Assessment of Low-Ag Solder Alloys for General Lead-Free Assembly

Reading time ( words)

Editor's Note: This article originally appeared in the February 2012 issue of SMT Magazine.The electronics industry has been transitioning from Sn37Pb to lead-free Sn3Ag0.5Cu over the last 10 years. One challenge with this transition is the use of 3% silver in lead-free SnAgCu solder causing rapid price increases in lead-free solder alloys due to the limited global availability of silver metal. Based on this, there has been activity and interest in transitioning to low-silver (≤1.2wt%) lead-free solder alloys in BGA/CSP component spheres, wave solder and rework alloys and, now, solder paste. In terms of the global solder market per year, approximately 180,000 tonnes of solder is used with approximately 20,000 tonnes in solder paste and 160,000 tonnes for wave soldering. If we consider the approximate global silver metal usage per year as 20,000 tonnes, with a world capacity of 21,500 tonnes, there is only a spare capacity of 1,500 tonnes of silver per year which could be used for the solder market. This would be enough silver for a lead-free solder alloy up to approximately 1wt% Ag.With Sn3Ag0.5Cu (SAC305), due to the limited availability of silver, and with precious metals such as silver and gold being an easy target for investment/speculative buying, the cost of silver has increased considerably in recent years from around $250,000 per tonne in 2005 to $1,000,000 per tonne in 2011. Based on this, there have been numerous requests from customers for a low-silver solution.Koki developed a low-silver alloy (Koki S01X7C) which contains only 0.1Ag in a lead-free alloy formulation Sn0.1Ag0.7Cu0.03Co. This helped to reduce the cost considerability compared with Sn3Ag0.5Cu as the final alloy price is based mainly on tin metal rather than silver as shown in Figure 1.Figure 1: Cost breakdown of Sn3Ag0.5Cu versus low-silver lead-free solder alloy (Sn0.1Ag0.7Cu0.03Co) showing lower overall solder price based on less silver used.

Low-silver lead-free alloys have been developed by many solder manufacturers to address the cost and silver supply issue, but they have been mainly found to have reduced thermal cycling reliability performance versus Sn3Ag0.5Cu. Also temperature differences between low-silver alloys and Sn3Ag0.5Cu need to be considered when transitioning over, as the melting point of the low-silver alloy tends to increase. Sn3-4Ag0.5Cu lead-free alloys have a melting point of approximately 217ºC compared with 225ºC for Sn1Ag0.5Cu and 227ºC for Sn0.7Cu. SMT Solder Paste Alloys

Their higher melting temperature of low-silver alloys compared with Sn3-4Ag0.5Cu lead-free alloys is particularly a factor in reflow soldering where reflow profile temperatures and times tend to be higher than during wave soldering. There are some low-silver solder paste alloys used today, typically for smaller cell-phone-type products, where temperatures during reflow are not excessive. Some of the low-silver solder pastes which could be used include Sn1Ag0.5Cu, Sn0.3Ag0.7Cu, Sn0.1Ag0.7Cu, Sn0.1Ag0.7Cu0.03Co, Sn0.7Cu, and Sn0.7Cu0.05Ni.

Wave Solder Alloys

For wave soldering, the use of low-silver solders is more of a consideration due to increased solder usage in wave machines versus SMT. Alloys which are used or considered include Sn0.7Cu0.05Ni, Sn1Ag0.7Cu, Sn0.7Cu0.3Ag and Sn0.7Cu0.3Ag0.03Ni (JEITA) alloys, Sn0.1Ag0.7Cu0.03Co.

As the melting point for the low-silver alloys is higher, this becomes a concern for wave soldering thicker boards, as the use of higher pot temperatures can cause potential component and board issues so higher pot temperatures are limited which would reduce wave holefill.

Hand Solder Rework

Alloys which are used or considered for hand soldering would include those low-silver alloys which are used in the solder pastes. These would include Sn1Ag0.5Cu, Sn0.3Ag0.7Cu, Sn0.1Ag0.7Cu, Sn0.1Ag0.7Cu0.03Co, Sn0.7Cu, Sn0.7Cu0.05Ni alloys.  Again, potentially higher soldering iron tip temperatures exist with the low-silver-based alloys, especially for thicker, more thermally-demanding boards.

BGA/CSP Rework

Alloys which are used or considered for BGA/CSP rework would be those low-silver alloys which are used in the solder pastes. These would include Sn1Ag0.5Cu, Sn0.3Ag0.7Cu, Sn0.1Ag0.7Cu, Sn0.1Ag0.7Cu0.03Co, Sn0.7Cu and Sn0.7Cu0.05Ni alloys. With the higher melting point of these low-silver-based alloys, higher board and component temperatures would be encountered similar to than encountered during SMT reflow, so excessive temperatures cannot be used or else there will be component or board damage.

Wave Rework

For alloys for use during mini-pot/solder fountain rework some of the key criteria involve obtaining good holefill and low copper dissolution and would be based on the alloys used for wave soldering. Some of the low-silver alloys considered based on low copper dissolution would include Sn0.7Cu0.05Ni, Sn0.7Cu0.3Ag0.03Ni (JEITA alloy), and Sn0.1Ag0.7Cu0.03Co.

As these are higher in melting point than Sn3-4Ag0.5Cu, with the use of higher pot temperatures, higher board and component temperatures would be encountered with a trade-off based on obtaining good wave holefill but not getting component or board damage or copper dissolution during wave rework.

BGA/CSP Sphere Alloys

With the need for improved mechanical drop shock reliability, low-silver sphere alloys, such as Sn1Ag0.5Cu, have seen increasing use. This has been offset for certain applications by the need to have improved ATC reliability comparable with Sn3Ag0.5Cu solder spheres with investigations occurring in this area to develop low-silver solder sphere alloys for improved mechanical shock and thermal cycling. Assembly of low-silver solder component sphere alloys with Sn3Ag0.5Cu has also required the need to increase minimum solder peak temperature to obtain good soldering between the Sn3Ag0.5Cu solder paste and the Sn1Ag0.5Cu component solder sphere.

Standards Affected by Use of Low-Silver Solder Alloys

When we consider assembling electronics boards with low-silver lead-free solders, we must also consider the processing temperatures and standards needed for low-Ag solder alloys with the following sections discussing the affected standards.

Reflow Assembly Standards

Different reflow profiles are needed for tin-lead, lead-free (high-silver) and lead-free (low-silver) processing with the guidelines indicated below:

  • Tin-lead: Time over 183ºC: 30 to 90 sec, peak: 205 to 215ºC;
  • Lead-free (Sn3-4Ag0.5Cu): Time over 217ºC: 30 to 90 sec, peak: 235 to 260ºC; and
  • Lead-free (Sn0.3Ag0.7Cu): Time over 227ºC: 30 to 90 sec, peak: 240 to 270ºC

In addition to processing temperatures, we must also consider component temperature rating reflow standards which include IPC/JEDEC J-STD-020. The current J-STD-020 standard for Sn3-4Ag0.5Cu (MP: 217ºC) indicates components rated to 3x 245ºC to 3x 260ºC peak (based on package thickness/volume) with an additional 1x 260ºC for area array rework (BGA/CSP) parts not rated to 260ºC. The time above 217ºC is 60 to 150 seconds.

For Sn0.3Ag0.7Cu (MP: 227ºC) this could be changed to 3x 255ºC to 3x 270ºC peak with an additional 1x 270ºC for area array rework. The time above 227ºC is 60 to 150 seconds.

Wave assembly standardsFor wave soldered through-hole components, JEDEC JESD-B106-D standard is used which indicates for tin-lead first-pass wave soldering the component should be able to withstand 260ºC for 10 seconds. For lead-free Sn3-4Ag0.5Cu first-pass wave soldering the component should be able to withstand 270ºC for 7 seconds. The standard also indicates an optional lead-free through-hole rework component temperature rating of 270ºC for 15 seconds.

For Sn0.3Ag0.7Cu this could be changed with first-pass wave soldering the component being able to withstand 280ºC for 7 seconds with an optional lead-free Sn0.3Ag0.7Cu through-hole rework component temperature rating of 280ºC for 15 seconds.

For wave-soldered components which are immersed in the wave, for example bottom side SMT components, JEDEC JESD22-A111 standard is used which indicates the component should be able to withstand 260ºC for 10 seconds during tin-lead wave soldering. This standard has not been updated for lead-free wave soldering. For Sn3-4Ag0.5Cu and Sn0.3Ag0.7Cu wave soldering we could use the following component temperature ratings:

  • 270ºC for 10 seconds (Sn3-4Ag0.5Cu).
  • 280ºC for 10 seconds (Sn0.3Ag0.7Cu).

Board Temperature Ratings

Base Materials Specification for PCBs (IPC4101) indicates that time to delamination of the board laminate at 288ºC (T288) should be at least 5 minutes for lead-free Sn3-4Ag0.5Cu solder. For Sn0.3Ag0.7Cu we could use a time to delamination of 10 minutes at 288ºC for Sn0.3Ag0.7Cu or a higher temperature than 288ºC for 5 minutes.

The laminate decomposition temperature (Td) for lead-free Sn3-4Ag0.5Cu should be at least 325ºC. For Sn0.3Ag0.7Cu we could use a laminate decomposition temperature 335ºC for Sn0.3Ag0.7Cu.

Component and Board Warpage Standards

The amount of warpage which can be allowed during assembly for boards and components needs to be minimized to prevent such defects as head-in-pillow.

There has been work by JEITA and JEDEC to update component coplanarity specifications to include coplanarity requirements during SMT reflow with the maximum package warpage during lead-free Sn3-4Ag0.5Cu reflow up to 260ºC peak temperature being 3 to 6 mils dependent on the component ball pitch from 0.4 mm to 1.27 mm. For Sn0.3Ag0.7Cu, we could use a criteria of 3 to 6 mils warpage for Sn0.3Ag0.7Cu up to 270ºC peak temperature.

For board warpage, IPC standards refer to maximum board flatness of 7.5 mil/inch for tin-lead or lead-free Sn3-4Ag0.5Cu at room and reflow temperatures. There is no indication of maximum board flatness for low-silver solder alloys such as Sn0.3Ag0.7Cu. In addition, the current IPC board standards do not scale correctly at this time with package size and I/O count so that board flatness specifications need to be adjusted with lead-free Sn3-4Ag0.5Cu and Sn0.3Ag0.7Cu solder updates needing to be incorporated into the IPC standards.

Manufacturing Tests on Low-Silver Solder Alloys

In addition to component and board temperature and warpage specifications we should also consider what manufacturing tests would be needed for the assembly of low-silver lead-free solder. Tests would include SMT assembly, wave assembly and rework (BGA/CSP, mini-pot/solder fountain, hand-soldering evaluations). The INEMI Alternative Alloy group is working on a document to address this.

Reliability Tests on Low-Ag Alloys

There is a relatively long history of testing of low-silver solder alloys. Some of the early work dates back to 1992 with the work of the UK Department of Trade and Industry Project on lead-free solder project involving Sn3.5Ag, Sn0.7Cu, Sn2Ag alloys. Based on this work, Nortel manufactured Meridian telephones with Sn0.7Cu alloy in addition to conducting investigations with Sn2Ag based on lower silver use in the lead-free alloy.

In 2006, with the movement to lead-free soldering, many component suppliers moved from Sn3-4Ag0.5Cu to Sn1Ag0.5Cu BGA/CSP component solder spheres for improved mechanical drop shock resistance. Because of the proliferation of low-silver BGA/CSP component spheres, the iNEMI Alternative Alloy group started a program in 2009 looking at thermal cycling reliability investigations on low-silver alloys mainly for BGA/CSP solder spheres with Sn3Ag0.5Cu as solder paste. iNEMI Alternative Alloy Project is looking at 12 lead-free alloys with tin-lead solder as the control. The project is mainly looking at the impact of silver content on reliability together with the impact of common dopants, such as nickel.

For wave soldering in 2007, iNEMI conducted investigations on low-Ag alloys including Sn3Ag0.5Cu, Sn0.7Cu0.05Ni, Sn0.3Ag0.7CuBiX (X = dopant alloy). In the same year JEITA conducted investigations on low-silver alloys mainly for wave with Sn0.3Ag0.7Cu and Sn1Ag0.7Cu as the main alloys were under investigation due to reduced cost of Sn3-4Ag0.5Cu wave-soldered consumer type products. Sn0.3Ag0.7Cu and Sn1Ag0.7Cu alloys were recommended for wave soldering by the JEITA group with Sn1Ag0.7Cu recommended for increased ATC thermal cycling requirements.

For wave rework in the same year, iNEMI looked at investigations on Sn0.7Cu0.05Ni alloys for wave rework for thicker server boards to reduce copper dissolution with Sn3Ag0.5Cu still used for SMT and wave.

Other examples of low-Ag solder testing in industry include copper erosion testing of Sn0.1Ag0.7Cu0.03Co (Koki S01X7C) versus Sn3Ag0.5Cu (SAC305) using copper samples. In these test the (Cu-Co)6Sn5 intermetallic compound (IMC) layer formed with Sn0.1Ag0.7Cu0.03Co alloy acting as a barrier for copper erosion compared with Sn3Ag0.5Cu alloy.

In addition IMC formation testing was performed with Sn3Ag0.5Cu, Sn0.1Ag0.7Cu0.03Co, Sn0.3Ag0.7Cu and Sn0.7Cu0.05Ni solder during reflow. The slope of the graph increased with increased reflow time with IMC layers for Sn3Ag0.5Cu and Sn0.3Ag0.7Cu growing rapidly. The IMC layer growth was low for Sn0.1Ag0.7Cu0.03Co and Sn0.7Cu0.05Ni based solder with Sn0.1Ag0.7Cu0.03Co having the lowest IMC growth rate as indicated in Figure 2.

Figure 2: IMC growth rate during reflow soldering for Sn3Ag0.5Cu, Sn0.3Ag0.7Cu, Sn0.1Ag0.7Cu0.03Co, and Sn0.7Cu0.05Ni lead-free solders.

With assembly of chip components using these alloys during reflow following by thermal cycling, for 2012R [0805] soldered chip resistors, solder joint cracking occurred early with Sn0.3Ag0.7Cu (JEITA alloy) and Sn0.7Cu0.05Ni alloys. Sn0.1Ag0.7Cu0.03Co had low solder joint cracking similar to Sn3Ag0.5Cu as shown in Figure 3.

Figure 3: ATC testing from -40ºC to +125ºC for Sn3Ag0.5Cu, Sn0.3Ag0.7Cu, Sn0.1Ag0.7Cu0.03Co, and Sn0.7Cu0.05Ni lead-free solders soldered to 2012 [0805] chip resistors.

In terms of reflow investigations with different peak reflow temperatures and time, good solder joints formed at a peak temperature of 240ºC on soldered 2012 [0805] chip components with Sn0.1Ag0.7Cu0.03Co solder paste. Reflow profiles at or above 240ºC peak with time over melting point of 30 seconds were recommended. Evaluations on a mobile phone product with Sn0.1Ag0.7Cu0.03Co soldered QFN, BGA and chip components using visual, cross-section and X-ray showed no issues.

Reliability Standards

In addition to component and board standards, reliability standards for low-silver solder alloys should be standardized. Reliability test standards based on those developed for tin-lead and lead-free Sn3Ag0.5Cu solder would be used for low-Ag lead-free solders.

IPC Solder Products Value Council (SPVC) has developed a basic material properties document for alternative lead-free low-silver solder alloys which indicates that test data is needed for:

  • Alloy composition;
  • Differential scanning calorimetry (melting point);
  • Wetting balance (solderability);
  • Copper dissolution test;
  • Coefficient of thermal expansion (CTE);
  • Tensile testing (ultimate tensile stress, etc.); and
  • Dynamic modulus testing (elastic modulus, etc.).

The iNEMI Alternative Alloy group generally agrees with the IPC SPVC document with the following comments:

  • The Basic Material Property Testing document should not have copper dissolution and creep tests included at this time. 
  • Note in the document that copper dissolution and creep test methods are being developed and will be included in a future revision.
  • iNEMI Alternative Alloy group recommends working with the IPC SPVC to establish a group to develop standard tests for copper dissolution and creep.

There is ongoing work to develop a joint IPC/iNEMI document on basic material properties for alternative low-silver solders.

iNEMI also has recommendations on mechanical testing and ATC testing for alternative low-silver lead-free alloys. The recommended tests address the need for standard data regarding:

  • Thermal fatigue of SMT joints;
  • Mechanical shock of SMT joints; and
  • Through-hole (TH) joints, preconditioned pull tests only.

The tests include:

  • ATC: 0ºC to 100ºC, -40ºC to 125ºC according to IPC 9701 standard.
  • Mechanical Shock: according to JEDEC standard JESD22-B111.
  • Pin-pull testing: Pin pull testing of conditioned/aged TH soldered components (pin-pull force measured versus pin wetted length).

The testing described in the iNEMI reliability document is for general alloy reliability assessments. These tests do not replace those performed by alloy producers and their customers to qualify alloys for specific product use. In addition, electronics manufacturers are encouraged to perform application-specific testing and further engineering, as appropriate, prior to the use of any solder alloy for a given product.

Conclusions and Future Work

Based on a review of the status of low-silver lead-free solders, solders can be developed which have reduced solder material cost reduction with reduced risk of precious metal price fluctuations in addition to good mechanical property test data compared with Sn3Ag0.5Cu.

Due to the generally higher melting point of low-silver lead-free solders, apart from reliability and process testing, we need to consider updating component and board temperature rating standards including J-STD-020 (SMT Components), JEDEC JESD B-106D (Wave and Wave Rework Components), JEDEC JESD22-A111 (Immersion Waved Soldered Components), IPC 4101 (Board Laminate Rating), JEDEC and JEITA Component Warpage Standards (During Reflow) and IPC Board Warpage Standards (During Reflow).

The industry needs to work on developing and updating test methods and standards for low-silver solders, including:

  • Physical property data;
  • Reliability data;
  • Manufacturing tests (SMT, wave and rework); and
  • Temperature-related standards for board and components as previously mentioned.

Note: The information discussed in this article was presented at the IPC Conference on Reliability in Irvine, California in November 2011. About the author: Jasbir Bath is a Consulting Engineer for Christopher Associates/Koki Solder in the Americas. He is the owner of Bath Technical Consultancy which provides consulting and training services in the EMS industry. Bath is an iNEMI consultant working on the INEMI lead-free rework optimization project. He was the Corporate Lead Engineer with Solectron Corporation and Flextronics International for 10 years with a role involving tin-lead and lead-free solder process development.