Homepage » Allgemein »

From board assembly to semiconductor

Allgemein
From board assembly to semiconductor

Technological advances often necessitate the need to re-evaluateexisting manufacturing processes and inevitably require fundamental changes to meet challenges. The use of solder spheres is an accepted method of creating the electrical interface between die and substrate for device families based on flipped chips. Now, from the SMT environment, paste printing moves toward the advanced-packaging technology, offering benefits in terms of cost and volumes. The otherissue is now, how can we eliminate lead from components, i.e. from terminations, pads and solder balls.

Stan Renals, Steve Dowds, The Indium Corporation

Driven by the demand for a greater count of I/O connections between mated parts, sphere utilization in assembly is becoming extremely complex and increasingly expensive. Motivated by the need to overcome these limitations, users are now testing various assembly methods to optimize their particular processes. These procedures include some alternatives:
• Spheres placed on flux deposited via printing or pin-transfer process, reflowed in a convection oven and generally cleaned if underfill is to be applied at a later stage of production.
• Solder paste/flux paste stencil-printed onto the component pads prior to sphere placement, reflowed in a convection oven, then cleaning has to be a consideration.
• An alternative reflow method using a small ND:YAG laser, which simply melts each individual sphere. Although a reliable process, this method is not generally used for high-volume applications.
• Solder jetting which converts molten solder directly into solder bumps. This method cannot satisfy volume-production requirements and is not considered feasible at this stage.
After evaluating these options, users firmly consider solder bumping as an alternative to sphere placement, but there are serious limitations in terms of the height-to-pitch ratio. As a rule of thumb, the maximum bump height that can be achieved is 50% of the pad-pitch size. However, wafer bumping of semiconductors using solder paste is now a recognized method of production. The cost of bumping paste is far less than spheres, and methods of production can be enhanced by doubling or tripling the throughput of current technology since tooling costs no longer inhibits the process window.
The material cost issue
The material cost per pad/bump of paste becomes significantly less than that of spheres as the geometrical size of the bumps decrease. A number of print thickness and associated values have been studied, and the most common values used are lay downs of 15cm2/gram and a print thickness of 4mil (100µm). This assumption uses 63/37 bumping paste valued at euro 0.10/gram. If we use 30 mil (760µm) square pads, assumed to be the equivalent to 30 mil (760µm) spheres at a cost of 0.10 per 1000 spheres. The printing process will produce 2586 bumps/gram versus 1000 spheres at euro 0.10 per thousand. If the hypothesis is valid that the sphere price remains con-stant for all sizes, once we get down to 10 mil (254µm) diameter, more than 23,000 pads canbe printed per gram of paste. This equationdoes not even consider that as sphere-size requirement decreases, the cost of smaller size spheres increases.
Although utilizing spheres has distinct advantages, more than 50% of the material/process costs are associated with sphere placement. It is estimated that the printing cost is 50% of the sphere-placement cost, because multiple lines can be printed at greater speeds than placement. It can be argued that paste waste has not been taken into consideration, or that the yields of these different processes are not comparable. Currently sphere yields can be as high as 98%. Although bumping-paste yield-data is limited, figures as high as 97% have been recorded at users in Asia. As this process receives more exposure, yields equaling or even exceeding current technology can be anticipated.
Depending on the sphere size and placement equipment being used, average-tooling costs can vary from euro 10,000 to 20,000. Stencil printing used in surface-mount technology (SMT) can be utilized to perform the same function with associated costs at only a fraction of tooling costs. In addition, the time required to change equipment associated with each sphere size is eliminated. Production teams have recognized the cost benefits in switching to paste bumping and improved yields. Also, line engineers see the added value by eliminating flux-paste/solder paste printing before the sphere placement. This means an entire process step can be removed.
Solder-bumping methods employed now to manufacture flipped chip joints are emerging as a cost-effective solution at the integrated-circuit (IC) level. This technology has been available to conventional SMT manufacture for more than 20 years, and obviously is a natural progression to the IC level. The main breakthrough has come from a clean process featuring tighter tolerances as well as wafer handling and support.
Focus on alloys
The materials used for flip-chip in package (FCIP) are generally high melting-point alloys, such as 97Pb3Sn or 95Pb3Sn. Such a paste ensures that the solder joints will not re-melt during a subsequent assembly process using 63Sn37Pb solders. Flip-chip on board (FCOB) applications use eutectic or near eutectic solders such as 81Pb19In, which gives better fatigue performance and compatibility with a NiAu-platted pad on a substrate.
The alloy used in BGA and CSP processes, especially in the ceramic column grid-array (CCGA) or ceramic ball-grid array (CBGA), is typically 90PbSn. The semiconductor die is mounted onto an array package using eutectic alloy. The higher melting point ensures a stand-off in any subsequent assembly process. Plastic ball-grid array (PBGA) technology uses eutectic or silver-bearing alloys of 62Sn36Pb2Ag.
Lead-free – a specialty
The pending WEEE directive in Europe (see lead-free special in EPP Europe #9/10, 2002), together with the Japanese home-electronics recycling law which came into effect in April 2001, have combined to give the industry an impetus towards going Pb-free. This movement to rid electronics of lead includes components as well as solder paste. Particularly in Japan, and to a lower extend in European companies, Pb-free adoption is advancing quickly, driven by legislation, environmental concerns, application advantages, etc. Many selection criteria used when choosing a Pb-free alloy for soldering also apply when considering components, including solderability, toxicity, stability, availability, reliability and compatibility with existing flux and process technologies.
The industry is compelled to use tin as the base metal for a Pb-free solder since it provides reasonable wettability, is non-toxic, easily available, and melts at an appropriate temperature. In addition, tin is very stable unlike zinc that reacts with both acids and bases. However, since pure tin has a relatively high melting point of 232°C, it must be doped to produce an alloy with a lower melting point and improved wettability. In functional terms, the Pb-free solder paste choice comes down to tin-bismuth-silver (SnBiAg) and tin-silver-copper (SnAgCu).
SnBiAg has a lower melting point (205 to 210°C, depending upon composition), and exhibits better wettability than other alternatives, and it generally appears to be the best choice. Unfortunately, this alloy is sensitive to Pb-contamination – an addition of only 1.0% of lead will reduce the joint fracture strength by 80%. (T.J. Biaggio, K. Suetsugu, T. Okamura: Challenges and Solutions for Lead-free Soldering of Large PCBs). Therefore, the maximum allowable Pb limit is 0.2% for this alloy. Another objection often raised against Bi-containing alloys is that there are concerns about Bi availability. It is certainly true that bismuth is often produced as a by-product of Pb-mining, but the electronics industry only accounts for 0.5% of the global Pb-consumption. Even if the entire electronics industry stopped using Pb now (which it won’t), it would barely make any difference to the global Pb production.
Another factor to consider is that the industry will be switching from alloys which are 36 to 40% Pb to alloys which are 90%+ Sn. So, Sn would in any case mostly replace the lead. In our opinion, Bi availability is not a real concern. However, as long as Pb is in the system (on the board or in component terminations), reliability may be an important concern. Once Pb is out of the system entirely, SnBiAg alloys may become preferred, due to the lower melting point and superior wetting of this alloy when compared to SnAgCu. SnAgCu is not sensitive to Pb contamination; lead will not reduce tensile strength in the joint by anything like this amount. This means that components with SnAgCu can be used in conjunction with any of the alloys, both Pb-containing and lead-free, without any concerns about reliability. In addition, SnAgCu has been recommended by IPC, NEMI in the USA, Brite-EURAM in Europe and JEITA in Japan, and seems likely now to become the most popular replacement for SnPb.
There are many proposed compositions in existence, but it is fair to say that there are two primary options: 95.5Sn3.8Ag0.7Cu and 95.5Sn4Ag0.5Cu. The former is subject to a number of patents, and the latter combination is ‘open’ (there is no patent, due to ‘prior art’). In practice, the two formulations are so close related that they are practically the same alloy. There is no difference in material properties between the two, and furthermore the tolerance on minor constituents in alloys is typically ±0.2%. With this in mind, although one may order an alloy with a nominal composition of 3.8Ag0.7Cu, one may actually receive 4Ag0.5Cu, or vice versa. To ensure material-purity levels, users should verify that their material suppliers have all the appropriate licenses. If licenses are present, then patents will not be a concern.
ZUSAMMENFASSUNG
Zwischen den Kontaktpads der umgedrehten (flipped) Chips und der Substrate setzt man Lotkugeln als leitende Verbindung zwischen den unterschiedlichen Kontaktflächen der Bauteile ein. Neben fertigen Zinnbällchen (Solder Spheres), deren Aufbringen auf solch ein Kontaktarray sehr aufwendig sein kann, kann man per Schablone Lotpaste direkt auf die Kontaktpads der einzelnen Dies auf dem Wafer drucken und diese dann zu Lothöckern (Solder Bumps) umschmelzen. Das geht schnell und ist zudem sehr kostengünstig, auch im Bleifrei-Prozeß.
RÉSUMÉ
Entre les pads de contact des flip-chips retournés et les substrats, on utilise des boules de soudure pour établir la liaison conductrice entre les différentes surfaces de contact des composants. Outre les petites boules d’étain finies (solder spheres) dont la mise en place sur une telle surface de contact peut être très complexe, il est également possible d’imprimer par sérigraphie de la pâte à braser directement sur les différents dies et de les transformer ensuite en bosses de soudure (solder bumps) par fusion. Le procédé est rapide et très économique, même dans la production sans plomb.
SOMMARIO
Tra i pad di contatto dei chip voltati (flipped) e dei substrati si impiegano palline di saldatura come collegamento conduttivo tra le diverse superfici di contatto dei componenti. Oltre alle palline di stagno ultimate (Solder Spheres), la cui applicazione su di un tale array di contatto può essere molto dispendiosa, mediante una mascherina è possibile applicare direttamente la pasta di saldatura sui pad di contatto dei singoli componenti, e procedere successivamente con la stampa sul Wafer e quindi alla saldatura a fusione (Solder Bumps). Ciò è molto rapido e inoltre molto conveniente dal punto divista dei costi, anche nel processo senza piombo.
Acknowledgements: Reflow Soldering Processes and Troubleshooting Dr. Ning-Cheng Lee of The Indium Corporation Solder bumping and paste/flux choices feed cost-efficient wafer treatment
DirEKt solder-ball placement technology provided by paste print through stencil (courtesy of DEK)
Placed but not reflowed solder ball – eutectic PbSn, 300 microns diameter (Courtesy of PacTech)
Current Issue
Titelbild EPP EUROPE Electronics Production and Test 11
Issue
11.2023
READ
Newsletter

Subscribe to our newsletter now

Webinars & Webcasts

First hand technical knowledge

Whitepapers

Find all current Whitepapers here

Videos

Find all current videos here


Industrie.de Infoservice
Vielen Dank für Ihre Bestellung!
Sie erhalten in Kürze eine Bestätigung per E-Mail.
Von Ihnen ausgesucht:
Weitere Informationen gewünscht?
Einfach neue Dokumente auswählen
und zuletzt Adresse eingeben.
Wie funktioniert der Industrie.de Infoservice?
Zur Hilfeseite »
Ihre Adresse:














Die Konradin Verlag Robert Kohlhammer GmbH erhebt, verarbeitet und nutzt die Daten, die der Nutzer bei der Registrierung zum Industrie.de Infoservice freiwillig zur Verfügung stellt, zum Zwecke der Erfüllung dieses Nutzungsverhältnisses. Der Nutzer erhält damit Zugang zu den Dokumenten des Industrie.de Infoservice.
AGB
datenschutz-online@konradin.de