Dynamics of secondary contamination from the interaction of high-power laser pulses with metal particles attached on the input surface of optical components

S. G. Demos; J. C. Lambropoulos; J. C. Lambropoulos; R. A. Negres; M. J. Matthews; S. R. Qiu

doi:10.1364/OE.27.023515

1. Introduction

Laser-induced damage on optical components exposed to high-average-power or peak-intensity laser pulses is a well-recognized issue that affects the operational parameters as well as the cost of operation of such systems. It is generally accepted that the origin of damage initiation is either defects incorporated during the material manufacturing process (such as crystal growth, bulk material manufacturing, surface polishing, coating deposition, etc.) or contamination from optics handling and/or exposure to the laser operation environment. The detrimental effect of contamination on the laser-damage resistance of optical surfaces has long been recognized [1] but has received more attention during the last 20 years with the development of high-power laser systems designed for the investigation of inertial confinement fusion and fundamental laser–matter interactions under extreme conditions [2–15]. These studies provided definite evidence that contamination induced from handling and environment must be minimized if not eliminated; in fact, with the significant advancement in the optics manufacturing processes, the relative contributions of defect-induced versus particle-induced, contamination-driven damage have balanced out if not reversed.

Metallic particles are commonly found contaminants on surfaces of optical components in high-energy laser systems. Researchers have made great efforts to understand the impact of these contaminants on their laser performance. A study by Matthews et al. [16] investigated the morphological evolution of particles of several different relevant materials placed on the input (front) surface of antireflective (AR)-coated fused-silica test optics under successive 351-nm, ∼9 J/cm², nanosecond-pulsed laser exposures. These materials, either opaque or transparent, were relevant to the operation environment of the National Ignition Facility (NIF) laser; therefore, they can be considered as potential sources of contamination. The results suggested that for the metals studied (aluminum and stainless steel), the initial laser pulse caused melt flow associated with strong localization and adhesion to the silica surface. This type of secondary contamination can lead to Fresnel diffraction of incident light and damage induced on the exit surface of the optic [17]. In addition, metallic particles cause surface pitting with pit depths as deep as 600 nm, which can lead to increased scattering of the propagating laser beam [18].

For the case of high-reflectivity dielectric coatings, contaminant particles are often responsible for limiting their lifetime and performance [19]. Qiu et al. [20] examined different protective capping layers as the means to circumvent the potential negative impact of such contaminant particles. The post laser exposure (1053-nm, 14-ns pulses) surface morphology indicated that the plasma formed was causing “burning” of the mirror surface and the recoil of the molten particle along the mirror surface. Furthermore, it was suggested that the recoil momentum of the partially melted particle because of the ejected plasma is particle shape dependent [21]. Additional studies focused on the interaction of particles attached on the output (exit) surface of optical components. These studies involved time-resolved imaging of the dynamics of the interaction of laser pulses with contaminants and modeling of the energy-coupling mechanisms [22], the dynamics and temperature of the formed plasma [23], and the ensuing surface modifications [24]. In this geometry, the momentum transferred to the particle from the formation of plasma facilitates its ejection from the surface, which enabled quantification of the kinetic properties of the particle.

The knowledge attained from this previous work is directly applicable to the present study, which examines the dynamics of the interaction of microscale, nominally spherical metal particles attached on the input (front) surface of optics. However, the momentum attained by the particle is thrusting the particle against the surface. The resulting response of the particle is therefore nontrivial, including the mechanism of secondary contamination; previous work has provided only the phenomenology of the final modifications, while the intermediate steps were speculative and qualitatively described. The present work involves time-resolved microscopic shadowgraphy with adequate spatial and temporal resolution to resolve details of the dynamics of plasma formation, shock-wave expansion, particle ejection, and secondary contamination by small molten droplets that separated from the original particle. This work has been designed to complement prior recent work [16–24]. For this reason, the experiments were performed using stainless steel (SS) and titanium (Ti) particles as model contaminant on bare fused-silica surfaces as well as high-reflectivity multilayer dielectric coatings under single-pulse, nanosecond laser excitation at 355 nm and 1064 nm, respectively.

2. Experimental arrangement

The basic experimental system used in this work has been described in detail elsewhere [25]. Several adaptations were introduced to perform this work. Figure 1 shows a schematic depiction of the system including the two excitation geometries used in insets (a) and (b). The pump laser was operating at 355 nm, producing ≈8-ns (FWHM) pulses, or 1064 nm, producing ≈10-ns pulses. A different excitation geometry and substrate were used with each excitation wavelength. Specifically, excitation at 355 nm was used in combination with stainless-steel particles (316L alloy) dispersed on the input surface of a 5-cm-round, 0.5-mm-thick commercially available silica substrate that was positioned vertically. The pump beam in this case was directed at 36° with respect to the normal to the sample’s surface (z axis) on the substrate impinging directly on the particles. Stainless-steel particles were procured with nominally spherical shapes and an average diameter of 27 μm; the observed distribution of sizes spanned from about 15 μm to about 45 μm. In addition, titanium particles dispersed on the surface of an ~7-μm-thick multilayer dielectric high reflector at 45° and p polarization were studied under excitation at 1064 nm, where the pump beam was directed at 45° with respect to the normal to the sample’s surface. The ${SiO}_{2} / {HfO}_{2}$ multilayer dielectric coatings were deposited on a 5-cm-round, 10-mm-thick commercially available BK7 and optimized to provide reflectivity of >99.5% at 1053 nm. The Ti particles were nominally spherical, having an average diameter of 45 μm, but smaller sizes, similar in diameter to the stainless-steel particles, were selected to be exposed to the pump pulses. The beam profile of the pump laser impinging on the surface of the substrate was nearly flattop (with ~25% local intensity variations) and had an elliptical shape (because of the angle of incidence of the laser beam) with a minor axis of about 315 μm. The pump laser fluence was about 12.5 ± 2 J/cm² under 355-nm excitation and about 17.5 ± 2 J/cm² under 1064-nm excitation, both of which are relevant to the operational fluences used in large-aperture laser systems.

Fig. 1 Schematic diagram of the experimental system. Insets show the photoexcitation geometry for (a) stainless-steel (SS) particles located on uncoated fused-silica substrate and (b) titanium (Ti) particles located on a dielectric high reflector. CCD: charge-coupled device; HR: high-reflectance.

Download Full Size | PDF

Two identical microscope systems providing up to 50 × optical magnification were positioned orthogonally to each other and used to image the area containing the particle along the surface of the sample, referred to as transmission-view (TV) microscope, and normal to the surface, referred to as side-view (SV) microscope. Time-resolved images were acquired using pulsed illumination obtained from the probe laser operating at 532 nm, producing 180-ps (FWHM) pulses. The output of the probe beam was split to illuminate the particle parallel and orthogonally to the substrate surface, making it possible to acquire dynamic images of the particle’s response to the laser pulse at predetermined delay times with respect to the time of peak intensity of the pump pulse. The probe laser fluence was of the order of 1 mJ/cm² and had no impact on the behavior of the particles under exposure to probe pulses alone. In this manuscript, we define the x–y plane as the substrate surface, with an x and y axis along the horizontal and vertical directions, respectively; the z axis is normal to the substrate.

3. Experimental results

Figure 2(a) shows a side-view image of the location of the SS particle obtained at about a –3-ns delay (peak-to-peak time delay of the probe and pump pulses as recorded using a fast photodiode and a 15-GHz oscilloscope) under exposure to a fluence of ≈12 J/cm² by the pump laser pulse. The pulse is propagating toward the left-hand side at 36° with respect to the orthogonal direction to the surface. This image is the ratio of the transient image (acquired at an ≈–3-ns delay) to the image acquired prior to irradiation of the particle by the laser pulse. This image normalization method helps suppress spatial inhomogeneity of the probe-pulse illumination and increases contrast of the feature of interest. Figure 2(b) is the same image as in Fig. 2(a) with outlined (by dashed lines) features of interest. Specifically, feature 3 is the SS particle, about 18 μm in diameter, which maintained its original position on the surface after this short delay time (the particle did not have time to react). A plasma plume was generated, however, and is observed in the image as a darker feature (2) along with the shock wave (1) at its very early stage as it expands outward from the particle. Since the particle is only partially illuminated by the pump pulse, plasma is forming only within this exposed area and subsequently expands. The shock wave is launched from this partial area of plasma formation and has not yet reached the interface with the substrate, as is indicated by the outline of the shock (1) in Fig. 2(b).

Fig. 2 Side-view images of the location of the SS particle (18-μm diameter) acquired at about a −3 ns delay under 355-nm laser exposure of ≈12 J/cm², capturing the position of the shock wave (1) at different stages of its expansion along with the plume (2) and the particle (3). (a) and (c) are different events, (b) is the same as (a) with the features of interest outlined by dotted lines. The laser illuminates the particle from the right-hand side.

Download Full Size | PDF

The image shown in Fig. 2(c) was obtained from a different but practically identical event at a slightly longer delay and captures the shock wave immediately after it had reached the interface with the substrate. The shock wave is subsequently expanding outward as indicated by the arrows in Fig. 2(c). However, the initial movement of the shock wave (and the plume) toward the surface brings into the surface vaporized material contained within the plume. In addition, this mechanism is associated with exposure of the surface to the high temperature of the plume, which leads to modifications that are referred to as “plasma burning” in [19] and deposition of plume material onto the surface. This represents the first mechanism of secondary contamination of the surface from the interaction of the particle with the laser.

The images shown in Fig. 3 are complementary to those shown in Fig. 2 and were recorded by the TV microscope through the transparent substrate. The image in Fig. 3(a) was obtained at about a –4-ns delay and shows the particle, about 17 μm in diameter, along with the generated plume and shock wave. Figure 3(b) is a digital magnification of the image shown in Fig. 3(a) to help better visualize the features of interest including the particle (3), the plume (2), and the shock wave (1). The pulse propagates toward the surface at 36° with respect to the orthogonal direction (viewer), illuminating the particle from the left-hand side. As a result, both the plume and shock wave are formed and launched toward the left-hand side. As demonstrated in Fig. 3(c), acquired at about a 0-ns delay, the shock wave continues to expand around the particle. The average speed of the shock wave is of the order of 10 km/s. This initial response demonstrated that a significant amount of laser energy was absorbed by the particle (andthrough the interaction of the laser with the plasma), leading to the ejection of evaporated material. This in turn leads to the transfer of momentum (as discussed in detail in [22]) that exerts a force on the particle toward the surface. In addition, the presence of vaporized material from the particle suggests that there should also be liquid material formed on the particle surface. As the particle begins to move, the liquid material can detach from the particle, leading to additional secondary contamination of the surface. This process is exemplified by the images shown in Figs. 4–6.

Fig. 3 Transmission-view images of the location of the SS particle (17-μm diameter), acquired at about a –4-ns delay, capturing the asymmetric expansion of the shock wave (1) and the plume (2) along the substrate surface, as well as the particle (3). (a) and (c) are different events, (b) is a digital magnification of (a) with the features of interest outlined by dotted lines. The laser illuminates the particle from the left-hand side.

Download Full Size | PDF

Fig. 4 Transmission-view image of the location of the SS particle (25-μm diameter) acquired at about a 125-ns delay showing the movement of the particle from its original position (indicated by the solid red line) and the onset of separation of liquid droplets. (b) is the same as (a) with the features of interest outlined. The laser illuminates the particle from the left-hand side. The particle moves to the right and away from the substrate while the droplets move toward the substrate.

Download Full Size | PDF

Fig. 5 Three examples (a, b, and c) of the motion of SS particles at a 1025-ns delay as captured by the SV microscope (a-i, b-i, and c-i) and the TV microscope (a-ii, b-ii, and c-ii) along with the final TV images (a-iii, b-iii, and c-iii). The laser beam is at 36° with respect to the z axis along the x–z plane. The inset shows the direction of ejection of each particle.

Download Full Size | PDF

Fig. 6 Side-view, time-resolved images acquired at about a 514-ns delay, capturing the ejection of three different SS particles (similar events) from the surface and the separation of liquid material that is carried onto the surface.

Download Full Size | PDF

Figure 4(a) shows the normalized TV image of the particle (about 25 μm in diameter) at a 125-ns delay. During this delay time, the particle moved slightly to the right; as a result, the ratio image shows the area no longer occupied by the particle as a bright feature (on the left-hand side). On the other hand, the area that represents the shift in the particle is a darker feature (on the right-hand side). To better highlight this movement of the particle, Fig. 4(b) shows the same image with the initial location of the particle highlighted by a red circle, while the perimeter of the particle at a 125-ns delay is shown with dashed line. The contrast of the image shown in Fig. 4(b) is enhanced to better show the dark projections, indicated by arrows, located outside of the particle’s perimeter; these projections represent liquid droplets that start separating from the particle. These droplets are observed at the opposite side to that of laser irradiation, which is the direction in which the formed plume thrusts on the particle (and the generated droplets).

The representative images shown in Fig. 5 capture the motion of the particle at much longer delays, ~1 μs. Figure 5(a-i) shows the SV image of the particle at a 1025-ns delay. The laser beam is at 36° with respect to the z axis along the x−z plane. The particle is about 21 μm in diameter, and the image indicates that its motion along the z axis is about 13 μm. This motion results from the recoiling of the particle away from the surface after the momentum acquired initially thrusted it toward the surface. Figure 5(a-ii) shows the same particle at the same delay time as acquired by the TV microscope. The particle is observed to have moved about 42 μm along the x axis. This means that the particle has been ejected from the surface at an angle of about 73° with respect to the z axis at a speed of about 43 m/s. Figure 5(a-iii) shows the final image acquired at the end of the process, where only the contamination by liquid droplets that have separated from the particle is visible on the substrate surface. Comparison of the transient and final images allows one to better understand this secondary contamination process. Specifically, the droplets in Fig. 5(a-ii) located between the initial position (bright feature after normalization) and the transient position of the particle at a 1025-ns delay exhibit no change in position in the final image. Therefore, these droplets were attached on the surface at an earlier time. The double droplet shown in Fig. 5(a-ii) by an arrow with asterisk is split into two droplets in the final image [Fig. 5(a-iii), also denoted with an asterisk]. In addition, the second droplet shown with a plain arrow in the transient image is no longer visible in the final image. These observations indicate that the droplets located ahead of the particle are still in transit and some can eventually land on the surface.

A second example is shown in the next set of microscope images. Specifically, the particle shown in Figs. 5(b-i) and 5(b-ii) is about 32 μm in diameter and has shifted by about 17 μm along the z axis and 27 μm along the x axis at a 1025-ns delay. This correlates to an ejection angle of about 58° with respect to the z axis at a speed of about 32 m/s. Furthermore, the final image [Fig. 5(b-iii)] shows a trail of contamination droplets that were left on the surface after the particle was ejected.

A similar scenario arises from the third example captured by the images shown in Figs. 5(c-i)–5(c-iii). The particle in this case is spheroidal, having short and long semi-axes of about 28 μm and 54 μm, respectively. The particle is ejected from the surface with a speed of about 34 m/s at about 56° with respect to the z axis. The TV image acquired at 1025 ns [Fig. 5(c-ii)] shows a number of liquid droplets that have detached from the particle. The final image [Fig. 5(c-iii)] shows that several droplets located hundreds of micrometers from the particle’s original position (the most distant indicated by arrows) have contaminated the surface.

The above results suggest that the contamination of the surface by liquid droplets occurs in two separate phases: The first phase is very early in the particle ejection process (before the 1025-ns delay), where a significant fraction of the melted part of the particle is deposited on the surface. The second phase occurs after the ejection of the particle, leaving a trail of droplets along the path of the particle. We postulate that the second mechanism of contamination by droplets results from the rotation of the particle after its ejection, which facilitates further separation of the remaining melted material on the particles that are in part directed toward the surface. To better understand the first phase, additional experiments were performed at shorter delays to capture the separation of the melted material during the particle ejection process.

Figure 6 shows transient images acquired at a 514-ns delay for three typical examples. These images were normalized by division of the transient image by the final image (after the particle was removed). The particles were relatively smaller than in previous events, with diameters of about 14 μm, 21 μm, and 18 μm, respectively. These images show the trail of molten material separating from the particle located between the surface and the particle during its detachment from the surface. This behavior may be attributed to the fact that the original momentum imparted on the particle is toward the surface, while the melted material was formed on the side of the particle exposed to the laser beam. While the recoiling of the particle causes its ejection from the surface, melted material separates from the particle and continues its trajectory toward the surface. The images captured at 500-ns delay shown in Fig. 6 demonstrate that melted material has separated from the particle. At least a portion of this melted material eventually reaches the surface near the original position of the particle.

The above results, incorporating Figs. 2–6, were obtained using SS particles dispersed on a bare silica substrate under 355-nm excitation at 36°. Figure 7 shows representative examples using 1064-nm excitation at 45° on Ti particles attached to high-reflectivity (at 45°), multilayer dielectric coatings. The fundamental difference between the excitation geometry in these two configurations, depicted by insets (a) and (b) in Fig. 1, is that the reflection of the laser light on the mirror leads to exposure of the particles from two orthogonal directions (a) via direct exposure and (b) by the reflected light impinging on the side of the particle. Thisdifference has three important implications: First, the total laser fluence irradiating the particle is greatly enhanced by nearly twice that of the nominal laser fluence. Second, a larger part of the particle is exposed to the laser beam, including the area near the point of attachment of the particle to the coating surface. Third, the direction of the (total) momentum transferred to the spherical particle (from the direct irradiation and the irradiation by the reflected beam) is roughly along the surface of the mirror, thereby setting the particle in motion close to the surface. As a result, a larger part of the particle’s surface liquefies, including the area close to the surface of the mirror. In addition, the motion of the particle closer to the surface of the mirror further enhances contamination of the mirror by the liquid droplets produced on the particle itself.

Fig. 7 Transmission-view, time-resolved images representing characteristic examples of the dynamics of the interaction of Ti particles attached on the input surface of a 45° multilayer dielectric coatings at delays of (a-i) –3 ns, (b-i) 114 ns, and (c-i) 500 ns along with the corresponding final (static) images (a-ii), (b-ii), and (c-ii), respectively.

Download Full Size | PDF

The above considerations are supported by the exemplary results shown in Fig. 7, capturing the transient and final images for three case examples. All images shown in Fig. 7 are normalized by the image acquired before the arrival of the pump pulse while the pump pulse is propagating from the left side (as shown in Fig. 5). In the first case example, the transient image was acquired at a −3-ns delay [Fig. 7(a-i)] and captures the contour of the shock wave produced from the exposure of the ≈25-μm-diam particle to the 1064-nm laser pulse. In comparison to the analogous image shown in Fig. 3(c) for the SS particle attached on the transparent substrate, the shock wave formed on the mirror is considerably more asymmetric with respect to the position of the particle. This is not surprising since the two-sided exposure of the particle (from direct and scattered beams) discussed above supports plasma formation on the left (as viewed) side of the particle, which becomes the primary direction of shock-wave propagation. The results of this interaction are captured in the final image [Fig. 7(a-ii)] showing extensive contamination of the surface of the dielectric coatings by liquid droplets produced by the particle. Similarly, Fig. 7(b-i) shows the transient image acquired at a 114-ns delay, while Fig. 7(b-ii) shows the final image of this site revealing the induced contamination. The diameter of the Ti particle in this case is ≈22 μm, and its displacement from the original position in the transient image suggests that its speed is ≈80 m/s. The outline of the particle in the transient image Fig. 7(b-i) is barely visible as a result of the large amount of liquified material (droplets) produced that has separated from the particle. The final image Fig. 7(b-ii) shows contamination of the surface with a very significant amount of liquefied material from the particle that was attached to the surface. A similar result is demonstrated by the third case example shown in Fig. 7(c-i) (transient image) and Fig. 7(c-ii) (final image). The particle in this case was ≈24 μm and its ejection speed was estimated to be ≈90 m/s. The above results are representative of numerous measurements that all provided a very consistent account of the dynamics of the interaction of the Ti particles with the 1064-nm laser pulse.

4. Modeling

In this section we attempt to obtain a practical approximation of (a) the amount of melted material produced during exposure of the metal particles to laser pulses and (b) the pressure applied on the substrate during recoil ejection of the particles. We first proceed to provide some estimates of the thermal response of the metallic particles based on heat conduction. A detailed description of the energy transfer by a laser beam illuminating a metal particle has been considered in [26]. The interaction occurring in the context of this work is dynamic, initially involving exposure of the particle to the laser beam (that can be described as in [26]) followed by the formation of the expanding plume before the peak of the laser pulse (as demonstrated in Figs. 2 and 3). This gives rise to energy being rapidly transferred away from the particle via the ejection of evaporated material (to form the plume) and screening of the laser beam by the formed plasma. It is therefore difficult to estimate what portion of the laser pulse energy is deposited in the particle in the form of heat. To obtain insight into this process, we consider the particle heating process itself. The thermal diffusion characteristic times (τ_th) are found from

τ_{th} = R^{2} / (4 α),

where R is the particle radius (14 μm for SS and Ti particles) and α is the thermal diffusivity (3.5 × 10⁻⁶ m²/s for SS and 9.3 × 10⁻⁶ m²/s for Ti.) We find that the thermal diffusion times are 14 μs for the SS and 5.2 μs for the Ti particles. These times are significantly longer than the laser pulse widths (of the order of 10 ns). Therefore, relative to the time needed to heat the entire particle, we conclude that we can assume that the particles are heated by the laser beam instantaneously.

We can also assess the extent of near-surface melting by using the relation between the location-dependent transient temperature rise T(x,t) near a surface, bounding a semi-infinite solid resulting from an instantaneously deposited heat flux E_s,

T (x, t) - T_{initial} = [E_{s} / \sqrt{π k ρ c t}] \exp (- x^{2} / 4 α t),

where k is the thermal conductivity (13.4 W/m.K for SS and 22 W/m.K for Ti), T_initial is the initial temperature, t is the delay time, and x represents the depth of the melt layer. This solution neglects the temperature dependence of the thermal properties, and the size (2R) of the particles. It is only applicable when the depth x is much smaller than the particle size (2R). Using data for stainless steel, we show in Fig. 8 the computed thickness of the melt layer into a surface exposed to different heat flux varies as a function of the delay time from the termination of the laser pulse. Estimates for Ti particles exhibit similar trends. These depths into the particle surface are smaller than the particle size, thereby justifying the approach of the surface bounding a semi-infinite solid.

Fig. 8 The dependence of the melt layer thickness into a stainless-steel surface as a function of delay time following the instantaneous energy deposition of different heat flux values (as noted on the curves) associated with energy absorption by a ~10-ns laser pulse.

Download Full Size | PDF

The results shown in Fig. 8 suggest that during the early delay times extending to ≈1 μs, during which the processes of particle ejection from the surface and droplet ejection form the particle take place, the melted layer is only a few micrometers in depth. This in turn limits the number and size of the droplets produced. It must be noted that the ejection of droplets and liquid material removes heat energy from the particle, which leads to its faster cooling.

This process also prevents a larger volume of the particle to be melted and as a result, the inner volume of the ejected particles remains in the solid state phase. Comparison of the particle size prior to its exposure to the laser pulse to that captured by the transient images (see Fig. 5) does not indicate a measurable change in the particle diameter. Based on the results shown in Fig. 6, this may indicate that only a small fraction of the laser energy is being absorbed by the laser beam and converted in to heating the particle.

We can also estimate an upper bound on the contact average pressure resulting from the particle recoil ejection from the fused-silica surface. We assume that the recoil event is equivalent to an average pressure pave, exerted by a spherical particle (radius R), in a Hertzcontact-like event, whereby the particle exerting kinetic energy (in terms of the normal velocity V)

E_{kinetic} = 1 / 2 m V^{2}

is equal to the elastic strain energy stored in the fused-silica surface

E_{kinetic} = E_{strain} ​ .

Under the conditions of the Hertz contact between a particle of radius R under a normal force F, in contact with a flat surface, the elastic strain energy caused by the elastic deformation is

E_{strain} = (8 / 15) E^{'} δ^{5 / 2} R^{1 / 2},

where E′ is the reduced elastic modulus E′ = E/(1–v²), with E the fused-silica Young’s modulus (72 GPa) and ν the Poisson ratio (0.17.) Here δ is the elastic deformation caused by the force F. On the other hand, the radius a of the area of contact is given by

a^{3} = (3 / 4) (F / E') R, or a = {(R, δ)}^{1 / 2}

and the average pressure is

P_{pave} = F / (π a^{2}),

while the maximum pressure P_max in the center of the contact region, is (3/2) P_pave. Therefore, knowledge of the ejected particle normal velocity V allows for the calculation of the strain energy E_strain and, therefore, the force of contact F and the radius a of the area of contact.

Using a velocity of 32 m/s at an angle of incidence of 36° for the ejected steel particles, we find that the recoil surface force is F = 0.15 N, and the area of contact has radius a = 2.7 μm with a surface deflection of 540 nm. The corresponding average contact pressure is then 6.2 GPa (approximately 8% of the elastic Young’s modulus) with the maximum pressure at 9.3 GPa. It is clear that these stresses are large, and related to induced subsurface damage, especially since it is well known that fused silica densifies at such levels of compressive stress. It is important to note here that the densification of fused silica is even further facilitated by large shear stresses in addition to pressure [1]. The maximum shear stress (σ_s) occurs at a depth of about a/2 below the surface, with a magnitude of

σ_{s ​_{\max}} \sim P_{\max} / 3 .

We can estimate, therefore, that a maximum shear stress of about 3 GPa occurs at a depth of about 1 to 2 μm below the surface.

The analysis above neglects any adhesion effects between the metallic spheres and the glass substrate. To assess the effect of adhesion, we use the Johnson–Kendall–Roberts (JKR) model or Derjaguin–Muller–Toporov (DMT) model of elastic contact [27–29]; therefore, the mechanical force F is enhanced by an adhesive contribution ΩγπR, where Ω is a dimensionless number in the range of 4 to 6 (depending on the JKR or DMT theories), γ is the surface adhesive energy, and R the particle radius. We can estimate an upper bound for γ by using the cohesive energy (i.e., critical energy release rate) of fused silica via $γ \leq K_{c}^{2} / E,$ where K_c is the silica fracture toughness $(0.75 MPa . \sqrt{m})$ and E is its Young’s modulus. We then find that the adhesive contribution to the force F is less than 0.0021 N, amounting to less than 2% of the mechanical force of 0.15 N found above. We therefore conclude that adhesive contributions are negligible.

Previous work related to the interaction of nanosecond laser pulses with metallic particles located on the exit surface of transparent substrates [22] considered the role of four different mechanisms that can introduce ejection of the particle following its interaction with the laser pulse and the role of attractive forces involved. These momentum transfer mechanisms include: (a) thermal expansion that pushes the particle surface into the substrate; (b) recoil momentum transfer resulting from the ejection of material from the particle surface; (c) confined plasma pressure; and (d) partially confined plasma pressure. Within the excitation geometry relevant to this work (i.e., particle is located on the input surface), only the first two mechanisms are applicable (as there is no plasma forming at the interface between the particle and the substrate). Within the first mechanism, the velocity attained by the particle is provided by:

V \approx \frac{β A I}{4 ρ C}

where β is the thermal expansion coefficient f the material, A is the material absorptivity, I is the laser intensity, ρ is the material density, and C is the specific heat capacity. This formulation suggest that the speed attained by the particle is on the order of 1 m/s or less.

The second relevant mechanism can be adequately described using the concept of momentum coupling, originally introduced to explain the interaction of plane metal targets with large-aperture laser beams [30]. Within this approach, the key parameter is the coupling coefficient C_m, which represents the ratio of momentum (ΔP) imparted to the particle and the incident laser energy (E), as C_m = ΔP/E. In this case, the velocity attained by the particle is provided by:

V = \frac{3 ζ C_{m} E}{4 ρ R},

where ζ is a shape-dependent coefficient (with ζ = 0.3 for a spherical particle). The estimated velocities in this case (assuming ρ = 8 g/cm² and C_m = 3 dyn/W) are of the order of 10 m/s, which is of the same order as the velocities observed in our experiments.

5. Discussion

The results discussed above suggest that there are three contamination mechanisms following the interaction of laser pulses with metallic particles attached on the input surface of optics. The first mechanism is related to the initial plume expansion toward the surface, which would leave a layer of contamination around the particle. The second mechanism is related to the liquid material formed on the particle that separates during the ejection of the particle from the surface. This material is subsequently deposited around the initial particle location and mostly on the side of the particle along the direction of laser irradiation. The third mechanism is related to droplets of liquid material that separate from the particle after its ejection. As a result, these droplets can be deposited at significant distances from the initial location of the particle.

The trail of the droplets deposited on the surface via the third mechanism allows one to appreciate the direction of propagation of the particles after their ejection from the surface. For nearly spherical particles, it was observed that the particles are ejected along (or close to) the plane defined by the direction of laser beam propagation and the orthogonal direction to the surface (along the x–z plane). This is exemplified by the images shown in Figs. 4(a) and 4(b). The image shown in Fig. 5(c), however, demonstrates that the particle was also moved away from the x–z plane, exhibiting movement also along the y axis. This particle is spheroidal in shape, where the long semi-axis is orthogonal to the direction of particle propagation along the x–y plane. This is not surprising since the motion of the particle is related to the momentum transferred by the plume [22]. Because the expansion of the plume is vertical to the surface, the attained momentum in the direction orthogonal to the elongated side of the particle is larger than that on the shorter side. This in turn pushes the particle asymmetrically along the direction vertical to the elongated surface. Consequently, the direction of particle ejection depends strongly on its shape, as previously discussed [21,22]. This can be particularly important for irregularly shaped contamination particles, especially those with extended, nearly flat surfaces. The effects described here can lead to thrusting of the particle closer to the surface and subsequently an extended (spatially) contamination by liquid droplets.

The inset shown in Fig. 5 depicts the direction of propagation of the particles for the three examples shown. It is observed that the angle of ejection from the surface is larger than the incidence angle of the laser beam. This indicates that the interaction of the particle with the substrate during the initial impulse of the particle against the surface involves a loss of momentum prior to ejection of the particle from the surface; this fraction of the initial impulse imparted onto the substrate can lead to plastic deformation and/or mechanical damage. This type of damage from contamination particles has been reported previously [16–21].

Overall, the results obtained using the Ti particles dispersed on the multilayer dielectric coating surface suggest a more-severe secondary contamination compared to the contamination induced by SS particles on bare silica. This is assigned to the excitation geometry, namely the fact that laser light reflected on the coating illuminated the particle from the side, thereby increasing the total exposure fluence on the particle and creating liquefied material over a larger part of its surface, including near its point of attachment on the coating surface.

The behaviors observed in this work are expected to be analogous to those occurring under a wide range of excitation conditions when the interaction of the laser pulse with the particle supports an ablation event. For example, the morphology of secondary contamination under ultrashort pulsed excitation [31] is similar to that observed with the nanosecond pulses used in this work and can be fully explained using the dynamic processes described here.

6. Conclusion

This work provides direct experimental measurements of the ejection of model (spherical) metal contamination particles located on the input surface of transparent and reflective optical elements following exposure to a single ns laser pulse. The results provide information on the dynamics and expansion geometry of the shock wave and plume as well as the speed and directions of the ejected particles. Furthermore, the time-resolved images provide a better understanding of the mechanisms for secondary contamination of the surface following the initial interaction of the particle with the laser beam. Specifically, data suggest that there are three distinct mechanisms of secondary contamination arising from (a) the interaction of the expanding hot plume with the cold surface that may be related to a thin film and nanoparticle deposition on the surface around the particle location, primarily in the direction toward the laser beam; (b) separation of liquefied material from the main particle as it rebounds from the surface, leading largely to material being deposited very close to the site of the particle; and (c) droplets of liquid material separating from the particle at later times (after its detachment from the initial point of contact), resulting in contamination farther from the initial site of the particle. These results enhance the existing understanding attained from previous studies that were based on post-mortem examination of laser-irradiated sites (as in [19–22]) by providing the underlying dynamics of distinct processes involved at different temporal regimes.

Funding

U.S. Department of Energy (DE-AC52-07NA27344); National Nuclear Security Administration (DE-NA0003856).

Acknowledgment

This report was prepared as an account of work sponsored by an agency of the U.S. Government. Neither the U.S. Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the U.S. Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the U.S. Government or any agency thereof.

References

1. I. A. Fersman and L. D. Khazov, “The effect of surface cleanliness of optical elements on their radiation resistance,” Sov. J. Opt. Technol. 37, 627–629 (1971).

2. M. D. Feit, A. M. Rubenchik, D. R. Faux, R. A. Riddle, A. B. Shapiro, D. C. Eder, B. M. Penetrante, D. Milam, F. Y. Genin, and M. R. Kozlowski, “Modeling of laser damage initiated by surface contamination,” Proc. SPIE 2966, 417–424 (1997). [CrossRef]

3. F. Y. Génin, M. R. Kozlowski, and R. Brusasco, “Catastrophic failure of contaminated fused silica optics at 355 nm,” Proc. SPIE 3047, 978–986 (1997).

4. F. Y. Génin, K. Michlitsch, J. Furr, M. R. Kozlowski, and P. A. Krulevitch, “Laser-induced damage of fused silica at 355 nm and 1064 nm initiated at aluminum contamination particles on the surface,” Proc. SPIE 2966, 126–138 (1997). [CrossRef]

5. D. M. Kane and D. R. Halfpenny, “Reduced threshold ultraviolet laser ablation of glass substrates with surface particle coverage: A mechanism for systematic surface laser damage,” J. Appl. Phys. 87(9), 4548–4552 (2000). [CrossRef]

6. S. Palmier, I. Tovena, L. Lamaignère, J. L. Rullier, J. Capoulade, B. Bertussi, J. Y. Natoli, and L. Servant, “Study of laser interaction with aluminum contaminant on fused silica,” Proc. SPIE 5991, 59910R (2005). [CrossRef]

7. J. Honig, M. A. Norton, W. G. Hollingsworth, E. E. Donohue, and M. A. Johnson, “Experimental study of 351-nm and 527-nm laser-initiated surface damage on fused silica surfaces due to typical contaminants,” Proc. SPIE 5647, 129–135 (2005). [CrossRef]

8. W. H. Gourdin, E. G. Dzenitis, D. A. Martin, K. Listiyo, G. A. Sherman, W. H. Kent, R. K. Butlin, C. J. Stolz, and J. A. Pryatel, “In-situ surface debris inspection and removal system for upward-facing transport mirrors of the National Ignition Facility,” Proc. SPIE 5647, 107–117 (2005). [CrossRef]

9. M. A. Norton, C. J. Stolz, E. E. Donohue, W. G. Hollingsworth, K. Listiyo, J. A. Pryatel, and R. P. Hackel, “Impact of contaminates on the laser damage threshold of 1ω hr coatings,” Proc. SPIE 5991, 59910O (2005). [CrossRef]

10. S. Palmier, J. L. Rullier, J. Capoulade, and J.-Y. Natoli, “Effect of laser irradiation on silica substrate contaminated by aluminum particles,” Appl. Opt. 47(8), 1164–1170 (2008). [CrossRef] [PubMed]

11. S. Palmier, S. Garcia, and J.-L. Rullier, “Method to characterize superficial particulate pollution and to evaluate its impact on optical components under a high power laser,” Opt. Eng. 47(8), 084203 (2008). [CrossRef]

12. Y. Ye, X. Yuan, X. Xiang, X. Cheng, and X. Miao, “Laser cleaning of particle and grease contaminations on the surface of optics,” Optik (Stuttg.) 123(12), 1056–1060 (2012). [CrossRef]

13. T. Jitsuno, H. Murakami, K. Kato, E. Sato, K. Mikami, S. Motokoshi, N. Miyanaga, and H. Azechi, “Recent progresses on insights of laser damage mechanisms and influence of contamination in optics,” Proc. SPIE 8786, 87860B (2013). [CrossRef]

14. X. Ling, G. Wang, Y. Zhao, X. Liu, and J. Shao, “Laser-induced damage of the optical coatings due to organic contamination in vacuum,” Appl. Surf. Sci. 270, 346–351 (2013). [CrossRef]

15. K. E. Gushwa and C. I. Torrie, “Coming clean: Understanding and mitigating optical contamination and laser induced damage in advanced LIGO,” Proc. SPIE 9237, 923702 (2014). [CrossRef]

16. M. J. Matthews, N. Shen, J. Honig, J. D. Bude, and A. M. Rubenchik, “Phase modulation and morphological evolution associated with surface-bound particle ablation,” J. Opt. Soc. Am. B 30(12), 3233–3242 (2013). [CrossRef]

17. F. O. Génin, M. D. Feit, M. R. Kozlowski, A. M. Rubenchik, A. Salleo, and J. Yoshiyama, “Rear-surface laser damage on 355-nm silica optics owing to fresnel diffraction on front-surface contamination particles,” Appl. Opt. 39(21), 3654–3663 (2000). [CrossRef] [PubMed]

18. E. Feigenbaum, S. Elhadj, and M. J. Matthews, “Light scattering from laser induced pit ensembles on high power laser optics,” Opt. Express 23(8), 10589–10597 (2015). [CrossRef] [PubMed]

19. C. J. Stolz, S. R. Qiu, R. A. Negres, I. L. Bass, P. E. Miller, D. A. Cross, J. A. Davis, S. Sommer, C. C. Widmayer, B. J. MacGowan, P. K. Whitman, and P. J. Wegner, “Transport mirror laser damage mitigation technologies on the National Ignition Facility,” Proc. SPIE 10691, 106910W (2018). [CrossRef]

20. S. R. Qiu, M. A. Norton, R. N. Raman, A. M. Rubenchik, C. D. Boley, A. Rigatti, P. B. Mirkarimi, C. J. Stolz, and M. J. Matthews, “Impact of laser-contaminant interaction on the performance of the protective capping layer of 1 ω high-reflection mirror coatings,” Appl. Opt. 54(29), 8607–8616 (2015). [CrossRef] [PubMed]

21. S. R. Qiu, M. A. Norton, J. Honig, A. M. Rubenchik, C. D. Boley, A. Rigatti, C. J. Stolz, and M. J. Matthews, “Shape dependence of laser–particle interaction-induced damage on the protective capping layer of 1ω high reflector mirror coatings,” Opt. Eng. 56(1), 011108 (2016). [CrossRef]

22. S. G. Demos, R. A. Negres, R. N. Raman, N. Shen, A. M. Rubenchik, and M. J. Matthews, “Mechanisms governing the interaction of metallic particles with nanosecond laser pulses,” Opt. Express 24(7), 7792–7815 (2016). [CrossRef] [PubMed]

23. C. D. Harris, N. Shen, A. M. Rubenchik, S. G. Demos, and M. J. Matthews, “Characterization of laser-induced plasmas associated with energetic laser cleaning of metal particles on fused silica surfaces,” Opt. Lett. 40(22), 5212–5215 (2015). [CrossRef] [PubMed]

24. R. N. Raman, S. G. Demos, N. Shen, E. Feigenbaum, R. A. Negres, S. Elhadj, A. M. Rubenchik, and M. J. Matthews, “Damage on fused silica optics caused by laser ablation of surface-bound microparticles,” Opt. Express 24(3), 2634–2647 (2016). [CrossRef] [PubMed]

25. R. N. Raman, R. A. Negres, and S. G. Demos, “Time-resolved microscope system to image material response following localized laser energy deposition: Exit surface damage in fused silica as a case example,” Opt. Eng. 50(1), 013602 (2011). [CrossRef]

26. C. D. Boley, S. A. Khairallah, and A. M. Rubenchik, “Calculation of laser absorption by metal powders in additive manufacturing,” Appl. Opt. 54(9), 2477–2482 (2015). [CrossRef] [PubMed]

27. K. L. Johnson, Contact Mechanics (Cambridge University Press, 1985).

28. K. L. Johnson, K. Kendall, and A. D. Roberts, “Surface energy and the contact of elastic solids,” Proc. R. Soc. Lond. A Math. Phys. Sci. 324(1558), 301–313 (1971). [CrossRef]

29. B. V. Derjaguin, V. M. Muller, and Yu. P. Toporov, “Effect of contact deformations on the adhesion of particles,” J. Colloid Interface Sci. 53(2), 314–326 (1975). [CrossRef]

30. C. R. Phipps Jr., T. P. Turner, R. F. Harrison, G. W. York, W. Z. Osborne, G. K. Anderson, X. F. Corlis, L. C. Haynes, H. S. Steele, K. C. Spicochi, and T. R. King, “Impulse coupling to targets in vacuum by KrF, HF, and CO₂ single-pulse lasers,” J. Appl. Phys. 64(3), 1083–1096 (1988). [CrossRef]

31. K. R. P. Kafka and S. G. Demos, “Interaction of short laser pulses with model contamination microparticles on a high reflector,” Opt. Lett. 44(7), 1844–1847 (2019). [CrossRef] [PubMed]

Dynamics of secondary contamination from the interaction of high-power laser pulses with metal particles attached on the input surface of optical components

Abstract

1. Introduction

2. Experimental arrangement

3. Experimental results

4. Modeling

5. Discussion

6. Conclusion

Funding

Acknowledgment

References

Cited By

Figures (8)

Equations (10)

Optics Express