The impact of O/Si ratio and hydrogen content on ArF excimer laser ablation of fused silica

The impact of stoichiometry and hydrogen content on the ArF excimer laser ablation characteristics of fused silica is investigated. Near-surface substoichiometic SiO x with x < 2 diminishes the ablation threshold as a result of increased absorption. The ablation rate is raised by an elevated hydrogen content. As conﬁrmed by mass spectrometric analysis, the laser-induced formation of substoichiometric near-surface layers within the ablation spot sustains absorption and ablation for consecutive laser pulses. [DOI


INTRODUCTION AND PROBLEM STATEMENT
Fused silica of high purity is one of the most important materials for a number of various optical components and devices.Due to its transmission characteristics, far ultraviolet or infrared laser wavelengths usually need to be applied for pure laser micro structuring without auxiliary means.In addition, high fluences are required for ablation, giving rise to thermally-induced disturbing effects such as the formation of melt or micro cracks.In order to achieve high-quality ablation at low fluence, different hybrid laser ablation methods are in hand, as for example laser-induced plasma-assisted ablation (LIPAA) [1,2], laser-induced backside wet etching (LIBWE) [3,4], laser etching at surface adsorbed layers (LESAL) [5,6] and laser-induced backside dry etching (LIBDE) [7].In the latter case, a thin coating, featuring a high absorption for the particularly used laser wavelength, is deposited onto the fused silica surface.For instance, the ablation threshold fluence is significantly decreased by the use of silicon monoxide (SiO) [8] or silicon suboxide (SiO x , where 1 < x < 2) [9,10] as absorbing coating.In addition to the deposition of a SiO or SiO x coating onto the surface, a near-surface layer within the fused silica silicon dioxide (SiO 2 ) bulk material can be chemically reduced to SiO x with the aid of hydrogenous atmospheric pressure plasmas [11].This technique represents an alternative solution for decreasing the required laser ablation threshold fluence [12,13].
Against this background, the influence of silicon suboxide layers of different composition on excimer laser ablation of fused silica was investigated in the present work in order to provide a comparison of the particular ablation thresholds and depths.Further, the chemical composition of such laser ablated fused silica surfaces was determined.

EXPERIMENTAL SETUP AND PROCEDURE
Investigations on the influence of different SiO x compositions on excimer laser ablation were carried out for three different samples and substrates, respectively, all from Heraeus GmbH: (i) fused silica Suprasil 2B (hereinafter called FS) with a defined stoichiometric ratio O/Si = x of 2, (ii) Suprasil 2A (O/Si = 2), coated with a 77 nm-thick SiO x -layer from Laseroptik GmbH with x ≈ 1 [14] (hereinafter called c-FS), and (iii) plasma treated Suprasil 2B (hereinafter called p-FS).For p-FS, a gradient oxygen-reduced near-surface layer within the SiO 2 bulk material, reaching its maximum substoichiometric O/Si-value of approx.1.88 at a sputter depth d s of 10 nm, was observed in previous work.The maximum depth d p of the plasma-affected bulk material is approx.100 nm [11,15].For the plasma treatment, an atmospheric pressure plasma jet source with an outlet nozzle diameter of 3 mm [16] was applied, using forming gas 90/10 (90 % N 2 , 10 % H 2 ) from Linde AG as process gas.The gas flow rate was 4.2 l/min.Plasma treatment was performed for 170 minutes at a plasma power 40 mW, applied to an area of 625 mm² (25x25 mm²) on the sample surface via moving the sample by a motorised xy-linear stage.The working distance from the plasma nozzle outlet to the sample surface was 8 mm.After preparing the samples, the transmission T was determined by the use of a spectrometer UV-VIS PE Lambda 19 from Perkin Elmer.
In order to determine the content of hydrogen as well as the stoichiometric ratio O/Si of the investigated samples, secondary ion mass spectroscopy (SIMS) was applied using a SIM spectrometer Quadrupol 4550 from Cameca.Here, the acceleration voltage of the caesium primary ions was 1 kV, where the investigated sputtered area was 0.36 mm² (600x600 µm²).SIMS was performed on the particular reference surfaces before laser ablation and on the bottom of the ablated spots after laser ablation.The initial calibration of the relative sensitivity factor (RSF) was carried out by applying three different standards with defined elementary concentrations, listed in Table 1.
Ablation experiments were performed employing an argon fluoride (ArF) excimer laser LPX 315i from Lambda Physik with an emission wavelength of 193 nm and a pulse duration of 20 ns.In order to achieve homogenous large-scale ablation, a non-imaging beam homogeniser was introduced to the laser beam path as shown in Figure 1.
The back focal length of this homogeniser, consisting of two perpendicularly-crossed cylinder micro lens arrays and a Fourier lens, was 100 mm.The homogeneous area projected onto the sample surface, i.e. the ablation spot area, was 1 mm².As determined by in-situ energy measurements during ablation, the fluences applied to the particular sample surface were in the range from 0.15 J/cm² to 5 J/cm².Front side ablation experiments were performed for the identification of the ablation threshold fluence and for the determination of the ablation depths.For this purpose a surface profiler Dektak 3030 Auto II from Veeco, Inc. with a vertical resolution of 2 nm was employed.

Hydrogen concentration of the different samples
Initially, the content of hydrogen within the investigated samples was determined by SIMS.For untreated fused silica Suprasil 2B, an averaged hydrogen concentration c(H) av of 0.88% was measured.For the Suprasil 2A substrate of the coated samples, c(H) av was 0.58%.A significant increase in c(H) av of 5.15% , i.e. approx.6-times higher with respect to untreated fused silica Suprasil 2B, was found in the case of plasma treated Suprasil 2B.This high value is most likely explained by diffusion of atomic hydrogen generated by dissociation of the used hydrogenous plasma working gas.This has already been observed in previous work [11].As a result of the different silicon suboxide compositions and the presence of hydrogen, the investigated samples feature quite differing transmissions T at the laser wavelength of 193 nm, which was applied for ablation.The particular values were T FS = 99.9%,T c−FS = 0% and T p−FS = 92.4%,respectively.Regarding these transmission values, one has to consider that different absorption-increasing mechanisms are involved.In the case of c-FS, absorption is mainly due to the SiO x layer whereas in the case of p-FS, the near-surface formation of SiO x as well as the implantation of hydrogen [17]- [19] into deeper regions of the fused silica bulk material contributes to modified optical characteristics of the bulk material.

Influence of the stoichiometry on ablation characteristics
The ablation threshold fluence F th was defined as the mean value in between the highest applied fluence without any material removal and the lowest one, where considerable ablation was observed.The measured ablation threshold fluences   F th for 3 and 10 pulses are nearly identical and lead to an average threshold fluence F thaver of 3.43 J/cm² for FS, 1.08 J/cm² for c-FS and 2,6 J/cm² for p-FS.The threshold for c-FS is given by that value, where the fused silica substrate and not only the coating is ablated.Figure 2 shows the dependency of the ablation depths (10 laser pulses per spot) on the laser fluence Φ for the three different investigated sample types.
For instance, the averaged ablation rate at approx.3.7 J/cm² (see Figure 2), where ablation took place for all sample types, amounts to 50 nm/pulse for FS, 100 nm/pulse for c-FS and 140 nm/pulse in the case of p-FS.Even though the lowest ablation threshold is found in the case of c-FS, the highest ablation rate was obtained on p-FS.Also for ablation at higher fluences, the highest ablation depth was found in the case of p-FS whereas c-FS and FS feature a comparable ablation behaviour as shown in Figure 3. Here, the particular ablation depths were plotted vs. the applied dose in order to provide a normalised and well-evaluable presentation.The dose is given by the product of the fluence Φ and the number of pulses.
The observed differing ablation characteristics for p-FS on the one hand and both c-FS and FS on the other hand can be explained by the above-mentioned plasma-induced significant increase in hydrogen within the glass bulk material, giving rise to an enhanced coupling of incoming laser irradiation.In contrast, the absorbing SiO x -coating on c-FS is removed after the first few laser pulses.
For a more detailed investigation of the influence of implanted hydrogen, SIMS was also performed at the ablation spot bottoms after applying 3 and 10 laser pulses at a fluence of approx.5 J/cm².For the evaluation, the near-surface layer with a thickness of 10 nm was not considered due to the fact that the measurement was distorted during the formation of a sputter equilibrium between the involved elements and surfaceadherent contaminations such as hydrocarbons.Even when excluding the first 10 nm for such analysis, the stoichiometric ratio O/Si was generally decreased compared to that of the non-irradiated material for all investigated samples.After ablation applying 10 laser pulses, the averaged O/Si ratio for a sputter depth range from 10 to 25 nm was 1.92 for both FS and c-FS.However, the most significant reduction was found in the case of plasma treated fused silica, shown in Figure 4, where the O/Si ratio after 10 laser pulses was 1.44 for the above-mentioned sputter depth range.This significant higher laser-induced chemical reduction could be explained by the increased presence of hydrogen, implanted by the plasma pretreatment.Such implanted hydrogen can be assumed to act as reaction partner for oxygen embedded in the SiO 2 network, where the reaction results from laser induced heating of a near-surface layer according to The increased effect for ablation with higher numbers of pulses as shown in Figure 4 can thus be explained by the higher dose applied leading to an increased accumulation of reaction products.
Improved ablation of glasses and glassy materials after the first pulse is a well-known phenomenon.It is commonly explained by a number of different, interacting effects such as roughening and crystallisation as a result of short-time melting and re-solidification of the glass substrate material [20].
In the case of LIBWE, an altered state of the absorption coefficient of the fused silica after laser ablation was reported.
The authors attributed this modification to the high temperature and the interactions between the fused silica and the pyrene/toluene absorber solution [21].The above-presented SIMS results provide a further aspect contributing to the explanation of this effect.Since even in the case of pure FS the formation of substoichiometric SiO x after the first laser pulse was verified it can be stated that the accompanying increase in absorption gives rises to enhanced coupling of incoming laser irradiation once the surface was ablated.To our best knowledge, this is the first work reporting this effect.

CONCLUSIONS
The investigated samples feature considerably different ablation characteristics.The lowest ablation threshold was found for coated fused silica (c-FS) and is easily explained by the high absorption of the applied SiO x -coating.With respect to pure fused silica (FS), the ablation threshold could be reduced by the presented plasma pre-treatment.For ablation at higher number of pulses and fluences, respectively, the highest ablation rate was found in the case of plasma-pre-treated fused silica (p-FS) which is most likely due to the implanted hydrogen and accompanying effects such as a continuous formation of a new SiO x -layer on the bottom of the ablation spot after each laser pulse.This assumption is supported by the presented SIMS analysis of the ablated surfaces.It was shown that for all investigated sample types such SiO x -layers of different stoichiometric composition result from laser ablation.

FIG. 4
FIG. 4 Stoichiometric ratio O/Si for p-FS samples after 3 and 10 pulse laser ablation including particular fitted progression vs. sputter depths.

TABLE 1
Elementary concentrations of the applied SIMS calibration standards.