Phase II: Optimizing Cleaning Energy in Batch and Inline Spray … · 2016-11-16 · Phase II:...

Phase II:

Optimizing Cleaning Energy

in Batch and

Inline Spray Systems

Steve Stach, Austin American Tech.Mike Bixenman, Kyzen Corp.

Outline

Project Review

Phase 1 Questions / Theory

Phase II Results

Phase II Conclusions / Tools

Future Work – Phase III

Introduction

The benefit of a well defined cleaning process:

improves manufacturing efficiencies

increases process yields

So, what constitutes an optimized process?

cleaning agent effective on wide range of soils

integration of machine with chemistry

mechanical design delivers chemistry at the heart of the residue

control and re-use of fluids

Challenges

Converge of circuit boards and die packaging technologies

higher performance electronic devices

Technical issues from a cleaning perspective:

low standoff

fine pitch solder bump arrays

ionics trapped underneath active components

spacing between conductors may pose risk of electromigration

Challenges

Statement of Problem

Staying ahead of the ever-advancing technology

curve:

industry challenged to improve cleaning processes

increased complexity of board and geometry

new solder paste and flux formulations

improved performance at lower cost

Mechanical and chemical energy

are the key variables to meeting

demands

Research Basis

Cleaning is becoming more market critical

Miniaturization converges circuit assembly and die packaging technologies

More assemblers cleaning “no-clean” because of performance issues

More assemblers cleaning “Lead-free” as residues are generally tougher & uglier than 63/37 leaded “no-clean”

Our Hypothesis – Understanding mechanical and chemical energy are the key variables to meeting demands

Research Objectives

Improve cleaning performance at reduced cost

New approaches to mechanical energy delivery in machine

design:

Energy targeted at the heart of the toughest residue

Advanced cleaning chemistry formulations:

lower operating temperature,

concentration, long bath life

Process optimization of machine

and chemistry parameters

Process speed & low process cost

Phase II Research Questions

What are the cleaning mechanisms?

What are the rates of the cleaning mechanisms?

How does the gap spacing affect the cleaning rate?

Is there any difference between lead-free and

leaded flux cleaning?

How does flux aging affect cleaning rate?

Can we write an equation to predict cleaning rate?

Project Review

Research Objective

Understand & design better cleaning fluids and “spray in air” de-fluxing systems to meet new market demands

Smaller, faster, more reliable, eco-friendly electronic products

Phase I (2004) – Develop theoretical model

Propose Rate equation for predicting cleaning performance

Define forces and how they are effected by spray nozzle design

Phase II (2005) – Design experiment to test phase I model

“Real time ” measurement of cleaning rate

Modified rate equation based on observed data

Comparison of lead-free vs. traditional leaded no-clean

Phase III (2006) - Optimize system variables

Multiple-lead-free fluxes, process timing, chemistry, machine and process variables,

Phase I: Rate Theory

“The over-all process cleaning rate is sum of the static rate plus the dynamic rate”

5 MPH + 55MPH = 60MPH

Static Rate + Dynamic Rate = Total Rate

Static Rate = Rate at which process will proceed on its own

Dynamic Rate = Rate increase attributable to machine

Phase I: Rate Theory

“The over-all process cleaning rate is sum of the static rate plus the dynamic rate”

5 Mils/Sec. + 55 Mils/Sec. = 60 Mils/Sec. Static Rate + Dynamic Rate = Total Rate

Static Rate = Rate at which process will proceed on its own

(solubility driven)

Dynamic Rate = Rate increase attributable to machine

(energy driven)

Solubility's Contribution

“Dissolve-it”

age old, tried and true

augmented with heat, blasting

and scrubbing

Rate of solubility

dependent on dissolution rate

temperature effect in

dissolving residue

concentration of solvent

needed to dissolve residue

Fluid Dynamics

Soluble

Very Soluble

Marginal Solubility

Dissolution Rate

Temp

The RS & RD Cleaning Rate Balance

Rp=Rs+RD

System

Design

Wash Rinse Dry

%Rs %Rd %Rs %Rd %Rs %Rd

Static

Immersion

100% 0% 100% 0% 100% 0%

3-D

Dish

Washer

70% 30% 70% 30% 50% 50%

2-D

Planarized

Batch

30% 70% 40% 60% 20% 80%

Inline

”Air Spray”

20% 80% 20% 80% 2% 98%

Process Parameters

Key to predicting optimized process performance

understanding the nature of the soil and the chemical

and physical needs for removing it

Static Rate – Driving forces

Driving Factor Range Typical Governing Equation

Solubility constant

of residue in

solvent

Pka

Insoluble - <.001

Marginal - .001-0.1

Soluble - > 0.1

[Anion] [Cation]

[Residue]

Temperature 0-100 C Temp./Solubility

coefficient

depends on

residue and

solvent

Dynamic Rate – Driving forces

Source of Surface

Energy

Range of Energy

Available

Governing

Equation

Capillary action 0-2”wc

(0-1.0 psii)

Pc = 2. γ / R

Gravity Flow 0-1”wc

(0-0.5 psii)

Pg = ρ.g.h

Impingement

Pressure

0-275”wc

(0-10 psii)

Pi = ½(ρv2)

Understanding Capillary Action

Interfacial pressure difference at equilibrium

10

1

psi

0.1

0.01

0 20 40 60

Gap/diameter, mils

Planar Cylinder

Think of surface tension as:

balloon of sorts surrounding the cleaning fluid

If it is thin and weak

cleaning fluid easily moves in and out of tight spaces

If the side wall is thick and strong

cleaning fluid will resist flow into tight spaces

Negative & Positive Surface Energy

Surface can attract or repel

Depends on surface/liquid interactions

Indicated by wetting angle

> 90° repulsion (-)

or < 90° attraction (+)

<90° >90°

Spray System Nozzle Design

Spray Type Typical

pressure @

2”,50psi man.

/Pressure

loss/in

Indicated use

Fan/Delta 2 psi /

~50%

drop/inch

Wide coverage,

overlap for high

impingement or

close work

distance

Conical 0.4 psi /

~75%

drop/inch

Widest coverage

area, lowest

kinetic energy,

flooding

applications

Coherent 10 psi /

~10%

drop/inch

Smallest coverage,

highest energy

density over

longest distance

Four variables determine

Impact pressure

Nozzle type

Distance

Manifold pressure

Fluid Density

Spray Nozzle Pressure Measured

at Board Surface

Manifold

Pressure

Flow: (gpm) Impingement (psi) @ Coverage width

(in)@

0.078”Coherent Jet 1” 2” 4” 1.5” 4.0”

30 psig 0.69 15 10 6.5 0.6 0.7

40 psig 0.82 17 12 8 0.6 0.7

50 psig 0.89 19 13 9.5 0.6 0.8

60 psig 0.97 20 15 11 0.6 0.8

F40-1.0 Fan Nozzle

30 psig 0.89 3.2 1.6 0.2 1.5 3.25

40 psig 1.06 4.4 1.8 0.3 1.7 3.60

50 psig 1.20 6.0 2.3 0.5 1.7 4.0

60 psig 1.30 7.2 2.5 0.5 1.8 4.0

Phase II Testing Outline

Test Parameters/Matrix Development

Test Vehicle Design & Prep

Development of Cleaning Analysis Recording

Lab “CARL”

Video Analysis

Testing Results

Conclusions

Our Phase II Variables

Nozzle Type

Coherent 0.078”

Material Variables

(2)Flux Types (use the most common) Indium SMQ 92J, & Multicore LF300

(1)Cleaning Chemistry Kyzen 4630

(1)Chemistry concentration 20%

(Vendor Recommended)

Process Variables

(2)Reflow Profiles (Vendor Recommended)

(1)Volume & Location of Residue Fill gap & clean excess

(1)Time Between Reflow & Clean Cool to RT not too exceed 1hr

(2)Wash temperatures 120˚F & 140˚F

(2)Spray Impingement Pressure 2psig & 10psig

(1)Location of spray Fixed, centered, adjacent

Design Variables

(1)Component dimensions 0.875”X0.875”

(2)Stand-off height 3mil, 4mil, & 5mil

(1)Surface materials glass

Phase II test groups

Sn/Pb T-20

Group 1 120F-16psii-3mil




Group 5 140F-16psi-4mil




3 samples per group X 16 groups = 48 total samples

Sample 1 aged 20mins, #2 – aged 70mins, #3 - aged 4hours

Sn/Ag/Cu LF300






Group 14140F-16psi-5mil



Test vehicle & preparation

•Flux

Glass slides with 3,4,& 5 mil gaps

Top slide .875” x .875” x

•Bottom glass

•Glue Dot –

•Heat or UV cured• Shim Shock

(removed)

•Top Glass

•Weight

•Top Glass

Hot Plate

•Bottom glass

•Thermocouple

CCD Camera

Computer/Video Recorder

MonitorGlass

Spacer

Flux

Spray

Gauge

Light

Pump Fluid Reservoir

Valve

Spray Chamber

Removable

force tool

Data Capture

Experimental set-up for Phase II – Spray cleaning optimization Study

Baseline / Control

Measures the ….

Solubility rate of the cleaning solution

Simulate static conditions

Dissolution rate of flux residue

Fixed variables

Time

Temperature

Cleaning concentration

Control Slide

Testing the Research Question

What is the cleaning rate improvement by

applying physical energy?

Fixed

Time, 2 minutes

Nozzle, Coherent

Location of spray: Fixed, centered, adjacent

Component dimensions

Surface materials: glass

Eutectic Solder Paste

Variables

Solder Paste Flux: Indium SMQ 92J

Cleaning concentration: 10 & 15%

Reflow, Ramp to spike: 218ºC

Flux volume: Fill gap & excess around slide

Time between reflow & clean: 17, 1hr, 4 hr

Wash temperatures: 120 & 140ºF

Spray Impingement Pressure: 2, 4, 10 psi

Standoff Height: 3, 4, & 5 mil

Eutectic Cleaning Soon after Reflow

Eutectic Cleaned 1-hour after Reflow

Eutectic Cleaning 4-hours after Reflow

Pb-Free Solder Paste

Variables

Solder Paste Flux: Loctite LF-300

Cleaning concentration: 10 & 15%

Reflow, Ramp to spike: 240ºC

Flux volume: Fill gap & excess around slide

Time between reflow & clean: 17, 1hr, 4 hr

Wash temperatures: 120 & 140ºF

Spray Impingement Pressure: 2, 4, 10 psi

Standoff Height: 3, 4, & 5 mil

Pb-Free Cleaned Soon after Reflow

Pb-Free Cleaned 1-hour after Reflow

Pb-Free Cleaned 4 hours after Reflow

Data Analysis Phase II

Larger Gaps Clean Faster?

Yes for 63/37 leaded 210C profile

Not so for Lead-free 240C profile

Leaded 210C Peak reflow Un-Leaded 240C Peak reflow

0

5

10

15

20

25

BT >0% BT 25% BT 50% BT %75 BT 100%

Breakthrough (%)

Tim

e (sec.)

3 Mil

4 Mil

5 Mil

Data Analysis Phase II

Break through time was a function of time between

reflow and clean

Leaded 210C Peak reflow Un-Leaded 240C Peak reflow

Data Findings

Cleaning Rate Slows with Time

& Greater Time/Temp Reflow Exposure

Original Flux solvents evaporate leaving tougher

solvent depleted areas

Observed Cleaning Propagation

Cleaning Channels develop in

“solvent rich” out gassing channels

Bottom glass

Jet

Flux

Residue

Turn up the Pressure?

Too much pressure redirects the spray 3 dimensionally

Optimum pressure impact Too high a pressure impact

spreads 2-D spreads 3-D

Development of New Phase II

Cleaning Rate Equation (linear)Phase I Proposed Rate Equation

Time to clean = TBr + (DFmax/(Rs + Rd))

Becomes

Time to clean (using linear rate) = TL

=TBr + (DFmax/(Rs + (Rch*%Tch) + Rco))

Where;

Time to clean the area = TA

Distance of flux to be cleaned = DFmax

Time to breakthrough = TBr

Static cleaning rate = Rs

Dynamic channeling cleaning rate = Rch

Dynamic concentric cleaning rate = Rco

Fractional % of time were channeling occurs = Tch

Development of New Phase II

Cleaning Rate Equation (area & volume)Time to clean area (using area cleaning rates) = TA

=TBr + (AFmax / (Rs + (Rch*%Tch) + Rco))

Time to clean volume (using volumetric cleaning rates) = Tv

=TBr + (VFmax / (Rs + (Rch*%Tch) + Rco))

Where;

Time to clean the area or volume = TA or Tv

Area or volume of flux to be cleaned = AFmax or AVmax

Time to breakthrough = TBr

Area or volumetric static cleaning rate = Rs

Area or volumetric dynamic channeling cleaning rate = Rch

Area or volumetric dynamic concentric cleaning rate = Rco

Fractional % of time were channeling occurs = Tch

Phase I Conclusions

Science of optimizing spray-in-air requires

accurate model to predict performance

All cleaning systems are governed by two fundamental

principles:

solubility rate of the cleaning solution for the residue

physical energy available in the cleaning system

maximizing the physical energy delivered to the surface

increases the dynamic cleaning rate

Understanding the static cleaning rate plus the

dynamic cleaning rate balance is key in predicting

optimization

Phase I Conclusions

Surface energy

Energy available at the cleaning surface to do the work

In spray-in-air system

Work of cleaning requires energy to displace a fluid across a distance to create the force sufficient to achieve the rate of cleaning

Low surface tension easily moves fluid in an out of tight spaces

Capillary forces work for and against since they work for wetting but inhibit rinsing

Understanding fluid potential and kinetic energy allows for nozzle and pump configuration that maximizes surface energy

Manifold efficiency can increase surface cleaning by as much as 25%

Phase I Conclusions

Design and layout of nozzles is an important step

in optimization

Conical and fan nozzles spread the spray to cover

larger areas

Coherent jets hold together longer giving

maximum energy transfer per unit area

Overlapping jets can be an effective strategy for

increasing surface energy density

Phase I Conclusions

Rate to chemical dissolution can be augmented

with various forms of physical assistance such

as heating, impingement, concentration and

time.

Contamination loading can slow the cleaning

process by shifting chemical equilibrium

Chemical dynamics steady state when make up

exceed soil load

Phase II Conclusions

The “CARL” tester provided excellent test bed to evaluate cleaning rates and mechanisms (30 captures per second)

There are two different dynamic cleaning mechanisms

Initial cleaning rates favored larger gaps

Rates beyond 0.5” under component varied depending on dynamic mechanism & redirection

Aging of the flux residue before clean significantly reduces the rate of cleaning for no-clean and lead-free

There is an optimal impingement pressure to clean a planar gap

The lead-free flux was harder to clean than the leaded no-clean

The phase I rate equation can be modified to include break-thru time, channeling and concentric cleaning mechanisms discovered in phase II testing

Follow on Research

Phase III will use the tools developed in Phase II to further

understand the “Lead-Free” cleaning process

New comparison of lead-free fluxes

(2)Water soluble & water +(2) chem. soluble

Evaluate affect of pre-soak on break-through time

Evaluate 1 & 2 mil gaps in “CARL”

New chemistries – Alkaline vs. Neutral pH vs. Water

Nozzle design evaluation

Characterize New Nozzle types in “CARL”

Wide angle fan, narrow angle fan, new coherent

Pressure vs. rate curves for different nozzle types

Special Thanks!

Matt Kincheloe, Robert Werkun - AAT

Ram Wissel - Kyzen Corporation

Indium Corp. & Multicore/Loctite for samples

Optimized Static and Dynamic Driving Forces for

Removing Flux Residue under Flush Mounted Chip Caps

SMTAI Technical Forum

Donald Stephens Convention Center

Rosemont, IL

September 24-28, 2006

Mike Bixenman, Kyzen Corporation

Steve Stach, Austin American Technology

Abstract

Removal of flux residue under highly dense chip caps

presents a difficult cleaning challenge. Chip caps are

flush mounted to the circuit card. Upon reflow, flux

residue fills the gap under the chip cap. Cleaning fluids

must wet, dissolve, penetrate the flux dam, and flow

under the component to adequately remove all flux

residues. Increased board density, miniaturization, and

Pb-free soldering magnify this problem. To address

this problem, process parameters in the form of

cleaning temperature, time, cleaning chemistry

concentration, and impingement energy must be

considered. This paper presents the results from a

designed experiment of an advanced cleaning fluid

combined with an optimized inline spray-cleaning

machine for removing flux residue under flush

mounted chip caps.

Introduction

An optimized cleaning process delivers the necessary

chemical and mechanical energy to clean the most

difficult and sensitive areas of the part being cleaned.

Understanding the balance between static chemical and

mechanical driving forces is fundamental to predicting

and optimizing process variables. The timing and

sequence of events in a cleaning process are critical.

Each section or step in the process requires careful

thought and understanding. As the gap from the board

surface to the bottom of the component decreases,

experience tells us that cleaning becomes more

difficult.

Cleaning under flush mounted chip caps, with narrow

spacing, represents a difficult cleaning challenge. Many

variables influence the process-cleaning rate. Research

data suggests four critical variables when cleaning

electronic circuit assemblies. Higher cleaning chemistry

concentration typically increases static cleaning (rate at

which the cleaning fluid dissolves flux residue without

agitation). Increased cleaning temperature typically

improves the dissolution rate. Increased time allows the

cleaning fluid to dissolve flux under tight standoffs until

break-through occurs, which allows the fluid to flow

under the part and dissolve remaining flux residue.

Mechanical impingement creates a driving force that

increases penetration rates and reduces the time needed

to clean under the component. A fifth variable must be

considered when cleaning under tight standoffs – surface

tension and capillary action. Lower surface tension

improves capillary action, which allows the cleaning

fluid to wet and penetrate at a faster rate. The purpose of

this designed experiment is to determine optimize

chemical and mechanical forces required to remove flux

residue under flush mounted chips caps.

Problem Statement

Chip caps, flush mounted to the board, create a flux dam

under the component during reflow. The flux dam seals

the underside of the component with a thick resinous

material that is difficult to completely remove. Devices

placed in tightly packed arrays further increase the

cleaning difficulty, as there is very limited access for the

cleaning fluid to reach the contaminant. This design

challenge requires both improved chemical and

mechanical technology. The chemical driving forces can

be improved by adding materials to increase the speed

of cleaning and by improving the wet-ability of the

material to penetrate under flush mounted devices.

The difficulty, some ingredients used to improve

cleaning speed can darken or etch solder joints or

affect part markings and labels. The mechanical

driving forces require optimization in nozzle design,

selection and positioning to address difficult cleaning

challenges on the board. Even with improved chemical

and mechanical forces, time is a critical factor. The

time the board is exposed to the cleaning material

(wash time), along with the time between re-flow and

cleaning (aging time) are very important variables.

Figures 1 and 2 illustrate examples of flush mounted

chip caps. Part positing on the board, number of reflow

cycles before cleaning, peak reflow temperature, and

the thickness of the solder paste all affect cleaning

efficacy. Chip caps positioned in series may increase

the difficulty of the cleaning fluid to penetrate under

the chip cap. The number of reflow cycles or

increased reflow temperatures, such as those used for

Pb-free increase cleaning difficulty. The thickness of

solder paste increases the level of flux residue under

the chip cap, which creates a more difficult cleaning

challenge (Figure 2)

Figure 1: Chip caps in Series

Figure 2: Solder Paste Thickness

Normal Print Heavy Print

Literature Review

Inline and batch spray-in-air cleaning systems reduce the

time to clean printed circuit assemblies. Fluid mechanics

suggests that the energy delivered to the surface is equal

to the mass times the velocity squared. Impingement

pressure at the cleaning surface is dependent on the

nozzle type and distance from the nozzle manifolds to

the surface of the part. Maximizing the physical energy

delivered at the gap under the component requires

optimal pressure that reduces bounce and improves

penetration.

Flux residues clean at different rates based on the flux

make-up, time after reflow, reflow temperature, and the

cleaning fluid design. Water-soluble flux residues

typically clean at a faster rate than do rosin flux

residues, which typically clean at a faster rate than low

solids synthetic flux residues. Flux residue becomes

more difficult to clean with the passage of time after

reflow. Higher reflow temperatures allow the lower

molecular weight solvent molecules to evaporate at a

faster rate, leaving higher molecular weight resin

molecules, which increases the difficulty of cleaning the

residue. Cleaning fluid designs either dissolve or react

with the flux soil, which influences the static cleaning

rate.

As the gap from the board surface to the bottom of the

component decrease, cleaning becomes more difficult.

The standoff height of flush mounted chip caps is

typically less than one mil. Longer cleaning time is

required to wet, dissolve and breakthrough the underside

of the component. The longer time to achieve

breakthrough correlates with lower belt speeds and

longer time under wash manifolds to removal 100% of

residue. Lower dynamic surface tension allows cleaning

fluids to penetrate and dissolve the flux at a faster rate,

which decreases time to achieve breakthrough.

Once cleaning fluids start to penetrate the gap under the

chip cap, experiments reveal one of two possible

mechanisms occurs. For hardened flux residues,

concentric cleaning removes the flux residue in ever-

increasing diameters driven by dissolution of the flux in

the cleaning solvent (Figure 3). Softer flux residue

cleans at a faster rate through channeling where the

cleaning fluid penetrates in rapidly developing channel

inside the reflowed flux mass (Figure 4). Several

variables point to the degree of difficulty for cleaning

flux residue. Higher reflow temperatures harden the flux

resins, increasing cleaning difficulty. Time after reflow

increases cleaning difficulty to a point. Conversely,

boards reflowed in nitrogen leave the flux residue soft

and improve the cleaning rate. Additionally, boards

cleaned soon after reflow improves the cleaning rate.

Figure 3: Concentric Cleaning (Hard Residue)

Figure 4: Channeling (Soft Residue)

The middle of the flux is softer and more easily

breached by pressurized cleaning solvent (Figure 3).

The soldering process causes the heat to drive the

solvent molecules toward the center of the flux

creating a solvent rich zone. This solvent rich zone is

softer and more easily penetrated and dissolved.

Conversely, a solvent depleted zone is created next to

the heated and exposed surfaces. When a zone is

solvent depleted it becomes harder and more

crystalline and thereby significantly more difficult for

cleaning agents to soften and dissolve.

Figure 5: Flux Degassing

Static Cleaning Fluid Design

Cleaning fluids vary in their design based on solvency,

saponification, wetting (surfactancy), inhibition, and

defoaming characteristics. The best cleaning fluids

optimize and build performance characteristics that

effectively accomplish several tasks in combination. To

clean under chip caps, critical design features must

improve the static cleaning rate, wetting, low surface

tension, low foam, and protection of the alloys, labels

and other board components.

The static cleaning rate requires materials that rapidly

soften and solubilize the flux soils soon after contact.

The challenge is to create a universal formulation that

works well on many flux residue types. There are well

over 200 commercial solder pastes, paste flux and wave

flux materials used by industry. Although the flux

characteristics do have commonality, there are

differences that vary the static cleaning rate. Cleaning

chemistry design firms who study the many soil types

and design universal cleaning fluids open the process

window and allow uses to select different flux types

without a major impact on the cleaning process.

Wetting and surface tension effects occur through

surface treatment that reduces the droplet size and allow

the cleaning fluid to move easily in and out of tight

spaces. Surface tension can be thought of as a balloon of

sorts surrounding the cleaning. Surface-active agents

create a thin and weak droplet that improves capillary

action needed to wet under chip caps at a more rapid

rate. Surfactant free cleaning agents form a large

droplet, which facilitates initial wetting of tight spaces.

Once the fluid makes its way under the component, high

surface tension affect cause the cleaning solution to

repel and prevents rapid movement under tight spaces.

Foam is a critical parameter when using high

impingement sources. Foam increases surface tension

and allows millions of air bubbles to entrain within the

cleaning fluid. Foam causes pumps to cavitate and kills

cleaning effectiveness under tight standoffs. Foam can

also be concentration dependent. When cleaning under

chip caps, users must understand the foaming

characteristics of the cleaning fluid. Does the cleaning

fluid foam at low concentrations? Cleaning fluids that

generate no foam at low or high concentrations are the

best choice when cleaning under low standoffs.

Saponification is a common method of cleaning flux

residue. Alkaline materials react with the resin structure

to form a water-soluble soap. Most aqueous cleaning

fluids use some form of saponification. To protect the

alloy and other soft metals, minor ingredients help to

prevent alkaline attack to the alloy. Dull solder joints

occur when the cleaning fluid leaches Sn/Pb from the

surface of the solder. Properly designed materials

produce solder joints that are not attacked and render a

bright and shinny solder joint.

Fluid Jet

Impingement

Mechanical Cleaning Design

Nozzles are used in air spray systems to create jets that

carry the energy to the surface of the part to be

cleaned. The kinetic energy of the jet at the surface of

the board is determined by the nozzle type (see figure

6) driving pressure (manifold pressure), position

(distance and angle of the Jet relative to the most

distant surface to be cleaned) and the type of nozzle

used.

Conical and fan nozzles spread the spray to cover

larger areas at the expense of reducing the mass per

unit and velocity of the jet. Coherent jets hold together

longer and thus deliver more energy over a greater

distance.

Figure 6: Comparison of Fluid Jets

Coherent, Fan, and Conical Jets

All jets will break-up and slow down over distance in

air. Coherent jets hold together longer giving the

maximum energy transfer per unit area at greater

distances. Overlapping jets can be an effective

strategy for increasing surface energy density as long

as the splash at the surface does not dampen the impact

force.

The inline cleaner design used for testing utilized two

banks of overlapping Delta Fan Jets set at a 15 degree

angle followed by three banks of vertical 0.060”

Coherent Jets on 0.5” centers. The manifold was set

4.0 inches off the belt. The manifold pressure was set

to 40 lbs/ in.2.

Methodology

The objective of the designed experiment is to

understand the time in the wash section of the cleaning

machine required to clean 1210 and 1825 chip caps on

the Kyzen test card (Figure 6). Six solder pastes will be

evaluated, three eutectic and three Pb-free. The solder

pastes are industry standard materials use by many

Class 3 and Class 2 board designs. After cleaning, the

components will be removed from the test board and the

percentage of flux level under the component will be

graded. The data will be analyzed quantitatively.

Figure 6: Test Board Design

The components were reflowed using a standard eutectic

Sn/Pb solder profile for the eutectic solder pastes and a

ramp-to-spike Pb-free profile for the Pb-free solder

pastes. The component placement positions component

where the leading and trailing gap is sandwiched in

between two, chips, one chip, and no chips. This

variable will not be broke out during the data analysis.

The design matrix (Figure 7) outlines the process

variables used to run the designed experiment. The

cleaning temperature for this experiment was fixed at

150ºF. Six boards were run for each solder paste in the

matrix. The cleaning fluid time, which correlates to soak

and impingement in the wash section, represents the

variable studied in this experiment. Previous designed

experiments found that the critical variable for cleaning

under flux mounted chip correlates positively to soak

and impingement time in the wash. The times studies

were 2 minutes, 3 minutes, 4 minutes, 6 minutes, and 8

minutes, which represents soak and impingement time in

the wash section. This variable measures the importance

of soak and impingement time to achieve 100% cleaning

under all chip caps on the board. The experiment also

studied the static cleaning rate to understand the

correlation that impingement spray pressure contributes

to the cleaning process. The board was immersed in the

wash tank for 10 minutes and run through the rinse and

dry section in the inline. This data point determines the

level of cleaning with no impingement applied from the

wash section.

The test matrix fixed the cleaning fluid at 15%

concentration and wash temperature of 150ºF. Data

from previous designed experiments suggests that

higher cleaning concentration marginally improves

cleaning performance. For example, if the

concentration were elevated to 20% concentration, the

data suggests that cleaning will improve, but the

improvement will not significantly reduce wash soak

and impingement time. The temperature was set at

150ºF since previous designed experiments correlate

higher temperature with improved cleaning under flush

mounted chip caps.

The chip caps were placed on the test board in a

horizontal and vertical direction. Two sides of the chip

caps were sealed, which allow the fluid to penetrate

from either the vertical position or the horizontal

position. Coherent nozzles were used in the wash

section, which form a circular pattern upon contact.

When using coherent nozzles, we would hypothesize

that component place would have no affect on cleaning

results. There are nine 1210 and 1825 chip caps placed

in both the vertical and horizontal directions.

The research hypotheses:

H1: The time in the wash zone correlates to improve

cleaning under flush mounted chip caps.

H2: Component placement in the vertical or horizontal

position exhibits a weak correlation to improved

cleaning from vertical or horizontal direction.

Figure 8: Chip Cap 1210 - Vertical Placement

0%

20%

40%

60%

80%

100%

Static 2 minutes 3 minutes 4 minutes 6 minutes 8 minutes

Wash Soak and Impingement Time (n=9)

% Clean

Eutectic 1

Eutectic 2

Eutectic 3

Pb-Free 1

Pb-Free 2

Pb-Free 3

Figure 7: Test Matrix

Data Analysis

The data reports the mean percentage of flux residue

remaining under nine 1210 and nine 1825 chips in the

vertical and horizontal positions. The designed

experiment tests the hypothesis that longer time in the

wash section shows a strong correlation to improved

cleanliness. Additionally, the designed experiment tests

the hypothesis that part positioning on the board shows a

weak correlation to part cleanliness.

Components were removed from all test boards, and the

flux residue viewed and graded by a single individual, to

score the percentage flux residue under the chip cap.

The data reports the mean value for nine sites on both

the 1210 and 1825 chip cap in the vertical and horizontal

position for the six test conditions (Figure 7).

Figure 8 illustrates the test data for the 1210 chip cap

placed in the vertical position. The data suggests a

positive correlation for improved cleaning from

increased time in the wash section. The solder pastes

selected represent the leading eutectic and Pb-free solder

pastes used by assemblers. Previous experiments

suggest that Pb-free cleaning is more challenging to

clean but the data shows similar performance

characteristics. To achieve 100% flux residue removal,

the part requires roughly eight minutes in the wash

section. The data support hypothesis one that infers that

longer time in the wash is needed to achieve 100%

cleaning under flush mounted chip caps.

Figure 9 reports the data for the 1210 chip cap placed in

the horizontal position. The data shows a weak

correlation to improved cleaning from component

positioning. The supports hypothesis two that infers that

cleaning will performance at a similar rate regardless of

positioning when using coherent nozzles. The circular

pattern of the nozzle suggests that the directional spread

of the cleaning fluid will be equally dispersed over the

surface of circuit card. The data shows remarkable

consistency to the level of flux residue found under the

chip caps.

Figure 9: Chip Cap 1210 - Horizontal Placement

0%

20%

40%

60%

80%

100%



% Clean

Eutectic 1

Eutectic 2

Eutectic 3

Pb-Free 1

Pb-Free 2

Pb-Free 3

Figure 10: Chip Cap 1825 - Vertical Placement

0%

20%

40%

60%

80%

100%



% Clean

Eutectic 1

Eutectic 2

Eutectic 3

Pb-Free 1

Pb-Free 2

Pb-Free 3

Figure 11: Chip Cap 1825 - Horizontal Placement

0%

20%

40%

60%

80%

100%



% Clean

Eutectic 1

Eutectic 2

Eutectic 3

Pb-Free 1

Pb-Free 2

Pb-Free 3

Figures 10 and 11 illustrate the data for the 1825 chip

cap, which represents the larger chip cap on the Kyzen

test board. Similar to the 1210 chip cap, cleaning

showed a strong correlation with the length of time in

the wash section. The data suggests a slight cleaning

improvement from parts cleaned in the vertical position.

Additionally, the data suggests a higher percentage of

flux residue removed at shorter wash times but a lower

percentage of flux residue removed at the eight minute

time window. The residue remaining under the 1825

chip cap was near the solder filet, which we believe is

due to flux degassing (see Figure 5).

The 1825’s cleaned faster than the 1210’s. This could

be due the larger volume of flux deposited, however

previous testing suggest this could be due to a higher

degree of de-gassing associated with the smaller flux

volume. In small volumes of flux they loose the

solvents faster and can thus require additional softening

time to begin active cleaning.

Recommendations

The data suggests that cleaning under flush mounted

chip caps is a difficult challenge. Many assemblers use

low residue no-clean flux and only inspect for flux

residue on the exterior of the chip cap. This is not the

case for Class 3 military and medical assemblers, who

require 100% of the flux residue removed. When total

flux removal under chip caps is a requirement, the

authors recommend that assemblers specify longer wash

sections when using inline-cleaning equipment. For

batch cleaning designs, longer wash cycles address the

issue of time. Inline cleaning machines wash sections

range from 18” to 60” in length. A machine with an 18”

section requires an extremely slow belt speed to achieve

eight minutes in the wash section. To assure 100%

cleanliness, longer wash sections should be specified

from the inline cleaning Machine Company.

The data suggests that nozzle section improves cleaning

under flush mounted chip caps. This experiment tested

coherent nozzles for two reasons. The impingement

force to the board is greater with coherent jets,

Secondly; coherent jets provide a circular direction that

provides impingement from all directions on the circuit

card. Past experiments also find success with high flow

delta fan nozzles. Delta fan-jets direct the fluid on a

perpendicular direction to the board surface. This could

reduce cleaning effectiveness on parts positioned in the

vertical direction. To address this concern, higher flow

nozzles are recommended when using delta fan-jets.

Past data suggests that the flooding action onto the

board surface improves penetration under flush mounted

parts. Additionally, past data suggests that a 10-15

degree manifold angle at the entrance and exit side of

the machine improves cleaning efficacy.

When selecting an engineered cleaning fluid, multiple

functionality is needed. The cleaning fluid should

exhibit a high static cleaning rate, which correlates with

the cleaning fluids ability to rapidly dissolve flux

residue. For cleaning under flush mounted chip caps,

low surface tension and capillary action are needed to

penetrate the gap at a more rapid rate. The cleaning fluid

must not foam, as this will increase surface tension and

reduce cleaning efficacy. The cleaning fluid should not

aggressively attack and dull solder joints. Chemical

attack onto the solder joint leaches small levels of the

solder alloy from the surface of the solder.

When cleaning under flush mounted chip caps, cleaning

temperature must be considered. Many of the advanced

cleaning fluids operate at lower operating temperature,

which reduces cleaning chemistry usage. When cleaning

under flush mounted chip caps, higher processing

temperatures softens the flux residue and cause the

residue to expand. This allows the cleaning fluid to

dissolve the residue at a more rapid rate, which reduces

the time required in the wash section.

Conclusion

Removal of flux residue from under flush mounted chips

caps is a difficult cleaning challenge. The designed

experiment tested cleaning efficacy as a function of time

and directional placement for removing flux residue

from under flush mounted chip chips. The data suggests

a strong correlation to time in the wash section to

achieve 100% cleaning of multiple chip caps placed

onto a printed circuit assembly. The data also suggests

that coherent jets cleaning well and offer flexibility in

placement of chip caps onto the board surface.

An optimized cleaning process requires the right balance

of static and dynamic cleaning forces. When using

inline-cleaning equipment, the length of the wash

section improves cleaning performance. The prewash

section wets the board with the cleaning fluid by

penetrating and softening the flux residue under tight

standoffs. The wash impingement section must break the

flux dam under the component to achieve flow under the

part. The data suggest that eight minutes in the wash

section is needed to achieve a process window that

produces 100% clean parts.

Follow on Research

Follow on research on a number of process variables is

needed to understand driving forces. Further testing is

needed to study the surface tension effects of the

cleaning fluid to part cleanliness. Does lower surface

tension equate to improve cleaning under the chip cap?

A designed experiment of similar cleaning fluids that

vary in surface tension affects adds to the body of data

and knowledge currently known on surface wetting.

Additional testing is needed to correlate part cleanliness

to the static cleaning rate holding the dynamic cleaning

rate constant. There are a number of engineered cleaning

fluids on the market, which complicate the user’s ability

to select the best product for the application. What are

the factors that make one cleaning fluid better than

competing offerings? Why do these materials offer

better cleaning under flush mounted chip caps?

Additional testing of temperature affects is needed to

determine the cleaning temperatures that remove 100%

of the flux residue under flush mounted chip caps. This

experiment fixed the temperature at 150ºF. Would a

lower temperature achieve similar results? This is an

important question since lower cleaning fluid

temperature improves cost of ownership.

Following on testing of nozzle design and flow is

needed. Does a Delta fan jet at high flow better than a

coherent jet? Do nozzle angles improve cleaning under

flush mounted chip caps? There is very little published

data studying this issue.

Authors

Mike Bixenman is the CTO of Kyzen Corporation.

Kyzen is the world market leader for engineered

cleaning targeted at electronic assembly and advanced

packaging process cleaning. Mr. Bixenman has over 16

years field experience in the design and processing of

cleaning fluids within high technology manufacturing

environments.

Steve Stach is the President and CEO of Austin

American Technology. AAT manufactures inline and

batch cleaning equipment targeted at electronic

assembly precision cleaning applications.

This paper is the third in a series of papers the studied

advanced processing of cleaning fluids. For copies of

previous submissions, please email either Kyzen

([email protected]) or Austin American

Technology ([email protected])

mailto:[email protected]

mailto:[email protected]

Phase II: Optimizing Cleaning Energy in Batch and Inline Spray … · 2016-11-16 · Phase II:...

Documents

Transcript of Phase II: Optimizing Cleaning Energy in Batch and Inline Spray … · 2016-11-16 · Phase II:...