서지주요정보
Effects of non-conductive film (NCF) resin formulation and bonding parameters on high-speed Cu pillar/Sn-Ag Micro-bump bonding = 비전도 접속 필름 레진 조성과 본딩 변수가 고속 구리 필라/주석-은 마이크로 범프 접속에 미치는 영향
서명 / 저자 Effects of non-conductive film (NCF) resin formulation and bonding parameters on high-speed Cu pillar/Sn-Ag Micro-bump bonding = 비전도 접속 필름 레진 조성과 본딩 변수가 고속 구리 필라/주석-은 마이크로 범프 접속에 미치는 영향 / Hyeong-Gi Lee.
발행사항 [대전 : 한국과학기술원, 2016].
Online Access 원문보기 원문인쇄

소장정보

등록번호

8029814

소장위치/청구기호

학술문화관(문화관)B1층 보존서고

DMS 16012

휴대폰 전송

도서상태

이용가능(대출불가)

사유안내

반납예정일

리뷰정보

초록정보

As pitches of the bumps gets finer and finer as forty micro-meters, many problems occur such as flux residue, voids, and underfill overflow. As a solution for these problems, Non-conductive Film (NCF) for TSV chip stacking application is the effective solution. To interconnect chips on substrates using pre-applied NCF, thermo-compression bonding is the most common method, however conventional ramp-up bonding process takes about 300 seconds for next bonding process due to precise heating and cooling process. Isothermal bonding using hot bonding tool is alternative method for mass-production of pre-applied NCF. However, curing behavior of NCF should be considered because thermo-compression bonding time of isothermal bonding is too short to cure the NCFs. Liquid epoxy amount of NCF resin was optimized because liquid epoxy was related to adhesion of NCF at the room temperature before curing. Optimized NCF that contained 50 wt% liquid epoxy showed highest adhesion and appropriate elongation properties. Cu pillar/Sn-Ag micro-bumps using both conventional single chip packages and WLPs showed excellent daisy chain resistances of $12 \Omega$ , bump contact resistances of $3 \Omega$ , and equivalent reliabilities. 2-phenylimidazole was selected for curing agent of fast-cure NCF because curing on- set temperature was higher than film coating and NCF lamination temperature. Curing speed of imidazole-NCF was 67.5 times faster than that of DICY-NCF at bonding temperature. Conventional ramp-up bonding and isothermal bonding method were used to analyze the effect of bonding method. The heating rate of isothermal bonding was higher than that of conventional ramp-up bonding and final joint gap decreased enough to achieve the stable solder joints with imidazole-NCF. Solder joint gap of imidazole-NCF was maintained after physical contact of solder because degree-of-cure of imidazole-NCF reached to 90 % at solder melting temperature. Isothermal bonding parameters were also investigated in terms of the bonding pressure, bonding temperature, and bonding time.

전자 패키지가 고밀도, 고성능, 작은 패키지 크기 등을 요구하게 되면서 3차원 패키징 기술이 필 요하게 되었다. Through Silicon Via (TSV) 기술은 3차원 패키징을 가능하게 하였으며 TSV의 접 합은 구리 필라와 주석 범프의 이중 범프가 주로 사용된다. 이들 범프는 플럭스와 언더필을 사용 한 리플로우 공정으로 접합되었으나 범프 피치가 작아지게 되면서 잔류 플럭스 존재, 언더필 기 공, 언더필 범람 등의 문제를 일으키게 되었다. 미세 피치에서는 비전도 접속 필름을 사용하여 문 제를 해결할 수 있으며 웨이퍼레벨 공정과 열압착 본딩 시간 감소를 통해 비전도 접속 필름 기술 의 상용화를 기대할 수 있다. 본 연구에서는 비전도 접속 필름을 사용하여 웨이퍼레벨 공정성과 열압착 본딩 시간 감소에 대한 연구가 진행되었다. 비전도 접속 필름을 라미네이션 한 후 다이싱 하는 공정에서 취약점이 나타났으며 다이싱 공정 조건 최적화로 문제 해결이 되지 않았고 비전도 접속 필름의 물성과 다이싱이 관련이 있음을 확인하였다. NCF의 액체 에폭시 함량이 늘어남에 따라 비전도 접속 필름의 상온 점착성과 연신율이 증가하였으며 액체 에폭시 50 wt% 함량의 NCF가 $0.31 kN/cm^2$ 의 가장 높은 점착성과 0.27 mm의 낮은 연신율을 가지는 것을 확인할 수 있 었다. 최적화된 비전도 접속 필름을 사용하였을 때 문제 없이 다이싱 공정을 진행할 수 있었으며 웨이퍼레벨 공정이 본딩과 신뢰성에도 문제를 일으키지 않은 것을 알 수 있었다. 본딩과 냉각 시 간을 합쳐서 300초가 넘는 기존의 열압착 본딩 공정 또한 등온 본딩 공정을 통해 개선되었다. 냉 각 과정이 없고 15초 이하의 짧은 본딩 시간을 가지므로 불안정한 범프 접합이 발생할 수 있다. 따라서 짧은 본딩 시간에도 높은 경화도를 가지는 이미다졸계 비전도 접속 필름을 제작하였으며 이론적으로 250도에서 1초 이내의 경화시간을 가지는 것을 계산할 수 있었다. 기존의 본딩 방법 과 등온 본딩 방법을 비교하면 승온 속도의 증가로 인해 본딩 후 최종 갭이 감소하는 것을 알 수 있으며 따라서 등온 본딩에서 갭 유지를 통해 좋은 범프 접합을 얻기 위해서는 경화 속도가 빠른 이미다졸계 비전도 접속 필름이 더 효과적인 것을 확인할 수 있었다.

서지기타정보

서지기타정보
청구기호 {DMS 16012
형태사항 viii, 93 p. : 삽화 ; 30 cm
언어 영어
일반주기 저자명의 한글표기 : 이형기
지도교수의 영문표기 : Kyoung-Wook Paik
지도교수의 한글표기 : 백경욱
학위논문 학위논문(박사) - 한국과학기술원 : 신소재공학과,
서지주기 Including references
QR CODE

책소개

전체보기

목차

전체보기

이 주제의 인기대출도서

Advantages of 3D packaging

3D chip stack in electronic packages

Applications ofTSV technology

Non-Conductive Films (NCFs) for TSV chip-stacking

Material properties of underfill and NCF

Conventional underfill processesand wafer-level processes usingNCFs

Thermo-compression bonding profile

(a) Conventional ramp-up bondingusingflip chip bonder,(b) Isothermal bondingusingflip chip bonder and thermo-compression bonder,and (c) Isothermal bonding using larger bonding tool

Comparison of bonding time for 4 dies

(a) Bad solderjoints presumably due to low degree-of-cure ofNCF,(b) Reliability problem due to solder wicking

Scope of the thesis

Conventional single chip bonding process using NCFs

Wafer-level chipbonding process using NCFs

Blade wafer dicing conditions for NCF laminated wafer

Wafer dicing failure with dicing conditions in Table 2-1.

Photos of(a) Cupillar/Sn-Ag micro-bump and (b) Si substrate with

(a) Adhesion strength and (b) elongation measurement oflaminated NCFson Si wafers

DSC curves ofNCF 1, NCF 2, and NCF 3

Adhesion measurement of3 types ofNCFsat the room temperature

Elongation measurement of3 types ofNCFs at the room temperature

Opticalimages of3 types ofNCFs laminated bare wafers after a dicing process

Opticalimages oftop surface near a scribing line after the NCF lamination on an entire wafer and dicing processes

(a) After dicing ofNCFs laminated wafer, (b) The tilted view ofa singulated chip

Conventional singlechip packages: Cu pillar/Sn-Ag micro-bumpjoints

WLPS: Cu pillar/Sn-Ag micro-bumpjoints as a function ofbonding pressures of (a) 20N, (b) 40N, (c) 60N, and (d) 80N

Joint gap of conventional singlechip packages and WLPs

Open circuit ratesin daisy chain resistance measurement

Bump contact resistances depending on bondingforce

Bump contact resistances of (a) conventional singlechip packagesand

Bump contact resistances of (a) conventional single chip packages and

Curing agents for epoxy resin

2-Phenylimidazole for epoxy resin

Curing mechanism of (a) DICY and (b) imidazole curing agen

Curing behavior ofDICY-NCF and imidazole-NCF

On-set and peak temperature ofNCFs

Structure of test vehicles

IR absorbance graphs ofNCFsin FT-IR spectroscopy measurement

Bonding profile and schematicimage of chip bonding

DSC dynamic scans ofDICY-NCFs with heatingrateof2,5,10,20,40

Degree-of-cure of DICY-NCF VS. temp.

Ozawa plotlogarithms of heating rate againstthe inverse of

DSC dynamic scans ofimidazole-NCFs with heating rate of2, 5,10,20,

Degree-of-cure ofimidazole-NCF VS. temp.

Ozawa plot logarithms ofheating rate against the inverse of temperature at constant conversion level generated from DSC dynamic scan (imidazole-NCF)

Dynamic and isothermal kinetic parameters at different degree-of-cure (imidazole-NCF)

Predicted isothermal curves ofdegree-of-cure atvarioustemp. from DSC dynamic scans ofDICY-NCF

Predicted isothermal curves ofdegree-of-cureatvarioustemp. from DSC dvnamic scansofimidazole.NCF

Predicted degree-of-cure ofDICY-NCF and imidazole-NCT at bonding

Solderjoint morphology after conventional ramp-up bonding using

Temperature profile of bonding process

Joint gap and joint resistance of samples bonded using DICY-NCF and

Degree-of-cure ofDICY-NCF and imidazole-NCF during chip bonding

Schematicimages of solderjoint morphology afterchipbonding

(a) Conventional ramp-up bonding, (b) isothermal bonding, (c) temperature profileof(a), and (d) temperature profile of(b)

(a) Gap size modeling, (b) viscosity depending on heatingrate, and (c)

Possible solderjoint morphologies using NCFs at fine-pitch TSV intercoruection

Test vehicles

Isothermal bonding

Variables of isothermal bonding process

Bonding profile: (a) bonding temperature and (b) bonding time

Minimum viscosity ofDICY-NCF and imidazole-NCF with various silica filler contents

Viscosity of (a) DICY -NCF and (b) imidazole-NCF with various silica

Cross-sectional view of solderjoint using DICY-NCF with (a) 0wt%,(b)

Contactand daisy chain resistance ofsamplesbonded with DICY-NCF and imidazole-NCF by conventional ramp-up bonding

Cross-sectional view of solderjoint using DICY-NCF with (a) 0wt%,

Contact and daisy chain resistance of samples bonded with DICY-NCF and imidazole-NCF by isothermal bonding

Cross-sectional view of solderjoints using DICY-NCF at(a) 40'C,(b)

Jointgap ofNCFsduring conventional ramp-up bonding

Degree-of-cure ofNCFsduring conventional ramp-up bonding

Cross-sectional view of solderjoint usingDICY-NCF at (a) 140 CC(b)

Joint gap of NCFs during isothermal bonding

Degree-of-cure ofNCFs during isothermal bonding

Cross-sectional view ofsolderjoints bonded with DICY-NCF at the

Joint contact resistances of solderjointsusing DICY-NCF and

Jointgap ofsolderjoints using DICY-NCF and imidazole-NCF bonded with various bonding pressures by isothermal bonding

Solder contact topad from cross-sectional imagesusing DICY-NCF

Cross-sectional view of solderjoints bonded with DICY -NCF at the

Joint contact resistances ofsolder joints using DICY-NCF and

Joint gap ofsolderjoints using DICY-NCF and imidazole-NCT bonded with various bonding temperaturesby isothermal bonding

Solder contact topad from cross-sectional imagesusingDICY-NCF

Heating rates of bonding profiles

Degree-of-cureofDICY-NCF and imidazole-NCF bonded with various

Sn-3A.g-0.5Cu solder wettability at various temperature [4]

Cross-sectional views of solderjoint bonded with DICY NCF at the

Joint contact resistance of solderjoints using DICY-NCF and

Joint gap ofsolderjoints using DICY-NCF and imidazole-NCT bonded with various bondingtimeby isothermal bonding

Solder contact topadfrom cross-sectional images usingDICY-NCF

Degree-of-cure ofDICY-NCF and imidazole-NCP bonded with various

Mechanism ofisothermal bonding usingimidazole-NCFs