Introduction to On-Wafer Characterization at Microwave Frequencies

Transcription

1 Introduction to On-Wafer Characterization at Microwave Frequencies Chinh Doan Graduate Student University of California, Berkeley

2 Introduction to On-Wafer Characterization at Microwave Frequencies Dr. Tariq Alam Senior Applications Engineer Cascade Microtech, Inc. presented by: Chinh Doan, University of California, Berkeley

3 Presentation Outline Microwave Probing Technology Air Coplanar TM Probes On-Wafer Calibration Methods SOLT, TRL/LRM, SOLR, LRRM On-Wafer Verification Methods Layout Rules Calibration and Measurement Software WinCal TM

4 What Are My Measurement Objectives? Innovating Test Now Determine S-parameters of on-wafer active devices between 500 MHz to 5 or 50 GHz or beyond Future? Wafer size (6 to 8 to 12 inch) Frequency range of interest Need for thermal Test automation for throughput

5 What Equipment Do I Need? Vector Network Analyzer Cables Probes Probe positioners Probe station Controller Contact Substrate Impedance Standard Substrate (ISS) Only Cascade Microtech provides the Total Measurement Solution from the Test Ports of the VNA down to the wafer level

6 Microwave Probing On-wafer fixturing needs: Electrically wide BW transmission lines low contact R Mechanically consistent probe shape placement durability Optimize for loss, impedance match, power and current handling capability, contact force, tip visibility...

7 Microwave Probe Transmission Line Contacts Poor Variable loop inductance prevents calibration (Requires repeatable transition) Better Long path to single ground contact limits bandwidth Best Precise line impedance right to the ground-signal-ground contacts

8 ACP Probe Technology K-Connector Hard Absorber Block Absorber Air Coplanar Waveguide Tip Soft Absorber Low-Loss Cable Low-loss, low density teflon dielectric coax Microwave absorber - consistent attenuation - termination of coaxial shield energy - provides rigidity 15 W CW at 10 GHz 5 A DC current

9 Low-loss ACP Innovating Test Low-loss and standard ACP Probes Application Noise Load Pull

10 Air-Coplanar Tips Innovating Test Precision tip fabrication for tight impedance control High tip visibility for consistent placement on pads BeCu tips for Au or Cu pads Tungsten tips for Al pads Clear view of contact point Preferred 23 contact angle Wide contact area

11 Installing and Using ACP Probes φ Use Cable strain relief on positioners Use Contact Substrate to planarize probes *definition: planarization - the ability to insure all contacts are at the same height

12 Using ISS Alignment Marks Innovating Test Internal Apex Initial contact Full skate and overtravel Used to set skate and probe separation

13 Maintaining ACP Probes Innovating Test Keep tips clean of dirt and debris Inspect and clean connectors Electrical verification -- Probe Test

14 Calibration

15 vv v Innovating Test How Do I Calibrate My VNA? a o b o Port 1 Forward Switch Perfect Reflectometer Error Adapter DUT [S] Reverse Port 2 a 3 b 3 Microwave Errors (Forward) Calibration Standards Directivity Port-2 Match Open Thru Port-1 Match Transmission Tracking Short Line Reflection Tracking Crosstalk Load Etc.

16 One-port Network Analyzer Model a O E R E D E S G DUT b O 1 E R = frequency response of measurement channel E D = Directivity of coupler E S = Port match Calibrate with three known reflection coefficients Short - open - load (SOL) Short - offset short 1 - offset short 2

17 Two-port Network Analyzer Model (Forward Model) E X E R S 21 E T a O DUT b 3 E D E S S 11 S 22 E L b O 1 S 12 E T = Models imperfections in transmission response E L = Models signal reflected back into DUT from P2 Calibrate with short-open-load on each port plus a Thru (SOLT) (uses 10 knowns)

18 VNA Calibration b O a O Ideal Network Analyzer a 3 b 3 DUT e 10 e00 e 11 S 21 e 22 e S S 12 S 22 e 23 e 33 e 32 One port or two port calibration standards N 1 N 2 N 3 Switch modelled and measured separately signals simultaneously measured two two-port error boxes

19 Calibrating the Probe Tips with Coplanar Waveguide Impedance Standards ISS GROUND GROUND GROUND GROUND SIGNAL SIGNAL SIGNAL GROUND GROUND Short Thru Loads GROUND GROUND Electrical behavior of standards standard dimensions probe pitch probe placement (use alignment marks) cal coefficients supplied with probe

20 SOLT Calibration Short L short Load L term Open (probes in air) C open Thru Delay Thru All standards must be perfectly known available on virtually every vector network analyzer (CalKit required) open has capacitance (often negative) short and load have inductance sensitive to probe placement mathematically overdetermined unpredictable behavior

21 TRL/LRM Calibration Thru Line(s) OR Reflect Thru-Reflect-Line requires least info about standards S-parameters referenced to line Zo reference plane at centre of Thru requires multiple probe spacings Design rules Zo is inherently complex at low frequencies not suitable for fixed spacing probes (e.g., probe card) Line-Reflect-Match Match referenced to Z match

22 Multi-line TRL Benchmark Cal Method Thru Reflect Line(s) Modified TRL algorithm developed by the U.S. National Institute of Standards and Technology Benchmark on-wafer calibration method Takes an optimal weighted average of all the line On-wafer standards (with DUT) preferred Renormalizes the S-parameter impedance to 50 Ohms

23 SOLR Calibration Short Load Short-Open-Load- Reciprocal Thru Open (probes in air) reciprocal Thru requires only S 12 = S 21 tolerant to lossy or highly reactive insertion standard Reciprocal OR convenient for use with fixed probe spacing in probe cards Does not require a custom Thru convenient for use when DUT terminals are orientated at 90 available in WinCal

24 SOLR for Right Angle Measurements 1.0 Carefully constructed right angle Thru standard Thru is non-ideal, large dip at 20 GHz Errors in standard cal s SOLR immune to Thru errors S21 [db] Orthogonal SOLT Orthogonal LRRM Orthogonal SOLR -0.5 Straight LRRM [GHz]

25 LRRM Line Cascade Microtech Calibration Research Line-Reflect-Reflect-Match Calibration Reflect Reflect (probes in air) Match like TRL, only Match acts as infinitely high loss line one transmission line standard only allows fixed probe spacing calibration Thru (line) delay, Match resistance must be known referenced to laser trimmed resistor required measurement of only one load standard load inductance compensation uses off-wafer standards (ISS) same standards as SOLT only - no need for cal kit available in WinCal

26 LRRM Calibration System drift baseline LRRM compares with system drift limit best fixed probe position calibration SOLT /LRM growing error w/freq possible CalKit error possible ref plane error

27 Popular Calibration Methods for Wafer Probing Z 0 Inherently Probe Card Absolute Reference Consistent Support Accuracy SOLT TRL LRM/ LRRM SOLR Trimmed Resistor No Fair Fair Transmission Best (if Lines Yes Poor Corrected) Trimmed Resistor Yes Fair Good Trimmed Resistor Yes Best Good

28 Calibration Verification

29 How Do I Know If My Calibration is Successful? GROUND GROUND GROUND SIGNAL SIGNAL SIGNAL GROUND GROUND GROUND Open stub Pad GROUND GROUND SIGNAL SIGNAL GROUND GROUND Line (delay)

30 Calibration Verification Standards Thru GHz 40.0 GHz 0.10 [db] Unity Gain 1 ps Line Phase Lag [GHz]

31 Calibration Verification Standards Reflect (probes in air) GHz GHz [db] [GHz] Unity Gain Negative Capacitance

32 Calibration Verification Standards Match 0 S21 of Standard Loads [db] HPC40 on HPC ISS [GHz] Measured S21 when probes placed on 50ohm loads

33 Independent Calibration Verification Standards G S G Open stub GHz GHz [db] [GHz] Linear Phase Lag

34 Common Calibration Errors Poor Calibration Verification/ Repeatability Inaccurate Probe Placement Inaccurate Probe Definitions Poor Probe Contact LRM Sensitivity to Differences in Load Standards Incorrect Models for Lines Calibration Drift Narrow Band Suck out Poor Phase Stability of Cable Dirty Connectors

35 Beware of an Unconnected Substrate GSG pads shield like CPW Fields terminate on backside of wafer on one side GS pads fringe to the ground plane or chuck Parasitic couplings to the conductors near DUT GSG shields magnetic and electric fields better than GS

36 Layout Guidelines

37 Device Characterization with Coplanar Waveguide Microwave Probes GROUND SOURCE GROUND SIGNAL GATE DRAIN SIGNAL GROUND SOURCE GROUND

38 Pad Size (Passivation Window) Recommended minimum pad size is 80um x 80um for ACP Probes when performing automated Smaller pad dimensions can be used for manual probing HPC Probe allows 40um x 70um manual probing Passivation height must be considered Pad height variation must not exceed 25um

39 Process Control Monitor Device Layout G G S S G G GSG Test Device Open Pads & Metal Shorted Metal (Remove Y pad ) (Remove Z metal )

40 PPR Corrected H 21 Measurement 0.25 µm CMOS Transistor 60 H As measured F T = 25 GHz Corrected for pad parasitics F T = 33 GHz FREQUENCY (GHz)

41 WinCal VNA Calibration and Measurement Tool

42 Calibration Features Calibration SOL, SOLT, SOLR, LRM, LRRM, TRL Stability tests

43 Setting Up a CalKit CalKit can be easily defined and downloaded to VNA Removes one of the most common sources of error

44 Measurement Features Tools Read/Save S- Parameters Pad Parasitic Removal Probe Test

45 Monitoring the Calibration Stability check checks system drift

46 Pad Parasitic Removal Measures intrinsic devices

47 Calibration Repeatability LRRM automatic calibration is very repeatable Operator dependent probe placement errors manual cal s are not as repeatable

48 Summary Microwave require careful calibration, verifications and attention to detail Many new applications to accommodate the varied needs of growing wireless and high-speed digital needs Let Cascade Microtech help you keep up with future innovations

49