Scholarly article on topic 'Solderjet Bumping as a Versatile Tool for the Integration of Piezoelectric Deformable Mirrors'

Solderjet Bumping as a Versatile Tool for the Integration of Piezoelectric Deformable Mirrors Academic research paper on "Materials engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Physics Procedia
OECD Field of science
Keywords
{"Deformable mirror" / unimorph / "Solderjet Bumping" / "Laser soldering"}

Abstract of research paper on Materials engineering, author of scientific article — Thomas Burkhardt, Matthias Goy, Marcel Hornaff, Michael Appelfelder, Claudia Reinlein

Abstract A deformable mirror (DM) is a device that aims to compensate laser-induced mirror deformation and thermal lensing in the optical system. The mounting of membrane based DM with screen-printed actuators is crucial, as stress may deform the membrane and change their characteristics (shape, piezoelectric deflection, natural frequency). We present the laser-based Solderjet Bumping (SJB) technique to assemble mounts for piezoelectric-activated DM. The discussed polymer-free joining offers advantages, such as improved temporal stability and low outgassing, over adhesive bonding. We evaluate the optimum number of solder joints with respect to resonance behavior by finite elements analysis and experimental measurements. Long-term evaluation over a period of more than four years shows no significant change of resonance behavior. Thus, we prove the SJB bonding technique to be stable for dynamic applications over several years, and consider it a versatile tool for integration of DM.

Academic research paper on topic "Solderjet Bumping as a Versatile Tool for the Integration of Piezoelectric Deformable Mirrors"

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Physics Procedia 83 (2016) 540 - 548

9th International Conference on Photonic Technologies - LANE 2016

Solderjet Bumping as a versatile tool for the integration of piezoelectric deformable mirrors

Thomas Burkhardf, Matthias Goya, Marcel Hornaff, Michael Appelfeldera, Claudia Reinleina'*

aFraunhofer Institute for Applied Optics and Precision Engineering, Albert-Einstein-Strasse 7,07745 Jena, Germany

Abstract

A deformable mirror (DM) is a device that aims to compensate laser-induced mirror deformation and thermal lensing in the optical system. The mounting of membrane based DM with screen-printed actuators is crucial, as stress may deform the membrane and change their characteristics (shape, piezoelectric deflection, natural frequency). We present the laser-based Solderjet Bumping (SJB) technique to assemble mounts for piezoelectric-activated DM. The discussed polymer-free joining offers advantages, such as improved temporal stability and low outgassing, over adhesive bonding. We evaluate the optimum number of solder joints with respect to resonance behavior by finite elements analysis and experimental measurements. Long-term evaluation over a period of more than four years shows no significant change of resonance behavior. Thus, we prove the SJB bonding technique to be stable for dynamic applications over several years, and consider it a versatile tool for integration ofDM.

©2016PublishedbyElsevierB.V. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Bayerisches Laserzentrum GmbH

Keywords: Deformable mirror, unimorph, Solderjet Bumping, Laser soldering

1. Introduction

Mounting is crucial to performance of piezoelectric actuated membrane mirrors as mount-induced moments and constraints change the mirror's characteristics such as flatness, piezoelectric stroke, and resonance behavior. Such deformable mirrors (DM) must have a high quality optical surface and offer large piezoelectric stroke. Further, their dynamic properties must be reliable. Only if these requirements are fulfilled, can DM be applied to compensate for wavefront aberrations in optical systems. The large stroke of DM relies on thin membranes with large aspect ratios; typical diameters of several tens of millimeters and thicknesses below 0.5 mm.

Unimorph deformable mirrors can be applied to compensate for atmospheric wavefront aberrations (Feinleib et al., 1974; Steinhaus et al., 1979; Ma et al., 2013) as well as aberrations in laser systems (Aleksandrov et al., 2007; Cheriaux et al., 2007; Zhi-Jun et al., 2009), e.g. thermal lensing (Primmerman et al., 1976; Aleksandrov et al., 2005). Unimorph mirrors with 214 actuators have been proposed by Ma et al. (2013) for atmospheric applications. Furthermore, thermal-piezoelectric deformable

* Corresponding author. Tel.: +49-3641-807-343 ; fax: +49 -3641-807-604 . E-mail address: claudia.reinlein@iof.fraunhofer.de

1875-3892 © 2016 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Bayerisches Laserzentrum GmbH

doi:10.1016/j.phpro.2016.08.056

mirrors (TPDM) with piezoelectric and thermal actuators as well as integrated sensors requiring 89 reliable contacts have been introduced for the compensation of thermal lensing by Reinlein et al. (2014).

The mirror's dynamic properties need to be reliable as current control strategies use the dynamic behavior of the actuated mirror membrane in order to improve the control performance in terms of control rate and deviation. It is therefore beneficial to identify a reliable technology for mounting and electrical contacting of DMs. Different approaches are worth discussing: clamping or material fit bonding such as adhesive bonding or soldering.

Clamping is a basic principle that relies on form and force fit joining using retainers, e.g. spring clips, snap rings, threaded retaining rings (Yoder, 2004). It could be realized by pressing the membrane's front surface against the flat surface of the mount either by a retaining ring, a flange type retainer, or by screws or springs on the rear surface. The point or line contact of such mountings might lead to potentially large contact stresses that could impair the functionality of the system due to surface deformation. Elastomeric mounting provides a means of distributing such preloading forces but introduces organic materials to the system.

Compared to mechanical clamping material fit joining allows further miniaturization of the joint and the minimization of mount-induced stresses. Bonding techniques using polymeric adhesives or soldering by metallic solder alloys can provide such material fit joints. Both processes could produce extensive plane or point-shapedjoining areas. Localizedjoints minimize stress but decrease mechanical stiffness and may therefore change the dynamic behavior of the assembly. Adhesive bonding seems quite obvious as it is a simple and inexpensive technique. Unfortunately it has certain drawbacks: low thermal conductivity of the adhesive, shrinkage, reduced long-term stability, moisture induced degradation, and eventually outgassing. Solderjoints provide a higher thermal conductivity, higher mechanical strength, and improved stability over time with respect to environmental influences. The all in-organic bonding materials prevent outgassing and offer significant advantages for many fields of application. Soldering, however, requires a solderable metallization on both components and the thermal reflow of the solder alloy. Laser-based soldering is well suited for the joining of optical components made of fragile and brittle materials such as glasses, ceramics, and optical crystals due to its localized and minimized input of thermal energy. In particular the limited heating of the local soldering is a key advantage over high temperature processing of a global reflow of the assembly. Such a reflow process could lead to significant deformation or disruption of the LTCC (Low Temperature Cofired Ceramics) membrane and depolarization ofthe piezoelectric elements.

We therefore propose a laser-based soldering technique, Solderjet Bumping (SJB), for mounting and packaging ofdeformable mirrors for both the attachment of the mirror membrane to a mount and the electrical contacting of the piezoelectric actuators to a printed circuit board. It is based on prior work for mounting of optical elements, e.g. lenses (Burkhardt et al., 2015a) and unimorph deformable mirrors (Reinlein et al., 2013). It simplifies the manufacturing process and the localized heating during soldering is beneficial for a low stress mounting of sensitive components such as the used mirror membranes. SJB bonds have been characterized with respect to their thermal and dimensional stability of assemblies joined by SJB, bridging large gaps of about 100 |im between the components. They exhibit a length change of ~25 nm and a tilt of ~1 arcsec when subjected to a temperature variation from 10°C to 40°C over a year, supporting the claim of long term stability (Lorenz et al., 2015). We have also demonstrated the stress relaxation in solderjoints by photoelasticity measurements of optical path difference (Burkhardt et al., 2015b). However, the capability of the SJB technique to influence DM dynamic properties, and the long-term stability of soldered DM mounts have never been addressed.

This study examines DM mounting and actuator contacting by SJB. The aim of this work is the determination ofthe necessary number of discrete solder joints in order to obtain membrane characteristics comparable to a fixed mount by analysis of their resonance behavior. Moreover, three pilot samples were set up to analyze the resonance behavior at the optimum number of solder joints in order to detect mounting-induced moments and constraints. In addition, we evaluate the long-term stability of SJB bonds by resonance measurements over a period of more than four years.

2. Experimental setup and procedures

2.1. Laser-based Solderjet Bumping

Joining, packaging, and assembly of sensitive optical components benefit from flux-free bonding techniques that prevent contamination of surfaces. Precisely controlled heating, e.g. laser-based, and solder reflow lead to minimized areas of thermal influence and are thus advantageous for high-precision joining. The localized heating avoids thermal dealignment during assembly and enables the sequential build-up of systems. Different techniques of heating solder alloys by laser irradiation are proposed using either thin film solder layers (Banse et al., 2005), solder pastes (Stauffer et al., 2005), solder preforms heated at thejoint (Hoult et al., 2003), or thejetting of laser-molten solder droplets (Beckert et al., 2010). Furthermore Stein et al. (2014) extend the use of this principle to high temperature joining/brazing using Cu89Snll alloy preforms and a novel ceramic nozzle type. A contact-free alternative, the Pick&Align soldering technique, has been suggested by Faidel et al. (2012) using integrated resistive layers.

The proposed SJB technique allows for a flux-free and contact-free bonding of optical components. In contrast to thin film soldering techniques, it is possible to bond materials with lower demands to surface quality - especially with respect to surface

figure - since small gaps can be bridged by the solder. Using this advantage, it is also feasible to adjust components prior to joining in six degrees of freedom and within complex 3D-integrated geometries. It uses spherical solder preforms of soft solder alloys in a diameter range of 40 цm to 760 цm. The jetting of liquid solder droplets (reflown by an IR laser pulse) from a placement capillary provides excellent thermal contact ofthe alloy with the components. The bond head integrates solder volume feeding, reflow, and application of the solder in a compact device allowing for a highly automated and flexible use. SJB allows the bonding of a broad range of materials in heterogeneous combinations. The formation of a metallic solder joint using components made of non-metallic materials requires a wettable metallization layer applied to the components; as provided by thin film (e.g. physical vapor deposition, PVD) or thick film (e.g. by screen printing of metal pastes) processes. Sputtered three layer systems using titanium adhesion layer, a platinum diffusion barrier, and a noble gold finish preventing oxidization and acting as a wetting surface, provide superb conditions for wetting of liquid solder droplets.

Using SJB we demonstrated sub-micron accuracy packaging of numerous micro-optical assemblies, including direct fiber coupling and the hermetic sealing of an endoscopic tip, (Beckert et al., 2010), the bonding of polarization maintaining fibers (Ojeda et al., 2013), the assembly of a compact and robust solid-state laser for the ExoMars mission (Ribes et al., 2015), and the low stress mounting of lenses (Burkhardt et al., 2015a). The assembly of a multi-beam deflection array for next-generation electron beam lithography outlines the features of this technique with respect to vacuum compatibility and highest component placement accuracy (Burkhardt et al., 2011 a).

2.2. Thermally-piezoelectric deformable mirror sample & fabrication processes

We introduce a deformable mirror sample as a simplified model of a TPDM, based on a design outlined by Reinlein et al. (2013) to reduce effort and cost for the presented investigations (see Fig. 1). The front surface of a TPDM is fixed to its mount and plated with galvanic copper thick-film. A dielectric high reflective thin-film is deposited on top of this copper layer after ultra-precision turning. The mirror mount is made from WCu or AlSi70 with a thermal expansion closely matched to the mirror membrane. Differences in thermal expansion may lead to tensile stress exceeding the rupture strength of the materials, causing cracks to form and propagate. However, the mount ofthe model used in this investigation is made from stainless steel as thermal loads will not be applied. In addition, the LTCC substrates are simplified and provide no thermal actuators and sensors. The dimensions (layers thicknesses and the relevant diameters) equal those ofthe TPDM setup.

Insertion holes Solder bumps

Fig. 1. Schematics of simplified mirror sample with detailed view of solderjoining geometry.

The model mirror substrate is a cup-shaped low-temperature co-fired ceramic (LTCC, DuPont 951) with a membrane thickness of 200 цm and a diameter of 44 mm. Additionally, the membrane is reinforced by an 600^m-thick LTCC annulus rim between the diameter 35 mm and 44 mm. A piezoelectric actuator (PZT disc, PIC151, PI Ceramic), with a thickness of 500 цm and a diameter of 35 mm, is adhesively bonded onto the rear face ofthe LTCC substrate. The front and the rear face ofthe PZT disc are coated with a copper-nickel alloy to provide conductive layers as electrodes. The front surface of the LTCC-mirror substrate is coated with titanium and copper (PVD), which enables both the possibility of generating a reflective surface (the copper serves as a starting layer for a galvanic copper thick-film) and the contact area for the solderjet process. The counterparts for the soldering process are the gold plated conical insertion holes (0 700 цm) located at the sloped inner ring area of the mount. The distance between mount and LTCC during the deposition of the solder drops is 100 цm that will be filled with the liquid solder. The mount itself has an outer diameter of 70 mm and an overall height of 9 mm. The inner diameter and the clear aperture are 50 mm and 36 mm, respectively.

2.3. Mirror characterization

The prediction of fabrication induced stress for (deformable) mirrors is made by Laser Doppler Vibrometry (LDV). LDV is a non-contact measurement of surface vibrations; both deflection and velocity. With the laser beam from the LDV directed at the mirror surface the vibrational amplitude and frequency can be calculated from the Doppler shift of the reflected laser beam

frequency. We use a Polytec OFV-511 fiber optical system for measurement of single-point vibrations, e.g. the mirror center to determine the first natural resonance; and a Polytec PSV-400 scanning vibrometer (SLDV) for mapping vibrations over a surface, e.g. to determine the higher modes of deflection. The single point LDV uses a sinusoidal electrical voltage with an amplitude ofl V and an offset of 1 Vto drive the piezoelectric actuator with a chirp signal between 1 kHz and 5 kHz. The SLDV applies a periodic chirp signal in the frequency range between 1 Hz and 20 kHz with an amplitude and an offset of 1 V, respectively, to the mirrors actuators. A frequency resolution of ~1.5 Hz is achieved with FFT over 12,800 lines. The lateral scanning point distance is 0.75 mm.

3. Numerical modeling and simulation

We use the finite elements analysis software CoventorWare 2010 to investigate the effects of number of solder joints on the resonance behavior and deflection shapes of the TPDM. The resonant deflection frequencies (and shapes) of an annular, point-wise fixed membranes are compared with an annular fixed membrane (reference). The objective is the identification of the required number of discrete solder bumps (or resulting solder joints). As quality criteria we chose the resonance frequency and deflection shape difference between the ideal annular fixed membrane and the discretely annular fixed membrane. Linear extruded brick elements and the Pave, QMorph algorithm are used for meshing of the model with element sizes of 100 цm in planar direction and the thickness of the individual layers. The full model is then solved by the MemMech solver. The material properties used for the finite element analysis are shown in Table 1.

Table 1. Material properties used for finite element analysis.

PZT LTCC Solder bump

PIC 151 DuPont 951 Sn3Ag0.5Cu

Density / kg-m3 7,760 3,100 7,400

Young's Modulus / GPa 59.4 130 44

Poisson Ratio / 1 0.34 0.27 0.34

We apply three solder bumps equispaced on a diameter of 40 mm and simulate the first nine resonance frequencies and deflection shapes of the mirror membrane. Afterwards we place another three bumps intermediate to these bumps but on the same radius. This has been repeated for 12, 24, and 48 bumps. Simulations are also drawn out to evaluate the influence of the bumps height. We simulate heights of 100 цm, 10 цm, and 0 цm (fixation at the bump diameter). As a result, the first resonance frequency increases with respect to the number of applied bumps converging to the resonance obtained of the ideal annular fixed membrane (Table 2). An increase in bump height lowers the resonance frequency due to a reduced mechanical stiffness of the system. However, the differences for different bump heights are below 3.5% and are therefore negligible. These results support the possible use of SJB for solderjoints bridging a gap of up to 100 цm between mirror substrate and mount.

Table 2. Simulation results of first resonance frequencies for different bump heights and number of bumps.

Number of bumps 0 ^m 10 ^m 100 ^m

3 bumps 1,897 Hz 1,881 Hz 1,831 Hz

6 bumps 2,450 Hz 2,434 Hz 2,382 Hz

12 bumps 2,648 Hz 2,635 Hz 2,589 Hz

24 bumps 2,731 Hz 2,723 Hz 2,693 Hz

48 bumps 2,761 Hz 2,755 Hz 2,737 Hz

A possible issue relating to the mounting of such membranes is a print-through of the solder joints - a surface deformation based on the mechanical fixation. Fig. 2 shows the deflection shape of a fixed mount (left column) and the residual of the deflection shape for the mounting with discrete solder bumps. The residual is the difference between deflection shape of fixed and discrete mount using normalized amplitudes. Mounting the mirror membrane using three and six bumps clearly shows the expected print-through of the discrete joints. Considering the first deflection shape of the 12 bumps model (shown with highlighted bump positions), the difference of deflection is negligible. Furthermore, this effect decreases with an increasing number of bumps (24 and 48). We therefore opted for the application of 24 bumps to mount the DM based on the small deflection shape difference to the fixed annulus, and the reasonable production effort.

Fig. 2. Deflection shapes as a function of the number of solder bumps. Absolute deflection is shown for fixed model and the residual of the deflection shape for discrete bump models (difference to the fixed model). Red indicates a displacement in positive direction while blue indicates a displacement in negative direction. The 12 bumps model is shown with highlighted bump positions.

4. Experimental results and discussion

The first step ofthe mounting process is the optimization ofthe laser soldering parameters with respect to maximization ofthe bond force. Using Sn3Ag0.5Cu solder spheres with a diameter of 760 ^m we find max. shear forces per single solder joint of 15N to 24N for the material combination of W85Cu, AlSi70 vs LTCC (Burkhardt et al., 2011b). Similar bond forces are achieved with stainless steel mounts. In the second step, we evaluate the number of solder joints on the resonance behavior ofthe membrane. The first nine resonance frequencies and resonant deflection shapes ofthe membrane are then measured for 3, 6, 12,

and 24 solder pads and compared to the results of numerical simulations shown in section 3. Three model mirror substrates are manufactured as described in section 2.2.

Each mirror membrane is characterized at four different boundary conditions: 3, 6, 12, and 24 solder bumps. The application of the soldering bumps is divided in four runs delivering three bumps at the two first runs, six bumps at the third run, and twelve bumps at the last run, respectively. The bumps ofthe first run were placed in an angle of 120 degrees around the center ofthe mirror. The dynamic membrane behavior is characterized after each ofthe listed mounting steps. The first resonance frequency at the membranes center is measured by the one channel vibrometer (Polytec OFV-511) and shown in Fig. 3. Single point vibrometry is used to keep the time between subsequent soldering processes to a minimum. With increasing number of solder joints the first resonance frequency increases and approaches asymptotically the frequency for the fixed annulus. The measurement results of the three samples are in accordance with the simulation results, albeit marginally lower. The general trend of increasing resonance frequency is clearly reproduced by the mirror models.

After completing the fabrication of the 24 solder bumps in the three model mirrors, we use the scanning vibrometer Polytec PSV 400 to determine the deflection shape of the higher resonance modes. We compare the deflection shapes calculated by CoventorWare with the measured surfaces and visually inspect them for solder bump print-through. Not all of the calculated deflection shapes can be measured as the 3rd and 5th recessive modes are suppressed by the dominant modes. Table 3 further lists the first three resonance frequencies (fh f2, fs) and figures of merit: quality factor (Q factor) and the ratio of the higher order resonance frequencies {f2/fi,fs/fi) ofthe mirror membranes depending on the numbers of bumps. The Q-factor is a measure for the damping of the first resonance frequency as it is calculated by the ratio of resonator's bandwidth relative to its center frequency. A high Q-factor implies low mechanical losses and high sensity in sensing applications. The mean deviation between calculated and measured frequency is 9% for three solder pads and it decreases to 3% for 24 solder pads. The numerical model shows f1 = 2,693 Hz for 24 bumps. The Q factor ofthe vibration moderately increases with the number of solderjoints achieving values between 67 and 126 in case of24 solderjoints.

Table 3. First resonance frequency and figures ofmerit for the number bumps used forbonding. Measurements for 6 and 12 bumps use LDV whereas 24 bumps use SLDV. For comparison: simulation models showf = 2,693 Hz for 24 bumps.

Solder Sample 1 Sample 2 Sample 3

bumps (2011 | 2016) (2011 | 2016) (2011 | 2016)

fi! Hz 2,289.3 2,257.4 2,339.2

6 Q factor 38.7 36.2 35.2

fil Hz 2,489 2,447.2 2,536.3

Q factor 40.4 37.9 38.7

fil Hz 2,606 2,606 2,603 2,612 2,609 2,668

Q factor 126.3 77.4 99.5 104 67.3 101

f2 / fi 2.04 2.01 1.99 2.02 2.06 2.05

fs / fi 3.46 3.51 3.43 3.41 3.6 3.47

We have done the first characterization ofthe sample mirrors in November 2011 and repeated them in March 2016. Table 4 and Fig. 4 therefore compare the results of the "as soldered" with the resonance behavior after long-term storage. The

comparison ofthe first resonance frequency/! between the measurements taken 2010 and 2016 (see Tab. 3) reveals marginal differences only. This indicates a mechanically stiff bond. However, we observe slight changes in the Q factor. It increases for sample 3, remains mainly constant for sample 2 while it decreases significantly for sample 1. The changes ofthe the Q factor go along with slight deviations ofthe ratio f2/fi and fs/fi. Although we cannot detect a clear tendency or correlation, we consider these changes small.

Table 4 lists the the computed resonance frequencies and the values obtained by the scanning vibrometer. A major similarity between the simulations and the measurements is observed. In general the calculated values by CoventorWare are higher than experimental values which could be attributed to the overestimation of the Young's modulus of one of the layers or thickness differences of those layers. The evident minor differences between the samples can be explained by different thicknesses ofthe adhesive layer between the LTCC mirror substrate and the PZT disc, the lateral position of the PZT disc, or the LTCC layer thickness itself.

Table 4. Resonance frequencies (by FEA simulation and measured by Scanning Vibrometer).

Mode Simulation fn /Hz Sample 1 fn /Hz (2011 | 2016) Sample 2 fn /Hz (2011 | 2016) Sample 3 fn /Hz (2011 | 2016)

1* 2,693 2,606 2,606 2,603 2,612 2,609 2,668

2nd 5,505 5,328 5,235 5,170 5,277 5,392 5,481

3rd 5,505 N/A N/A N/A N/A N/A N/A

4«> 9,564 9,028 9,130 8,941 8,911 9,401 9,250

5* 9,564 N/A N/A N/A N/A N/A N/A

6* 11,101 10,202 10,239 9,938 9,975 10,252 10,427

Figure 4 displays the frequency response of the three sample mirrors between 1 kHz and 20 kHz and compares the results of 2011 and 2016. In general, the first and the second resonant frequencies (the two peaks) remains stable between the samples and time. A closer look at the first resonance frequency (2.6 kHz) reveals that there is a maximum deviation of 0.46% between the sample mirrors and only a deviation of about 0.115% (averaged) between the measurements in 2011 and 2016. The broadening ofthe resonance peaks corresponds with the decreasing Q factor shown in Tab. 3. The very low changes ofthe first resonance frequency over the long time (more than 4 years) indicates a mechanically and long-term stable bond between the plated LTCC and the mount.

1.00.8' 0.6' 0.4' 0.2' 0.0'

.1«! I I I i I — LTCC PZT 1 2011 — LTCC PZT 1 2016

: 1th ¡6th

J J

Frequency/Hz 10000

Fig. 4. Frequency response ofsample 1 at the center ofthe membrane.

For additional visualization the measured deflection shapes of sample 1 are shown in Figure 5. The mode shapes show an expected behavior compared to the numerical model as shown in Fig. 2. Samples two and three (not pictured here) exhibit comparable mode shapes. No print-through of the solder joints is observed proofing SJB as a technique to mount membrane based deformable mirrors.

Fig. 5. Deflection modes of sample 1 measured by Scanning Vibrometer.

5. Conclusion

We evaluated the integration deformable mirror using the Solderjet Bumping technique to fixate a piezoelectric membrane to its mount by discrete solderjoints. A finite elements simulation model was set up to detect the influence of the number of solder joints on the resonance behavior. We demonstrated that the number of solder bumps influences the first resonance frequency and the deflection shape of the membrane. A number of 24 solder joints, equally spaced on the diameter of the membrane was evaluated as a reasonable compromise between high resonance frequency and manufacturing effort. Three samples were set up and characterized by laser Doppler vibrometry. In agreement with the simulation the first resonant frequency increases with the number of solderjoints approaching the theoretical resonance behavior of an annular fixed deformable mirror. The samples with 24 solder joints, equally spaced on the diameter of the membrane are also analysed by scanning laser-Doppler vibrometry to detect deflection mode shapes. It was further shown that print-through ofthe solderjoints can be prevented for higher numbers of joints. Measurements after more than four years show no significant change of the resonance behavior withing the confidence range ofthe measurements. We thus consider SJB an appropriate and long-term stable technique for the mechanical contacting of fragile optical membranes.

Acknowledgements

The presented work has been funded by the Federal Ministry of Education and Research (BMBF) within the project "Kompetenzdreieck Optische Mikrosysteme-KD OptiMi" (FKZ: 16SV5473). The authors would like to thank R. Schmidt, G. Leibeling, S. Müller, M. Scheler, T. Müller, and S. Schulze for sample manufacturing, preparations, coating, admeasurements.

References

Aleksandrov, A.G., Zavalova, V.E., Kudryashov, A.V., Rukosuev, A.L., Samarkin, V., 2005. Adaptive correction of a high-power titanium-sapphire laser radiation. J. Appl. Spectrosc. 72.

Aleksandrov, A.A., Kudryashov, A.V., Rukosuev, A.L., Cherezova, T.Yu., Sheldakova, Yu.V., 2007. An adaptive optical system for controlling laser radiation. J. Opt. Technol. 74.

Banse, H., Beckert, E., Eberhardt, R., Stöckl, W., Vogel, J., 2005. Laser Beam Soldering - a New Assembly Technology for Microoptical Systems, Microsystems Technologies 11, 186-193.

Beckert, E., Burkhardt, T., Hornaff, M., Kamm, A., Scheidig, I., Stiehl, C., Eberhardt, R., Tünnermann, A., 2010. Submicron accuracy optimization for laser beam soldering processes. Proc. SPIE 7585.

Burkhardt, T., Mohaupt, M., Hornaff, M., Zaage, B., Beckert, E., Döring, H.-J., Slodowski, M., Reimer, K., Witt, M., Eberhardt, R., Tünnermann, A., 2011a. Packaging Technology of Multi Deflection Arrays for Multi-Shaped Beam Lithography, Proc. 42nd International Symposium on Microelectronics IMAPS, 600-607.

Burkhardt, T., Bruchmann, C., Kamm, A., Hornaff, M., Beckert, E., Gebhardt, S., Müller, J., Hoffmann, M., Eberhardt, R., Tünnermann, A., 2011b. Smarte adaptiv-optische Mikrosysteme - Aufbautechnologie und thermomechanische Charakterisierung. Proceedings des Mikrosystemtechnik Kongress 2011, 587590.

Burkhardt, T., Hornaff, M., Kamm, A., Burkhardt, D., Schmidt, E., Beckert, E., Eberhardt, R., Tünnermann, A., 2015a. Low-strain laser-based solder joining of mounted lenses. Proc. SPIE 9574.

Burkhardt, T., Hornaff, M., Burkhardt, D., Beckert, E., 2015b. Optical path difference evaluation oflaser-soldered optical components. JEOS RP 10, 15059.

Cheriaux, G., Rousseau, J.-P., Burgy, F., Sinquin, J.-C., Lurpon, J.-M., Guillemard, C., 2007. Monomorph large aperture adaptive optics for high peak-power femtosecond lasers. Proc. SPIE 6584.

Faidel, H., Gronloh, B., Winzen, M., Liermann, R., Esser, F., Morasch, V., Luttmann, J., Leers, M., Hoffmann, D., 2012. Passive alignment and soldering technique for optical components. Proc. SPIE 8235.

Feinleib, J., Lipson, S.G., Cone, P.F., 1974. Monolithic piezoelectric mirror for wavefront correction, Appl. Phys. Lett. 25, 311-313.

Hoult, A. P., McLenaghan, A. J., Rathod, J., 2003. Advances in laser soldering using high power diode lasers. Proc. SPIE 4831.

Lorenz, H., Beckert, E., Schödel, R., 2015. Phase topography-based characterization of thermal effects on materials and joining techniques. Applied Optics 54 (1), 2046-2056.

Ma, J., Liu, Y., Hu, Y., Xu, C., Li, B., Chu, J., 2013. Low-cost unimorph deformable mirror with high actuator count for astronomical adaptive optics, Opt. Eng. 52.

Ojeda, J., Beckert, E., Burkhardt, T., Hornaff, M., Kamm, A., 2013. Experimental evaluation of the polarization crosstalk when soldering a polarization-maintaining fiber into a v-grooved substrate. IEEE T. Compon. Pack T. 3, 543-548.

Primmerman, C.A., Fouche, D.G., 1976. Thermal-blooming compensation: experimental observations using a deformable-mirror system. Appl. Opt. 15.

Reinlein, C., Appelfelder, M., Gebhardt, S.E., Beckert, E., Eberhardt, R., Tünnermann, A., 2013. Thermomechanical design, hybrid fabrication, and testing of a MOEMS deformable mirror. J. Micro/Nanolithogr., MEMS, MOEMS 12.

Reinlein, C., Appelfelder, M., Goy, M., Gebhardt, S.E., Gutzeit, N., 2014. Testing of thermally piezoelectric deformable mirror with buried functionality. Proc. SPIE 8978.

Ribes, P., Burkhardt, T., Hornaff, M., Kousar, S., Burkhardt, D., Beckert, E., Gilaberte, M., Guilhot, D., Montes, D., Galan, M., Ferrando, S., Laudisio, M., Belenguer, T., Ibarmia, S., Gallego, P., Rodriguez, J.A., Eberhardt, R., Tünnermann, A., 2015. Solderjet bumping technique used to manufacture a compact and robust green solid-state laser. Proc. SPIE 9520.

Stauffer, L., Würsch, A., Gächter, B., Siercks, K., Verettas, I., Rossopoulos, R., Clavel, R., 2005. A surface-mounted device assembly technique for small optics based on laser reflow soldering. Optics and Lasers in Engineering 43, 365-372.

Stein, S., Heberle, J., Gürtler, J.H., Cvecek, K., Roth, S., Schmidt, M., 2014. Influences of Nozzle Material on Laser Droplet Brazing Joints with Cu89Snll Preforms. Physics Procedia 56, 709-719.

Steinhaus, E., Lipson, S.G., 1979. Bimorph piezoelectric flexible mirror. J. Opt. Soc. Am. 69,478-481.

Yoder, P.R. Jr., 2004. Mounting optics in optical instruments. 2nd printing in SPIE press.

Zhi-Jun, R., Xiao-Yan, L., Mei-Biao, L., Chang-Quan, X., Xiao-Ming, L., Ru-Xin, L., Zhi-Zhan, X., 2009. Wavefront Correction of Petawatt Laser System by a Deformable Mirror with 50 mm Active Aperture. Chinese Phys. Lett. 26.