Photodegradation of the dye molecules can be measured with techniques that differentiate between polymer host and dopant. For example, the AF455 dye is well known to have a large two-photon absorption cross-section, which leads to molecular excitation, followed by relaxation through fluorescence (called two-photon fluorescence, or TPF). In studies that monitor the two-photon fluorescence signal from the AF455 dye in polymer under high laser intensities, it was found that the dye degrades and recovers [15]. So, there is at least one mechanisms of subsurface damage that has been demonstrated to originate in the modification of the dopant molecule.
The degradation and recovery of DO11 dye in PMMA was observed using amplified spontaneous emission as a probe of the dye molecule [12]. During photodegradation, the linear absorption spectrum showed satellite peaks that are reminiscent of dimer formation. That study concluded that photo-tautomerization followed by dimer formation was the cause of degradation, while the breakup of dimers was responsible for self healing. Interestingly, the molecule did not recover when the measurements were done in liquid solution [13]. This suggests that the polymer is involved in the healing process, presumably by limiting the volume of the phase space associated with degradation, thus making recovery more probable [14].
In degradation and recovery of AF455 as monitored with TPF, on the other hand, no appreciable changes in the linear absorption spectrum were observed. This suggests that the laser-induced dye modification is most pronounced in electronic transitions between excited states, which affect TPF but not linear absorbance.
Our observations demonstrate a synergism between the dye dopant and the polymer. Dye dopants, such as AF455, recover from degradation when placed in a polymer matrix, and, the polymer matrix “heals” more effectively in the presence of the dopant dye.
Photodegradation of the polymer in the presence of dye can take three routes: (1) The polymer chains can be broken, (2) charge can be ejected by the polymer and trapped in the polymer or the dopant, and (3) charge can be released by the dye and trapped in the polymer.
Similarly, photodegradation of the molecule can originate in: (1) broken bonds in the molecule or some other change in molecular structure such as isomerization, (2) ejection of charge in the molecule that is trapped in the polymer; and (3) charge ejected by the polymer and trapped in the molecules.
The effects of the laser on the polymer can be studied by observations of the physical damage to the polymer while spectroscopic techniques can be used to differentiate between contributions from the dye and polymer. The synergism between the dye and the polymer suggests mechanism [2]. However, given that the molecular dopants are known to plasticize the polymer host by lubricating the slipping between polymer chains [16], the presence of the dye may mediate the recovery of a broken polymer chain. Similarly, the presence of the dyes would mediate the shrinkage of holes caused by oblation. The support provided by the polymer by physically surrounding the dopants, on the other hand, may keep the molecular fragments localized, which mediates recombination and therefore promotes healing. More studies are required to pin down the mechanisms.
Summary of results
A summary of all of the above experiments is given to consolidate each conducted experiment to provide a better understanding of what the possible causes of femtosecond index modifications in crystalline and anamorphic materials. At the very least, the data provided will add to the science of index of refraction changes in bulk transparent materials. Below is a list of the key findings.
1. Optical Microscopy:
a. In 2005–2006 Nomarski DIC images provide evidence that the subsurface features are resulting from phase changes within the bulk transparent substrate that are attributed to index of refraction changes.
b. In 2007–2008 Nomarski DIC images result in total processed feature recovery in amorphous substrates.
c. Recovery in amorphous substrates with dopants is approximately 5 times faster (or greater) than non-doped substrates.
2. Diffraction Efficiency Experiments:
a. In the subsurface, only phase gratings are formed, leading to diffraction efficiencies of η ~1% for single pulse and ~30% for multi-pulse.
b. Diffraction efficiency decreases by approximately half after ~1 year in SLG.
c. Diffraction efficiency does not decrease in SiC samples.
3. Time Study:
a. Amorphous samples, doped and undoped, experience index modification recovery.
b. Crystalline samples do not experience any notable index modification recovery.
c. These observations proved to be valuable in the fact that they provide support of the TEM, X-ray, and Raman measurements [11] suggesting no chemical alterations in the SiC, but rather a lattice deformation.
Studies of photodegradation, healing of subsurface features and laser oblation of dye doped polymer show that a new synergistic effect between polymer and host may lead to an enhancement of the effects of self healing. We propose that the mechanism of refractive index change of the polymer is consistent with the production of trapped charges, based on the accelerated rate of healing in the doped versus the un-doped samples. Furthermore, the dyes may plasticize the polymer matrix, making it heal more quickly.
While permanent refractive index changes are desirable for making device structures, other application, such as electrooptic devices in polymer fiber waveguides, require that the material be robust to high intensity light [17]. In such applications, self-healing is an important and useful property for prolonging the device’s life.
In conclusion, index changes in doped amorphous substrates recover from a single femtosecond laser pulse approximately 5 times faster than that of undoped amorphous materials suggesting that electron-hole pairs are recombining after enough time has transpired. In addition, crystalline materials do not recover suggesting that their crystal lattice has been deformed permanently. More work in separating the effects of dopant and host are required to uniquely determine the mechanisms that are responsible for our observations.
Acknowledgement
Special acknowledgment is given to the US Air Force Research Laboratory for making this research possible; also to Dr. Christopher Brewer (WPAFB/AFRL/RXPJ) for his contribution to this research. Additionally, thanks to Dr. Mark Walker (WPAFB/AFRL/RXPJ/General Dynamics) for his contributions.
References and links
1. G. Petite, P. Daguzan, S. Guizard, and P. Martin, “Femtosecond history of free carriers in the conduction band of a wide-bandgap oxide,” in IEEE Annual Report Conference on Electrical Insulation and Dielectric Phenomena (IEEE, 1995), Vol. 15, pp. 40–44
2. G. L. DesAutels, C. D. Brewer, M. A. Walker, S. B. Juhl, M. A. Finet, and P. E. Powers, “Femtosecond micromachining in transparent bulk materials using an anamorphic lens,” Opt. Express 15(20), 13139–13148 (2007). [CrossRef]
3. S. S. Mao, F. Quére, S. Guizard, X. Mao, R. Russo, G. Petite, and P. Martin, “Dynamics of Femtosecond Laser Interactions with Dielectrics,” Appl. Phys., A Mater. Sci.Process. 79, 1695–1709 (2004).
4. J. W. Chan, T. Huser, S. Risbud, and D. M. Krol, “Structural changes in fused silica after exposure to focused femtosecond laser pulses,” Opt. Lett. 26(21), 1726–1728 (2001). [CrossRef]
5. J. Burton, L. Sun, M. Pophristic, S. Lukacs, F. Long, Z. Feng, and I. Ferguson, “Spatial Characterization of Doped SiC Wafers by Raman Spectroscopy,” J. Appl. Phys. 84(11), 6268–6273 (1998). [CrossRef]
6. R. Paschott, “Encyclopedia of Laser Physics and Technology,” http://www.rpphotonics. com/brillouin_scattering.html, (2008).
7. C. Kittel, Introduction to Solid State Physics, (Wiley and Sons, Inc.1969), 4th edition, pp. 673–675.</bok>
8. C. Palmer, Diffraction Grating Handbook, fourth edition, (Richardson Grating Laboratory, Rochester, NY, 2000) www.gratinglab.com.
9. S. Barden, J. Arns, W. Colburn, and J. Williams, “Volume-Phase Holographic Gratings and Efficiency of Three Simple Volume-Phase Holographic Gratings,” Publ. Astron. Soc. Pac. 112(772), 809–820 (2000). [CrossRef]
10. L. DesAutels, C. Brewer, M. Walker, S. Juhl, M. Finet, S. Ristich, M. Whitaker, and P. Powers, “Femtosecond Laser Damage Threshold and Nonlinear Characterization in Bulk Transparent SiC Materials,” J. Opt. Soc. Am. B 25(1), 60–66 (2008). [CrossRef]
11. L. DesAutels, C. Brewer, P. Powers, M. Walker, D. Tomlin, A. Fratini, S. Juhl, and W. Chen, “ “Femtosecond Index Change Mechanisms and Morphology of SiC Crystalline Materials,” Phys. Lett. A 373(5), 583–591 (2009). [CrossRef]
12. B. F. Howell and M. G. Kuzyk, “Amplified Spontaneous Emission and Recoverable Photodegradation in polymer doped with Disperse-Orange-11,” J. Opt. Soc. Am. B 19, 1790 (2002). [CrossRef]
13. B. F. Howell and M. G. Kuzyk, “Lasing Action and Photodegradation of Disperse Orange 11 Dye in Liquid Solution,” Appl. Phys. Lett. 85(11), 1901 (2004). [CrossRef]
14. N. B. Embaye, S. K. Ramini, and M. G. Kuzyk, “Mechanisms of reversible photodegradation in disperse orange 11 dye doped in PMMA polymer,” J. Chem. Phys. 129(5), 054504 (2008). [CrossRef]
15. Y. Zhu, J. Zhou, and M. G. Kuzyk, “Two-photon fluorescence measurements of reversible photodegradation in a dye-doped polymer,” Opt. Lett. 32(8), 958–960 (2007). [CrossRef]
16. F. Ghebremichael and M. G. Kuzyk, “Optical Second Harmonic Generation as a Probe of the Temperature Dependence of the Distribution of Sites in a Poly (methyl methacrylate) Polymer Doped with Disperse Red 1 Azo Dye,” J. Appl. Phys. 77(7), 2896 (1995). [CrossRef]
17. D. J. Welker, J. Tostenrude, D. W. Garvey, B. K. Canfield, and M. G. Kuzyk, “Fabrication and characterization of single-mode electro-optic polymer optical fiber,” Opt. Lett. 23(23), 1826–1828 (1998). [CrossRef]
18. J. Rogers, J. Slagle, D. McLean, R. Sutherland, B. Sankaran, R. Kannan, L. Tan, and P. Fleitz, “Understanding the One-Photon Photophysical Properties of a Two-Photon Absorbing Chromophore,” J. Phys. Chem. A 108(26), 5514–5520 (2004). [CrossRef]
