Your SlideShare is downloading. ×
0
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Micron Level Placement Accuracy for Optoelectronic Components
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

Micron Level Placement Accuracy for Optoelectronic Components

2,605

Published on

Presented at the ECTC 2009 in San Diego, Dan Evan and Zeger Bok present material and process consideration, 2 high accuracy case studies, the systems required to enable these processes and techniques …

Presented at the ECTC 2009 in San Diego, Dan Evan and Zeger Bok present material and process consideration, 2 high accuracy case studies, the systems required to enable these processes and techniques involved in the application. For complete download visit http://palomartechnologies.com/Applications/OptoelectronicPackaging.aspx

0 Comments
2 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
2,605
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
0
Comments
0
Likes
2
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. • Terms • Material & Process Considerations • High Accuracy Case Studies • Systems • Case Study 1: Multi-channel communication • Case Study 2: Wafer-Scale Eutectic Die to Wafer / P-Side Down Laser Attachment • Palomar’s Complete Solution
  • 2. The Solution: Model 6500 How Accurate is it? 75µ = Width of a Human Hair SMT – 40um Die – 25um 5um 7µ 2µ 1.5µ Red Blood Cell Bacteria Cell 6500 Placement Accuracy
  • 3. Term Definitions • High Placement Accuracy Geometry Z Translation dX, dY Rotation Tz Levelness Tx, Ty Y X Bondline dZ • Interconnect Method (In-situ / Offline) Eutectic Adhesive • Shift (Pre Cure – Post Cure)
  • 4. Flatness Cleanliness Material Considerations Symmetry Image Quality High Accuracy Attachment is run in production but requires close attention to detail and control of materials
  • 5. Substrate • Flatness (cleanliness) • Image Fiducial SUBSTRATE Fiducial Attach Material • Uniformity, shrinkage, Poor backside metallization symmetric application, curing stability Symmetric Epoxy Die • Flatness • Image Fiducial
  • 6. Part Geometry for 1 um alignment • 5um particle below one end with push edge outside of 1um tolerance Exaggerated 1.3 um Drawing 400 um 5 um 100 um Human Hair
  • 7. • Part Geometry for 1 um alignment 5um particle with push edge outside of 1um tolerance Particle between die
  • 8. Process Considerations 1) Image Recognition • Fiducial selection Two Point Refs / Look Down Substrate 2) Pick Strategy • Waffle, Gel, Double 3) Place Strategy Two Point Refs / Look Up Chip • Fixed Pattern vs. Pick die presentation format, collet, and Relative process are critical. Lookup refs remove pick error Place die based on one substrate fiducial 4) Curing or based on previously placed die. • Profile Pre vs. Post Cure Accuracy
  • 9. Systems Introduction
  • 10. 11 Model 6500 – Precision Eutectic Die Bonder Combined speed, accuracy, and compact footprint of the Model 6500 provide for high yielded throughput and optimized cost of ownership in a high accuracy assembly system. o Post eutectic placement attachment accuracy o ± 1.5 micron (3 sigma) o ± 0.1 degree post-placement rotation o Cycle time <7 seconds o Programmable pulse-heating o (500 C @ 100 C/sec ramp, +/- 2 C) o Waffle Pack, Gel Pak o <1 sq meter footprint o Work area 300 mm x 150 mm o 10-100 grams force, ± 1 gram (3 sigma) o 6 tool turret – “On the fly tool change” o Integrated data management and analysis
  • 11. 12 Integrated Data Management • Machine Calibration • Machine Reliability Availability and Maintainability Statistics (RAM STATS) • System Performance (Motion, Vision, ..) • Process Performance - Post Placement Accuracy Checking (PPAC) Dry PNP 1.2 um full scale PPAC 200 000 800 600 400 200 000 1 3 5 7 9 55 59 61 63 65 67 69 71 73 75 77 79 81 83 39 43 45 47 49 51 53 57 11 13 15 17 19 21 23 25 27 29 31 33 35 37 41 Samples Statements: 1 Fiducial 1 Fiducial 2
  • 12. Measurement Equipment Automate Optical Inspection ~0.5um Parts with built in measurement features such SEM ~0.1um as vernier calipers ~0.5um
  • 13. SEM Edge Alignment Measure
  • 14. 16 High Accuracy Attach Cases Application Attach Accuracy Multi-channel communication modules (Active Epoxy ±3-5 um Optical Cables) P-Side Down Laser (Singulated & Wafer Scale) Eutectic ±1.5-3 um LED Laser Print Head Epoxy ±3-5 um Laser Marker Head (VCSEL Laser) Epoxy ±2-3 um Lithography/Screen Interconnect Epoxy ±5 um Thru Via Die Stacking Misc ±3-5 um 3D MEMS Stacking Epoxy ±5-10 um
  • 15. Case 1: Multi-Channel Communication, Active Optical Cables Finisar Luxtera Zarlink
  • 16. Luxtera Blazar LUX5010 Multirate 4x10G Optical Active Cable Features Benefits •No optical interoperability issues •Easy maintenance - no need to clean Optical Active Cable: optics Closed Optical System •Eliminates costly fiber connectors Multi-Rate (1-10Gbps) One solution for multiple applications Single Laser for 4-Channels Better reliability •Standards compliant electrical interface (IB, FC, 10GbE) SFP+ Electrical Interface •Lower link costs: 8 CDRs eliminated (4 per end) •Lower link costs and better performance •Supports reach up to 300 meters for Single-Mode Optics QDR (4x10Gbps) •Eliminates need for EDC Components Hot-Pluggable •Field replaceable QSFP Form-factor •Provides more bandwidth density •Lower cooling costs and Low Power (0.5W/10G) simplified thermal management
  • 17. Active Optical Cable Tyco PARALIGHT Active Optical Cable Assemblies provide E/O Finisar brings serial active and O/E conversion built optical cable to 10G into connector applications
  • 18. Active Optical Cable Channel Electrical Input Light Channel Coupler Coupler Detector Source Fiber Electrical Output Active Optical Cable has an Electrical Input and Output All optical E-O, O-O, O-E elements are within cable
  • 19. 21 AOC Specifications (um) 5 4 3 2 1 VCSEL Array N VCSELs Gap X Linearity Line Die Gap Error in X ± 5 Die Linearity Error in Y ± 3
  • 20. allX allY Average 0.1 0.0 Wet Results Maximum Minimum Range 3StDev 3.9 -3.6 7.5 4.5 2.5 -2.7 5.2 2.6 X Errors Wet 5 4 3 2 5X 1 4X um 0 3X -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 2X -2 1X -3 -4 -5 Y Errors Wet 5 4 3 2 5Y 1 4Y um 0 3Y -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 2Y -2 1Y -3 -4 -5
  • 21. allX allY Average 0.5 0.0 Cured Results Maximum Minimum Range 3StDev 4.7 -2.2 6.9 4.2 2.4 -2.6 5.0 2.5 X Errors Cured 5 4 3 2 5X 1 4X um 0 3X -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 2X -2 1X -3 -4 -5 Y Errors Cured 5 4 3 2 5Y 1 4Y um 0 3Y -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 2Y -2 1Y -3 -4 -5
  • 22. Case 2: Wafer-Scale Eutectic Die to Wafer P-Side Down Laser Attachment 80/20 AuSN Attach
  • 23. P-Side Laser Attach P-Side Up example so stripe is visible. Y Edge Alignment +/- 3.0 um +/- 0.1 Deg X Stripe Alignment +/- 3.0 um W L InP Laser Diode Eutectically Attached (P-Side Down)
  • 24. Driving Accuracy for Laser Diode Placement Laser Diode Laser Diode Laser Carrier > 3um Laser Carrier > 3um Overheat + Focus Eclipse + Focus
  • 25. Systems & Software
  • 26. 28 Model 6500 WSP Precision Eutectic Die Bonder Combined speed, accuracy, and compact footprint of the Model 6500 provide for high yielded throughput and optimized cost of ownership in a high accuracy assembly system. o Post eutectic placement attachment accuracy o ± 3.0 micron (3 sigma) o ± 0.1 degree post-placement rotation o Cycle time 30 seconds o Programmable pulse-heating Pick Tool (500 C @ 65 C/sec ramp, +/- 2 C) o Waffle Pack, Gel Pak o Wafer Stage – Heated background temp o <1 sq meter footprint o Work area 300 mm x 150 mm o 10-100 grams force, ± 1 gram (3 sigma) o 1 Pulse heated tool optimized for Laser o Integrated data management and analysis
  • 27. Wafer Scale Packaging Eutectic Die Bonder Equipment Layout Pulse Heated Pick Tool with active cooling Gel pack load of Lasers Wafer Stage - Heated Lookup camera
  • 28. 30 Wafer Level Laser Diode Attach Load Materials Wafer Sub Lasers Pick Laser Lookup Ref Pulse Reflow Place & Hold Release Laser Repeat N.. Unload Materials Wafer/Lasers
  • 29. 31 Eutectic Process: Pulsed Heat Profiles Pulsed heat stage • Computer controlled Programmable Temp up to ‘Point and click’ profiling 500 C Parts at high Temp Temperature accuracy +/- for a limited 20C time Fast ramp (up to 65 C/s). No overshoot
  • 30. XY PLACEMENT (P-Side Measure) 3 3 2 2 1 1 0 d dX 0 X dY 1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15 -1 -1 -2 -2 -3 -3 P-Side accuracy measurements using a glass wafer Time consuming so correlated to N-Side measurements Instrument: Nikon VRM 3200
  • 31. XY PLACEMENT (N-Side Measure) 13.5 N-Side Accuracy Data, um 9.0 4.5 0.0 1 17 33 49 65 81 97 113 129 145 161 177 193 209 225 241 257 273 -4.5 -9.0 -13.5 Xlitho (PR) Yedge (M) N-Side accuracy measurements using a wafer N-Side allowable tolerance X=4.5um and Y=3.0um Instrument: Nikon VRM 3200
  • 32. Solder Age Affects on Shear Shear Shear Time Time Shear Average Shear Range Post‐bond Shear Time, [Hrs] [Hrs] Post‐bond Shear Time, [Hrs] [Hrs] [gram] 0 [gram] 0 0 24 48 72 0 24 48 72 24 24 Pre‐ 0 270 275 310 253 Pre‐ 0 36 128 128 87 bond  48 bond  48 24 217 210 241 193 24 53 21 90 38 Attach 72 Attach 72 Time  48 239 206 202 Time 48 6 7 119 [Hrs] 96 [Hrs] 96 72 192 189 72 24 31 Grid 1 2 Grid 1 2 270 275 310 400 400 217 253 300 300 239 210 241 128 128 200 36 200 206 193 87 202 100 53 100 192 90 189 21 0 119 38 0 0 6 0 7 24 24 31 24 48 48 0 72 0 72 24 24 48 48 72 72 Slight reduction in shear over time
  • 33. High Accuracy Attach Cases Application Attach Accuracy Multi-channel communication modules (Active Epoxy ±3-5 um Optical Cables) P-Side Down Laser (Singulated & Wafer Scale) Eutectic ±1.5-3 um LED Laser Print Head Epoxy ±3-5 um Laser Marker Head (VCSEL Laser) Epoxy ±2-3 um Lithography/Screen Interconnect Epoxy ±5 um Thru Via Die Stacking Misc ±3-5 um 3D MEMS Stacking Epoxy ±5-10 um 35
  • 34. Palomar’s Complete Solution Wire Bonders Die Bonders Integrated Assembly Lines Custom Handling Systems Process Development
  • 35. Contact & Resources • Present your packaging challenge and we will propose a solution at no cost • Download “Automated Eutectic Die Attach”, by Palomar Sr. Applications Engineer Zeger Bok • Visit Palomar’s Blogs: “Interconnection” and “Wire Bonders’ Speak” • Download Palomar System Data Sheets • Contact us directly
  • 36. Micron Level Placement Accuracy Case Studies for Optoelectronic Products Daniel D. Evans, Jr., Zeger Bok Palomar Technologies, Inc. 2728 Loker Avenue West Carlsbad, CA 92010 Phone: (800) 854-3467 E-mail: info@bonders.com Abstract Applications requiring ultra high placement accuracies of 1µm to 3µm are resurfacing in several optoelectronic applications such as Arrayed Laser Print Head assemblies, P- Side Down Laser Attachment applications, and Multi- Channel Optical Communication products. An overview of the technologies, placement accuracies, and attachment methods is presented for two cases. With placement accuracies for surface mount machines typically around 40µm, 10µm for die attach machines, and 1µm for ultra high accuracy placement machines, this paper will cover the differences in measurement, material, and process controls Figure 1 - Geometric Six Degrees of Freedom that are required to successfully achieve ultra high placement accuracies of 3µm. Interconnects are comprised primarily of two methods for optoelectronic assemblies: adhesive (epoxy) or metallurgical (eutectic solder). High Accuracy Die Attach Requirements and Challenges The application breakout given in Table 1 is useful to These attachment options can be either in-situ (serial) or explore general application, attachment, and accuracy batch (parallel). In-situ attachment is completed during the requirements for ultra high accuracy die attach. The main placement operation for each component and will have lower breakout is by application product or technology. For the throughput since the attachment time for each component is purposes of this paper, each of the applications explored is for added to the pick and place time. Batch attachment is specific attachment technologies (epoxy or eutectic) and completed after all components are placed so the actual general ranges of required placement accuracy. attachment can be completed as a parallel process. Batch attachment methods typically have higher throughput Table 1 - High Accuracy Application Overview compared to in-situ. CASE Attach Accurac Another consideration for the attachment method is its effect on placement accuracy. To understand the effect on VCSEL Array Epoxy ±3-5µm batch attachment methods, it is important to measure the pre- cure and post-cure accuracy of the components. In-situ P-Side Down Laser Eutectic ±3-5µm attachment methods can have higher placement accuracy, especially if the design does not include self-centering. LED Laser Print Head Epoxy ±2-5µm Material and Process Considerations Lithography/Screen Interconnect Epoxy ±3-5µm When approaching micron level placement accuracies, there are several important factors that need careful Thru Via Die Stacking Misc. ±3-5µm management and control: 3D MEMS Stacking Epoxy ±5-10µm • Substrate flatness, cleanliness, and fiducial clarity (in particular in case of edge alignment) Terms and Definitions • Die flatness, cleanliness, and fiducial clarity Pick and place is composed of geometric accuracy and • Attach material uniformity, shrinkage, symmetric interconnect method. To completely define placement application, and curing stability accuracy would require specifying all six (6) degrees of Measurement Considerations freedom, as shown in Figure 1. Most pick and place accuracy applications specify Z as a bond line and placement accuracy Measuring the actual placement accuracy down to 1-3µm as X error, Y error, and Theta-Z error. For the purposes of requires both well characterized equipment and measurement this paper, the specific requirements will be listed for each of methods [1]. Even if the equipment can support the the cases studied. resolution, the parts usually have imperfections beyond the
  • 37. targeted accuracy. Measurement methods must include ways One construction of a complete cable has an array of to handle these imperfections. VCSEL transmitters which require high accuracy placement The remainder of this paper will explore two cases for to allow better fiber optical coupling to the VCSEL lasers. high accuracy placement and attach. These assemblies generally require photodiodes in the same package as well but these are easier to optically couple and require less placement accuracy. Case 1: Multi-channel Communication Products The remainder of this case study will share geometry, (VCSELs) accuracy, and attachment requirements of the VCSEL arrays. Several new products related to optical communications such as Active Optical Cables use multiple channels of The VCSELs used in this study are 300µm square by transmit and receive pairs to produce the high bandwidth 200µm thick and are presented in 2x2 Gelpacks. The communication required for high performance computer substrate can be any material but more even materials will connections or digital audio/video connections. The basic produce more consistent results. The material in this study concept of an active optical cable link as shown in figure 2 is was a flexible circuit mounted to a PWB backing and an optical cable that contains electrical input and output. The presented in a common carrier in groups of 10 per carrier. input conversion from electrical to optical (E-O) and the The VCSELs are placed into 84-1LMIT1 epoxy which is output conversion from optical back to electrical (O-E) is deposited just prior to the pick and place process. included as an integral part of the cable so that the end user The arrangement of the VCSELs is shown in figure 4. sees only an electrical cable without the problems associated VCSELs 5 through 1 are placed left to right according to a to optical cable connections [2]. The various optical to specified pitch. VCSEL number 5 is the master VCSEL and optical (O-O) interfaces as well as the actual fiber are also all others are placed relative to it. The required accuracy for included in the cable. A complete cable as shown in figure 3 each VCSEL is ±5µm in X and ±3µm in Y from the target would include multiple channels for transmit and multiple location. channels for receive. Figure 4 – VCSEL Array, Gap X ±5µm, Gap Y ±3µm Placement results are evaluated pre-cured (wet) and post- cured to verify that shifting of the epoxy during cure does not adversely affect the results. Pre-cure results are shown in figures 5 and 6. Statistical Figure 2 – Active Optical Cable Schematic of one summaries of pre-cure results are given in Table 2. Pre-cure Transmission Path (Cables have Multiple Tx/Rx Pairs) results are within specification for both X and Y. Figure 5 – Pre-cured (Wet) X Error Results Figure 3 – Active Optical Cable Example looks like a USB Cable with Electrical Input and Electrical Output (Courtesy of Figure 6 – Pre-cured (Wet) Y Error Results Finisar)
  • 38. Table 2 – Pre-cure (Wet) Statistics case study will focus on pick and place and attachment X Y capability. Substrate time at temperature and its effect on Average 0.1 0.0 solder reflow and solder aging are explored as well. Maximum 3.9 2.5 Measurement methods for P-Side down laser attachment Minimum -3.6 -2.7 provide some challenges during process development and will Range 7.5 5.2 be covered as well. 3StDev 4.5 2.6 The placement requirements are based on positioning the Post-cured results are shown in figures 7 and 8. Statistical laser for optimal coupling of the laser optical output to summaries of post-cure results are given in Table 3. Post- coupling optics as shown in figure 9. The gray section cure results are also within specification for both X and Y. represents the wafer substrate area upon which the laser will be solder attached. The wafer substrate has fiducials which determine alignment targets and gold pads plated with 80/20 AuSn eutectic solder. The laser is 300µm long by 250µm wide by 125µm thick with backside (P-Side) plated gold pads for solder attach. The P-Side of the laser provides alignment features that are a combination of fiducials and the alignment point edge. Both are referenced during the alignment phase using an upward looking camera. Figure 7 – Post-cured X Error Results Figure 8 – Post-cured Y Error Results Figure 9 – P-Side Down Laser Alignment to Substrate Requirements Table 3 – Post-cured Statistics The alignment specifications are defined in figure 9 and X Y revolve around the P-Side intersection of the laser stripe Average 0.5 0.0 location in the X direction and the laser output edge to Maximum 4.7 2.4 substrate datum in the Y direction. Both X and Y directions Minimum -2.2 -2.6 have a ±3µm alignment tolerance. The angular alignment of Range 6.9 5.0 the laser edge is also defined as a ±0.1degree tolerance in 3StDev 4.2 2.5 Theta-Z. Case 2: Wafer Scale – Eutectic Die to Wafer P-Side Down Laser Attachment P-Side down laser attachment using eutectic solder is a mature process and has been used extensively for long haul and metro distance laser transmit modules [3]. The P-Side laser attach to heat spreader has traditionally been done using singulated heat spreaders with a pulsed heat in-situ reflow stage. The heat profile recipe is critical for proper solder phase in the final bond. Recent work in P-Side down laser attachment has expanded the technology to Wafer Scale – P-Side Down Laser Die to Wafer attachment. This technology shifts away from singulated submounts to bonding directly on a wafer or on substrates. Design considerations for component heat flow, power, and reliability are significantly affected by the specific geometry, materials, and interconnect process. This
  • 39. Figure 10 – Steady State Heat Wafer Stage and Pulsed Heat used for initial process development by measuring placement Pick Tool Optimized for Laser Die Eutectic Solder Attach accuracy through the backside of the glass. The P-Side Figure 10 shows a close up view of the Palomar measurement results in figure 12 show that all samples were Technologies Model 6500 with wafer stage, pulsed heat pick below ±3µm in the X and Y directions. tool, and lookup camera alignment algorithm that was used for execution of the tests. The pulsed heat tool is controlled through a temperature ramp profile as shown in figure 11. This is not the profile used for the study. While holding the laser in place on the wafer, the tool ramps from a background temperature to the reflow temperature, holds that temperature during the reflow time, and is then cooled down to the background temperature. Figure 12 – P-Side Measurement Results for two Glass Substrate Runs P-Side measurements are time consuming due to preparation of the sample materials and are nearly impossible to perform on actual wafer level substrates. To mitigate this problem, a correlation between P-Side and N-Side measurements was established which resulted in a more effective N-Side measurement specification of ±4.5µm in X and ±3µm in Y. The correlation was based on known misalignment errors between the lithography masks of both P- Side and N-Side of the wafer. All measurements could then be performed on a Nikon VMR3200 Automated Optical Inspection Microscope. Figure 11 – Generic Example Pulsed Heat Tool Profile The wafer level substrate was then built and measured using N-Side measurement techniques as shown in the chart The pulsed heat tool used to pick the laser die includes the of figure 13. Most samples were within specification for both following capabilities and benefits: X and Y dimensions. Device level testing is normally • Pulsed heat controller required to determine actual yield. • Temperature up to 600C • Temperature accuracy ± 20C • Fast ramp (up to 65C/s) - no overshoot • Parts at high temperature for a limited time • Programmable ‘point and click’ profiling The wafer assembly process steps include: • Load wafer onto steady state heated stage • Load laser die (pre-flipped with P-Side down) onto machine Figure 13 – N-Side Measurement on Wafer with Correlated • Repeat the following process for all bond sites: Tolerance of X ±4.5µm and Y ±3µm o Vision find wafer at next available site o Vision find N-Side of next available laser die Solder aging at temperature is typically not considered for and pick it singulated substrates since the solder is exposed to o Vision find P-Side of laser die on pick tool temperature for a short duration of time only. When moving using lookup camera to wafer level substrates however, solder can be exposed to o Align and place laser die to substrate site background temperatures over 200C for tens of hours o Apply bond force and initiate pulsed heat profile depending on the number of bond sites and the equipment o Release laser die upon completion of pulsed heat throughput. profile including forced cooling cycle to below Tables 4 and 5 provide a summary of the results from a reflow temperature study on the effects of exposure to time at temperature. A combination of time intervals at a background temperature of Since it is not possible to directly measure the P-Side 230C was used to evaluate both non-reflowed solder and interface of the completed assembly, glass substrates were reflowed solder in completed assemblies. The study involved
  • 40. waiting the pre-bond time intervals and then attaching a set of 4 laser die each. These laser die sets where then sheared after Case 1: Multi-channel Communication Products waiting the post-bond time intervals. This particular test (VCSELs) sequence allowed testing of solder exposure to temperature before laser die attach (0 to 72 hours), solder exposure to The VCSEL based application placed an array of 5 temperature after laser die attach, and a combination of both VCSELs into epoxy. Post-cure results better than ±5µm in X (0 to 96 hours = Pre-bond Attach Time + Post-bond Shear and ±3µm in Y were achieved. Time). Case 2: Wafer Scale – Eutectic Die to Wafer P-Side Down Table 4 – Solder Attach Average Shear Strength vs. Time Laser Attachment Shear The P-Side-Down 80/20 AuSn eutectic laser attach onto Post-bond Shear Time Shear Average Time wafer produced results of ±3µm in X and ±3µm in Y for P- [Hrs] [grams] [Hrs] Side measurements on glass substrate. N-Side measurements 0 24 48 72 0 on the wafer included shifts between N-Side and P-Side but 0 270 275 310 253 24 results still indicated ±3µm in X and ±3µm in Y when N-Side Pre-bond 24 217 210 241 193 48 measurements were correlated back to the P-Side. Attach The effect on solder quality of time at temperature was 48 239 206 202 72 Time studied and found to have some effect on average shear [Hrs] 72 192 189 96 strength. Additional studies could be completed to verify Grid 1 2 repeated results. Follow- on studies could be conducted on the reliability of laser devices as well. Table 5 – Solder Attach Range Shear Strength vs. Time Shear Post-bond Shear Time Shear Range [Hrs] Time [grams] [Hrs] Acknowledgments 0 24 48 72 0 Palomar Team members who contributed to equipment 0 36 128 128 87 24 and process development include Mike Artimez, Don Beck, Pre-bond 24 53 21 90 38 48 Tom Boggs, Bill Hill, Dan Martinez, Ricardo Saldana, and Attach Tim Hughes. 48 6 7 119 72 Time [Hrs] 72 24 31 96 References Grid 1 2 1. Bok, Z. et al, “Micron Level Placement Accuracy Case Studies for Optoelectronic Components”, The 41st Table 4 shows a reduction in average shear strength from International Symposium on Microelectronics, 270 grams to a minimum of 189 grams over all aging Providence, RI, November 2008 conditions compared to zero pre-bond attach time and zero 2. Dick Otte, Promex Industries, Chair of iNEMI post-bond shear time. Visual inspection of the sheared Optoelectronic Technical Working Group (OE TWG) samples showed no observable differences in solder joint Roadmap draft dated September 26, 2008, p. 16 quality at 60X magnification. 3. Weiss, S. et all, “Fluxless die bonding of high power Table 5 shows the effect on the range of shear strength for laser bars using theAuSn-metallurgy”, Electronic each data set. The data did not show continuing degradation Components and Technology Conference, 1997. of shear strength range for pre-bond attach time or post-bond Proceedings., 47th Volume , Issue , 18-21 May 1997 shear time exposure. Page(s):780 – 787 Summary and Conclusions Two separate optoelectronic application case studies have been reviewed. Both cases required ±5µm placement accuracy or better using adhesive or metallurgical attachment as shown in Table 6. Table 6 - High Accuracy Application Cases CASE Attach Accuracy VCSEL Array Epoxy ±3-5µm P-Side Down Laser Eutectic ±3-5µm

×