Laser-induced back-side wet etching of fused silica with an aqueous solution containing organic molecules

The laser-induced back-side wet etching of fused silica with aqueous solutions of pyranine (8-hydroxy-1,3,6-pyrenetrisulfonic acid trisodium salt) is reported. KrF and XeF excimer lasers were employed as light sources. Well-defined line-and-space and grid micropatterns, free of debris and microcracks, were obtained. Compared with other organic solutions, the aqueous pyranine etching medium etches more slowly but produces a higher quality etched surface. With the KrF laser, the etch rate ranged from 0.02 to 0.12 nm pulse^-1, depending on the dye concentration and the fluence of the laser. The etch rate decreased dramatically when the XeF laser was employed, which was partially attributed to the lower absorption efficiency of the aqueous pyranine solution at the XeF laser wavelength. Received: 20 November 2001 / Accepted: 21 November 2001 / Published online: 2 May 2002     

Laser-assisted etching of gallium arsenide in chlorine at 308 nm

Xenon chloride (308 nm) excimer laser-assisted etching of GaAs (100) in Cl₂ was demonstrated and characterized with respect to laser and gas parameters. The etch rate increased linearly with laser fluence from thresholds in the range of 50 to 75 mJ/cm² to the highest fluence studied, 650 mJ/cm². For a laser fluence of 370 mJ/cm², the etch rate varied with Cl₂ pressure reaching a maximum at a Cl₂ pressure of about 2 Torr. The etch rate decreased monotonically with Ar buffer gas pressure because of redeposition of GaCl₃ products into the etched channel. The redeposited GaCl₃ affected the etch rate and the etch morphology. The etch rate and morphology also varied with laser repetition rate. The mobility of chlorine on the surface also plays an important role in the etching mechanism.     

Backside etching of fused silica with Nd:YAG laser

The laser-induced backside etching of fused silica with gallium as highly absorbing backside absorber using pulsed infrared Nd:YAG laser radiation is demonstrated for the first time. The influence of the laser fluence, the pulse number, and the pulse length on the etch rate and the etched surface topography was studied. The comparable high threshold fluences of about 3 and 7 J/cm² for 18 and 73 ns pulses, respectively, are caused by the high reflectivity of the fused silica-gallium interface and the high thermal conductivity of gallium. For the 18 and 73 ns long pulses the etch rate rises almost linearly with the laser fluence and reaches a value of 350 and 300 nm/pulse at a laser fluence of about 12 and 28 J/cm², respectively. Incubation processes are almost absent because etching is already observed with the first laser pulse at all etch conditions and the etch rate is constant up to 30 pulses.The etched grooves are Gaussian-curved and show well-defined edges and a smooth bottom. The roughness measured by interference microscopy was 1.5 nm rms at an etch depth of 0.6 μm. The laser-induced backside etching with gallium is a promising approach for the industrial application of the backside etching technique with IR Nd:YAG laser.     

Using IR laser radiation for backside etching of fused silica

Laser-induced backside etching of fused silica with gallium as highly absorbing liquid is demonstrated using pulsed infrared laser radiation. The influences of the laser fluence, the pulse number, and the pulse length on the etch rate and the etched surface topography were studied and the results are compared with these of excimer laser etching. The high reflectivity of the fused silica-gallium interface at IR wavelengths results in the measured high threshold fluences for etching of about 3 J/cm² and 7 J/cm² for 18 ns and 73 ns pulses, respectively. For both pulse lengths the etch rate rises almost linearly with laser fluence and reaches a value of 350 and 300 nm/pulse at a laser fluence of about 12 and 28 J/cm², respectively. The etching process is almost free from incubation processes because etching with the first laser pulse and a constant etch rate were observed. The etched surfaces are well-defined with clear edges and a Gaussian-curved, smooth bottom. A roughness of about 1.5 nm rms was measured by AFM at an etch depth of 0.95 μm. The normalization of the etch rates with respect to the reflectivity and the pulse length results in similar etch rates and threshold fluence for the different pulse widths and wavelengths. It is concluded that etching is a thermal process including the laser heating, the materials melting, and the materials etching by mechanical forces. The backside etching of fused silica with IR-Nd:YAG laser can be a promising approach for the industrial usage of the backside etching of a wide range of materials. PACS 81.65.C; 81.05.J; 79.20.D; 61.80.B; 42.55.L     

In situ reflectivity investigations of solid/liquid interface during laser backside etching

In situ reflectivity measurements of the solid/liquid interface with a pump-probe setup were performed during laser-induced backside wet etching (LIBWE) of fused silica with KrF excimer laser using toluene as absorbing liquid. The intensity, the temporal shape, and the duration of the reflected light measured in dependence on the laser fluence are discussed referring to the surface modification and the bubble formation.The vaporisation of the superheated liquid at the solid interface causes a considerable increase of the reflectivity and gives information about the bubble lifetime. The alterations of the reflectivity after bubbles collapse can be explained with the changed optical properties due to surface modifications of the solid surface. Comparative studies of the reflectivity at different times and the etch rate behaviour in dependence on the laser fluence show that the in situ measured surface modification begins just at the etch threshold fluence and correlates further with etch rate behaviour and the etched surface appearance. The already observed surface modification at LIBWE due to a carbon deposition and structural changes of the near surface region are approved by the changes of the interface reflectivity and emphasizes the importance of the modified surface region in the laser-induced backside wet etching process.     

Laser etching of fused silica using an adsorbed toluene layer

A new method for laser etching of transparent materials with a low etch rate and a very good surface quality is demonstrated. It is based on the pulsed UV-laser backside irradiation of a transparent material that is covered with an adsorbed toluene layer. This layer absorbs the laser radiation causing the etching of the solid. The threshold fluence for etching of fused silica amounts to 0.7 J/cm². The constant etch rate of about 1.3 nm/pulse that has been observed in a fluence interval from 2 to 5 J/cm² is evidence of a saturated process. The limited thickness of the adsorbed layer causes the low etch rates and the rate saturation. The etched surface structures have well defined edges and low surface roughness values of down to 0.4 nm rms. PACS 81.65.Cf; 81.05.Kf; 79.20.Ds; 61.80.Ba; 42.55.Lt     

Chemical etching investigation of polycrystalline p-type 6H-SiC in HF/Na2O2 solutions

In this work, an experimental study on the chemical etching reaction of polycrystalline p-type 6H-SiC was carried out in HF/Na₂O₂ solutions. The morphology of the etched surface was examined with varying Na₂O₂ concentration, etching time, agitation speed and temperature. The surfaces of the etched samples were analyzed using scanning electron microscopy (SEM), energy-dispersive X-ray (EDX) Fourier transform infrared spectroscopy (FT-IR) and photoluminescence. The surface morphology of samples etched in HF/Na₂O₂ is shown to depend on the solution composition and bath temperature. The investigation of the HF/Na₂O₂ solutions on 6H-SiC surface shows that as Na₂O₂ concentration increases, the etch rate increases to reach a maximum value at about 0.5 M and then decreases. A similar behaviour has been observed when temperature of the solution is increased. The maximum etch rate is found for 80 °C. In addition, a new polishing etching solution of 6H-SiC has been developed. This result is very interesting since to date no chemical polishing solution has been developed on the material.     

Laser etching of transparent materials at a backside surface adsorbed layer

The laser etching using a surface adsorbed layer (LESAL) is a new method for precise etching of transparent materials with pulsed UV-laser beams. The influence of the processing parameters to the etch rate and the surface roughness for etching of fused silica, quartz, sapphire, and magnesium fluoride (MgF₂) is investigated. Low etch rates of 1 nm/pulse and low roughness of about 1 nm rms were found for fused silica and quartz. This is an indication that different structural modifications of the material do not affect the etching significantly as long as the physical properties are not changed. MgF₂ and sapphire feature a principal different etch behavior with a higher etch rate and a higher roughness. Both incubation effects as well as the temperature dependence of the etch rate can be interpreted by the formation of a modified near surface region due to the laser irradiation. At repetition rates up to 100 Hz, no changes of the etch rate have been observed at moderate laser fluences.     

Microstructuring of fused silica by laser-induced backside wet etching using picosecond laser pulses

The laser-induced backside wet etching (LIBWE) is an advanced laser processing method used for structuring transparent materials. LIBWE with nanosecond laser pulses has been successfully demonstrated for various materials, e.g. oxides (fused silica, sapphire) or fluorides (CaF₂, MgF₂), and applied for the fabrication of microstructures. In the present study, LIBWE of fused silica with mode-locked picosecond (t_p = 10 ps) lasers at UV wavelengths (λ₁ = 355 nm and λ₂ = 266 nm) using a (pyrene) toluene solution was demonstrated for the first time. The influence of the experimental parameters, such as laser fluence, pulse number, and absorbing liquid, on the etch rate and the resulting surface morphology were investigated. The etch rate grew linearly with the laser fluence in the low and in the high fluence range with different slopes. Incubation at low pulse numbers as well as a nearly constant etch rate after a specific pulse number for example were observed. Additionally, the etch rate depended on the absorbing liquid used; whereas the higher absorption of the admixture of pyrene in the used toluene enhances the etch rate and decreases the threshold fluence. With a λ₁ = 266 nm laser set-up, an exceptionally smooth surface in the etch pits was achieved. For both wavelengths (λ₁ = 266 nm and λ₂ = 355 nm), LIPSS (laser-induced periodic surface structures) formation was observed, especially at laser fluences near the thresholds of 170 and 120 mJ/cm², respectively.     

Spontaneous growth of arsenic oxide micro-crystals on chemically etched MnAs surfaces

Micrometer-size crystals are observed to grow spontaneously on chemically etched MnAs surfaces. The wet chemical etching leaves a nearly exclusive pile of amorphous arsenic on the surface when the MnAs layer is etched incompletely. Using Raman spectroscopy, we identify that these micro-crystals are the arsenolite crystal of arsenic oxides. The manganese in the MnAs layer is oxidized by the hydrogen peroxide in the etch solution to MnO₂, which then works as the catalyst for the rapid oxidation and crystallization of the amorphous arsenic. Received: 14 November 2002 / Accepted: 18 November 2002 / Published online: 29 January 2003 RID="*" ID="*"Corresponding author. Fax: +49-30/20377-515, E-mail: takagaki@pdi-berlin.de     

248 nm laser etching of GaAs in chlorine and ozone gas environments

KrF laser etching of GaAs in Cl₂ and O₃ gas ambients by direct laser illumination is reported. The etch depth per pulse in Cl₂ was found to be linear versus the laser fluence on the sample in the 0.2–1.1 J/cm² range. It increased as a function of the Cl₂ pressure up to 6 Torr and slightly decreased for pressures above this value. It also decreased as a function of the laser repetition rate. Very smoothly etched surfaces were obtained after irradiation using the Cl₂ and O₃ etching gases. Auger analysis of the etched GaAs surfaces shows almost no traces of chlorine after etching in Cl₂, whereas a thick oxide layer of about 1500 Å thickness was found after etching in ozone.     

Modifying of etching anisotropy of silicon substrates by surface active agents

The influence of alcohol additives on etch rate anisotropy of Si(hkl) planes has been studied. The etching processes were carried out in 3 and 5 M KOH aqueous solutions saturated and non-saturated with alcohols. Isopropanol, 1-propanol and tert-butanol were examined. It has been showed that the etching process cannot be controlled only by the surface tension of the solution. Saturation of the etching solution with alcohols modifies etch rate anisotropy, lowering the ratio of the etch rate of (110) and vicinal planes to the etch rate of (100) plane. The morphology of Si(hkl) planes etched in 3 M KOH solution saturated with tert-butyl alcohol has been studied in detail. Smooth (331) and (221) planes have been achieved in this solution. The (100) plane turned out to be densely covered by hillocks, opposite to the (100) plane etched in weak-alkaline solution saturated with isopropanol. To explain this phenomenon, the mechanism of hillocks formation on Si(100) surface has been proposed.     

Thermal and carrier transport originating from photon recycling and non-radiative recombination in laser micromachining of GaAs thin films

Coupled thermal and carrier transports (electron/hole generation, recombination, diffusion and drifting) in laser photoetching of GaAs thin film is investigated. A new volumetric heating mechanism originating from SRH (Shockley–Read–Hall) non-radiative recombination and photon recycling is proposed and modeled based on recent experimental findings. Both volumetric SRH heating and Joule heating are found to be important in the carrier transport, as well as the etching process. SRH heating and Joule heating are primarily confined within the space-charge region, which is about 20 nm from the GaAs surface. The surface temperature rises rapidly as the laser intensity exceeds 10⁵ W/m². Below a laser intensity of 10⁵ W/m², the thermal effect is negligible. The etch rate is found to be dependent on the competition between photovoltaic and photothermal effects on surface potential. At high laser intensity, the etch rate is increased by more than 100%, due to SRH and Joule heating. Received: 24 January 2002 / Accepted: 11 April 2002 / Published online: 10 September 2002 RID="*" ID="*"Corresponding author. Fax: +1-310/206-2302, E-mail: xiang@seas.ucla.edu     

Effects of halogenated organic solvents on laser-induced backside wet etching of fused silica

Laser-induced backside wet etching of fused silica using a solution of pyrene dissolved in halogenated and non-halogenated solvents is presented. A significant influence of the solvent used on the etch rate and the etched surface appearance was ascertained. The etching of uniform and smooth surfaces with rates of ∼0.1 nm/pulse for laser fluences below 500 mJ/cm² is observed only for halogenated solvents. Furthermore, reduced threshold fluences, only small incubation effects, and a constant etch rate in dependence on the pulse number were found. The experimental data suggest an additional etch process at low laser fluences characterized by the very low etch rate and the smooth etching observed only with halogen-containing solvents. The generation of halogen radicals/compounds close to the heated surface due to the decomposition of the solvent causing the attack of the surface seems the most probable mechanism. PACS 81.65.Cf; 81.05.Kf; 79.20.Ds; 61.80.Ba; 42.55.Lt; 68.45.Da     

Laser projection-patterned etching of (100) GaAs by gaseous HCl and CH3Cl

Laser projection-patterned etching of GaAs in a HCl and CH₃Cl atmosphere performed using a pulsed KrF-excimer laser (=248 nm, =15 ns) and deep-UV projection optics (resolution 2 m) is reported. The etching process carried out in a vacuum system having a base pressure of 10^–6 mbar is shown to result from a purely thermochemical reaction. Etching takes place in two steps: (i) between the laser pulses, the etchant gas reacts with the GaAs surface-atomic layer to form chlorination products (mainly As and Ga monochlorides), (ii) local laser surface heating results in the desorption of these products allowing further reaction of the gas with the surface. The influence of the etching parameters (laser energy density, gas pressure and pulse repetition rate) on the etch rate and the morphology of the etched features was studied. Etch rates up to 0.15 nm per pulse, corresponding to the removal of 0.5 GaAs molecular layer, are achieved. The spatial resolution of the etching process is shown to be controlled by the heat spread in the semiconductor and by the nonlinear dependence of the etch rate on the surface temperature. As a result, etched features smaller or larger than the projected features of the photomask are achieved depending on the laser energy density. Etched lines having a width of 1.3 m were obtained at low fluences by the projection of 2 m wide lines onto the GaAs surface.     

Chemical etching of silicon by CO2-laser-induced dissociation of NF3

The pulsed infrared laser dissociation of NF₃ is reported for the first time, and is used to investigate silicon etching. The role played by collision-enhanced multiple-photon absorption and dissociation is considered, with data on the nonlinear decrease of the absorption cross-section with increasing pulse energy and increasing pressure presented. Using an experimental arrangement in which the laser beam is focussed parallel to the surface, the dissociation process induces spontaneous etching of silicon. Fluorinecontaining radicals diffuse from the focal volume to the surface where a heterogeneous chemical reaction occurs. Etching was monitored by use of a quartz-crystal microbalance upon which a thin film of amorphous silicon was deposited. For a surface with no previous exposure to the photolysis products, dissociation causes the formation of a surface layer prior to the onset of etching. X-ray photoelectron spectroscopy demonstrates this to be a fluorosilyl layer possessing a significant concentration of SiF₃ and SiF₄. In contrast, a surface already thickly fluorinated does not form a thicker layer once laser pulsing commences again. In this case, etching starts immediately with the first pulse. The etch yield dependencies on several parameters were obtained using silicon samples possessing a thick fluorosilyl surface layer. These parameters are NF₃ pressure, laser wavenumber, pulse energy, buffer gas pressure, and perpendicular distance from focal volume to surface. Modeling of the etch yield variation with perpendicular distance shows the time-integrated flux of radicals impinging on the surface to be inversely proportional to the distance. Attempts at etching SiO₂ under identical conditions were unsuccessful despite the evidence that thin native oxide films are removed during silicon etching.     

Influence of resists on reactive ion etching

In the course of plasma etching we can observe a loading effect, i.e. the etch rate depends on the size of the etched surface exposed to the plasma. This phenomenon was explained according to Mogab by the plasma active etch species depletion via a rapid etch reaction. But there exist more coomplicated systems, for example SiO₂-photoresist SCR17-CHF₃, where the SiO₂ surface can be etched and a polymer layer can grow on the photoresist surface. The etching of SiO₂ is also influenced by different resists in the case of differences in their chemical structure. The degree of electrode coating with a resist influences both the etch rate of the masking layer. This may be used for the control of the etching selectivity in the SiO₂-resist system independently of other process parameters.The author is grateful to Mr. Z. Pokorný for his help in preparing the SiO₂ layers used in all experiments.     

Laser-induced etching of titanium by Br2 and CCl3Br at 248 nm

A quartz crystal microbalance (QCM) has been used to study the KrF* excimer laser-induced etching of titanium by bromine-containing compounds. The experiment consists of focusing the pulsed UV laser beam at normal incidence onto the surface of a quartz crystal coated with 1 m of polycrystalline titanium. The removal of titanium from the surface is monitored in real time by measuring the change in the frequency of the quartz crystal. The dependence of the etch rate on etchant pressure and laser fluence was measured and found to be consistent with a two-step etching mechanism. The initial step in the etching of titanium is reaction between the etchant and the surface to form the etch product between laser pulses. The etch product is subsequently removed from the surface during the laser pulse via a laser-induced thermal desorption process. The maximum etch rate obtained in this work was 6.2 Å-pulse^–1, indicating that between two and three atomic layers of Ti can be removed per laser pulse. The energy required for desorption of the etch product is calculated to be 172 kJ-mole^–1, which is consistent with the sublimation enthalpy of TiBr₂ (168 kJ-mole^–1). The proposed product in the etching of titanium by Br₂ and CCl₃Br is thus TiBr₂. In the etching of Ti by Br₂, formation of TiBr₂ proceeds predominantly through the dissociative chemisorption of Br₂. In the case of etching with CCl₃Br, TiBr₂ is formed via chemisorption of Br atoms produced in the gas-phase photodissociation of CCl₃Br.     

Laser-assisted etching of diamonds in air and in liquid media

2 O, (CH₃)₂SO). Diamond samples are virtually transparent at this wavelength, and the coupling of laser radiation to diamond is via the formation of a thin graphitized layer at the diamond surface. The etching rate in liquid media is slightly higher than in air at otherwise equal conditions and is as high as 50 μm/s for etching with a scanning laser beam. Raman spectra measurements carried out on diamond samples etched in air show the presence of glassy carbon on the surface, whereas for samples etched in a liquid the diamond peak at 1332 cm^-1 dominates with significantly lower intensity of the glassy carbon peak. Electroless copper deposition on the laser-etched features is studied to compare the catalytic activity of the diamond surface etched in air with that etched in liquids. Possible mechanisms responsible for the observed difference both in the structure of the etched area and in the electroless Cu deposition onto the surface etched in various media (air or liquids) are discussed. Received: 2 August 1996/Accepted: 7 January 1997