1655_05.pdf

SECTION 5

Specific Areas of Cleaning

© 2001 by CRC Press LLC

CHAPTER 5.1

Surface Cleaning, Particle Removal

Ahmed A. Busnaina

CONTENTS

IntroductionAdhesion ForcesParticle Removal

Hydrodynamic RemovalLaminarTurbulent

Ultrasonic RemovalMegasonic RemovalBrush CleaningParticle Removal Mechanism

LiftingSlidingRolling

Chemical CleaningReferences

INTRODUCTION

Surface contamination by small particles and other contaminants is a major problem inmany industries, such as semiconductor, storage, imaging, aerospace, pharmaceutical,automotive, food, and medical equipment. Contaminant particles range in size from sev-eral hundred microns to less than 0.1 �m. Surface contamination can result from particledeposition in the manufacturing environment as well as particle generation by the manu-facturing process or process tool. Improving the clean environment further or isolatingproducts can only solve part of the contamination problem. There is always a need foreffective and economical techniques for surface cleaning a variety of substrates.

Adhesion of small particles to substrates presents a serious problem to many indus-tries. Particulate surface contamination is one of the reasons for yield problems in theseindustries. The adhesion forces of these particles are greatly affected by many of theprocesses that the substrate may go through, such as cutting, polishing, etching, rinsing,


and drying, that follow the particle deposition. The adhesion forces to be considered in theprocess include van der Waals, electrostatic forces, and chemical bonds. Chemical bondsare usually orders of magnitude larger than van der Waals bonds.

As the size of the circuit line width approaches 0.1 �m in the semiconductor industry,the situation will become much more serious with respect to very small particles. There isa need for efficient and reliable particle removal techniques capable of removing very smallparticles without causing surface damage. Many studies have been conducted using vari-ous methods to detach particles from surfaces.18 –45 In this chapter the most common andwidely used particle removal (surface cleaning) techniques will be reviewed.

ADHESION FORCES

Adhesion forces are the forces responsible for adhering a particle to a surface. It isimportant to know the adhesion forces for particles that need to be removed to ensure thatthe removal force applied is sufficient for the particle removal. Adhesion forces depend onthe particles and substrate material and the medium they are in (water, air, etc.). The adhe-sion force is also a function of size; it is proportional to the radius of the particles. The rea-son smaller particles are more difficult to remove than large particles is not because theadhesion force is larger. On the contrary, it is smaller. But since the removal forces applieddepend on the area (R2) of the particles (such as removal using hydrodynamic, megasonic,etc.) and particle mass (R3) (such as removal using centrifugal, gravity, vibration, etc.) theforce that can be applied to a particle decreases much faster than the adhesion force.

Adhesion forces have been categorized by Krupp1 into three classes:

Class I: Includes van der Waals and electrostatic forces which act in the periphery ofthe adhesive area as well as in the contact area.2 –5

Class II: Includes various types of chemical bonds as well as intermediate bonds(hydrogen bonds). Chemical bonds are usually an order of magnitude strongerthan van der Waals bonds.

Class III: Includes sintering effects such as diffusion and condensation and diffusivemixing. These forces are usually known as interfacial reactions.

Class I forces, i.e., van der Waals and electrostatic forces, are the major contributors toparticle adhesion. Dry uncharged surfaces in contact with dry uncharged particles willexperience van der Waals and electrostatic double-layer forces as the only adhesion forces.Charged particles and surfaces will introduce an additional force (electrostatic imageforce). Wet surfaces, on the other hand, can shield these forces and thus reduce them sig-nificantly. The van der Waals forces can be reduced by a factor of two and the electrostaticforces can be more or less eliminated. Visser6 –8 has published several papers on particleadhesion. A detailed review of the adhesion forces and particle–surface interaction is pre-sented in other references.9 –15 Class II forces (chemical bonds and intermediate bonds) canalso occur on silicon substrates following certain conditions of treating the substrateschemically followed by rinse-and-dry processes. Adhesion forces resulting from chemicalbonds are usually orders of magnitude larger than van der Waals adhesion forces.

The van der Waals force (an intermolecular adhesion force acting between moleculesthat arises because of the polarizability of the molecules) is the dominant adhesion force forsmall particles (less than 50 �m). It arises from the short-period movement of the electronsin the atoms or molecules giving rise to momentary areas of charge concentrations calleddipoles.14


The van der Waals force (vdW) is given as

Fvdw = (1)

where A is the Hamaker constant, R is the radius of the spherical particle, and z is the sep-aration distance between the particle and the substrate. The average separation distance(z) between the two surfaces is taken as 4 Å (for smooth surfaces). For the ideal case inwhich both the spherical particle and surface are not deformed, the vdW is proportionalto the radius of sphere as shown in Equation 1. However, when deformation occurs, themagnitude of the adhesion force will also depend on the contact area between the particleand the surface. When a sphere and a flat substrate come into contact with each other,the attractive force deforms the interface and a circular adhesion area is formed. The totaladhesion force consists of two additive components, the force acting between the adher-ents before deformation, and the force acting on the contact area due to the deformation,F(vdW · deform).

12,15

Fvdw = (1 ) (2)

Krishnan and Busnaina showed that the adhesion-induced deformation is a plastic in thecase of polystyrene latex (PSL) particles on silicon and that it may take up to 70 h to reachequilibrium.12,15 Adhesion-induced deformation will occur when the particle is softer thanthe substrate, or vice versa. Electrostatic force constitutes the main force of attraction forparticles larger than 50 �m in diameter. For dry particle–substrate system the electrostaticforce becomes important. The presence of electrostatic charge can drastically alter the totaladhesion force. Zimon13 reported that the force of adhesion could be increased by a factorof two when the net number of unit charges per particle on 40 to 60 �m particles increasesfrom 700 to 2500. Most particles carry some electric charge, and some may be highlycharged. Particles at low humidity were found to retain their charge and are held to sur-faces by an attractive electrostatic force.14 A charged particle experiences an electrostaticforce in the vicinity of a charged surface or other charged particle. The charge on a particlecan be negative or positive, depending on whether the particle has an excess or deficiencyof electrons. If a charged particle carrying a charge Q comes in contact with an unchargedplate, the charged particle induces an equal and opposite charge on the surface; this isknown as the coulombic or the electrostatic image force (Fcl) and is given as:2–4

Fcl = (3)

where D is the particle diameter, and z0 is the separation distance. This electrostatic forcewill deteriorate with time because of the dissipation of the charge. When moisture is pres-ent in the air medium, condensation can take place between the particle and substrate. Thecapillary condensation gives rise to a capillary force (FC). Which is given by:2 –4

FC = 4 πr �LV (4)

where �LV is the surface tension (for the liquid–vapor interface), and r is the radius of thespherical particle. Equation 4 is applicable to smooth surfaces and represents the maximumforce that could be experienced as a result of capillary condensation. However, the force of

Q6(D+z0)

2

a2

R zA R6 z 2

A R6 z 2


Figure 1 The effect of relative humidity on the adhesion force (PSL particles/silicon).

adhesion approaches the values predicated by Equation 4 only at relative humidity near100%.16 The capillary force is made up of two components: surface tension at the perimeterof the meniscus and the capillary due to the difference in pressure between the liquid andvapor phases. The existence of tension in a liquid–gas interface causes a difference inhydrodynamic pressure across the interface if the interface is curved. The capillary forcedepends on several parameters such as particle size, the surface tension of the condensedfluid, the wettability (contact angle) of the substrate surface. Capillary forces are propor-tional to particle size and the adhesion of large particles is found to increase with the rela-tive humidity of the air.

Equation 4 applies only to perfectly smooth spheres and in a saturated atmosphere.Zimon13 found that, in air with relative humidity near 100%, the majority of particles areheld with forces less than those predicated by Equation 4. Kordecki and Orr17 observed thatcapillary condensation begins to appear at relative humidity above 50%. Luzhnov18

reported that adhesion due to capillary force occurs when the relative humidity exceeds70%. The same effect was also reported by Zimon,13 who concluded that at relative humid-ity of 50%, and particularly at humidity below 50%, capillary forces have no effect on theadhesion force. But all agreed that at relative humidity between 50% and 65%, the capillaryforce starts to have an effect on the total adhesion force. Zimon went on to conclude thatbetween 70% and 100% relative humidity, the capillary force dominates the other adhesionforces and should be the only adhesion force considered.

Busnaina and Elsawy19 showed that the effect of relative humidity on the adhesion andremoval for the 22 �m PSL particles on silicon substrates was very significant. The removalof PSL particles was very low at high and low relative humidity. The lowest adhesion force(highest removal efficiency, 49%) occurs at 45% relative humidity as shown in Figure 1.

PARTICLE REMOVAL

Hydrodynamic Removal

Hydrodynamic removal is the removal of a particle by using the inertia of a movingfluid on a particle. This can be done using a jet, overflow, or spin rinse, or using many other


applications. The hydrodynamic force on a particle is applied through the drag and the liftforce the fluid applies on any body in its path. The hydrodynamic forces depend on thecross-sectional area of the particle, fluid velocity, and density. The force is directly propor-tional to these parameters.

Drag on a sphere in a uniform flow can be obtained by utilizing available experimen-tal data.2 –4 For low Reynolds number flows Oseen’s approximation can be used. Submicronparticles exist in the boundary layer and usually in that part known as the viscous sublayer.To calculate the correct drag on the sphere, the sphere is broken into small discrete cylin-ders. The velocity across the cylinder is arrived at using boundary-layer analysis. For partsof the sphere that lie within the viscous sublayer, standard law of the wall is used. By usingthis velocity and expressions for the drag coefficient, a local drag force is found. These localdrag forces are then summed over the sphere producing a total drag force. The hydrody-namic lift force is calculated in the same manner.2 –4,20

Visser,21 in addition to his later work in theoretical aspects of adhesion, conductedexperiments in 1970 concerning particle removal. The apparatus consisted of two concen-tric cylinders, the outer one was fixed and the inner one was capable of rotating at a maxi-mum of 5000 rpm. The adhering system involved 0.21 �m carbon black particles depositedon cellulose film on the inner cylinder. Visser21 assumed the criterion of 50% removal as ameasure of the adhesion force.

Musselman and Yarbrough22 used a model of viscous drag from a high-velocity sprayto predict the drag force on particles at different spray nozzle pressures. They explain thedifficulties in hydrodynamic drag removal due to “particle hideouts” in the boundarylayer. Although free stream velocities may be substantial, the local fluid velocity at the par-ticle is small because of its proximity to the wall. Both the turbulent and laminar boundarylayers cause this problem. Drag force on the particle was calculated by a summation of thelocal drags at different heights on the sphere. Musselman and Yarbrough predicted thedrag vs. particle size at different nozzle pressures.

Kurz et al.5 used a rotating disk (silicon wafer) to generate hydrodynamic force toremove 1-�m or larger particles. They used PSL spheres on bare silicon in deionized wateras the medium. Removal rates above 90% were reported for particles larger than 2.0 �m.Taylor, Busnaina, and co-workers2 –4 measured the removal force (using hydrodynamicdrag and lift forces) of submicron particles on silicon substrates and correlated it with thetheoretical adhesion force. The results indicate that the theoretical adhesion force (using theHamaker equation) was in agreement with the experimental measurements. Most of thehydrodynamic removal was effective at removing micron-size particles or larger. The effi-ciency of submicron particle removal has been shown to be small.2 –4

There are three hydrodynamic factors acting on the adhering sphere; a Saffman20 liftforce, Stokes drag, and turbulent bursts. Saffman lift results from the gradient in the shearflow. Drag originates due to a pressure difference across the sphere. In the case of a veryslow flow around a sphere, Stokes drag provides adequate formulation. Turbulent burstsare present in turbulent flows and act to move fluid very rapidly from one section of theflow. They are influenced by vortex patterns among other things. The bursting activity isnot yet wholly understood.

Laminar

Rizk and Elghobashi provide these formulas:Stokes drag:

FD · 3 π � d V


Saffman lift:

FL = 1.615�d2 ( ) V

where V is the relative velocity between the sphere and the fluid.

Turbulent

In their analysis of particle detachment, Cleaver and Yates23 specify two types of fluidforce working on the contaminant. The first is the drag force, FD, given by

F0 = 8ρν2( )2

The other force acting against adhesion is the Saffman lift force and it is given by

FL = 10.1ρν2( )3

where d is the particle diameter and U* is the shear velocity. By defining a Reynolds num-ber based on the shear velocity as

Re* =

it can be seen that for Re* � 1 the drag force dominates the lift force. As Re* increases thelift becomes a more dominant factor.

Cleaver and Yates also contend that the turbulent bursts are a main mechanism in par-ticle removal. They present an equation for the time between bursts or burst time, tb,

tb = 75

While the authors present an in-depth discussion about the role of bursting and the param-eters involved in bursting, they do not actually propose a bursting force. Cleaver and Yatesstate that the lift force is due to the bursting activity. The impulse force generated by theburst is left undefined.

Ultrasonic Removal

Ultrasonic cleaning, where transducer frequencies operate between 25 and 200 kHz, iswidely used in many aerospace, automotive, electronic hardware, medical industries. Inreality, any frequency higher than 17 kHz is considered ultrasonic. However, because thedominant cleaning mechanism is different at low frequency (25 to 200 kHz) as comparedto frequencies higher than 360 kHz, the lower frequency cleaning is known in the industryas ultrasonic cleaning while the high frequency is known as megasonic cleaning. Theboundary between the two has not been clearly defined yet, but it lies somewhere between200 and 360 kHz.

νU*

dU*

ν

dU*

ν

dU*

ν

1 duν dx

12


Ultrasonic cleaning has been an accepted cleaning method for decades and still provesto be valuable today. A typical ultrasonic cleaning system consists of piezoelectric trans-ducers attached to the bottom of a tank. These transducers typically vibrate at a single fre-quency. The vibration energy generated by these transducers is transmitted to the cleaningsolution, creating longitudinal pressure waves. The key to ultrasonic cleaning is the phys-ical implosion of gas or vapor bubbles in a cleaning solution. The major drawback to ultra-sonic cleaning is the damage to a surface caused by cavitation. Cavitation forms when thetensile strength of a liquid is exceeded as a result rapid alternation between positive andnegative pressure of the sound wave propagating through the liquid. The cavitation bub-bles are formed when the pressure is in the negative area. There are two types of cavitiesformed: transient and stable. The transient cavities occur at low frequency, change in size,and implode, whereas stable cavities change very little in size and do not implode.36 –38 Thetransient cavities undergo a series of expansion and contraction within the fluid until theyreach a critical size. At the critical size the bubble implodes, creating a high-velocity jet andincreasing the local temperature to well above thousands of degrees Kelvin. It has been cal-culated that the velocity of the jet and the local temperature during the implosion can reachup to 130 m/s and 3000 K, respectively. These implosions remove particulate or film con-taminants on the surfaces to be cleaned. These implosions can also cause damage by creat-ing erosion.32 –34 Many studies on cavitation impact have been done and found thatcavitation damage can be minimized by changing the liquid property, such as temperatureand gas content, as well as the transducer frequency sweep. Cavitation is the dominantcleaning mechanism in ultrasonic cleaning, but it is not the only one. Acoustic streaming,which is the dominant cleaning mechanism in megasonic cleaning, is equally as importantto particle removal in ultrasonic cleaning. Figure 2* shows the effect of ultrasonic cleaningtime and temperature on the removal of submicron PSL particles using 68 kHz frequencyand 400 W of input power. The figure shows that the optimum temperature occurs at anintermediate value in the considered range used and that the cleaning time is optimumafter 15 min of cleaning.

Megasonic Removal

Megasonic cleaning, where transducer frequency typically operates between 360 and1200 kHz, is widely used in many semiconductor (wafer cleaning) and the hard disk indus-tries. Cavitation implosion does not occur in megasonic cleaning and therefore the domi-nant cleaning mechanism is the acoustic streaming. The acoustic streaming removesparticles by exerting a hydrodynamic removal force (drag and lift) but at much highervelocity near the surface as compared with typical hydrodynamic cleaning using a jet orspin rinse.

Recently, megasonic cleaning at high frequencies near 1 MHz has gained attention asan efficient, although poorly understood, Si surface cleaning technique. Olaf24 made earlyobservations of sonic cleaning of glass surfaces in the range from 15 kHz to 2.5 MHz.Rosenberg25 used ultrasonic cleaning for removing contaminant films and concluded thatthe removal was due to cavitation. McQueen26,27 recognized the importance of acousticstreaming in decreasing the boundary layer thickness, based upon his studies of removingsmall particles from surfaces. Megasonic cleaning applications were first described indetail by RCA engineers.28,29 Kashkoush and Busnaina and co-workers30 –38 studied ultra-sonic and megasonic particle removal, focusing on the effects of acoustic streaming.Removal percentage increased with power. Their results also indicated different removalefficiencies for PSL, silica (SiO2), and silicon nitride (Si3N4) particles. Greatly enhanced

* Chapter 5.1 Color Figure 2 follows page 104.© 2001 by CRC Press LLC

Figure 4 The effect of cleaning time on megasonic cleaning using 760 kHz frequency, 41.5°C, and345 W of input power.

particle removal efficiency on Si from megasonics in SC-1 and SC-2 solutions (hydrochlo-ric acid and hydrogen peroxide mixtures used in the RCA process for silicon wafers) wasreported by Syverson et al.40 Again, removal increased with increasing power, up to a max-imum tested value of 150 W. Wang and Bell41 performed experiments using megasonics forcleaning after RIE planarization. Of the parameters they tested, power had the greatestinfluence on the results. Cleaning improved with increased power up to the maximumtested value of 300 W,41 another result consistent with what was observed by Kashkoush,Busnaina, and Gale.30 –38

Megasonic power exerts a greater influence on particle removal efficiency than doessolution temperature, both in water and in SC-1 solution. Removal efficiency increaseswith increasing power up to an intermediate point above which it decreases slightly. Indeionized (DI) water, removal efficiency decreased slightly at temperatures above 50°C,whereas in SC-1 solution it was generally highest at temperatures above 50°C. AlthoughSC-1 removes particles more efficiently than DI water, particularly at lower megasonicpowers, it was still possible to achieve 100% removal in DI water under the proper condi-tions. SC-1 solutions, which are significantly more dilute in NH4OH content than the stan-dard 5:1:1 recipe, work well in the presence of megasonic energy. Particle removalefficiency decreases when the ammonia content is decreased slightly from the 5:1:1 ratio,but increases again as ammonia content is further decreased. The efficiency then remainshigh even for R as low as 0.01. Figure 3* shows the effect of cleaning time and temperatureon megasonic cleaning of polished thermal oxide wafers using silica slurry using 760 kHzfrequency and 640 W of input power. Figures 4 through 6 show the effect of input power,cleaning time, and temperature on the removal of submicron silica slurry particles usingmegasonic cleaning at 760 kHz. Figure 4 shows the effect of cleaning time using 41.5°C and345 W of input power. The figure shows that the optimum cleaning time is 20 min afterwhich the cleaning efficiency goes down. This is due to particle redeposition on thesubstrate.

Figure 5 shows the effect of input power using 41.5°C and 20 min of cleaning time. Thefigure shows that the optimum power is about 500 W. This is due to the fact that highpower generates many more bubbles that interfere with the acoustic streaming, therebydecreasing the removal efficiency. Figure 6 shows the effect of the cleaning temperature


Figure 5 The effect of input power on megasonic cleaning using 760 kHz frequency, 41.5°C, and20 min of cleaning time.

Figure 6 The effect of cleaning temperature on megasonic cleaning using 760 kHz frequency, 20min of cleaning, and 345 W of input power.

using 20 min of cleaning and 345 W of input power. The figure shows that the optimumcleaning temperature is higher than 40°C.

Brush Cleaning

Brush cleaning works by using a soft brush that engulfs a particle and removes it byapplying a torque through the brush rotation. The pressure helps the brush engulf the par-ticle and the rotation applies the torque that overcomes the adhesion moment and removesthe particle.

Brush cleaning is widely used in the industry, especially following chemical–mechan-ical polishing processes of silicon or metal substrates. There are few scientific publishedstudies on the effectiveness of brush cleaning in removing small particles adhered by vander Waals forces or chemically bonded. There has also been some recent work on cleaningoxide silicon wafers using PVA (polyvinyl acetate) brush, DI water, basic chemistry, or sur-


factants.43 Brush cleaning can be effective if applied properly by optimizing the water flow,the rotational speed, and brush pressure. Using chemistry during brush cleaning canenhance particle removal.42 Research shows that the brush pressure is one of the mostimportant parameters in removing particles.42 The pressure helps the brush engulf the par-ticles to be removed. The rotation of the brush applies the torque that will overcome theadhesion moment and remove the particles.

Complete removal using the brush scrubber with DI water is achieved in cleaning ther-mal oxide silicon wafers dipped in STI silica slurry by Busnaina et al.43 They showed thatintermediate brush pressure, speed, and time gave the best overall particle removal effi-ciency. High pressure and long cleaning time will cause scratches (more defects) in the sub-strate. Figure 7* shows the effect of cleaning time and brush speed on brush cleaning usinga PVA brush at 40 psi pressure between the brush and the substrate. The figure shows thatthe optimum cleaning time is longer than 30 s at 40 psi brush pressure.43

Particle Removal Mechanism

This section discusses the removal mechanisms of particles. Three different mecha-nisms may contribute to particle removal: lifting, sliding, or rolling. Consider deformedPSL particles on silicon substrates; the magnitude of the adhesion force is several orderslarger than removal forces.12,15,45 –47

Lifting

Particles will be removed from the surface if the lift force acting on particles is largerthan the adhesion force.

FL � Fa (5)

Sliding

Particles will also be removed by sliding if the drag force,45 lift force, and adhesionforce satisfy the following equation:

FD � �(Fa � FL) (6)

where κ is the coefficient of friction. The ratio of drag force over adhesion force, RS, isdefined to judge whether detachment by instantaneous sliding occurs or not (if it isassumed that the lift force is very small and can be neglected compared with the adhesionforce).

RS � FF

D

a (7)

If RS �, particles will be removed by sliding.

Rolling

Hubbe48 evaluated the torque balance on a spherical particle in contact with thesurface. Sharma et al.49 further included a factor of 1.399 in Equation (8) since the dragforce and the hydrodynamic torque on a particle near the wall could be substituted by an


Figure 8 The lift force FL, drag force FD, and adhesion force FA acting on a particle in a shear flow.

effective force at a distance of 1.399R from the surface. When large deformation (a/R 0.1)and the lift force are considered, the torque balance equation about point O can bedescribed as follows (as shown in Figure 8.):

(1.399R � �) FD � (Fa � FL )a (8)

where a is the contact radius, and � � R � (R2 � a 2)0.5 is the relative approach between theparticle and the substrate.

The particle will be removed instaneously when the removal forces are applied, if theremoval force overcomes adhesion force Fa. The ratio of the hydrodynamic rolling momentto the adhesion resisting moment, RM, is given by (neglecting a very small lift force):

RM �FD (1.3

F9

a

9aR � �) (9)

When RM 1, particles are removed by the drag force instantaneously.The relationship between RS and RM is given by

RRMS � 1.399

Ra (10)

Chemical Cleaning

Chemicals such as detergents, surfactants, etchers, etc. are often used to enhance phys-ical cleaning. The chemical can be used to increase the wetting, wash and dissolve organiccontaminants, or change the charge on particles to make them more repulsive to facilitatethe removal process. In this chapter only chemicals used to remove particles are discussed.

Basic chemistry is often used in particle removal from silicon and metal substrate. It isusually accompanied by megasonic cleaning or overflow rinse. Basic chemistry has beenused to remove silica, alumina, PSL, and silicon nitride particles.30 –40 The basic chemistry isused mainly to increase the repulsive charge (� potential) between the particle and the sub-strate. The � potential of the particle and the substrate has to be known to use the proper


solution pH that provides the maximum repulsion. A number of investigators have stud-ied the effects of � potential on deposition of particles onto surfaces.50 –57 Marshall andKitchener44 examined the deposition of carbon black particles from dilute aqueous sus-pensions onto glass. They observed that deposition was greatest when the � potentials ofthe particles and the substrate were of opposite sign. Hull and Kitchener50 studied deposi-tion of PSL particles onto a rotating disk. They found that when the particles and the sub-strate have opposite charge, deposition followed the expected diffusion-limited behavior.However, when the particles and substrate had like charges (repulsive interaction) therewas considerably less deposition. Ali55 measured the � potentials of a number of differentparticle types in semiconductor processing liquids with an emphasis on applications toultrafiltration of semiconductor chemicals. Albaugh and Reath56 correlated particle countsin a process bath with surface counts following deposition from the bath onto hydrophilicwafers. They demonstrated the strong influences of pH and ionic strength on deposition.

The dependence of � potential on pH plays a significant role in surface cleaning. �potential decreases as pH increases; it is typically positive at low pH, and negative at highpH. The point at which the � potential of a solid surface is zero is referred to as its isoelec-tric point or point of zero charge (pzc). The pzc of different solids depends upon the H andOH� ion concentrations in the solution, and therefore occur at different pH values(pH � �log[H ]). At high pH, the particle can release H ions into solution, resulting in anegative charge for the particle. At the pH of water, a silicon surface with a native oxide hasa negative � potential (the pzc for a hydrophilic silicon substrate is approximately 2.6).55

Thus, negatively charged particles will be repulsed from a hydrophilic substrate at this pH,and even more strongly at higher pH. In DI water, silica and PSL particles are both nega-tively charged, whereas silicon nitride (Si3N4) particles typically carry a positive charge.57

Thus, silica and PSL will be repelled from a wafer surface in water while silicon nitride willbe attracted. The reduction of � potential at high pH contributes to the success of the SC-1solution (pH � 11) as a particle removal chemistry. Figure 9 shows the � potential (particlecharge) of colloidal silica particles as a function of pH.37,38

Ionic strength of the liquid also affects electrical double layer interaction. When ionicstrength is high, the Debye length decreases and the strength and range of double layer

Figure 9 The ζ potential (particle charge) of silica particles as a function of pH.


interactions are significantly reduced.57 Thus, where repulsion between particles and sur-faces is expected, an increase in ionic strength will increase deposition. Similarly, low ionicstrength gives rise to a thicker double layer and more repulsion between the particle andthe substrate.

REFERENCES

1. Krupp, H., Adv. Colloid Interface Sci., 1, 111–140, 1967.2. Taylor, J., Measurement of Detachment Forces for Submicron Particles on a Silicon Wafer, M.S.

thesis, Clarkson University, Potsdam, NY, December 1990.3. Taylor, J., Busnaina, A.A., Kern, F.W., and Kunesh, R., in Proceedings, IES 36th, New Orleans, LA,

April 23–27, 1990, 422–426.4. Busnaina, A.A., Taylor, J., and Kashkoush, I., J. Adhesion Sci. Technol., 7, 5, 441, 1993.5. Kurz, M., Busnaina, A.A., and Kern, F.W., in Proceedings, IES 35th, Anaheim, CA, May 1–5, 1989,

340–347.6. Visser, J., Adv. Colloid Interface Sci., 2, 331–363, 1972.7. Visser, J., Adv. Colloid Interface Sci., 15, 157–169, 1981.8. Visser, J., Adv. Colloid Interface Sci., in Surface and Colloid Science, E. Matijevic, ed., Vol. 8, John

Wiley and Sons, New York, 1976.9. Hamaker, H.C., Physica, 4, 1937.

10. Tabor, D., J. Colloid Interface Sci., 58, 1977.11. DeMejo, L.P., Rimai, D.S., and Bowen, R.C., J. Adhesion Sci. Technol., 2, 331–337, 1988.12. Rimai, D.S. and Busnaina, A.A., J Particulate Sci. Technol., 13, 249, 1995.13. Zimon, A.D Adhesion of Dust and Powder, Plenum Press, New York, 1969.14. Hinds, W.C., Aerosol Technology, John Wiley & Sons, New York, 1982.15. Krishnan, S., Busnaina, A.A., Rimai, D.S. and DeMejo, D.P., The adhesion-induced deformation

and the removal of submicrometer particles, J. Adhesion Sci. Technol., 8 (11), 1357–1370, 1994.16. Davies, C.N., Aerosol Science, Academic Press, New York, 1966.17. Kordecki, M.C. and Orr, C., Jr., Arch. Environ. Health, 1, 7, 1960.18. Luzhnov, Yu., M., Research in Surface Forces, Consultants Bureau, New York, 1971.19. Busnaina, A.A. and Elsawy, T.M., The measurement of particle adhesion forces in humid and dry

environments, in Adhesion Society Proceedings 21st Annual Meeting, Feb. 22–25, Savannah, GA,1998.

20. Saffman, P.G., J. Fluid Mech., 22, 385–400, 1965.21. Visser, J., J. Colloid Interface Sci., 34, 1970.22. Musselman, R.P., and Yarbrough, T.W., J. Environ. Sci., 51–56, 1987.23. Cleaver, J. and Yates, B., J. Colloid Interface Sci., 44, 1973.24. Olaf, J. , Acustica, 7, (5), 253, 1957.25. Rosenberg, L.D., Ultrasonic News, Winter, 16, 1960.26. McQueen, D.H., Ultrasonics, 24, 273, 1986.27. McQueen, D.H., Ultrasonics, 28, 422, 1990.28. Schwartzman, S., Mayer, A., and Kern, W., RCA Rev., 46, 81, 1985.29. Mayer, A. and Schwartzman, S., J. Electr. Mater., 8, 855, 1979.30. Kashkoush, I., Busnaina, A., Kern, F., and Kunesh, R., in Particles on Surfaces 3: Detection, Adhesion,

and Removal, K. L. Mittal, Ed., Plenum Press, New York, 1991, 217–237.31. Kashkoush, I., and Busnaina, A., in Proceedings, IES 38th Annual Meeting, San Diego, CA, May

6–10, 1991, 861–867.32. Kashkoush, I. and Busnaina, A., Particulate Sci. Technol., 11, 11, 1993.33. Busnaina, A. and Kashkoush, I., Chem. Eng. Commun., 125, 47, 1993.34. Kashkoush, I. and Busnaina, A., in Proceedings, IES 40th, 1993, 356.35. Kashkoush, I., Ph.D. thesis, Clarkson University, Potsdam, NY, 1993.36. Gale, G., Busnaina, A., and Kashkoush, I., in Proceedings, Precision Cleaning ‘94, Rosemont, IL,

May 17–19, 1994, 232–253.


37. Busnaina, A.A., Kashkoush, I.I., and Gale, G.W., J. Electrochem. Soc., 142, 8, 2812–2817, 1995.38. Gale, G.W. and Busnaina, A.A., J. Particulate Sci. Technol., 1995.39. Hirota, Y., Appl. Surf. Sci., 60, 619, 1992.40. Syverson, W., Fleming, M., and Schubring, P., in Second International Symposium on Cleaning

Technology in Semiconductor Manufacturing, Electrochemical Society Proceedings, PV92-10, 1992, 10.41. Wang, P. and Bell, D., in Third International Symposium on Cleaning Technology in Semiconductor

Device Manufacturing, Electrochemical Society Proceedings, PV94-7, 1994, 132.42. Roy, S.R., Ali, I., Shinn, G., Furusawa, N., Shah, R., Peterman, S., Witt, K. and Eastman, S., J.

Electrochem. Soc., 142, 1, 216–226, 1995.43. Busnaina, A.A., Moumen, N., and Piboontum, J., Contact post-CMP cleaning of thermal oxide

wafers, in Proceedings of the VLSI Multilevel Interconnection Conference (VMIC), Santa Clara, CA,February 8–12, 1999.

44. Marshall, J.K. and Kitchener, J.A., J. Colloid Interface Sci., 22, 342, 1966.45. Zhang, F. and Busnaina, A., Particle adhesion and removal in chemical mechanical polishing

(CMP) and post-CMP cleaning, Electrochem. Solid-State Lett., (in press), 1999.46. Zhang, F. and Busnaina, A.A., The effect of particle adhesion on chemical mechanical polishing

(CMP) removal rate an post-CMP cleaning, in Adhesion Society Proceedings, 21st Annual Meeting,Panama City, FL, February 21–24, 1999.

47. Zhang, F. and Busnaina, A., The role of particle adhesion and surface deformation in chemicalmechanical polishing, Electrochem. Solid-State Lett., 1, (4), 1998.

48. Hubbe, M.A., Colloid Surf., 12, 1984.49. Sharma, M.M., Chamoun, H., Sarma, D., and Schechter, R., J. Colloid Interface Sci., 149, 1992.50. Hull, M. and Kitchener, J.A., Trans. Faraday Soc., 65, 3093, 1969.51. Clint, G.E., Clint, J.H., Corkill, J.M., and Walker, T., J. Colloid Interface Sci., 44, 121, 1973.52. Ruckenstein, E. and Prieve, D., J. Chem. Soc. Faraday II, 69, 1522, 1973.53. Prieve, D. and Ruckenstein, E., J. Colloid Interface Sci., 60, 337, 1977.54. Brouwer, W. and Zsom, R., Colloids Surf., 24, 195, 1987.55. Ali, I., Electrokinetic Characteristics of Particulate/Liquid Interfaces and Their Importance in

Contamination from Semiconductor Process Liquids, Ph.D. thesis, University of Arizona,Tucson, 1990.

56. Albaugh, K.B. and Reath, M., in Proceedings, Microcontamination 1991, San Jose, CA, October16–18, 1991, 603.

57. Riley, D.J. and Carbonell, R.G., J. Colloid Interface Sci., 158, 259, 1993.


CHAPTER 5.2

Cleaning Metals: Strategiesfor the New Millennium

Carole LeBlanc

CONTENTS

IntroductionGoalsEarly ObservationsCleaning Test PlanTest ResultsInterpreting the Results

Surface Analysis TechniquesFluorescenceFTIR SpectrometryGoniometryGravimetric AnalysisMicroscopy with Photographic CapabilitiesOSEE

Cleaning Test DatabaseCleaning SteelCleaning AluminumMore about Metal CleaningAcknowledgmentsReferences

Editor’s Note: The following study outlines the approach used by the Surface CleaningLaboratory at the University of Massachusetts Lowell to develop alternative cleaningapplications. The study is a good example of a success story outlining the benefits of devel-oping practical, industrially oriented studies in the academic community. On a more uni-versal scale, the approach and findings provide valuable and logical guidelines for theindividual manufacturer. In the following study, results and trends refer to testing per-formed by the Surface Cleaning Laboratory during the period 1994 through 1999 — B.K.


INTRODUCTION

The passage of the Toxics Use Reduction Act (TURA) in 1989 by the Massachusetts leg-islature marked the creation of the Surface Cleaning Laboratory (SCL) at the University ofMassachusetts Lowell. Fully operational since 1994, SCL is the research and testing facilityof the Toxics Use Reduction Institute (TURI). That same year, the U.S. EnvironmentalProtection Agency Toxic Release Inventory (TRI) revealed that many of the state’s indus-tries such as metalworking and electronics use several hazardous chemicals recognized fortheir excellent solvating powers.

GOALS

The objective of the laboratory is to develop and promote safer alternatives to thesehazardous materials, primarily organic and chlorinated solvents used to clean metal sur-faces (Figure 1), without causing economic hardship or a loss in cleaning performance.Specifically, reductions in ozone depletion, global warming, and volatile organic com-pound (VOC) emissions are sought as well as decreases in exposures to flammable, car-cinogenic, and other toxic substances.1 Termed environmental indicators, these factors areimportant not only to workers directly involved in surface cleaning but to the communitiesin which the processes are conducted and, ultimately, the consumer. The Institute deter-mines the potential chemical damage associated with any given cleaning solvent by usingthe state-of-the-art databases of its Technology Transfer Center (TTC).

EARLY OBSERVATIONS

Prior to beginning a project, a brief questionnaire (Figure 2) is filled out so that testingcan be tailored to meet the needs of the individual firm. Designed with templates from bothindustry and government, this format also assists plant personnel involved with cleaningto become familiar with all aspects of the process.

Figure 1 Metal-cleaning solvents assessed for replacement.

Trichloroethylene,Trichloroethane

TCE, TCA

23%

GlycolEther3%

Toluene /Heptane

6%

Other7%

Acetone2% Aerosol

5% IneffectiveDetergents

7%

Alcohols 12%

Dichloro-fluoroethane

4%Mineral Spirits

8%NMP*

4%UnspecifiedDegreaser

7%

Perchloroethylene'Perc'14%

*N-Methyl Pyrrolidone


Figure 3 presents an overview of business sectors where evaluations have been con-ducted during the past five years.

A further breakdown of surface substrates and common soils (i.e., surface contami-nants) under investigation is found in Figures 4 and 5, respectively. As indicated in Figures4 and 5, over 80% of the evaluations have involved metal cleaning applications.

CLEANING TEST PLAN

Prior to the search for safer and greener chemical cleaners, solvent cleaning was oftenperformed via single-species vapor degreasing. In addition to immersion, vapor

Figure 2 Cleaning test request form.

1. Please print or type. Be as thorough as possible.2. Attach MSDS of present relevant chemistries.3. Do not send any samples/parts without first contacting SCL.Test to be witnessed? No Yes

DESCRIBE THE PART/PRODUCT TO BE CLEANEDMaterials of construction: Metal Plastic OtherPlease specify type:List percentages cleaned (if more than one substrate):(for example, 60% of parts are aluminum; 40% are 304 stainless steel)Surface (circle two): Rough or Smooth Hard or SoftApprox. size (dimensions in inches):Geometry: Simple (e.g., flat) OR Complex (contains inaccessible areas)Gram weight: Min. Max.What is this part/product used for?

DESCRIBE THE CURRENT CLEANING PROCESSContaminants to reduce or eliminate (circle all that apply) Oil Grease Wax Flux Dirt SaltsCombination (describe): Other:Are samples of contaminants available? No Yes (if available, attach MSDS)

Manufacturing step immediately before cleaning:Manufacturing step immediately after cleaning:

Number of parts cleaned per week (or shift, etc.): per batch:Equipment in use (circle all that apply): Vapor degreaser Agitation/air sparging unitImmersion/soak/dip tank Ultrasonics Pressure spray washer (approx. psi)Other: Specify vendor if possible:

Cleaning Chemical(s): (attach MSDS) Concentration: %Time: min. Temp. deg.F Water source, if applicable: DI (deionized)/Tap

Rinse Cycle, if any: Time: min Temp.: deg.F Water source: DI (deionized)/TapDrying Cycle, if any: Method Time: min. Temp: deg.FAny problems with present cleaning system?

After cleaning, parts are (circle one): Used Immediately OR StoredIf stored, How: How long:What is the purpose of cleaning (i.e., desired product specifications)?Methods employed for evaluating cleanliness: None Visual Microscopic UVOther performance test, if any (please describe):Comments or Areas of Concern:Return any samples/parts? No Yes


degreasers were employed because of the superior ability of traditional solvents to dissolveorganic matter. In vapor-phase cleaning, the final “rinsing” of the surface was accom-plished by the mere condensation of solvent vapors.

Unlike these chlorinated organic solvents, environmentally friendlier aqueous (i.e.,water-based) detergents may not depend on their penetrability for their cleaning efficiencies,especially for the removal of petroleum-based surface debris. They rely instead on a numberof chemical processes such as solubilization, wetting, emulsification, deflocculation, seques-tration, and saponification.2 To complete these tasks, aqueous cleaners can be complex mix-tures of surfactants, emulsifiers, and other additives in an alkaline (pH � 7.0) formula.

The basic outline of a cleaning test plan is shown is Figure 6. Because of their decreasein chemical energy relative to chlorinated organic solvents, several aspects of an aqueouscleaning system must be optimized. These include but are not limited to agitationmethods, cleaner concentration, temperature, and cycle time as reflected in Phase II ofFigure 6.

Figure 3 Evaluation by business sector, 1994–1999.

Paper1%

Medical7%Recycling

3%

General Mfg18%

Plastics2%

Electronics9% Aircraft

4% Consulting2%

Adhesive12%

State/Military5%

Optical3%

Metalworking34%

Figure 4 Substrate categories tested, 1994–1999.

Non-Steel Metals29%

Plastics Related4%

Electronics Related 4%

Glass3%

Liquid*1%

Tex-tile1%

Other1%

*Analysis only

Steel57%


As this Figure 6 illustrates, metal surfaces initially cleaned are flat coupons (Phase I)matched to the parts’ materials of construction, followed by the eventual cleaning of actualproduct (Phase IV). Phase IV is concerned with the significance of part shape (screw con-figurations and blind holes and so forth) on metal cleaning.

Note that in this model, mechanical energies (Phase III) are critically examined onlyafter the proper cleaners, concentrations, and temperatures (Phase II) are ascertained.Other testing protocols exist that may be as effective; it is essential that the experimentaldesign be logical and consistent.

Figure 5 Surface contaminants studied vs. number of cleaning tests performed, 1994–1999.

Adh

esiv

es +

Asp

halt

Buf

fing

Com

poun

d

Car

bon

Bla

ck

Ure

than

e C

oat

Mac

hini

ng F

luid

Dir

t

Oil

Gre

ase

Met

als

Dus

t

Fin

gerp

rints

Lubr

ican

ts

Rus

t

Flu

x

Sea

lant

Wax Ink

Rus

t Pre

vent

er

Stic

kies

Pre

clea

n

* N

one

200

180

160

140

120

100

80

60

40

20

0

Num

ber

of T

ests

*Refers to applications related to surface cleaning

Figure 6 Phases of an aqueous-based surface cleaning test.

I. Brainstorm Compatibility and "Lift" Studies

IV. Actual Product Cleaning Studies III. Mechanical Energy Studies

V. Pilot Plant / Scale-up Feasibility Studies

Geometries and sizes of parts important to cleaningefficiency

Production volume or throughout dictated by

II. Temperature and Concentration StudiesChemical field may be narrowed / changed from Phase IHelps to 'scope' project more efficiently

Determine substrate surface / chemical cleanerreactivity issues (use MSDSs, Technical dataSheets, etc.)Monitor the effect of drops of selected detergentconcentrates on grossly contaminated couponsover time (Hansen method if no coupons available*)

ambient conditions; chemical ( no mechanical) energysubjective; visual

scientific study, may employ a variety of analytical tools forcleanliness evaluation

Follow chemical manufacturers' recommendationsfor both parametersEqualize timeMinimize same source agitation*

*chemical comparison tool;first use of mechanical energy; first round of scientific trials:gravimetric analysis

Duplicate optimal Phase III cleaning conditions

Number of chemical cleaner candidates further decreasesfrom Phase II

MANUFACTURING PROCESSDRYING

Duplicate optimal Phase III cleanliness testing

Application-specificEconomically-sensitiveTraining-dependentSpace-limiting


TEST RESULTS

The increased dependence on mechanical energy of replacement cleaners during labo-ratory testing is depicted in Figure 7. On average, in the author’s experience, three tests areconducted per company, each test involving experimentation with as many as six chemicalcleaners, over a period of 2 to 8 weeks.3

While Phase V (see Figure 6) deals with metal cleaning as it relates to the entire manu-facturing process and drying as the rate-limiting cycle, it is equally important to considerrinsing requirements in the selection of chemical cleaners and equipment.4 Figure 8 out-lines laboratory test results for both rinsing and drying regimes.

During the course of an investigation, companies receive the latest edition of“Industrial Cleaning Survey: Directory of Vendors,” which lists the suppliers of cleaningchemicals, equipment, and related items worldwide. Vendors are responsible for the infor-mation contained in the catalog, which is maintained on file by responding to theInstitute’s Vendor Survey Questionnaires. This publication prepares staff for any time-sen-sitive purchases that may be required due to process changes. Example entries are foundin Table 1.

Figure 7 Cleaning mechanisms used with alternative processes.

Ultrasonics28%

Spray Wash2%

Manual15%

Lift 'Peel' Test*4% Immersion

2%

Other2%

TankAgitation

41%

*Analysis only

Air Sparging4%

Extracting / Blasting2%

Figure 8 Rinsing and drying methods.

RINSING DRYING

Tap &Rinse Aid

3%

Tap & DIWater39%

Tap Water54%

DI Water4%

Oven

16%

Laminaire

Hood

27%

Infrared

7%Heat Gun

6%

Ambient Air

5%

ChemicalDisplacement 1%

Air Knife

38%


Table 1 Typical Entries in Cleaning Vendor Directory,Technical Report 15

Product(Address,Phone, Industrial Contaminant Physical KnownContact) Classification Applications Removal Compatibility Properties Dimensions Components Options Cost

Chemical Alkaline, Metal Buffing, High/low pH 9.0 N/A 50% NaOH N/A Per cleaner x water-based finishing/ lapping pressure VOCs 0 3% Surfactant lb

degreaser fabrication compounds spray FP N/AVP N/A

Cleaning Manual parts Automotive/ Greases/ All aqueous/ 4� (h) � Oil Perequipment z washer machine shop lubricants most semi- 3� (w) � skimmer unit

aqueous 4� (l) 150 lb


Possible funding sources for the project and the potential appointment of a Universityof Massachusetts graduate student intern to the in-house, scale-up portion (Figure 6, PhaseV) of the project are also discussed with individual companies. In addition to laboratorypublications, periodicals, and referrals to conference proceedings, other technology trans-fer assistance is provided through cleaning and degreasing workshops held on campusand in off-campus training sessions.

INTERPRETING THE RESULTS

Surface Analysis Techniques

If the purpose of solvent substitution testing is to identify a safer, greener cleaningprocess that is at least as proficient as its solvent-based counterpart, then benchmarkingbecomes necessary. Benchmarking is achieved by conducting scientific comparisons of var-iously treated surfaces via acceptable analytical methods. Depending on the application,ASTM (American Society for Testing and Materials) and other standards such as militaryspecifications may also be employed. Figure 9 details the surface inspection tools used bySCL and a brief description of the laboratory’s major techniques follows.

Fluorescence

Some contaminants, in particular lubricants, naturally fluoresce. Examination underblack light reveals the location and extent of this type of surface contamination. Artificialfluorescence is possible with the addition of chemical tags, similar to those used in foren-sics. This is a limited application, however.

FTIR Spectrometry

Fourier transform infrared spectroscopy correlates vibrational energy to the molecularsignature of a compound. Similar to other high-tech methods such as GC (gas

Figure 9 Types of surface analyses conducted vs. number of cleaning trials performed, 1994–1999.

200

150

100

50

0

Bla

ck L

ight

Cha

ract

eris

tic T

ests

Con

tact

Ang

le G

onio

met

ery

Fla

me

Test

FT

IR

Gra

vim

etric

Pee

l Tes

t

Vis

ual 1

/94-

5/95

Wat

er B

reak

20

86145

OS

EE

1/94-5/95

6/95-11/97

Mic

rosc

opy

46


chromatography), the curves generated in this analytical technique are both quantitativefor species identification (the placement of the curve on the electromagnetic spectrum) andqualitative for amounts (the area under the curve). A relatively expensive instrument, anFTIR spectrometer requires special training and care in sample preparation. Not all con-taminants can be analyzed this way and interpretation of graphs can be difficult because ofthe presence of interfering peaks. It may be used in clean rooms or disk drive manufacturewhere the origins of contamination may be entirely unknown and the amounts of contam-ination very low.

Goniometry

Like optically stimulated electron emission (OSEE) described below, laser or opticalcontact angle goniometry is the measurement of a secondary effect to extrapolate surfacecleanliness. A small drop of deionized (DI) water is placed on the substrate of interest. Alight is shown to reflect the interface of the droplet with the surface. Usually, the higher thecontact angle (that is, the height of the bubble), the greater the contamination. Conversely,water dropped on a clean surface generates a much smaller, flatter contact angle. An exam-ple of this effect is noticeable after waxing and then washing a car; the remaining wax actsas a contaminant and the residual water on the surface of the car “bubbles up.” The tech-nique is limited since only the cleanliness under the tiny drop is measured so that severalreadings must be taken. Flat surfaces are more conducive to accuracy.

Gravimetric Analysis

Properly employed, gravimetric analysis can be the most inexpensive and revealing ofall surface measurement techniques. Ideally, the part or test coupon is weighed a total ofthree times with the same analytical balance and under the same atmospheric conditions.Weights are taken (1) before artificial contamination, (2) after artificial contamination, and(3) after cleaning. These tests should be duplicated a number of times to ensure repro-ducibility of results. Percent soil removal and standard deviations can then be calculated.Some difficulty may arise in arriving at a precontamination weight under actual plant/pro-duction settings, although estimates may be possible. Care must be taken in selecting near-identical substrate pieces and applying the contaminant in a consistent manner. Theseproblems are largely avoided using test coupons.

Microscopy with Photographic Capabilities

From SEM (scanning electron microscopy) with magnifications as high as several thou-sand that provide actual surface morphology to light microscopy with magnifications aslow as decimal fractions, there is a magnification range to suit almost every surface clean-liness application. Parts cleaning, as opposed to precision cleaning, can be adjudicated witha stereoscope and magnifications well under 1000�. Computer software packages areavailable that “count” the soil load per photographic frame and store the information todisk for a permanent record.

OSEE

Optically stimulated electron emission or photo electron emission (PEE) is based on theprinciple that metals and certain surfaces emit electrons upon illumination with ultravio-let (UV) light. These electrons can be collected, measured as current, converted to a


voltage, and digitally displayed. A surface contaminant will either enhance or attenuate thissignal, depending on it own photoemissive nature. While OSEE will not identify a contam-inant, it is a good comparative tool to determine the degree of contamination. The methodis best suited for thin films (oils, etc.) and not particulate matter (dust, for example).5

Figure 9 confirms a trend toward more scientific surface evaluation. As the number oftests for which the evaluation laboratory was able to conduct gravimetric analysisincreased, visual inspection as the sole means to measure surface cleanliness decreasedproportionately.

Other similar, subjective inspection methods, for example, wipe glove and water breaktests, are considered unacceptable by SCL for most modern metal-cleaning applications.

Cleaning Test Database

To arrive at meaningful data in a more timely fashion, the laboratory developed a data-base of trial outcomes, based on 5 years of predominantly metals (86%) cleaning, search-able by four fields: surface substrate, surface contaminant, chemical cleaner, and cleaningequipment.

To demonstrate the use of this program, known as the Effective Test Conditions (ETC)Database, the information in Table 2 was obtained as a result of the inquiry, “What recom-mendations can be made to a manufacturer of opto-mechanical devices to clean lensesdealing with stain sensitivity?”6

CLEANING STEEL

Steel is defined as “an iron-based alloy, malleable in some temperature ranges as ini-tially cast, containing manganese, usually carbon, and often other alloying elements.”7 Incarbon steel and low-alloy steel, the maximum carbon is approximately 2%; in the high-alloy version, this concentration is about 2.5%. Carbon steel has no minimum quantity forany alloying elements other than manganese, silicon, and copper. It contains only incre-mental amounts of substances other than those mentioned and sulfur and phosphorus. Thedifferentiating line between low- and high-alloy steels is generally 5% metallic alloys.

Steels may be separated from cast irons, the large family of cast ferrous metals (con-taining at least 2% carbon, plus silicon and sulfur with or without other alloying elements)and low-carbon pure iron. In very low-carbon steels, the manganese content is the primarydifference. Steel usually contains at least 0.25% manganese and ingot iron much less. Cold-rolled sheets are milled from a hot-rolled, pickled coil that has been given substantial cold

Table 2 Partial List of Successful Glass-Cleaning Tests Conducted by SCL

SCL Number Substrate Soil Mechanism Cleaner Mfg95-409-01-2 Glass Wax Ultrasonics Oakite95-409-02-2 Glass Wax Immersion Alconox95-409-03-2 Glass Wax Ultrasonics Alconox96-435-01-8 Glass Rosin Ummersion Occidental97-550-01-3 Glass Grease Spray Alconox


reduction at room temperature. This results in a product requiring further processing butwith improved characteristics and uniformity.

Of the many different kinds of steel, the majority of tests conducted by the laboratorywith this substrate were categorized as stainless steel (in a few cases, companies wereunaware of the steel classification of parts from suppliers). These steels contain 12 to 30%chromium as the alloying element and usually exhibit passivity in aqueous settings.

Some trends in conducting effective cleaning trials on steel substrates in the laboratoryare revealed in Table 3.

In two of the above case studies, several potential replacement cleaners performed pro-ficiently under almost identical operating conditions. This suggests that cleaning practi-tioners should source equipment and chemicals separately. A chemical cleaner proffered bya cleaning equipment vendor may reflect an economic partnership, rather than the optimalselection for a particular application.

CLEANING ALUMINUM

The silvery-white, ductile metallic element aluminum is used to form many hard, lightalloys. Nevertheless, fewer tests were performed by SCL on this metal than on steel (seeFigure 4). Cleaning aluminum substrates can be challenging because of its proclivity toetch. Etching occurs when some of the metal is dissolved, along with the contaminant, as aconsequence of cleaning. Properly conducted gravimetric analysis can prove most usefulduring cleaning trials on aluminum surfaces since cleaned weights may be less than origi-nal (i.e., precleaned) weights. Table 4 illustrates this effect.

The assumption is made that no cleaner is capable of removing more than 100.00% ofa soil. In this application, all three soils are present on the surface of aluminum parts. Acleaner must be found to remove the contaminants without damaging the substrate. In thefirst trial, Cleaners B, C, and D were relatively successful on the viscous and difficult-to-remove lubricant mix. In subsequent testing on the easier-to-remove vanishing oil anddrawing compound, however, Cleaners A, B, and C pose moderate (light-gray-shadedarea) to substantial (dark-gray-shaded areas) etching risks. This is especially true if not allcleaning-cycle durations and temperatures can be accurately monitored at all times.Cleaner D is the best selection since it removed the vast majority of the three contaminantswithout etching in any of the test replicates.

MORE ABOUT METAL CLEANING

Cleaning of other metallic substrates, including brass, bronze, copper, gold, molybde-num, nickel and nickel alloys, silver, tin, and titanium, was tested in the search for safer,greener chemical solvents. Some of those results are generalized in Table 5.

In closing, tests conducted at the laboratory involve all aspects of the cleaning process.They confirm that the future of metal cleaning depends upon establishing industry stan-dards or a ranking system for the energy and water efficiency of related equipment as wellas a more complete understanding of the environmental and health consequences of newlydeveloped chemical cleaners.8


Table 3 Successful Cleaning, Rinsing and Drying Stages with Analyses (Representative Steel Samples)

Cleaner Types and Cleaning Methods, Rinsing Methods, Drying Methods,Steel Part and Concentrations Temperatures, Temperatures, Temperatures, SurfaceContaminants (vol %) and Cycle Times and Cycle Times and Cycle Times Analyses

*316 Stainless 5% aqueous to Ultrasonics (25 kHz) Tap Water at 140°F Convection oven Visualsteel heat neutral pH at 140°F for or DI water at at 100°F or IRexchangers 10 min ambient temp heat lamp for

Fingerprints and for 2 min each 30 min eachlight oils

Stainless steel 4–10% several Ultrasonics (40 kHz) Tap Water at Convection oven Photo-pump seals aqueous to or spray wash 130–150°F or at 140–145°F micrography

Coolant and alkaline pH at 110–150°F DI water at for 30–60 minmetal fines for 5–15 min ambient temp. or air knife at

for 2–5 min each ambient temp.for 2 min

Carbon steel 100% terpene/ Immersion/soak at None Air Dry at ambient Customerjet engine parts semiaqueous ambient temp. for temp. overnight performance

Rust preventative and hydrocarbon 2 min testand quenching oil

**1010 and 1020 3–5% several Ultrasonics (40 kHz) Tap water with/ IR heat lamp for FTIRcold rolled steel aqueous to or tank agitation without rust 1 min

Lubricant and alkaline pH at 100–130°F for prohibitor atmetal chips 2–5 min 120° for 0.5 min

Unspecified steel 5% aqueous to Immersion/agitation Tap water at 120°F Air dry at ambient Customerrachet handles alkaline pH at 150°F for 5 min for 0.5 min temp. overnight performance

Dirt, metal fines, testgrease, buffing compound

*Alloy designation (refers to steel’s state of composition, annealing, hardness, etc.).

**Customer-reported grade designation.


ACKNOWLEDGMENTS

The author wishes to express gratitude to Jason Marshall of the Toxics Use ReductionInstitute’s Surface Cleaning Laboratory for his vital contributions to graphics, databases,and the bench chemical testing that form the cornerstones of this chapter.

Table 5 Examples, Evaluation of Other Metal-Cleaning Applications

Surface Present Recommended Metal Part Contaminants Chemical Cleaner Chemical Cleaner

Copper tubes Machining oil TCE Alkaline aqueousNickel engine parts Oil, grease, wax Acetone SemiaqueousInconel turbine blades Penetrating oil None (new system) Alkaline aqueous

Table 4 Gravimetrically Based Contaminant Removal RatesReveal Potential for Surface Damage Due toChemical Etching

Trial I Percent Soil Removal, Lubricant Mix

Aqueous Cleaner A B C D

Coupon #1 87.15 96.04 99.26 98.60

Coupon #2 91.99 99.77 93.17 99.58

Coupon #3 78.57 100.00 94.97 99.25

Average 85.90 98.60 95.80 99.14

Std deviation 6.80 2.22 3.13 0.50

Trial II Percent Soil Removal, Vanishing Oil


Coupon #1 100.32 100.77 101.26 99.79

Coupon #2 100.54 100.05 101.24 99.68

Coupon #3 100.76 99.72 99.70 99.16

Average 100.54 100.18 100.73 99.54

Std deviation 0.22 0.54 0.89 0.33

Trial III Percent Soil Removal, Drawing Compound


Coupon #1 100.76 99.58 99.36 99.62

Coupon #2 100.57 100.00 99.18 99.72

Coupon #3 100.50 99.17 99.77 99.89

Average 100.61 99.58 99.44 99.74

Std deviation 0.13 0.41 0.31 0.14


REFERENCES

Note: Additional details may be obtained by contacting the laboratory Web site,http://scp.rti.org/lab.htm.

1. LeBlanc, C., The Toxics Use Reduction Act of 1989: Lessons Learned, in Annual ConferenceProceedings: Precision Cleaning, Cincinnati, OH, April 1997.

2. Grace Metal Working Fluids, Aqueous Cleaning Handbook, Lexington, MA, 1995.3. Marshall, J., Toxics Use Reduction Institute, Surface Cleaning Laboratory Notebook, University of

Massachusetts, Lowell, 1998.4. McLaughlin, C. and Zisman, A., The Aqueous Cleaning Handbook, Morris-Lee, New Jersey, 1998.5. Green Seal, Degreasing Agents: Proposed Standard for Aberdeen Proving Ground, MD,

Washington, D.C., 1999.6. Toxics Use Reduction Institute, Environmental Strategies for the Next Millennium: Laboratory

Services Dedicated to Surface Treatment, Cleaning and Analysis to Ensure Product Excellence,University of Massachusetts, Lowell, June 1999.

7. Boyer, H. and Gail, T., Metals Handbook, American Society for Metals, 1985.8. Kanegsberg, B. and LeBlanc, C., The cost of process conversion, in Annual Conference Proceedings:

CleanTech, Chicago, IL, May 1999.


http://scp.rti.org/lab.htm

CHAPTER 5.3

Very High Performance,Complex Applications

Barbara Kanegsberg

CONTENTS

IntroductionExamples of High-Precision Cleaning

Aerospace and Related IndustriesHigh-Precision Navigation SystemsOptics

Removal of Blocking AgentsAdditional Issues and Suggestions

Cleaning for Biomedical ApplicationsCleaning Issues, Biomedical ApplicationsCase Studies

Cleaning Electronics AssembliesProduct DesignSoils (Fluxes et al.)Build ProcessCleaning Process

Issues and ConclusionsReferences

INTRODUCTION

Defining high precision cleaning is difficult. Precision cleaning has been defined ascleaning of products of perceived high value, submicron level particulate removal, orcleaning products where the results of improper cleaning could be catastrophic. Precisioncleaning has also been described as cleaning something that did not look particularly soiledin the first place.1 On the other hand, those involved in processing optics could argue thatoptics may be embedded in heavy pitch and wax, which then have to be completelyremoved.


At one point, this author explained that everyone knew the difference between highprecision cleaning and ordinary cleaning. High precision cleaning was simple to define:that would be my manufacturing process. Ordinary cleaning, on the other hand, encom-passes everyone else’s manufacturing process.2 The author went on to explain that under-standing as much as possible about the manufacturing process in question, and thencollaborating with others who were also faced with stringent cleaning requirements, is cru-cial to advancing the science and art of contamination control.

Over perhaps a quarter of a century, aerospace applications were highly dependent onsolvents, notably the ozone-depleting chemicals (ODCs), CFC 113 (commonly referred toby the trade name Freon) and 1,1,1-trichloroethane (TCA). In the ODC era, given the timerequired to validate new processes, many high-precision, high-value processes changedonly to develop more and more steps. Over the years, if there were problems at a particu-lar stage that could conceivably be traced to contamination, the engineer in charge mightrecommend additional cleaning with CFC 113, TCA, or isopropyl alcohol (IPA), with someIPA and perhaps acetone for drying.

Over the past decade, increasingly stringent environmental regulations coveringvolatile organic compounds (VOCs), air toxics, and ODCs have driven many processchanges. The process of change has produced a number of false starts and inefficientresults. On the bright side, where process development has been approached with an eyeto better performance rather than simply coping with the regulation of the moment, moreefficient processes with fewer steps and lower usage of cleaning agents have been adopted.

EXAMPLES OF HIGH-PRECISION CLEANING

In this chapter, three examples will be discussed: aerospace applications, biomedicalapplications, and cleaning of electronics assemblies. Certainly, examples of high-precisioncleaning abound and are discussed in other chapters. For example, cleaning of motion pic-ture film, another critical cleaning application with unresolved issues, is discussed in theoverview of cleaning equipment (Chapter 2.1).

Aerospace and Related Industries

Considering issues of product reliability, military requirements, competition, andcostly testing that may take years to complete, aerospace has been understandably conser-vative in adopting new processes. At the same time, aerospace has been fearlessly inven-tive in evaluating and eventually choosing from among the range of new solvents,mixtures, and cleaning techniques. Cleaning techniques adopted include supercritical CO2,aqueous cleaning, cosolvent systems, proprietary blends, and acetone and other low-flash-point cleaning systems. In LOX systems (systems that will be exposed to liquid oxygen),the issue has been to find cleaning agents that pass stringent tests to prevent fires or explo-sions during use. For LOX cleaning, a range of cleaning agents may be allowed for initialcleaning, then, at the final stage, a level of defined but noninterfering residue is tolerated.

The following example is one in which a particular application was modified first byadopting a cosolvent system (solvent cleaning followed by rinsing in another solvent), andthen by evaluating a solvent. The study is presented to show the logic of the approach usedin a given situation, not to indicate that this is the desirable solution for all aerospace-related applications.


High-Precision Navigation Systems*

This example is based on experiences at Litton Guidance and Control Systems Divisionin modifying processes for a beryllium-based instrument. As with many such applications,process modification was required for environmental/regulatory considerations, notablythe phaseout of ODCs. Because new cleaning agents and processes have been developed,process modification has been multistep and is ongoing.

Navigation systems consist of gyroscopes and/or accelerometers along with the sur-rounding electronics. While cleaning problems associated with electronics may seemdaunting, cleaning is a much more complex and diverse problem for high-accuracy instru-ments than it is for typical electronics assemblies.3 In electronics assembly, there are rela-tively limited materials of construction and configurations, and there are widely acceptedindustry standards.

Definition and control of residue, residue of soil and of the cleaning agent itself, aremajor issues. The rosin mildly activated (RMA) flux was known to produce an inertresidue; very small amounts of this residue, should they occur, would not be acceptable.While low-solids fluxes have been implemented in some hand-cleaning operations in otherhigh-precision builds, process control and reliability testing were judged impractical forthis application. One reason is that in some cases, the sealed system is in an atmosphere ofa polyhalogenated flotation fluid; as such, there must be no residue of cleaning agent. Useof water is a controversial issue in this and in some other high-precision applications. Manyof the components are water sensitive and difficult to dry, so aqueous cleaning is unac-ceptable at many stages.

Even under clean room conditions, building precise instrumentation inherently gener-ates testing, rebuild, and, doing so, generates an array of soils including:

• Greases• Oils• RMA flux• Water-soluble flux• Fingerprints• Particles• Polyhalogenated flotation fluids

In assembling classic gyroscopes and accelerometers, one finds a much larger range ofmaterials of construction than are found in the typical electronics assembly. For inertialnavigation systems, an array of materials of construction is used in instrument build. Anentire list of materials of construction would require half a dozen pages. Some representa-tive materials of construction are indicated in Table 1, along with potential cleaning andcontamination problems. It is important to note that for Litton’s applications, it is notenough to consider each individual material separately. The entire assembly or subassem-bly must be evaluated to avoid potential issues of cleaning agent residue, galvanic interac-tion, outgassing, and product deformation. For this reason, the feedback of experiencedinstrument assemblers was crucial.

The other issue is determining how “clean is clean,” or “how clean is clean enough.” Whenthe product may be expected to perform continuously and reliably for a quarter of a century,addressing these ultimately unanswerable questions becomes increasingly important.

* This subsection is based on a paper presented at CleanTech ‘99, a conference sponsored by the CleaningTechnology Group, Witter Publishing Corp., May 1999, Rosemont, IL.3


Table 1 Representative Materials of Construction, Gyroscopes, Process Concerns

Materials of Construction Primary Concerns

1. Beryllium, aluminum Oxidation, erosion, cleaning agent2. Stainless steel, other ferrous metals Corrosion erosion, cleaning agent3. Complex assemblies containing 1 and/or Same as 1 and 2, plus galvanic interactions

2, and combinations of magnesium,gold, tungsten carbide, copper,Hy-MU-80 (nickel alloy)

4. Sapphire, specialized glasses Cleaning agent residue, subtle surface changes5. Kapton, plastic coatings, coated and Solid cleaning agent residue, outgassing of

uncoated epoxies vapors, softening, deformation, solubilization ofmaterial of construction

6. Flotation fluid Becomes a soil under test, rework conditions;residue undesirable, especially if reacts withcleaning agent (nucleophilic substitutionreaction, SN2, with alcohols)

7. Complex assemblies, many materials of Same as 1–5, plus entrapment of cleaning agent;construction (1–5), complex soils reactivity of cleaning agents with nonmetals

and residue of soils; situation exacerbated insealed systems

Whereas in the electronics world there are established cleaning standards, for precisioncleaning the goals are often pragmatic, based on expected use of the product; a combina-tion of analytical testing of residue with performance is often used. The goal is to minimizecontamination and residue to its lowest level. Those actually using inertial navigation sys-tems, particularly aircraft pilots, have expressed enthusiastic support for aggressive con-tamination control.

Process modification was multistep and involved a team approach by in-house engi-neering and production staff as well as ongoing collaboration with manufacturers of clean-ing agents and cleaning equipment. Initially, an array of cleaning agents and cleaningsequences were evaluated in-house. Some more-promising cleaning sequences includingvarious hydrocarbon and d-limonene (orange terpene) blends followed by self-rinsing orrinsing in IPA or perfluorinated materials, were then tested using the facilities and person-nel of a major cleaning agent manufacturer, in cooperation with a number of other clean-ing agent manufacturers. Results were evaluated visually and by analytical techniques,Fourier transform infrared spectroscopy (FTIR) and electron spectroscopy chemical analy-sis (ESCA). Some in-house residual gas analysis (RGA) was also incorporated in the study.Surprisingly, given the stringent process requirements, at least half a dozen promisingcleaning sequences were found. The most promising cleaning sequences were then testedagainst the most exacting critics of all: experienced production assemblers.

Based on assembler input, the cleaning sequences were refined and implemented. Theapproach adopted was a cosolvent system consisting of initial cleaning with a hydrocarbonblend containing various alcohols followed by two to three rinses with IPA. This new processallowed elimination of TCA cleaning; some perfluorinated material continues to be used asa final rinse to assure thorough removal of fluorolube. Overall, the number of process stepswas reduced; often an 18-step process was reduced to four to six steps. Even the processintroduced initially required refinement, because IPA was found to react with beryllium peri-odically; intermittent residues were found. Eventually, IPA was replaced with volatilemethyl siloxanes (VMS). At the time, the new processes represented the best option toreplace ODCs.


The cosolvent processes, while allowing replacement of TCA, were far from optimal.The subassemblies are very complex, with close tolerances and blind holes. While thecleaning agent can be removed with careful process control, extreme and constant vigilanceis required to assure that no cleaning agent residue is left. The hydrocarbon blends arecostly, and some of the operators found the odor to be disagreeable. The search for a morereliable process continued.

After extensive testing, a specific n-propyl bromide (nPB) formulation was adopted toreplace some of the multistep cleaning. The VG formulation has been implemented forcleaning during instrument build where cleaning activity similar to TCA is required. Thecleaning agent was chosen for a number of reasons, including solubility parameters, lowresidue, liquid- and vapor-phase cleaning characteristics, and compatibility with materialsof construction.

Operations engineers conducted the primary evaluation of the VG cleaning agent. Thematerials and processes group conducted some preliminary evaluations and found it to bevery effective for removal of a wide range of soils, including flux, without the requirementfor rinsing with another cleaning agent.

While nPB was adopted because of its aggressive solvency, which is similar to 1,1,1-trichloroethane, there is always a balance between solvency and compatibility. Berylliummetal, used in some of Litton’s inertial navigation systems because of its relatively highratio of strength to weight, is also a relatively reactive metal, forming beryllium oxide. Toput it into perspective with more commonly used metals, beryllium can be thought of asrather temperamental aluminum. Beryllium coupons supplied by Litton were tested by thecleaning agent manufacturer to determine the impact of exposure to the VG formula. After24 h of exposure at the boiling point, beryllium submerged in the solvent showed no signof discoloration or tarnish.5 It should be emphasized that the results of these studies shouldbe used to indicate the approach to be used in evaluation, not to recommend a specificprocess or product. Having indicated the results of these studies, it should be noted that,with reactive metals and aggressive cleaning agents, each formulation and each applica-tion should be considered as a separate case.

Given the aggressive nature of nPB, collaboration with the manufacturer is crucial. Inthe case of Litton, the industrial hygienist provided training, site testing, and recommen-dations for appropriate engineering controls and personnel protection to minimize expo-sure via inhalation or skin adsorption. At Litton, a combination of carbon tubes andcolorimetric indicator tubes was used. Employee exposure by inhalation of nPB in all theapplications measured was below 10 parts per million (ppm). Area samples, used in somecases to measure the concentration of nPB at specific locations in the workroom to predict“worst-case” potential exposures, were also below 10 ppm. The observed exposuredepends on the type of cleaning, cleaning action, ventillation, and worker education andawareness. In this case, activities included ultrasonic cleaning, spray and flushing systems.As with compatibility, worker exposure is very application and site specific and should beevaluated on a case-by-case basis.

Because of changing regulations and given the ongoing development of new productsand cleaning methods, process modification and process improvement are often bestachieved in a multistep manner. This has certainly been the case at Litton. The interimprocess eliminated Class I ODCs but was cumbersome to use. Adding a more aggressivecleaning agent with a higher evaporation rate along with reevaluating the necessity of var-ious cleaning and rinsing steps has significantly simplified the process. In addition tocleaning capability and reliability, the new process has provided improved processing timeand lower cleaning agent usage. Litton’s processing time has been reduced by over 40%and cleaning agent usage has been reduced to essentially one third (Tables 2 and 3).

Process modification would not have been successful without the input, cooperation,


Table 2 Summary: Litton Process Modification

Interim Process Simplified, Modified Process

Hydrocarbon blend (two used, depending VG, self-rinseon soil to be removed)

Volatile methyl siloxane rinse Perfluorinated B, particulate removalPerfluorinated APerfluorinated B, particulate removal

Table 3 Summary: Benefits, Process Modification

Cleaning Cleaning Solvent UsageTypical Process Time (min) Steps (gal/week)

Interim, with 35 4 45hydrocarbon blends

Replacement, with VG 20 3 15

and collaboration of the production assemblers. For many high-precision processes, theend product is produced on a very small scale and may be based on hundreds of assemblyand cleaning steps. Automated cleaning processes may be impractical. In such cases, buildand cleaning processes are highly specialized, involving the input of skilled and experi-enced assemblers.4 Particularly in situations with repeated hand-cleaning processesinvolving skill and judgment, it is important to involve those who will work with the alter-native cleaning processes on a day-to-day basis. Often, the assemblers themselves are ableto detect potential problems that are not picked up even with sophisticated analytical test-ing. For example, in flux removal, a pilot project was undertaken in which individuals per-forming overhaul and rework and hand-cleaning operations evaluated severalhydrocarbon blends and orange terpenes. While initial laboratory-scale evaluation indi-cated that all the products provided for pilot test cleaned equivalently, the techniciansreported subtle differences. Some said they could see differences in cleanliness, and sawproblems with certain of the cleaning agents. More-detailed, costly testing including sur-face analysis and outgassing confirmed that, indeed, certain cleaning agents were leavinga previously undetected residue. Attempting to conduct all possible analytical testing ini-tially would have been time-consuming, costly, and probably unproductive. In general,where the production people have been involved at various stages in new process devel-opment, a much more robust process has resulted.

Optics

Optics covers a range of applications from eyeglasses and contact lenses to an array ofspecialized sensors and components of sophisticated devices. The substrate is sometimes aplastic, but is more often than not a specialized glass. This section deals primarily withhigh-precision glass optics. Some of the considerations apply to plastics. With plastics, onemust also factor in issues of compatibility with some of the more aggressive solvents.

Processing of optics is shrouded in mystery. Competition-sensitive issues and militaryconcerns result in secrecy, vagueness, and lack of communication. While many of theseconcerns are no doubt justified, the result has been an array of complex, Byzantineprocesses that are difficult to control and troubleshoot. Because of the many fabricationsteps involved, processes are often performed for traditional reasons; it is often impractical


to take the time and effort to justify a particular operation. We know things have to be veryclean, and we often perform complex laboratory analysis. In the end, it comes down tomaintaining the process so that the next fabrication step can be successfully completed andthat the final device operates acceptably. Optics processing inherently develops into atightly woven sequence of processes. A supposedly minor change to one process, perhapsmandated by some regulatory requirement, may impact other parts of fabrication such thatextensive modification and reevaluation are needed.

In processing optics, the substrate is first machined and grossly shaped or sliced.Various surfaces are then polished with specific slurries, and the surfaces may be etchedand/or acid-treated. Finally, specific coatings are applied, often by vacuum deposition. Atvarious stages, cleaning has to occur in such a manner as to remove the soils without dam-aging the surface, which has just been carefully polished, treated, or coated. Cleaning mustbe accomplished without redepositing soil, adding residue of cleaning agent, depositingparticles, or changing the surface in some undesirable manner.

With optics and with other sensitive materials it should be remembered that waterquality and composition can impact surface properties. Contaminants can be deposited onand react with the surface. In addition, ultrapure deionized (DI) H2O can leach materialsout of surfaces and alter the surface.

A vast, often fanciful assortment of materials, which will be referred to as blockingcompounds, is used to hold the substrate in place for polishing, or to protect certain por-tions of the optical system from treatment. Blocking compounds may include such materi-als as:

• Pitch (asphaltum)• Soft wax• Beeswax• Rosins• Nail polish• Epoxies• Thermoplastics• Mixtures of plastics, rosin, and waxes of variable composition

The materials are chosen for a variety of reasons, often relating to the forces needed for pol-ishing. In some cases, optics may be set into a 3 mm or more thick base of an organic-basedblocking compound on a larger base plate. This mosaic-like object is then polished in a spe-cific slurry. At the end, the blocking compound and polishing compound must be removed.Classically, TCA was used for many optics operations. The optics could be soaked in TCA,either at ambient or at elevated temperature; the blocking compound would dissolve, leav-ing a light residue. Final cleaning would then be accomplished by vapor-phase cleaning.With the advent of the ODC problem and the TCA phaseout, a number of other chemicalsand processes have been used with varying success. Issues involve:

• Inadequate solvency• Cleaning agent residue• Scratching of the substrate• Etching of the substrate

Some blocking agents currently in use are summarized in Table 4. Some details follow. In addi-tion, it might be pointed out that, in general, particularly for rinsing and final cleaning ofoptics, there is no subsitute for high-quality, low-particulate cleaning agents. Where possible,


Table 4 Summary: Some Considerations, Examples of Newer Deblocking Compounds

General Performance Regulatory/SafetyDeblocking Agent Characteristics Concerns Issues

Chlorinated solvents • Liquid- and vapor- • High boiling point • Engineeringphase cleaning (PCE) may be controls required for

• Self-rinsing result in personnel• Aggressive solvency components protection• Rapid process damage • Air toxics

• Must be tested • Solventwith individual containment,application to reportingassure no surface requiredchanges

nPB • Liquid- and vapor- • Must be tested • VOCphase cleaning with individual • Low ODP

• Similar performance application to • Engineeringto TCA assure no surface controls required

changes for personnelprotection

• Regulatory statusmay change,pending evaluation

d-Limonene (orange • Good solvency for • Must be tested • VOCterpene) many blocking with individual • Distinct odor can

compounds application to be an issue• Liquid-phase assure no surface

cleaning changes• Some formulations • Leaves significant

can be rinsed with residue for mostwater high-end

applications• Residue from

additives mayinterfere withsubsequentapplications

• Rinsing withsolvent or waterrequired

• Some blends canoxidize

Esters (e.g., ethyl • Good solvency for • Must be rinsed • VOClactate, di-basic many blocking • Distinct odor canesters) alone or compounds be an issuein blends • Liquid-phase cleaning

• Formulations availablefor aqueous rinsing

N-Methylpyrrolidone • Moderate to good • Must be rinsed • VOCsolvency for many • Often somewhatblocking compounds longer processing

• Liquid-phase cleaning time


Table 4 Summary: Some Considerations, Examples of Newer Deblocking Compounds (Cont’d)

General Performance Regulatory/SafetyDeblocking Agent Characteristics Concerns Issues

• Often less prone toleaving residue thand-limonene or esterblends

Proprietary blends, • Moderate to good • Require rinsing • May containwater-soluble solvency for many significant amountsorganics blocking compounds of VOCs

• Appropriate mixedwaste streamhandling needed

Acetone • Good to aggressive • Exceedingly rapid • VOC exemptsolvency evaporation • Not an air toxic

• Can be used as • Very low flash pointheated liquid vapor • If heated, must bephase (with proper used in speciallyequipment) designed

• Useful in combination equipmentwith other deblockingagents, includingaqueous

Alcohols (IPA, • Good to aggressive • Solvency often • VOCsmethyl alcohol) solvency limited for • Low flash point

• Cold cleaning blocking agents • If heated, must be• Can be used as of interest used in specially

heated liquid vapor designedphase (with proper equipmentequipment)

• Useful in combinationwith other deblockingagents, includingaqueous

Hot water • Low solvency, acts by • Limited solvency • Disposal of wastemelting • May require streams

• Useful with soft wax multiple rinses• Rinse agent • Must control

water quality

Cold shock • Nonchemical • May damagesubstrate

Aqueous/surfactant • Moderate to good • May require • Disposal of wastesolvency for many multiple rinses streamsblocking compounds • Additives may • Additive packages

• Liquid-phase cleaning produce subtle may havesurface changes, significant VOCs ornot immediately require specialevident handling


electronics grade, HPLC grade, or the equivalent should be used. Certification of desiredproperties by the supplier and some assurance that the formulation will not be changedwithout notification are crucial.

Removal of Blocking Agents

Chlorinated, brominated solvents: Perchloroethylene (PCE) and other chlorinated andbrominated solvents have been tested, with varying success. PCE has an inherent limita-tion in that the higher boiling point may produce bubbling of the blocking compound. Thebubbling, boiling action may jar the optics, effectively undoing the previous polishingprocess. nPB has also been used effectively in some applications, and it has the advantageof allowing for final cleaning in the vapor phase. Many of the chlorinated solvents are cov-ered by a federal NESHAP (National Emissions Standard for Hazardous Air Pollutants), sothat careful record keeping and reporting are required. As of the time of writing, nPB isacceptable for use pending evaluation by the U.S. EPA SNAP group. With all aggressivesolvents, the changing regulatory picture and, in some areas, local restrictions on smog-producers (VOCs, or volatile organic compounds) or air toxics, may require that alterna-tive processes be developed.

Acetone, alcohols, low flash-point blends: Acetone, IPA and methyl alcohol are widely usedin fabrication of optics. Acetone has a certain appeal, particularly in areas of poor air qual-ity, in that is VOC exempt. In addition, many low-flash-point solvents are of low cost andcan be obtained in high quality. The authors cannot stress enough that low-flash-point sol-vents can be used safely and with confidence in appropriately designed equipment. There havebeen semidisastrous attempts to heat low-flash-point solvents or to adapt existing equip-ment in-house. Even though equipment for low-flash-point solvents is costly, the decreasein solvent usage and in disposal costs may result in a short payback period.

Hot water or cold shock: Either approach depends on physical rather than chemicalremoval of the blocking agent. In the case of hot water, one is simply melting the wax orrosin. While heat is important in many cleaning processes both to boost solvency and topromote melting, one must be aware of the inherent limitations of heat in the absence ofsolvency. Cold shock, either by freezing the optics or by using a CO2 snow gun, has hadlimited success in releasing the optics via changes in thermal expansion or by physicallycracking the blocking material. More often than not, however, particularly in productionsituations, the process has resulted in damaged substrate.

Aqueous/surfactant: Water-based cleaning is becoming increasingly popular because ofsafety and environmental concerns. Adopting water-based cleaning agents often requiresa change in the blocking agent. Heavy pitches and very hard waxes are often difficult if notimpossible to remove with water-based cleaning agents. In addition, aqueous/surfactantblends require increased cleaning action such as mechanical agitation or spray, ultrasonics,additional heat, and careful rinsing. Where agitation is used, it is important to protect thesubstrate. Rather than struggle through evaluation of a succession of aqueous cleaningagents and processes, it is often more productive to change the blocking agent to one morereadily removed with water-based cleaners.

Effect of additives: Glass is soft. In modifying older, solvent-based processes, one mightbe aware that additives both in aqueous cleaning agents and water-soluble organics havethe potential to produce subtle surface changes. These changes may not be immediatelyapparent but may manifest as some other problem many steps later in the process.

Additional Issues and Suggestions

Deblocking is a common problem in optics. There are additional issues that might beconsidered.


Cleaning vs. surface modification: Many optics preparation processes are old and steepedin the rich tradition of the company. (Translation: no one knows why they work, but if oneattempts to change anything, the process fails, and the person who made the change isblamed.) Part of the problem relates to subtle changes in the surface of the substrate. Thismeans that cleaning problems may become entangled with problems of surface prepara-tion. For example, optics are often cleaned in mixtures of very strong acids and salts. In oneinstance, in an attempt to reduce usage of a chromic acid mixture, solvents were tested forcleaning. The solvents were judged unacceptable on the grounds of a change in contactangle measurement. One might suspect, however, that the acid was modifying the surfaceand so changing the contact angle. The problem was not one of cleaning but probablyrather one of surface modification.

Troubleshooting: Optics processing is notorious for intermittent, inexplicable problems.With so many steps in the fabrication process, precise control of every factor in every oper-ation is difficult. If one is in the oh, so fortunate position of being put in charge of trou-bleshooting an optics fabrication process, it is often difficult to determine the source of theproblem. One must become a detective. Suppose that a coating process is suddenly nolonger successful. One would, of course, look at the coating operation itself and factors inthe immediate vicinity of the suspect problem including:

• Analytical testing to determine the nature of the contaminant (coated surface,surface prior to coating)

• Coating equipment operation• Personnel changes or dissatisfaction• Chemical changes• Water purity, if applicable• Clean room conditions• Surface preparation• Storage conditions prior to coating

If careful examination of these factors does not yield the source of the problem, oneneeds to venture farther afield, considering factors mentioned above as they apply to ear-lier operations, as well as:

• Earlier cleaning, polishing, and etching steps• Changes in any chemical• Changes in any process• Equipment condition• Recent equipment repair or overhaul• New lot or new source of substrate• Modification in initial slicing or polishing operations

Above all, see if the investigation can be turned into a team effort, involving all of thepeople who might be influenced or who might be the source of the problem. It can be a realchallenge to keep communication open and maintain a no-fault atmosphere, even withinthe facility. Of course, it is always possible that someone is purposely breaking clean roomdiscipline (a term the author finds to be unproductive and unprofessional). It is more likelythat someone or several people have made changes that they honestly felt were the equiv-alent to the status quo, or that they just were not aware of. It can be a real challenge to per-suade people to communicate. Many people who become involved in high-technology,


clean room operations like optics fabrication do so because they would prefer to workquietly, out of the public eye. Some new idea may have been quietly implemented which,while wonderful in isolation, does not quite mesh with the whole operation. The importantthing is to keep the atmosphere open, productive, and as free as possible from finger-point-ing. This is a real art, and it is an art the author is still developing. Side benefits includemore-up-to-date, well-documented processes in that production workers may confidewhat they are actually doing. They may also have some wonderful ideas for improvement.Quality circles and teams may seem a generation out of date, but honest communicationand respect for the ideas of others can really pay off.

Teaming ideally involves the vendor community as well. If subcontractors areinvolved, it is necessary to find out as tactfully as possible whether or not they have madechanges. The same holds for cleaning agent and cleaning equipment vendors and for anyoutsourced items such as maintenance of the clean room or of the water system.

Maintaining and improving the process: With optics fabrication, as with all multistep,high-precision processes, maintaining the process is crucial. This involves:

• Using the highest quality (chemicals, equipment, water, disposables, substrates)• Documentation of chemical quality through on-site testing and/or vendor certi-

fication• Control of disposables, storage conditions• Maintaining records of product and chemical lots• Documentation of water quality• Clean room control• General thorough record keeping• Employee education• Valuing the workers, listening to their ideas

In essence: invest in quality, trust nothing, strive for consistency, but do not become atyrant. Someday, we’ll all get this right!

Cleaning for Biomedical Applications*

Consider the phrases high-precision cleaning, high-value cleaning, or critical cleaning.The applications that immediately come to mind include wafer fabrication, computer sub-assemblies, precision optics, microelectronics, and high-accuracy inertial navigation sys-tems. The biomedical field brings an entirely new dimension to the concept of criticalcleaning applications, but it is an area that has not typically been discussed.

Cleaning and manufacturing for biomedical applications, in the context of this discus-sion, include

• Implants (metals and plastics) for long-term use in humans• Catheters and other devices for use in animal experimentation• Surgical instrumentation, which may include long, and extremely fine-bore

tubing

* This subsection is based on a paper presented at CleanTech ‘97, a conference sponsored by the CleaningTechnology Group, Witter Publishing Corp., April 1997, Cincinnati, OH.6


• Assorted plastics and metal disposable and nondisposable materials• Subassemblies for clinical instrumentation, including automated equipment for

clinical laboratory testing (blood tests, etc.)

The biomedical community is justifiably concerned about quality processes, but eachgroup tends to work in isolation and secrecy that surpasses that of many military applica-tions. There is an overwhelming reluctance to discuss either successes or difficulties in apublic forum.

A number of factors promote secrecy. It is difficult and costly to bring a new product tothe marketplace, or to change the manufacturing process for an existing product. The man-ufacturer faces not just the hurdles of coping with strict and often conflicting local andnational air quality, water quality, and employee safety regulations, but there are also theissues of biocompatibility, pyrogens, bacterial growth, and other harmful residues. In addi-tion, with biomedical applications there may be environmental regulatory concerns as wellas the challenge of working with the Food and Drug Administration (FDA).

Gaining FDA approval is often a major hurdle. Some companies reportedly fear thatemerging FDA guidelines will render their complex, costly testing inadequate. Therefore,even groups doing careful, thoughtful testing may be reluctant to present findings lest theycall attention to themselves by the FDA or by any other regulatory agency.

The FDA does not regulate or specify cleaning per se. However, the FDA does regulatethe effect of the manufacturing process on the finished device. This would then be part ofthe 510K Premarket Notification. One FDA spokesperson notes that while the 510K formatis standard, because there must be tens of thousands of different devices, across-the-boardtesting requirements would be difficult to define.

Because companies are fearful of incurring increased inspection by either the competi-tion or by government agencies, they keep both their successes and failures to themselves.In fact, some individuals declined to discuss issues involved in manufacturing biomedicalinstrumentation, even anonymously, on the grounds that their phraseology would be rec-ognized. The author would like to thank those colleagues, clients, and associates through-out the United States who shared their concerns and success stories. It should be noted thatin the following examples and case studies some names of companies and of individualshave been pointedly omitted to assure candor and to avoid competition-sensitive issues.

Cleaning Issues, Biomedical Applications

Many in the biomedical community developed cleaning processes using ODCs orother classic solvents. As in the inertial navigation world and other high-value-addedapplications, they knew the product was clean based on years of operational experience.With new processes, the issue becomes more complex.

Some issues of importance to the biomedical community in modifying their cleaningprocess are indicated in Table 5. Many of these factors are also important to other criticalcleaning applications. Manufacturers in the biomedical field have been justifiably focused onbiocompatibility. The Association for the Advancement of Medical Instrumentation (AAMI)has produced guidelines for reprocessing devices.7 Some groups also adapt the approachfor use in manufacturing of components. Given the range of new cleaning manufacturingprocesses being implemented, there may also be a need to address issues related to processcontrol and to impact of the process on quality and consistency of the finished device.Additional cleaning issues are discussed in Chapter 5.5 by Albert.


Table 5 Examples, Potential Contamination Problems, Biomedical Applications

Potential Contamination Potential Problems Possible Solutions

Soils (greases, oils, • Toxicity, biocompatibility • More aggressive cleaningpolishing compounds) • Increase in bioburden agent

• Blockages • Enhance cleaning action• Mechanical malfunctions (turbulation, ultrasonics)

• Monitor soil loadingCleaning agent residue • Toxicity • Avoid cleaning agents with

from solvating agent or • Variable changes in bioburden significant nonvolatileaqueous/saponifier • Lot-to-lot variability residue

• Mechanical malfunctions • Provide more rinsing thanis typically used

• Monitor the process toavoid excess carryover ofcleaning agent

Solvent or rinse agent • Variable residue on component • Choose a solvent notresidue • Outgassing leading to toxicity; readily adsorbed by

damage to components in materials of constructionsealed devices • Provide adequate drying

• Materials compatibility issuesMicrobial contamination • Unsuitability for use in sterile • Choose a more aggressive

situations solvent• Mechanical breakdown of • More aggressive cleaning,

clinical equipment longer cleaning• Clean at a higher

temperature

Case Studies

One company, which produces a variety of plastic and metal components for biomed-ical applications, switched from TCA to PCE in a contained solvent system from Pero. Theselection process involved many months of comparison studies and testing includingbioburden testing. Determination of bioburden involves contaminating the product withmicroorganisms, cleaning, then testing to determine the amount of remaining microorgan-isms relative to the control or reference cleaning method.

For this particular group, given the configuration and spacing of components, entrap-ment of solvent is an issue. While cleaning in IPA or with IPA azeotropes has been used formany biomedical applications, in this case, low-flash-point solvents were not removed rap-idly or completely enough to allow adoption of a flammable liquid/vapor-phase system.Even using high-temperature cleaning with PCE, specially designed fixtures were crucialto assuring adequate cleaning and drying. While this success story utilizes a system fromPero, it should be noted that other contained solvent systems (e.g., from Serec, BaronBlakeslee, Branson, Durr, Hyperflo, Unique, and Tiyoda) are being evaluated and/oradopted.

Aqueous cleaning has always been a popular option, particularly where compatibilityissues make solvent cleaning unwise. Aqueous cleaning can be successfully adopted wherecareful attention is paid to process design.8 The force of action of water may be enough toclean some soils of concern. Where companies have been accustomed to using aqueouscleaning, extending the use of aqueous cleaning has been a relatively simple, successful


process. One company eventually successfully implemented both solvent cleaning toreplace TCA and aqueous cleaning to replace CFC 113. The group found that, initially, itwas more difficult to ramp up the new aqueous process because blind holes in some of thedisposable plastics components made drying a limiting part of the process. Their dryingtechnique had to be improved.

Another group makes sealed implantables for relatively long-term application inhumans. The device must of necessity be extremely compact, so all components are veryclosely spaced. They use surface-mount electronics, and have been replacing CFC 113 forcleaning rosin flux. They also have an array of bonding problems, including bonding ofplastics to metals. If there is any flaw in surface preparation, subsequent bonding may notbe successful. Perhaps even more potentially distressing, any flaw in the surface can resultin subsequent adhesive failure in a warm, saline environment. One of the engineers notesthat, “like everyone else, they tried orange terpene alone, but were not successful.”

This maker of implantables has adopted cold cleaning with IPA cleaning, and aqueouscleaning prior to coating as an interim measure. IPA alone is not efficient for removing mostrosin fluxes. The process is admittedly inefficient and labor-intensive. Their cleanlinessstandards are

• Ionograph measurements• Coating adhesion at a subsequent step (i.e., the parylene sticks)

The group would like to adopt aqueous cleaning for environmental reasons. Theyexpect that some solvent usage may be required. However, they are concerned about thecapital outlay and chemical-handling issues associated with use of low-flash-point sol-vents.

In another example, a new process provided performance superior to the existing ones.A manufacturer of titanium bone replacement implants and ultrahigh-molecular-weightpolyethylene (UHMWPE) cartilage replacement implants wished to replace cleaningprocesses utilizing HCFC-141b and trichloroethylene. They found that a stabilized nPB for-mulation met the company cleaning requirements, including:

• Acceptable removal of buffing compound• Low concentration of retained solvent in the UHMWPE• Reduction of bacterial spore count in the UHMWPE by 50% or greater

Minimum solvent outgassing is required. Retained solvent was measured by gas chro-matography/mass spectrometry. Results for nPB were compared with HCFC 141b andtrichloroethylene. After 24 h of drying at ambient temperature with good airflow, 30 ppmof nPB were detected. By comparison, the level of retained solvent was over five times ashigh for HCFC 141b and over 15 times as high for PCE. After 96 h of drying, the level ofretained nPB decreased to 3.7 ppm. Measurements made at 106 h were 4.4 ppm for HCFC141b and 27 ppm for PCE. Bioburden studies indicated spore count reduction of 68 and73% after cleaning with nPB. Details of these studies have been reported.

Yet another group has had good success in using HCF XM, an HFC blend from DuPont,to clean plastic and metal surgical devices. HFC has low surface tension and providesincreased wettability along with compatibility with a wide range of materials. The blendwas chosen because it has a somewhat higher polarity and therefore a greater solvencyrange than the HFC alone. The engineer in charge of the project notes that one of themajor challenges is that cleanliness is not defined for the medical industry; the FDA hasno information as to how clean is clean. The process had to be developed and justified to


company management based on what would be logical standards considering the end-useapplication. In this case, residual particulate material of greater than 5 µm is consideredcrucial. HFCs (as well as HFEs and many VMS) are particularly effective in removingparticles.

Cleaning Electronics Assemblies*

Cleaning of electronics assemblies is a relatively well studied area that has been thetopic of a number of books,8,11 trade associations (e.g., IPC), and conferences (e.g., IPC,Nepcon). Because of the number of studies defining materials of construction andrecommending cleaning agents and cleaning equipment, some consider electronics clean-ing to be a fait accompli, needing little or no additional study. However, as designs change,environmental restrictions increase, and performance and economic demands become everhigher, many manufacturers continue to be faced with issues of cleaning and contamina-tion control. Some groups have chosen to outsource the entire design and assemblyprocess. However, the basic issues of design, manufacture, and cleaning remain.

In electronics assembly, industry has successfully varied a number of parameters toimprove manufacturing capabilities. The approach could be used in other industries. It ispossible to modify:

• Product design/materials of construction• Soils• Build process• Cleaning process

Products can be modified for easier assembly. Unfortunately, many design engineers,often with good reason, take the approach of automatically attempting to pack 10 lb of stuffin a 5-lb bag. There may be good reasons, such as some technical requirement for a com-pact or miniaturized product. However, the more closely spaced are the components, themore difficult the product is to clean.

Product Design

The effect of component density on ability of a particular cleaning system to workeffectively was illustrated by W. Machotka, C Knapp, and B. Kanegsberg in a study atLitton Industries in which leadless, 1-in. square component simulators were placed at vary-ing standoff distances.10 That is, components were spaced at 0.003, 0.005, 0.008, and 0.015in. from the surface of the board. The component simulators were dipped in RMA flux.Parts were charred by heating immediately before cleaning to simulate soldering. Theywere then cleaned in actual production batch and in-line cleaning systems using a varietyof cleaning agents. Efficiency of removal of RMA flux was estimated gravimetrically..Cleaning equipment and cleaning agents are identified only by code. This experiment wasnot completely balanced in that the cleaning agents tested in a particular piece of equip-ment depended on what was being used at the time in a particular production situation. Inthat sense, the options tested were considered acceptable, but all available options had not

* Significant portions of this section have been adapted from “Successful Cleaning/Assembly Processes for Smallto Medium Electronics Manufacturers,” a tutorial presented by the author at Nepcon West ‘98, Anaheim, CA,March 3, 1998.10


Table 6 Flux (%) Removed Relative to Standoff

Equipment Cleaning Agent 0.003 in. 0.005 in. 0.008 in. 0.015 in.

In-line a 1,1,1-Trichloroethaneazeotrope 98 100 100 100

In-line b HCFC 88 100 100 100In-line c d-Limonene A, semiaqueous 79 100 100 100Batch a d-Limonene A, semiaqueous 69 100 100 100In-line d Hydrocarbon blend,

semiaqueous 56 100 100 100In-line e Aqueous/saponifier A 79 93 100 100Batch b Aqueous/saponifier A 65 96 94 100In-line f Aqueous saponifier B 71 100 100 100In-line g

(older model) Aqueous/saponifier B 26 38 63 73

Notes: Five samples were run per test, Batch systems were manually operated; in-line systems wereautomated.

been tested by the production facility.As indicated in Table 6, below a 5 mil (0.005 in.) standoff, it is difficult to clean with less

aggressive solvents, semiaqueous, and aqueous methods. In addition, automated systemstend to be more effective in soil removal. Results provide additional evidence that systemdesign can influence cleaning; the very old cleaning system was not nearly as effectiveas the newer designs. This author has empirically observed that 5 mil tends to be theborderline level between relatively straightforward and relatively difficult cleaning appli-cations for all manner of assemblies, not just electronics.

Properly motivated design engineers, on being presented with such cleaning informa-tion, might modify the assembly to allow easier fluxing and defluxing. It is typically help-ful to involve the design engineers in process modification plans, keeping them on anyteams or at least keeping them up-to-date on proposed changes. Conversely, thoseinvolved in assembly process development would do well to look at the next generation ofassemblies to determine if they can be cleaned using the methods under consideration.

Soils (Fluxes et al.)

In electronics assembly, solder flux is the primary soil. Flux is, in a sense, a clean-ing/surface modification agent in that it facilitates soldering by preventing buildup ofoxides. However, once used it must itself be cleaned. Many manufacturers have made sig-nificant strides in manufacturing by modifying the soil, in this case, the flux. Choices in fluxinclude:

• Rosin based• Water soluble• Low solids, the so-called no-clean fluxes

All have their advantages and difficulties.Rosin-based flux is formed of pine tree sap with additives, including activators, some

of which are acids. RMA flux is classically used in military and other high-end applications.Rosin-based fluxes have had their problems. As naturally occurring materials, they canshow wide variations in soil from lot to lot, because, for one thing, groves of pine trees vary


in composition of the sap. Flux residue can be a problem, particularly if the flux cures aftersoldering and prior to cleaning. This most often happens with significant delay prior to thecleaning process. In addition, many of the polar additives are difficult to remove; they mayleave complex mixtures known as the infamous white residue. Because flux is a mixture ofpolar and nonpolar components, cleaning agents with a wide solvency spectrum are pre-ferred, or azeotropes, such as IPA/cyclohexane azeotrope, have proved effective.

Flux residue of any sort is not considered desirable. Residues can form crystals (dendriticgrowth) that can impair product function or even result in product failure. Dendritic growthis a sort of microscopic stalactite or stalagmite that appears on assemblies after cleaning,often after aging in a humid atmosphere. Acid or salt residue can be corrosive, damagingmaterials of construction. While manufacturers claim that they want no detectable residue,it has become widely accepted that rosin fluxes may leave a residue that is less damagingto the product than are some other types. Rosin flux, when heated in the soldering opera-tion and cured due to delayed cleaning or further heating with inadequate cleaning, mayleave trace residues under components. However, these residues are often hard and jewel-like, and they appear to be nonreactive. One would suspect that such flux residues form asort of artificial amber. Amber is petrified tree sap, and, in fact, several jewelers have qui-etly confided that realistic, artificial amber can be produced by heat-curing rosin flux.

Other fluxes are widely used because of pressures to avoid organic solvent cleaning orto avoid cleaning all together. Organic acid (OA) fluxes or water-washable fluxes can bevery effective and can obviate the need for solvent cleaning. They are meant to be cleanedwith water-based cleaning agents, not with solvents.

While some organic acid fluxes are synthetic, others are based on lemon juice or applejuice. This has caused some confusion among manufacturers who confuse lemon-basedflux with d-limonene (orange-peel-derived) cleaner. Changes in the soldering procedureare needed, and care must be taken in the cleaning process to avoid residues that can pro-duce dendritic growth or other interfering or corrosive residue.

Low-solids or no-clean fluxes are designed with minimal solids and minimal residue toavoid the need for cleaning altogether. Control of the overall assembly and soldering processis typically more exacting with no-clean fluxes. An inert atmosphere may be required, andthere is generally a much narrower process window. Because of the high level of process con-trol, the most successful initial implementation of low-solids fluxes occurred in very largescale manufacturing facilities where ongoing training and process control could occur.

Because solder joints may not be as aesthetically pleasing with no-clean processes, it isoften necessary to modify in-house or customer requirements from visual requirements tofunctional requirements. In addition, there are applications where even the small amountof residue left by low-solids fluxes is unacceptable. Because of this, examples of cleaningthe no-clean flux are increasingly seen. Sometimes, only a very dilute aqueous-based clean-ing agent is needed, or water without additives may be acceptable. It should be pointed outthat some low-residue fluxes are more amenable to being cleaned with water than are oth-ers. Some no-clean flux residue can result in product degradation in a humid atmosphere,a distinct disadvantage if the product is to be used in New Jersey or Texas in July. It is thepreference of this author to use a cleanable no-clean flux where possible.

While most manufacturers think of flux residue as the primary contaminant, other con-taminants can cause problems. Often, contamination is introduced by a components sup-plier and may not be detected until some problem arises after final assembly. Otherresidues include oils, sulfur-based compounds from machining fluids, and metal orceramic particles. Determining residue of flux and other contaminants can involve a hostof visual, microscopic, and analytical testing such as surface insulation resistivity and ionchromatography.


Build Process

It should be noted that it is also possible to modify the build process. Alternative sol-ders and soldering techniques are being developed. Epoxies may be used, or laser ablationmay replace soldering.

Cleaning Process

The type of cleaning process depends on multiple factors including the type of flux, thedesign of the assembly, expected product end use, customer requirements, worker prefer-ences, cleaning agent/cleaning equipment costs, and the local mix of safety and environ-mental regulations.

It should also be remembered that, for many high-end applications, electronics assem-bly is often much more than defluxing. An assembly may contain mixed OA and rosinfluxes, machining, oils, and lapping compounds. Subvendors may change the process,resulting in changes in soil residue and in cleaning agent residue.

An array of cleaning processes has been used with electronics assemblies including no-clean, water, aqueous/saponifier, water with large amounts of organic additives includingnonlinear alcohols and unidentified proprietary additives, semiaqueous, cosolvent, classicchlorinated solvents, brominated solvents, engineered solvents alone and as blends, andflammable solvents.

For every successful application of a given flux and cleaning process, an unsuccessfulone could be cited. For all the standardization, cooperative testing, and understanding ofmaterials compatibility, cleaning electronics assemblies is a very site-specific effort. Itshould also be remembered that eliminating electrostatic discharge (ESD) is important notonly to avoid assembly failure but also to achieve contamination control.

With no-clean fluxes the process window is not as large as it is with rosin or organicacid fluxes. Where no-clean flux was sucessful in a hand-soldering operation, the manu-facturing engineer noted a number of critical factors that basically add up to understand-ing what is being soldered:

• Condition and materials compatibility of the components and bare boards• Cleanliness of alloys and base materials (essential for adequate wetting)• End use of the product• Signal-to-noise requirements• Component and design requirements

The company worked both with clients and materials suppliers to choose the propermaterials. Issues of the composition of various alloys and thickness of application over thebase metal can affect solderability and shelf life. The engineer emphasized that it is a con-stant educational process to make vendors and customers knowledgeable.12

ISSUES AND CONCLUSIONS

The world is demanding increasingly high standards for high-value processes.Increased miniaturization and expected longevity and reliability (e.g., in pacemakers)implies designs that will be more difficult to assemble, increasingly difficult to clean, andwhere residual soils can be catastrophic. Standards are needed in clean rooms, chemicals,and process performance. Standards are set, without necessarily knowing whether thosestandards are relevant, sufficient, or overkill.


Fear is probably one of the biggest deterrents to progress in precision cleaning. Fearresults in unrealistic, dogmatic processes that experienced assemblers simply ignore. Fearresults in ever-increasing process steps, which are not only inefficient but may actually pro-duce contamination through excess product handling. Fear blocks communication, includ-ing communication with the workforce, the engineers, with subcontractors and vendors,and with those involved in other, often seemingly unrelated applications.

Some general keys to successful process implementation include:

• Testing the actual process in the proposed equipment• Comparing results with the control method• Providing appropriate automation• Involving the production team (including management and assemblers) in

process development and decision making• Providing for more than one cleaning option• Assuring thorough rinsing• Providing for rapid drying• Educating (not just training) the production people• Understanding that process optimization is an ongoing issue, one of continuous

improvement, with potential high production rewards11

High-value components may involve processes that are so specialized that standardsmay be nearly impossible to define. Certainly, understanding the performance of relatedprocesses and even seemingly unrelated product lines can be helpful in avoiding problems.In all, the best guidelines remain:

LogicChoice of quality cleaning agents and processesDocumentation of ongoing processesAn educated production forceGood communication

REFERENCES

1. LeBlanc, C., Toxics Use Reduction Institute, personal communication.2. Kanegsberg, B., Cleaning options in the high precision cleaning industry: overview of contami-

nation control working group XIII, in Proc., The 1993 International CFC and Halon AlternativesConference, October 20–22, Washington, D.C., 1993, 943.

3. Carter, M., Andersen, M.E., Chang, S., Sanders, P.J., and Kanegsberg, B., Cleaning high precisioninertial navigation systems, a case study and panel discussion, in Proc., CleanTech ‘99, May 18–20,Rosemont, IL, 1999, 294–301.

4. Kanegsberg, B., Abbink, B., Dishart, K.T., Kenyon, W.G., and Knapp, C.W., Development andimplementation of non-zone depleting, non-aqueous high precision cleaning protocols for iner-tial navigation subassemblies, in Microcontamination ‘93 Proc., 1993.

5. Shubkin, R., Albemarle Corporation, Test Results, Beryllium Compatibility, 1998.6. Kanegsberg, B., Cleaning for biomedical applications, in Proc., Precision Cleaning ‘97, Cincinnati,

OH, April 15–17, 1997.7. Arscott, E. et al., Validating Reusable Medical Devices: An Overview, Medical Device &

Diagnostic Industry, January, 1996.8. Cala, F.R. and Winston, A.E., Handbook of Aqueous Cleaning Technology for Electronic Assemblies,

Electrochemical Publications, 1996.


9. Shubkin, R.L., A highly effective solvent/cleaner with low ozone depletion potential, in PrecisionCleaning ‘97 Conf. and Proc., April, 1997 (studies performance with ABZOL (TM) VG.

10. Kanegsberg, B., Successful Cleaning/Assembly Processes for Small to Medium ElectronicsManufacturers, tutorial, Nepcon West ‘98, Anaheim, CA, March 3, 1998.

11. Tautscher, C.J., The Contamination of Printed Wiring Boards and Assemblies, Omega ScientificServices, Bothell, WA, 1976.

12. Kanegsberg, B., Choosing the process and tracking success—case studies, in Proc., Nepcon West‘97, Anaheim, CA, 1997.


CHAPTER 5.4

Cleaning Solutions in the SemiconductorWafer Manufacturing Process

Mahmood Toofan and John Chu

CONTENTS

IntroductionBasic Operations in Wafer Fabrication

Photoresist ChemistryPhotolithography and Masking ProcessRadiation-Sensitive PolymersComparison of Positive and Negative ResistsNegative-Acting PhotoresistsPositive-Acting Photoresists

SolventsSensitizersAdditives

Photoresist Performance FactorsResolutionAdhesion CapabilityExposure Speed and Sensitivity

General Wafer Cleaning TechniquesFEOL Cleaning ProcessesCleaning Process OptimizationTemperature EffectUltrasonic and Megasonic Effect

FEOL Cleaning ProcessesSulfuric-Peroxide Chemistry

Sulfuric Acid and Ammonium Persulfate ChemistryRCA Chemistry

Quaternary Ammonium Hydroxides/Choline–Surfactant ChemistryTMAH ChemistryOzone–Water Mixtures

BEOL Cleaning Processes


Chemistry of Positive Photoresist StrippersNMP-Based StrippersNon-NMP-Based Organic Strippers

Chemistry of Negative Photoresist StrippersChemistry of Post-Plasma-Etch-Polymer Removers

HA/Amine ChemistryHF/Glycol Chemistry

Challenges of the Future TechnologyCopper InterconnectsLow-k Dielectric Material

References

INTRODUCTION

In semiconductor device manufacturing, silicon wafers are processed to fabricate verylarge-scale integration (VLSI) or ultra-large-scale integration (ULSI) circuits. Since the earlystages of the semiconductor wafer processing in the 1960s, significant improvements andadvancements have been made in chip manufacturing. However, the chemistry of wafercleaning material and basic cleaning operations have remained fundamentally unchanged.During recent years, the geometry of the microcircuits, the diameter of the silicon wafers,and the processing equipment and methods have been significantly improved andupdated. In early stages, simple immersion tanks of cleaning solutions were employedwith manual agitation. Today, more-advanced cleaning solutions are applied on sophisti-cated wet benches and spray tools with automated chemical delivery systems and roboticarm movements for displacement of wafers.

In the past three decades, wafer fabrication technology has made significant advance-ments in terms of density of microcircuits, reduction of feature size, and increase in waferdiameters. Diameters of the wafers have increased from 2 to 3 in. to 8 to 12 in. On the otherhand, as the diameter of the wafers has increased, the geometry of microcircuits and inter-connects have been reduced from 6 to 8 �m (106 m) to 0.2 to 0.3 �m. These dramatic changesin wafer processing technology require more precise cleaning solutions with ultrahighpurity and advanced cleaning formulations that otherwise would not be use.

The cleanliness of the wafer surfaces and the purity of cleaning chemicals used inwafer fabrication processes are essential requirements to yield improvement in microelec-tronic device manufacturing. To meet the required specifications of sub-half-micron sub-strate geometry in wafer processing, the surface cleaning chemistry must meet stringentquality of clean room packaging, filtration, and ultrahigh purity of sub-ppb (part per bil-lion) ionic contamination. Trace ionic impurities, such as sodium or potassium cations andchloride anions and particulates, are especially detrimental if present on wafer surfacesduring thermal processing.

BASIC OPERATIONS IN WAFER FABRICATION

Wafer fabrication is the series of processes used to create the semiconductor devices on asilicon wafer surface. The polished silicon wafers with blank surfaces undergo hundreds ofprocess steps and end up producing hundreds of thousands of chips with multiple anddiverse functions. The designs of the devices and circuits are based on different transistorstructures. Among the major structure designs, bipolar and MOS (metal oxide semiconductor)


Figure 1a

Figure 1b

transistors are the most widely manufactured and used with numerous variations.Furthermore, there are several choices of processes and materials available to create each indi-vidual layer of any particular device structure.

Regardless of the process diversity and hundreds of varieties of process options, only fourmajor operations are performed during the fabrication process. These major operations arelayering, patterning, doping, and heat treatment.

Layering is the operation used to add thin layers of materials to the surface of the wafer.The layers are added to the surface in multiple major techniques: growing a silicon oxide orsilicon nitride layer on the wafer using a thermal process, and chemical vapor deposition(CVD). Rapid thermal operation (RTO) or rapid thermal process (RTP) technology is a natu-ral choice for the growth of oxides used in MOS devices. Other techniques such as evapora-tion, physical vapor deposition (PVD), spin-on deposition, and sputtering are also used toadd layers on the wafers.

Patterning is the series of steps that results in the removal of selected portions of theadded surface layers (Figures 1a and b). After removal, a pattern of the layer is left on thewafer surface. The material left or removed may be in the form of a hole in the layer or just aremaining island of the material.

The patterning process is named photomasking, photolithography, or microlithography.Photolithography is a multistep pattern transfer process similar to stenciling or photography.In photolithography the required pattern is first formed in photomasks and transferred intothe surface layers of the wafer through the photomasking steps. The polymeric materials usedin photolithography to transfer patterns to the wafer are called photoresist. Figure 2 showsthe ten-step process of pattern transfer to the wafer surface using photomasking process for anegative acting photoresist.

The foregoing was a brief description of some of the basic operations in wafer fabrication.


Figure 2 Pattern transfer process.


Table 1 Photoresist Components and Their Functions

Component Function

Polymer Changes structure due to reaction with radiation energy (polymerization orphotosolubility)

Solvent Used as thinner to allow application of a thin-film layer of the spun materialSensitizers Control modification of chemical reaction when exposed to lightAdditives For special purposes

Since the focus of this chapter is a review of cleaning technology in semiconductor waferfabrication, the emphasis of the following sections will be on the chemistry of photoresist,its cleaning solutions, and cleaning processes in wafer operations, especially advancedcleaning methods for film removal, such as photoresist strippers and post-plasma-etchpolymer removers.

For additional information on fabrication processes, readers are encouraged to refer toMicrochip Fabrication by Peter Van Zant.1

Photoresist Chemistry

Photoresists have been used in the printing industry for over a century. In the 1920sphotoresists found wide range of application in the printed circuit board industry. Thesemiconductor industry adopted this technology for wafer fabrication in the 1950s.Photoresists specifically designed for semiconductor use were first developed by theEastman Kodak Company. In the late 1950s it introduced Kodak Photo Resist (KPR), KodakMetal Etch Resist (KMER), and Kodak Thin Film Resist (KTFR)—negative photoresists. Ataround the same time, the Shipley Company introduced a line of positive-acting photoresists.Since that time, some other companies also have entered the market with photoresistsdesigned to keep pace with increasing demand in the industry for printing narrower linesin fabrication of fine-geometry integrated circuits. Today, different manufacturers offer awide range of products designed to match a variety of applications.

Photoresists are used in the masking process for patterning the wafers in the process ofphotolithography. Other terms used in industry for these steps are photomasking, masking, ormicrolithography.

Photolithography and Masking Process

Photolithography is one of the most critical operations in semiconductor manufactur-ing processes. It is the patterning process that sets two-dimensional horizontal patterns onthe various parts of the circuit design on the wafer. The photoresist materials perform thefunction of transforming a two-dimensional circuit design into a three-dimensional electriccircuit. The photoresist materials used in photolithography are generally formulated frompolymeric materials with photosensitive additives. Most photoresist materials consist offour basic ingredients, each having a different function. Table 1 shows the basic compo-nents of photoresists.

Radiation-Sensitive Polymers

The photosensisitive ingredients of the photoresist material are special polymers.Polymers are macromolecules containing carbon, hydrogen, and oxygen atoms that are


formed by repeated patterns of their monomers or simple molecules. Most plastics are aform of polymer. Photoreactive polymers are radiation sensitive and react with some typeof light energy, ultraviolet (UV) or laser. Those photoresists that contain these type of poly-mers are called optical resists. Other resists respond to X-ray radiation or e-beams, which arei-line or j-line resists.

Comparison of Positive and Negative Resists

Up to the mid-1970s, negative resist was dominant in the masking process. The adventof VLSI circuits and image sizes in the 2- to 5-�m range strained the resolution capabilityof negative resists. Positive resists had been around for over 20 years, but their poorer adhe-sion properties were a drawback and their superior resolution capability and pinhole pro-tection were not needed.

By the 1980s, positive resist became the resist of choice. The transition was not easy. Toswitch a fabrication line from negative to positive resist requires changing the polarity ofthe masks or reticles from clear field to dark field. Unfortunately, it is not a simple matterof reversing the fields in the mask-making process. The dimensions have to be adjusted toaccommodate the different characteristics of the positive resist. The determination of thecorrect mask of reticle dimensions is a lengthy procedure.

Positive resists have a higher aspect ratio compared with negative resists. In otherwords, have a better resolution capability and can resolve smaller geometry such as wirelines and via openings.

Another problem with negative resists is oxygenization. This is a reaction of the resistto oxygen in the atmosphere, and can result in a thinning of the resist film by as much as20%. Positive resists do not have this property. Cost is always an important consideration.Negative resists sell for about one third of the cost of positive resists.

Developing characteristics differ between the two types of resists. Negative resistsdevelop in readily available solvents and possess wider developer process latitude. Positiveresists require carefully prepared developer solutions and temperature control of the process.

The next-to-last step in the masking process is photoresist removal, which can take placein chemical solutions or in plasma systems. Generally, the removal of positive resists is easierand takes place in chemicals that are more environmentally sound. While positive photoresistsare the resists of choice for fabrication areas processing state of the art circuits, there are manylines still producing devices and circuits with image sizes greater than 5 �m. A great many ofthese lines use negative resists. Table 2 shows a comparison of properties of the two resists.

Table 2 Comparison of Negative and Positive Resists

Parameter Negative Positive

Aspect ratio (resolution) HigherAdhesion BetterExposure speed FasterPinhole count LowerStep coverage BetterCost HigherDevelopers Organic solvents AqueousStrippers

Oxide steps Acid AcidMetal steps Chlorinated solvent Simple solvents

compounds


Negative-Acting Photoresists

Negative photoesists are normally based on polyisoprene-type polymers. Polyisoprenepolymers naturally occur in rubber material. The Hunt Corporation developed the first syn-thetic polyisoprene polymer structure (Figure 3).

Before exposure to the light, the negative resist polymers exist in their unpolymerizedcondition (under which the polymers are not chemically linked to each other). When thephotoresist is exposed to proper light or energy, the polymers become cross-linked or, inchemical term, polymerized. This process may also be achieved when the photoresistmaterials are exposed to heat and/or visible light. To prevent this deterioration, the pho-toresist material is normally packaged in amber glass bottles or dark color, brown or blackplastic packaging. During the application process, to prevent accidental exposure, pho-tomasking and resist processing areas use yellow filters or yellow lighting.

Depending on the response of the photoresists to the type of energy or radiation, pho-toresists are normally referred to by their general category, such as UV, deep UV, X-ray,I-line, etc.

Positive-Acting Photoresists

Positive-acting photoresists are based on the phenol-formaldehyde polymer, alsocalled phenol-formaldehyde novolak resin (Figure 4). The novolak resin within the unex-posed photoresist is relatively insoluble. After exposure to the proper radiation energy, thephotoresist converts to a more soluble state. This reaction is called photosolubilization.Table 3 contains a list of commonly used photoresist polymers used for the photolithogra-phy process in the semiconductor industry.

Figure 3 Chemistry of negative photoresist (isoprene monomer).

Figure 4 Phenol-formaldehyde novolak resin structure.


Table 3 Commonly Used Photoresist Materials

Sensitivity ExposureResist Polymer Polarity (C/cm2) Source

Positive Novolak (M-cresoformaldehyde) � 3–5 � 10�5 UVNegative Polyisoprene � UVPMMA Poly-(methyl methacrylate) � 5 � 10 �5 E-beamPMIPK Poly-(methyl iso-propenyl ketone) � 1 � 10 �5 E-beam/deep UVPBS Poly-(butene 1-sulfone) � 2 � 10 �6 E-beamTFECA Poly-(trifluoro-ethyl chloroacrylate) � 8 � 10 �7 E-beamCOP Copolymer-(cyano ethyl acrylate-

(PCA) amido ethyl acrylate) � 5 � 10 �7 E-beam/X-rayPMPS Poly-(2-methyl pentene-1-sulfone) � 2 � 10 �7 E-beam

Source: Van Zant, P., Microchip Fabrication, McGraw-Hill, New York, 1990 with permission.

Solvents

The largest ingredient by volume in photoresist composition is the solvent. It is the sol-vent that converts the solid resist material to a liquid and allows the liquid photoresist tobe applied to the wafer surface as a thin layer by spinning. Photoresist is analogous topaint, which is composed of the coloring pigment and polymer dissolved in an appropri-ate solvent. It is the solvent that allows the application of the paint onto a surface in a thinlayer. For negative photoresist, the solvent is an aromatic hydrocarbon such as xylene.

In positive resist, a variety of solvents are used depending on the type of polymer. Themost commonly used solvents are ethoxyethyl acetate (EEA), 2-methoxy propyl acetate(propylene glycol monomethyl ether acetate, PGMEA), and ethyl lactate (ELS).

Sensitizers

Chemical sensitizers are added to the resists to cause or control certain reactions of thepolymer. In negative resists, the untreated polymer responds to a certain range of the UVspectrum. Sensitizers are added to either broaden the response range or narrow it to a spe-cific wavelength. In negative resists, a compound called bisaryldiazide is added to thepolymer to provide light sensitivity. In positive resists, the sensitizer is O-naph-thaquinonediazide.

Additives

Various additives are mixed with resists to achieve particular results. Some negativeresists have dyes that are intended to absorb and control light rays in the resist film.

Photoresist Performance Factors

The selection of a photoresist starts with the dimensions required on the wafer surface.The resist must first have the capability of producing those dimensions. Beyond that, itmust also function as an etch barrier during the etching step, a function that requires a cer-tain thickness for mechanical strength. In the role of etch barrier, it must be free of pinholes,which also requires a certain thickness. In addition, it must adhere to the top wafer surfaceor the etched pattern will be distorted, just as a paint stencil will give a sloppy image if itis not taped tight to the surface. These, along with process latitude and step coverage capa-


bilities, are resist performance factors. In the selection of a resist, the process engineer oftenmust make trade-off decisions between the various performance factors.

Resolution

The smallest opening of space that can be produced in a photoresist layer is generallyreferred to as its resolution capability. The smaller the line produced, the better the resolu-tion capability. Generally, smaller line openings are produced with thinner resist film thick-ness. However, a resist layer must be thick enough to function as an etch barrier and to bepinhole-free. The selection of a resist thickness is a trade-off between these two goals. Atpresent, a more-advanced photoresist is the 0.193-�m resist with a resolution sensitivity of193 nm. There is also a more-advanced line of positive resists in the UV range with an expo-sure wavelength as low as 154 nm.

The capability of a particular resist relative to resolution and thickness is measured byits aspect ratio (Figure 5). The aspect ratio is calculated as the ratio of the resist thickness tothe image opening. Positive resists have a higher aspect ratio compared with negativeresists, which means that for a given image-size opening, the resist layer can be thicker. Theability of positive resist to resolve a smaller opening is due to the smaller size of the poly-mer. It is similar to the requirement of using a smaller brush to paint a thinner line.

Figure 5 Aspect ratio.

Aspect Ratio = WT

Resist

W

T

Figure 6 Advanced aspect ratio.


Advanced photoresists can generate a via opening with an aspect ratio of up to 10:1 incopper interconnect technology. Figure 6 shows some advanced device structure withhigh-aspect-ratio via.

Adhesion Capability

In its role as an etch barrier, a photoresist layer must adhere well to the surface layer totransfer the resist opening faithfully into the layer. Lack of adhesion results in distortedimages. Resists differ in their ability to adhere to the various surfaces used in chip fabrica-tion within the photomasking process; there are a number of steps that are specificallyincluded to promote the natural adhesion of the resist to the wafer surface. Negative resistsgenerally have a higher adhesion capability than positive resists.

Exposure Speed and Sensitivity

The primary action of a photoresist is the change in structure in response to an expos-ing light or radiation. An important process factor is the speed at which that reaction takesplace. The faster the speed, the faster the wafers can be processed through the maskingarea. Negative resists typically require 5 to 15 s of exposure while positive resists take threeto four times longer.

The sensitivity of a resist relates to the amount of energy required to cause the poly-merization or photosolubilization to occur. Further, sensitivity relates to the energy associ-ated with specific wavelength for the exposing source. Understanding this propertyrequires a familiarization with the properties of the electromagnetic spectrum (Figure 7).Within nature, are a number of different types of energy: light, short and long radio waves,X-rays, etc. In reality they are all electromagnetic energy (or radiation) and are differenti-ated from each other by their wavelengths, with the shorter-wavelength radiations havinghigher energies.

Common positive and negative photoresist responds to energies in the UV and deepultraviolet (DUV) portion of the spectrum (Figure 8). Some are designed to respond to par-ticular wavelengths (peaks) within those ranges. Resist sensitivity refers to the specific wave-lengths to which the resist reacts. This property is also called the spectral response characteristicof the resist. Figure 9 is the spectral response characteristic of a typical production resist.The peaks in the spectrum are regions (wavelengths) that carry higher amounts of energy.

GENERAL WAFER CLEANING TECHNIQUES

Impurities on the surface of the silicon wafers come from various sources at differentstages of the manufacturing process. These impurities must be removed following each

Figure 7 Electromagnetic spectrum.

Name

Wavelength(cm)

GammaRays X-Rays

Ultra violet(UV)

Infrared(IR)

Short RadioWaves

BroadcastRadioWaves

10-11 10-6

10-4

10-3 102 10410-8

Visible


Figure 8 UV and visible spectrum.

250 500 1000

DUV

4 - 100

100 - 250

UV

250 - 400

Near UV

400 - 700

Visible

Wavelength (nm)

Figure 9 Exposure response of a positive photoresist.

process step to keep the substrates clean for the next process. Depending on the type of sur-face contamination or impurities, different cleaning solutions and techniques need to beapplied.1

Wafer cleaning solutions, depending on the process and their functions, are classified as

1. Cleaning solutions of bare silicon and oxidized wafers involving premetalprocesses, or so-called front end of line (FEOL) cleaning processes;

2. Cleaning solutions of postmetal processes used in different stages of metalliza-tion, or so-called back end of line (BEOL) processes.2


Surface conditioning or FEOL is a premetal process and normally uses acids and oxidantsto precondition and clean the wafer surfaces. Postmetal processing, or BEOL, whichincludes photoresist stripping, post-plasma-etch residue removing, and post-CMP (chem-ical mechanical polishing) slurry removal, uses sophisticated solvent formulations that arenot the traditional cleaners. These two processes can be distinguished as residual or con-tamination cleaning vs. bulk material removing, such as resist stripping.

The solvent formulations for bulk material removal or BEOL processes are either semi-aqueous solvent formulations or aqueous solutions of more advanced cleaning solutionsthat will include corrosion inhibitors and wetting agents for protection of fine geometry inmetal layers, and high aspect ratio via openings.

Depending on whether a semiconductor device is DRAM (dynamic random accessmemory) or logic, a wafer can undergo 25 to 30 different steps, including ash processingfollowed by a cleaning step using one of two chemistries.

FEOL Cleaning Processes

In FEOL cleaning, a universal method is treatment by sulfuric acid and hydrogen per-oxide (piranha) following an ash process, which takes place in either a spray tool or animmersion wet bench. There is little single-wafer processing application of the spray chem-istry in FEOL processes. Other cleaning chemistries used in FEOL processes include RCAclean, choline chemistry, and quaternary ammonium hydroxide or tetramethyl ammoniumhydroxide (TMAH) based chemistry in either spray or immersion equipment. These willbe discussed in more detail shortly.

Cleaning Process Optimization

Depending on the chemistry of the cleaning solution involved, the type of wafers, andthe process conditions, such as process temperature, and process time, and the equipmentused may vary from one fabrication site to another. Process engineers optimize their clean-ing process to achieve high yields and low defects.

Temperature Effect

An important factor in cleaning wafers is the bath temperature of the processing mate-rial. In an FEOL cleaning process such as RCA clean chemistry the chemical is normallyheated to an optimum temperature (typically 55 to 60°C), to achieve the best results.Photoresist developers and edge bead removers (EBR) in BEOL process are normallyapplied at ambient temperature. These materials only dissolve the uncured or soft-bakedphotoresist, which is not cross-linked or polymerized. In the resist stripping process, how-ever, the cross-linked and polymerized resist may require heated stripping solutions and alonger time for complete removal of the hardened photoresist. If plasma-treated photore-sist residues (post-plasma polymers) are not oxygen plasma ashed, they are even more dif-ficult to clean and will need more aggressive solutions at higher temperatures. In any case,whether an FEOL or BEOL process is contemplated, the temperature of the immersion tankor a spray tool is optimized and preset to safe operational conditions.

Ultrasonic and Megasonic Effect

For certain cleaning applications such as metal-lift-off processes or stripping of an ion-implanted and deep-UV-cured photoresist, without a plasma ashing, ultrasonic or mega-


Figure 10 Megasonic tank configuration. (After Kern, W., Ruzyllo, J., and R. Novak, Proc.Electrochem. Soc., 90-9, 5–15 and 67–68, 1989.)

sonic agitation may be necessary for complete dissolution. In FEOL cleaning using RCAclean process, megasonic energy has shown significant improvement on cleanliness of thewafers and particle removal efficiency of the solution. Figure 10 shows the configuration ofa megasonic wafer-cleaning tank.

FEOL CLEANING PROCESSES

Sulfuric-Peroxide Chemistry

Early cleaning processes of silicon wafers involved using concentrated inorganic acidssuch as boiling nitric acid, aqua regia, concentrated hydrofluoric acid, and mixtures ofphosphoric, acetic, and sulfuric acids. Mixtures of sulfuric acid and hydrogen peroxide, orso-called piranha solutions, are still being used in FEOL wafer cleaning applications. Interms of sulfuric chemistry and acid-to-peroxide mix ratios, process engineers use theirown selections. Instead of having a premixed, stabilized mixture, one can prepare twochemicals on site, as needed in different ratios.

Oxidizing agents such as mixtures of sulfuric acid and chromic acid were also used asa general-purpose glass cleaner or silicon wafer surface cleaner. This type of cleaners, how-ever, caused ecological toxic pollution and waste disposal problems.

Sulfuric Acid and Ammonium Persulfate Chemistry

Hydrogen peroxide (H2O2) is essentially unstable and readily disassociates to waterand oxygen at elevated temperatures.

H2O2 → H2O � 1/2 O2

An alternative oxidizing agent used in wafer cleaning is ammonium persulfate (AP) in sul-


furic acid mixtures. AP is added to sulfuric acid baths in a concentrations of 40 to 80 g/l.Since AP is less reactive at room temperature than H2O2, it is and therefore safer to storeand has a longer shelf life. Being less reactive, AP ensures a more stable and steady releaseof oxygen to the cleaning bath and more consistent and stable bath life to the chemical.Sulfuric acid–AP mixture is used in FEOL cleaning process, for general wafer cleaning andalso in resist stripping processes of nonmetalized wafers.

RCA Chemistry

The first systematically developed cleaning process for unprocessed or oxidized siliconwafers, called RCA clean, was developed at RCA and published in 1970. The RCA cleaningprocess involved a two step process of peroxide treatment with a high-pH alkaline solution(normally ammonium hydroxide mixed with hydrogen peroxide) as RCA1 or SC-1, fol-lowed by a treatment with a mixture of hydrochloric acid and hydrogen peroxide as SC-2.3

The solutions are made using ultrafiltered deionized (DI) water, electronic-gradeammonium hydroxide (29% wt/wt% as NH3), electronic-grade hydrochloric acid (37wt/wt%), and high-purity unstabilized hydrogen peroxide.

In the first treatment step, the wafers are exposed to a hot mixture of water-dilutedhydrogen peroxide and ammonium hydroxide. This procedure was designed to removeorganic surface films by oxidative breakdown and dissolution to expose the silicon or oxidesurface for concurrent or subsequent decontamination reactions. In this treatment, metalimpurities such as copper and zinc are dissolved and removed by a complexing agent ofammonia, for example, in forms of [Cu (NH3)

4]2� amino complex.The second treatment is designed to remove alkali ions, cations such as Al3�, Fe3�, and

Mg2� that form water-insoluble hydroxides in SC-1 ammonia solution.The volume ratios for the RCA standard clean 1, SC-1 clean used in first treatment step

are H2O:H2O2 (30%):NH4OH (29% as NH3) as 5:1:1 and the volume ratios for the RCA stan-dard clean 2, SC-2 clean used in second treatment step are H2O:H2O2:HCl as 6:1:1.

The processing temperature should be kept at 75 to 80°C to activate the mixture suffi-ciently without causing thermal decomposition due to higher temperatures.

The original RCA cleaning processes was based on a simple immersion technique.Several different improved techniques have been introduced over the years. More advancedautomated wet-bench immersion systems for large-scale production are now available andbeing offered in the industry by equipment manufacturers such as FSI International.

Quaternary Ammonium Hydroxides/Choline—Surfactant Chemistry

Among the other alternative alkaline cleaning solutions that have been studied on sil-icon wafers, choline (2-hydroxyethyl trimethyl ammonium hydoroxide),5 a strong basewith a chemical formula [N(CH3)3CH2CH2OH]OH, and pKb � 5.06, which is free of metalion, has drawn much attention.5–7 Immersion of HF-etched silicon wafers in a choline solu-tion followed by a water rinsing has shown very clean results.

Muraoka et al.5 have developed several techniques to clean silicon wafers usingcholine. They have reported that dilute aqueous solutions of choline with selected nonionicsurfactants can remove heavy metals from the silicone wafer surface and prevent replatingof these metals from solution on the wafer. Harri and Hockett6 compared wet cleaninginvolving choline with that of other solutions and concluded better electrical propertieswith the choline process.


TMAH Chemistry

Another strong base that is relatively stable at ambient temperature and is also free ofmetal ion is TMAH, N(CH3)4OH. TMAH is widely used as a positive photoresist developerin a relatively low concentration (2.5 wt%) in aqueous solution with surfactants. TMAH-based photoresist developers have replaced the traditional alkaline developers because oftheir low ionic impurities and high polymer dissolution capability.

In recent years TMAH-based formulations in organic solvents have been used for pos-itive photoresist stripping applications.8 –10 Even though the TMAH-based strippers arenow commercially available and are being used in the industry, they have shown somedrawbacks and do not provide a robust resist stripping process. A major disadvantage ofTMAH-based strippers is the high pH value of the product that makes it corrosive to sen-sitive metals, specially Al and Al alloys.

Aluminum reacts with alkaline solution in aqueous media, which results in etching ofthe aluminum lines, especially in the submicron geometry.

2Al � 2OH� � 2H2O → 2AlO2� � 3H2 (g)

Another disadvantage of TMAH strippers is the instability of the quaternary ammoniumhydroxide at elevated temperatures. An independent laboratory study of a commerciallyavailable TMAH/NMP-based stripper (NMP � N-methylpyrrolidone) using protonnuclear magnetic resonance (NMR) spectroscopy and gas chromatography/mass spec-trometry (GC/MS) has indicated the following: at high temperatures, TMAH disassociatesto trimethylamine and methyl alcohol in aqueous media.

N (CH3)4 OH → N(CH3)3 (g) � CH3OH (g) at t � 80°C

The bath life of a typical TMAH/NMP stripper is approximately 2 to 4 h at 85°C.At ambient temperatures, however, the TMAH-based developers are widely used in

the industry for positive photolithography process.

Ozone–Water Mixtures

Another oxidizing agent that has been historically used for wastewater treatment,drinking water sterilization, and in swimming pools is ozone in water. In recent yearsozone has been introduced into microelectronic industry in both wafer cleaning (FEOL)applications and in photoresist residue removal (BEOL) processes.

Ozone has basically the same role in oxidizing and cleaning organic residues as H2O2

has in the RCA clean. Ozone and H2O2 decompose virtually the same way:

H2O2 → O� � H2O

O3 → O� � O2

The biggest advantage of ozone over RCA clean is that ozone leaves no harmful decompo-sition residues or by-products. It is partially soluble in water, especially at lower tempera-tures. Generally, ozone is about ten times more soluble in water than oxygen. The lower thewater temperature, the higher the ozone solubility. The half-life of ozone in high-purity DIwater is about 20 min.


In recent years ozone chemistry has been receiving considerable attention because ithas a potential to be used in both FEOL and BEOL and to eliminate or reduce the usage oforganic solvents. In a study comparing the effectiveness of ozone chemistry and modifiedRCA clean, FSI International used its centrifugal spray processors to clean silicon wafers.The wafers were then examined for their metal ionic contamination and change in particlesbefore and after the cleaning process. Residual metals following the cleaning processes weremeasured using both secondary ion mass spectroscopy (SIMS) and total reflection X-rayfluorescence (TXRF) methods. Results of those studies are presented in Tables 4 and 5.4

BEOL CLEANING PROCESSES

Bulk material removing processes normally take place following three major processsteps in wafer fabrication:

1. Photolithography process (photoresist stripping and edge bead removal)2. Metal and oxide etched by plasma or reactive ion (polymer removing)3. CMP, chemical mechanical polishing (slurry removal process)

The major photoresist residues resulting from the photolithography process are normallycleaned using a photoresist-stripping process involving organic solvents.

The EBR process utilizes the following solvents:

1. PGMEA based2. Ethyl lactate based3. Organic solvents as ketones and esters (acetone, MEK, MIBK, and nBAc)4. Environmentally preferred alternatives (VOC exempt)

Table 4 SIMS Analysis of RCA Clean vs. Ozone Clean (atoms/cm2)

Relative Elemental Abundance

F Na K Cu Mg Al Ca

Clean starting wafersModified RCA clean 140 225 65 �20 22 1575 113Ozone clean 200 105 25 �20 10 1064 98

Contaminated starting wafersModified RCA clean 145 330 160 �20 N/A N/A N/AOzone clean 250 125 20 �20 N/A N/A N/A

Table 5 TXRF Analysis of RCA Clean vs. Ozone Clean (atoms/cm2)

Mn Fe Zn Br Cr Co Cu

Clean starting wafersModified RCA clean �0.4 0.3 0.2 1.4 �0.6 �0.4 0.2Ozone clean �0.4 �0.3 �0.2 �0.3 �0.6 �0.4 �0.2

Contaminated starting wafersModified RCA clean �0.4 �0.3 �0.2 1.8 �0.6 �0.4 �0.2Ozone clean 0.5 �0.3 �0.2 �0.3 �0.6 �0.4 �0.2


After plasma-etch of metal and oxide, the tough sidewall polymers are cleaned byadvanced formulations of post-etch polymer removers, or by ashing with oxygen followedby wet clean.

After CMP, the slurry particles are rinsed away with water-based, diluted solvents. Inmany instances, scrubbing with a brush and/or megasonic agitation is needed to dislodgethe much heavier slurry particles.

CHEMISTRY OF POSITIVE PHOTORESIST STRIPPERS

NMP-Based Strippers

Among other aprotic solvents, N-methyl pyrolidone (NMP) has attracted particularattention in photoresist stripping formulations. In some applications, pure NMP is used forstripping soft-baked resists. For relatively hard and cross-linked resists, a more aggressivealkaline mixture with high-pH values is needed to mix with NMP. Organic amines havebeen shown to possess the desired characteristics when mixed with the aprotic solvents.

During the past two decades the NMP/amine-based strippers have dominated thepositive resist stripping market because of their low toxicity and their resist cleaning effi-ciency. NMP/amine solutions are not only effective in cleaning hard-baked cross-linkedresists at elevated temperature but they are also 100% water soluble and biodegradable,which makes these formulations particularly popular, as opposed to the more toxic phe-nolic or chlorinated solvents.

For more-advanced applications, different wetting agents or nonionic surfactants, suchas poly-alkylene glycol (ethylene glycol or propylene glycol), are added to the stripper for-mulations.11 –15

Non-NMP-Based Organic Strippers

Other aprotic solvents such as tetra hydrothiophrene 1,1-dioxide (Sulfolane), dimethylsulfoxide (DMSO), dimethyl formamide (DMF), and dimethyl acetamide (DMAC) are alsoused in different stripping formulations.11,12

Several other stripping solutions using dibasic esters, alcohols, ketones, glycol ether, orother organic solvents have also been reported and are being used. Aqueous-based(water/surfactant) strippers using dibasic esters as active ingredient with neutral pH val-ues have also shown favorable results in stripping soft-baked and hard-baked bulkresist.13 –14

CHEMISTRY OF NEGATIVE PHOTORESIST STRIPPERS

Negative photoresists are polymerized rubber and are normally soluble in aromatichydrocarbons or phenolic solvents. The most commonly used solvent/thinner for negativephotoresist is xylene mixture. In the early 1960s, the first organic stripper, containing chlo-rinated aromatic hydrocarbon as solvent and an alkylbenzene sulfonic acid as surfactant,was introduced to the market by a company called Industri Chem. This formulation uti-lized phenol to create a water-rinsable solution; as such, the first organic resist stripperJ-100 was born.

Stripping photoresist by J-100 requires a heating bath in the range of 90 to 120°C, fol-lowed by a series of post-strip rinse solutions and a DI water rinse and spin dry. Since theintroduction of J-100, a number of suppliers have developed similar products, some


designed for direct water rinse. Other manufacturers have offered similar products contain-ing dodecylbenzene sulfonic acid, phenol, and chlorinated benzene solvents.15 –17 These prod-ucts were successfully used as workhorse strippers for two decades. However, in the 1970sthe environmental concerns over the toxic ingredients in these formulas led to the devel-opment of nonphenolic and nonchlorinated solvents with fewer waste disposal difficulties.

CHEMISTRY OF POST-PLASMA-ETCH-POLYMER REMOVERS

Tough sidewall polymers are created during the plasma etch process (Figure 14). Thecross-linked and hardened polymers need to be cleaned by advanced chemical formula-tions. These advanced formulations are normally referred to as post-plasma-etch-polymerremovers. In general, after plasma etch, the wafers are normally ashed in oxygen plasmaashers, in which the majority of the photoresist material is oxidized and removed from thesurface of the wafer. However, the organometallic polymer formed on the sidewall of themetal layer (Figure 11) or inside the via polymer openings (Figure 12) does not react andremains. If the polymer is not removed properly, the residue will cause failure in connec-tions, and subsequently the device will fail the electrical test.

A number of post-etch-polymer removers have been introduced to the market bychemical manufacturers and have been used since the early 1990s. EKC Technologies firstintroduced hydroxylamine (HA) chemistry for post-etch-polymer removal applications.After that, Ashland Chemical Co. and a number of other chemical manufacturers followedthe lead and offered various formulations. Among the many products available commer-cially, the following basic formulations have been used most frequently:

Hydroxyl amine chemistryHF/glycol chemistryOther organic alternative solvents

Figure 11 SEM images of sidewall polymers on metal lines, before and after clean.

Figure 12 SEM images of via polymers, before and after clean.


HA/Amine Chemistry

Hydroxylamine, NH2OH free base, commercially available as 50% by weight in water,is a strong reducing agent and a weak base with pH about 8.0 to 8.5. Commercially usedfor polymer removal application, EKC-265 manufactured by EKC Technology and ACT-935 manufactured by Ashland Chemical are HA and organic amine (diglycolamine ormonoethanolamine) mixtures. Although these products are widely used in the industry forpost-plasma-cleaning applications, the products have a number of disadvantages anddrawbacks. HA/amine mixtures are not good resist strippers and do not strip the pho-toresist if it is not ashed. The products have a short shelf life and are not stable at elevatedtemperatures. To increase their efficiency and improve corrosiveness, the manufacturersadd chelating agents such as catechol up to 5 wt%. Despite problems associated with HAchemistries, most DRAM manufacturers have been using HA/amine-type products formetal and via polymer cleaning applications. The metal stacks on those wafers are typicallyAl-Cu/Ti/TiN and the dielectric layers on vias are silicon dioxide. HA/amine productswith some process modifications can be adapted to those cleaning processes with fewprocess difficulties.18

In new technologies of copper metallization and with increasing applications of dif-ferent low-k dielectric material, usage of HA chemistry has become very limited and, there-fore, more-advanced formulations have become necessary for those processes. HA/aminemixtures are corrosive to copper and dissolve copper layers to form a water-soluble com-plex of Cu(NR3)4. Other sensitive metal alloys are also being used, such as tungsten (W), invia interconnects as tungsten plugs, which are also susceptible to amine corrosion. Figure13 presents a via structure with exposed W openings, before and after a cleaning processwith a noncorrosive polymer remover.

HF/Glycol Chemistry

Other formulations used in cleaning post-plasma polymers are hydrofluoric acid (HF)and ethylene glycol (EG) mixtures. Low concentrations of HF in EG are not corrosive tosensitive alloys, have strong residual cleaning capability and can be used as an alternativeto HA chemistry. However, HF attacks the silicon oxide layer on the wafer and hence is notsuitable for via cleanings. For metal cleaning with controlled temperature and process lat-itude, HF/EG mixtures can be used successfully with limited oxide loss. Ashland ChemicalCo. manufactures different products of HF/EG mixtures for various cleaning or oxide etch-ing applications. Another similar product that is also commercially available from ACSI(Advanced Chemical Systems International), another specialty chemical company, is calledNOE (natured oxide etchant), which is buffered HF with ammonium bifluoride in a poly-glycol mixture.

Figure 13 Exposed W via with sidewall polymer on Al metal, before and after clean.


CHALLENGES OF THE FUTURE TECHNOLOGY

The real challenge for the future of wafer cleaning technology is the integration of cop-per interconnects and new low-k dielectric material in wafer manufacturing. As we enterthe Third Millennium, major advancements and dramatic changes are taking place inmanufacturing and applications of electronic products. Computer manufacturers are moti-vated to come up with more powerful systems having more complex power transistors insmaller sizes.

Copper Interconnects

In chip manufacturing, a dramatic shift from aluminum to copper interconnects is tak-ing place. IBM was the first to produce products with 100% copper wires with substantialshipments of the Power PC 750 microprocessor starting in late 1998. Other major memoryand DRAM manufacturers are in various stages of making this transition from Al to Cuprocesses. Semiconductor equipment manufacturers such as Applied Materials and

Figure 14 SEM images of sidewall polymers on via and metal lines, before and after polymerremoval process. (Courtesy of Silicon Valley Chemlabs, SVC, Inc.)


Novellus and R&D institutions such as SEMATECH, a consortium of major semiconductormanufacturers in Austin, Texas, are following the lead to implement copper as replacementfor aluminum.

Chip performance is the motivation for this transition since copper has significantlybetter conductivity (60 to 70%) than aluminum. There is a continuous improvement in tran-sistor performance, which is about 20 to 30% per technology generation. As the perform-ance is improving, the geometry of interconnects including the wires is shrinking about30% per generation to provide a constant cost per circuit reduction. As the result, the prob-lem appears that increasingly power transistors must be wired together with thinner metallines (wires) that are getting smaller and less conductive, creating limited chip perform-ance. The shift from Al to Cu, which is 70% more conductive than aluminum, is helpingsolve this problem.

Integration of copper technology requires a completely different line of equipment andmaterial, including cleaning chemicals. Low-k dielectric material is a major part of thisemerging technology transition. CMP slurries and polishing material and post-CMP clean-ers are also part of this transition.

Low-k Dielectric Material

In recent years several different low-k dielectric materials have been developed andintroduced to the market by chemical manufacturers. These products are known as tradenames HOSP, FLARE, HSQ, FOX, and SiLK. Allied Signal has offered HOSP and FLARE,Dow Corning offers HSQ and FOX, Dow Chemical manufactures SiLK. These products,having lower dielectric constant k compared with silicon dioxide SiO2, replace the tradi-tional spin-on glass (SOG) or other thermally grown oxide material.

These, together with other newly invented low-k material, present challenges to thecleaning chemicals. The traditional solvent and amine chemistry may interact with thesematerials and change their low-k property. They could have been left behind on the wafers,causing contamination and via poisoning. Fortunately, new chemistry has been inventedto clean these low-k materials without the associated adverse effects of the traditional poly-mer strippers.

REFERENCES

1. Van Zant, P., Microchip Fabrication, A Practical Guide to Semiconductor Processing, 2nd ed., McGraw-Hill, New York, 1990.

2. Braun, A.E., “Photoresist Stripping”, Semicond. Int., Oct. 1999.3. Kern, W., Ruzyllo, J., and Novak, R.E., Semiconductor cleaning technology/1989, Proc.

Electrochem. Soc., 90-9, 5–15 and 67–78, 1989.4. Kern, W., Ruzyllo, J., and Novak, E., Cleaning technology in semiconductor device manufactur-

ing, Proc. Electrochem. Soc., 92-12 11–27.5. Muraoka, H., Hiratsuka, H., and Usami, T., Abstract 238, The Electrochemical Society Extended

Abstracts, 81-2, October 1981, 570.6. Harri, A., and Hockett, R.H., Semicond. Int., 8, 74, 1989.7. Gould, G., and Irene, E., J. Electrochem. Soc., 135, 1535, 1988.8. Haq, N., et al., U.S. patent 4,744,834.9. Steppan, H. et al., U.S. patent 4,776,892.

10. SVC Labs, Technical Note, Instability of TMAH Based Strippers.11. Ward, I. et al., U.S. patent 5,554,312.


12. Sizensky, J. et al., U.S. patent 4,617,251.13. Sahbari, J. et al., U.S. patent 5,741,368.14. Sahbari, J. et al., U.S. patent 5,909,744.15. Schwartzkopf, F. et al., U.S. patent 5,308,745.16. Corbey, W. et al., U.S. patent 3,673,099.17. Thomas, E. et al., U.S. patent 4,791,043.18. Lee, W. et al., U.S. patent 4,824,763.


CHAPTER 5.5

Biomedical Applications: AnalyticalCharacterization for Biocompatibility

David E. Albert

CONTENTS

IntroductionCleaning ProcessChoice of Disinfecting Chemicals

AlcoholsAldehydesHalogens and Halogen-Containing CompoundsQuaternary Ammonium CompoundsBiguanide DisinfectantsPhenol and Phenolic Derivatives

Analytical TestingChromatographic AnalysisSpectrophotometric Analysis

InfraredUltraviolet/Visible Spectroscopy

ConclusionReferences

INTRODUCTION

In the United States alone, nearly 9 million people work in the health care professions.1

Consequently, when proper infection control measures are not followed, the risk of diseasetransmission to the community results. Infection control does not begin and end with ani-mate objects but includes inanimate objects as well. Environmental surfaces may becomecontaminated by human pathogens. Extensive environmental contamination has beendemonstrated in rooms housing patients with antibiotic-resistant Staphylococcus aureus andEnterococcus.2 Microbes (bacteria, fungi, and viruses) can be carried from one person toanother on the surface of any equipment that is shared between them unless the equipmentis decontaminated between use. They can also be carried on the skin surface,


which is the reason hand washing between examining patients is important. Microbes gainaccess to the body through open wounds, inhalation of infected secretions, or by close con-tact with mucous membranes. The process by which microbes are passed from one infectedperson to cause infection in another is known as “cross-infection.” Although most institu-tion-acquired infections result from a patient’s endogenous flora or person-to-persontransmission, contaminated surfaces have been linked to nosocomial infections.3 Theassurance of proper cleaning, disinfection, and sterilization of instruments and medicaldevices in health care practice is a vital issue for all infection control programs. Obviously,there is a clear need for safe and effective cleaners and disinfectants worldwide.Decontamination of medical equipment, devices, and instruments involves the destructionor removal of any organisms present to prevent infecting other patients or hospital staff.Cleaning, disinfection, and sterilization are all procedures that are used in the decontami-nation process. Decontamination reduces the risks of cross-infection and helps maintainthe useful life of equipment, while helping control hospital-acquired infection.

The issues of cleaning medical instruments and devices have become very contro-versial, and to add fuel to this controversy is the ever-increasing practice of hospitals andhealth care facilities to reprocess disposable medical devices for reuse.4 The dual pressuresof cost containment and demand for health care services are driving this practice.Consumers are demanding more and better services while the federal government hasreduced Medicare funding by $44 billion as a consequence of the Balanced Budget Act of1977. According to 29 CFR 1910.1030(d)(4)(ii), OSHA requires that equipment and sur-faces be cleaned and disinfected after contact with blood or other potentially infectiousmaterial (OPIM).5 OSHA continues to require the use of a tuberculocidal disinfectant forthe decontamination of blood or body fluids. According to OSHA, decontamination isdefined as the use of physical or chemical means to remove, inactivate, or destroy blood-borne pathogens on a surface or item to the point where they are no longer capable of trans-mitting infectious particles and the surface can be rendered safe for handling, use ordisposal. Inherent in this definition is the idea that decontamination includes inactivationof all pathogens (bacteria, fungi, viruses) capable of producing disease and not just thosefound in the blood.

For disinfection and sterilization purposes, reusable patient care equipment will beclassified and processed according to recommendations of the Centers for Disease Controland the Association for Practitioners in Infection Control Guidelines on the Selection andUse of Disinfectants.6,7 Before implementing any cleaning, disinfection, and sterilization ofpatient care equipment, the appropriate category under which the article to be cleaned, dis-infected, or sterilized must be classified.

Class I—Critical. Equipment in this category includes any instrument that will beintroduced into the patient’s bloodstream, through the patient’s skin, or into other nor-mally sterile areas. Examples include surgical instruments, implanted devices, cardiaccatheters, pacemakers, and so forth. Sterility is required for these instruments.

Class II—Semicritical. Equipment in this category includes any instrument that willcome into contact with intact mucous membranes and does not penetrate body surfaces.Such instruments include noninvasive endoscopes, endotracheal tubes, MacGill forceps,oropharyngeal airways, endotracheal tube stylets, anesthesia masks, Ambubag masks,thermometers, laryngoscope blades, and so forth. Sterility is not essential. However, at aminimum, a high-level disinfection procedure that can be expected to destroy vegetativemicroorganisms, most fungal spores, tubercle bacilli, and small, nonlipid viruses is recom-mended.

Meticulous physical cleansing followed by an appropriate high-level disinfection treat-ment gives a reasonable degree of assurance that the items are free of pathogens.


Table 1 Instruments, Levels of Disinfection, Recommended Procedures

Level of Disinfection/Instruments Sterilization Procedure

Class I—Critical Sterility required Moist heat, dry heat, orIncludes all invasive ethylene oxide

instruments (e.g.,surgical instruments,intravenous catheters,implanted devices, etc.)

Class II—Semicritical High level required, must be Moist heat, 100°C for 30 min,All instruments that disinfected between aqueous 2% glutaraldehyde

contact mucous patients for 20–30 min, 1:10membranes (e.g., dilution of bleach for 20 minendotracheal tubes,endoscopes,airways, etc.)

Class III—Noncritical Low level required Chemical disinfectants includeethyl or isopropyl alcohol(70–90%), phenolic germicidaldetergent solutions, iodophoresfor intermediate level, anexposure time of at least 10min is required, for low level,all the above plus quaternaryammonium compounds

Class III—Noncritical Equipment in this category comes in contact with patients andtheir intact skin and rarely, if ever, is implicated in the transmission of disease. Items in thiscategory include crutches, bed boards, blood pressure cuffs, stethoscopes, and so forth.Routine cleansing with soap and water and an EPA-approved disinfectant is sufficient toreduce the number of microorganisms on the surface of this equipment. Alcohol may alsobe used to clean the surface of items if they are not visibly soiled. Table 1 gives examples ofinstruments, levels of disinfection, and the procedures recommended.

The process of cleaning and disinfection/sterilization must be properly validated. Toensure this process, the following issues must be considered. All surfaces of the devicemust come into contact with the cleaning and disinfecting agents to reduce the bioburdenadequately. The device materials should be compatible and unaffected by the process. Afterprocessing the device must function as designed by the manufacturer. The cleaning, disin-fection, and sterilization processes must allow for complete removal of the cleaning anddisinfecting agents.

Once the cleaning and disinfecting process has been implemented and validated, itbecomes necessary to demonstrate that chemicals used in this process are not left on themedical device or equipment at a level that may cause harm to patients. To demonstratethat this is indeed the case, numerous analytical procedures have been developed and usedto measure residual levels of disinfectants and chemical sterilants.

Sterilization validation must demonstrate that all device surfaces have a sterility assur-ance level (SAL) of at least 106 and can be maintained.8

This chapter will first examine the various kinds of chemical disinfectants used in themedical device industry and then explore various analytical techniques to measure theirpresence or absence from these medical devices and equipment.


CLEANING PROCESS

The way equipment is cleaned and stored may present issues that impact the successof infection control practices. The strategy for cleaning instruments and medical devicesessentially is dictated by the inaccessibility of the surfaces to be cleaned. The efficacy ofcleaning and disinfecting agents is contingent upon direct contact with the pathogens.

Organic material (bioburden) such as blood, pus, feces, and tissues, when developedby a surface not only prevents direct contact with the disinfectant, but can inactivate thedisinfectant as well. Chlorine and iodophores are particularly susceptible to such inactiva-tion.9 A perfect example of a very difficult convoluted instrument to clean is the flexiblescopes (i.e., endoscopes, bronchoscopes, and sigmoidoscopes). Flexible scopes are nor-mally classified as minimally invasive or semicritical instruments that touch only mucousmembranes.10 As such, the procedure for flexible scopes is to clean them in preparation forusing high-level disinfection. As with sterilization, the preliminary procedures needed toprepare an item for high-level disinfection are paramount for a successful result. Exposingall surfaces of the instrument or device to the cleaning agent or agents is imperative if thecleaning process is going to be successful. Proper preparation of the scope begins before iteven gets to the CS (central services) department or area. After use on the patient, the scopemust be flushed out immediately with an enzymatic solution.11 The outside must also bewiped off with the solution to remove gross soil. Transportation of the scope must be doneas quickly as possible to prevent drying. If the soil dries on the instrument or device, thena lengthy soaking procedure will be needed to loosen this material. Once the instrumenthas reached the decontamination area, it should be flushed with an enzymatic solutionfrom the distal to the proximal end. This would be a reverse flow in the instrument andwould loosen soil that tends to build up in layers. By flushing in the opposite direction,there is a greater chance that soil will lift off. A cleaning brush must then be applied to theinternal surfaces. This brush must be a good fit for the channel, not too small or too large,without flat spots or missing bristles, and should soak in solution before use.

To design effective and competitive cleaning products, producers combine several dif-ferent chemicals. Surfactants form the base of many cleaning products, removing dirt andsoil through a physicochemical process.12 These products contain hydrophobic andhydrophilic components that work together to loosen absorbed or chemically bound enti-ties. The old saying, “You can clean without sterilization, but never sterilize without clean-ing,” is true. Every effort must be made to reduce bioburden to as low a level as possible.The cleaner items are when entering the sterilizing solution or sterilizer, the better thechances of successful sterilization. Soaps and detergents are only mildly microbicidal.Their use aids in the mechanical removal of microorganisms by breaking up the oily filmon the skin (emulsification) and reducing the surface tension of water so it spreads andpenetrates more readily.

Detergents may be anionic or cationic.12 Anionic (negatively charged) detergents, suchas laundry powders, mechanically remove microorganisms and other materials but are notvery microbicidal. Cationic (positively charged) detergents alter membrane permeabilityand denature proteins. They are effective against many vegetative bacteria, some fungi,and some viruses. However, endospores, Mycobacterium tuberculosis, and Pseudomonasspecies are usually resistant. They are also inactivated by soaps and organic materials likeexcreta. Cationic detergents include the quaternary ammonium compounds (Zepharin,Diaprene, Roccal, Ceepryn, and Phemerol). Zepharin may be used to disinfect instruments(20-min soaks) but have no effect against the tubercle bacillus and are inactivated by soaps.

Cleaning agents can be made even more effective by the addition of enzymes.13

Enzymes are protein biological catalysts that are extremely efficient and speed up normal


biochemical reactions. Proteases are the most widely used enzymes in the detergentindustry. They hydrolyze proteins and break them down into more soluble polypep-tides or free amino acids. As a result of the combined effect of surfactants and enzymes,stubborn stains can be removed. The inefficiency of nonenzymatic detergents in removingproteins can result in permanent stains due to oxidation and denaturing caused by bleach-ing and drying. Blood, for example, will leave a rust-colored spot unless it is removedbefore bleaching. While protein stains can be easily digested by proteases, oily andfatty stains were problematic and continued to be a problem. However, in the late 1980s,genetically engineered proteins (lipases) were developed and added to detergent for-mulations.13

CHOICE OF DISINFECTING CHEMICALS

Several factors dictate how and with what chemicals a device is to be processed toensure that it is clean, decontaminated, and/or sterile. For those devices that do not comeinto contact with bodily fluids, cleaning only may be sufficient. Decontamination shouldbe carried out in accordance with the manufacturer’s instructions. Decontamination agentsmust be compatible with the article to ensure that decontamination does not change theproperties or damage the item. Product deterioration (deterioration of materials) and con-sequently function are extremely important considerations.16 Various changes, dependingon how the disinfection process is carried out, can occur such as weak spots, changed mate-rial (becomes more brittle), reduced performance, and an accumulation of biological mate-rial, which can interfere with performance (e.g., balloon catheters). Chemical treatmentsand detergents may remove some of the nonpolymer components of the plastic (plasticiz-ers, antioxidants, and fillers) and may alter the polymers.17,20 Prolonged exposure to 70%alcohol can disrupt adhesives, damage seals, and denature some plastics.

After cleaning, the item may be disinfected or sterilized. Items compatible with thehigh temperatures and pressures comprising steam sterilization cycle should be sterilizedin an appropriate autoclave. For items where exposure to high temperature is detrimentalto the device, it may be possible to sterilize using ethylene oxide. The manufacturer mustbe consulted before either method of sterilization is used. Those items which cannot besterilized may be disinfected using either low-temperature steam or immersion in liquidchemical disinfectants. Consideration should be given to the requirements for the safe useof such chemicals as alcohols, and aldehydes such as glutaraldehyde and peracetic acid.The following agents are commonly used as disinfectants, antiseptics, and sterilants in themedical community.

Alcohols

Aliphatic alcohols are antimicrobial in varying degrees by denaturing proteins.12

Ethanol (ethyl alcohol) in 70% concentration is bactericidal in 1 to 2 min at 30°C, butless effective at lower and higher concentrations. Ethyl alcohol, 70%, and isopropanol(isopropyl alcohol, IPA), 90%, are at present the most satisfactory general antiseptics forskin surfaces. They may be useful for sterilizing instruments but have no effect on spores,and better agents are available for this purpose.21 Alcohols are often combined withother disinfectants, such as iodine, mercuirals, and cationic detergents for increased effec-tiveness.


Aldehydes

Formaldehyde in a concentration of 1 to 10% effectively kills microorganisms and theirspores in 1 to 6 h. It acts by combining with and precipitating protein. It is too irritating foruse on tissues, but it is widely employed as a disinfectant for instruments.22 Formaldehydesolution USP contains 37% formaldehyde by weight, with methyl alcohol added to preventpolymerization, is extremely active, and kills most forms of microbial life.

Glutaraldehyde has been available for more than 30 years and its use was first regu-lated by the EPA. Standards of care for the use of glutaraldehyde products are availablefrom specialty organizations such as the Society of Gastroenterology Nurses Association(SGNA) and the Association of Operating Room Nurses (AORN), which have significantexperience and interest in the safe and effective use of liquid chemical disinfectants/steri-lants.23 These guidelines all recommend a 20 min soak in a 2% glutaraldehyde solution aftera meticulous manual precleaning with an appropriate detergent. When used with theguidelines and precautions available, glutaraldehyde is an appropriate and effective high-level disinfectant.24

Glutaraldehdye as a 2% alkaline solution in 70% isopropanol (pH 7.5 to 8.5) serves asa liquid disinfectant for most optical and other instruments and for some prosthetic mate-rials. It kills viable microorganisms in 10 min and spores in 3 to 10 hours, but the solutionis unstable, and tissue contact must be avoided. It must be rinsed from instruments andother items with sterile water before use. It is the “cold” sterilant of choice for lensed instru-ments.

Halogens and Halogen-Containing Compounds

Elemental iodine is an effective germicide and is most often used as a skin disinfec-tant.25 Its mode of action is not definitely known but does combine with cell protein and isan active germicidal agent with a moderate activity against spores. Iodine tincture USPcontains 2% iodine and is the most effective antiseptic available for intact skin.26 Its princi-pal disadvantage is the occasional dermatitis that can occur in hypersensitive individuals.

Iodine can be complexed with polyvinylpyrrolidone to yield povidone-iodine USP, aniodophore. This is a water-soluble complex that liberates free iodine in solution. Likeiodine, povidone-iodine is widely employed as a skin antiseptic and used as a surgicalscrub.26 Povidone-iodine solutions can become contaminated with Pseudomonas and otheraerobic Gram-negative bacteria. Iodine and povidone-iodine both are rarely used to disin-fect instruments or medical devices.

Chlorine has been used for many years as a disinfectant and has generally been asso-ciated with treatment of swimming pool water and water supplies. It exerts its antimicro-bial action in the form of undissociated hypochlorous acid (HOCl) and acts by oxidizingthe cell membrane.25 Hypochlorous acid is formed when chlorine is dissolved in water.Chlorine concentrations of 0.25 parts per million (ppm) are effectively bactericidal formany microorganisms except for mycobacteria.27 Organic matter greatly reduces theantimicrobial activity of chlorine.

Instrument disinfectants containing chlorine dioxide, known as Tristel, Dexit, andMedicide, are commercially available. These products contain two components, a base andan activator, requiring addition and dilution in accordance with the manufacturers’instructions. Freshly prepared chlorine dioxide is highly effective and rapidly destroys bac-terial spores.28 Sporicidal activity is maintained for 7 to 14 days provided the disinfectantis stored in sealed containers with minimal head space above the solution. Unfortunately,


chlorine dioxide is also more damaging to instrument and processor components thanaldehydes. A discoloration of the black plastic casings of flexible endoscopes has beenreported, but this change may be only cosmetic.

Quaternary Ammonium Compounds

Cationic surface-active agents (quaternary ammonium compounds) such as benzalko-nium chloride USP and cetylpyridinium chloride USP are both detergents and bactericidalagents.26 These agents are bactericidal probably by altering the permeability characteristicsof the cell membrane.29 Benzalkonium chloride (Zepharin) may be used to disinfect instru-ments. A 20-min soak is recommended. However, their toxicity seems to be somewhathigher than that of the biguanides described below.

Biguanide Disinfectants

All of the biguanides have a wide spectrum of antibacterial activity against both Gram-positive and Gram-negative bacteria.30 They have a high kill rate, but toxicity towardsmammalian cells is very low and irritancy is so insignificant that the biguanide disinfec-tants can be used on the sensitive mucosal surfaces.

Chlorohexidine is a bisdiguanide antiseptic that disrupts the cytoplasmic membrane,especially of Gram-positive organisms. It is employed as a skin cleanser, as a constituent ofantiseptic soaps, and as a mouthwash for combating plaque-inducing bacteria.Chlorohexidine is rarely used on inaminate objects and is neutralized by alcohol, thereforethe two agents should not be used in combination.31

Another biguanide-type disinfectant that has recently been used in contact lens solu-tions is polyaminopropyl biguanide (PAPB). This disinfectant appears to be very effectivewith a wide spectrum of antibacterial activity against both Gram-positive and Gram-neg-ative bacteria. Toxicity toward mammalian cells is very low and irritancy is so insignificantthat the biguanide antiseptics can be used on sensitive mucosal surfaces and the eye.

Phenol and Phenolic Derivatives

Phenol (5 to 10%) was the first disinfectant commonly used.12 However, because of itstoxicity and odor, phenol derivatives are now generally used instead.32 These includeorthophenylphenol, hexachlorophene, and hexylresorcinol.33 Phenol denatures protein.Concentrations of at least 1 to 2% are required for antimicrobial activity, whereas a 5% con-centration is strongly irritating to tissues. Therefore, phenol is used mainly for the disin-fection of inanimate objects and excreta.

Table 2 presents a summary and comparison of various liquid disinfectants used in themedical community.

ANALYTICAL TESTING

Although the retention of functionality is recognized as one of the most-pressing issuesto deal with in a decision to reuse, there is a scarcity of detailed scientific studies in the lit-erature that have investigated the deterioration of materials and function in reused dis-posables.


Table 2 Liquid Disinfectants Used in the Medical Community

Class Recommended Use Method of Action Advantages

Alcohols Cleaning some Changes protein Fairly inexpensive and(70% IPA or instruments structure of readily available.ethanol) Cleaning skin microorganism.

Presence of water assistswith killing action

Chlorine Spills of human body Free available chlorine Kills hardy virusescompounds fluids combines with contents (e.g., hepatitis)

Bactericidal—Good within microorganism, Kills a wide range ofFungicidal—Good reaction by-products organismsSporicidal—Good at cause its death Inexpensive

�1000 ppm Need 500 to 5000 ppm Penetrates wellRelatively rapid

microkillAldehydes (e.g., Bactericidal—Good Coagulates cellular Nonstaining, relatively

glutaraldehyde) Fungicidal—Good proteins noncorrosiveTuberculocidal— Usable as a sterilant

Excellent to plastics, rubber,Virucidal—Good lenses, stainlessSporicidal—Good steel, and other

materials thatcannot beautoclaved

Iodophors Disinfecting some Free iodine enters Kills broad range of(iodine semicritical microorganism and organismswith carrier) medical equipment binds with cellular Highly reactive

Bactericidal—Very components Low tissue toxicitygood Carrier helps penetrate Kills immediately

Fungicidal—Excellent soil/fat rather than byVirucidal—Excellent Need 30 to 50 ppm prolonged period of

Probably acts by disorder of timeof protein synthesis dueto hindrance and/orblocking of hydrogenbonding.

Phenolic Bactericidal— Overt protoplasmic poison Nonspecificcompounds Excellent Disrupts cell walls concerning

Fungicidal—Excellent Precipitates proteins bactericidal andTuberculocidal— Low concentrations fungicidal action

Excellent inactivate essentialVirucidal—Excellent enzyme systems

Quaternary Ordinary Affects proteins and Acts as a detergent toammonium housekeeping cell membrane of loosen soilcompounds (e.g., floors, microorganism Rapid action(QUATS) furniture, walls) Releases nitrogen and Colorless, odorless

Bactericidal— phosphorus from NontoxicExcellent cells Highly stable

Fungicidal—GoodVirucidal—Good (not

as effective asphenols)


Table 3 Thermal Analysis Techniques for Materials

Test Technique

Melting point DSCDegree of crystallinity DSCGlass transition temperature DSC, TGAComponent quantification TGA

Medical devices are commonly made of polymers or plastics of varying density. Low-density plastics are less resistant to heat than high-density ones. However, the propertiesof plastics are also obtained from additives that may stabilize the material to heat and lightor reduce costs (fillers), enhance beneficial properties such as abrasion, resistance andstrength, or provide lubrication, flexibility, or antifungal properties. Some of these addi-tives are susceptible to reprocessing and reuse and can be leached out, or their compositioncan be altered through exposure to light.20

Mechanical testing can be used to help determine the functional integrity of devices. Itmay be difficult to simulate in an experimental setting the various stresses a device may beexposed to in a clinical environment. Even if all the variables related to stress and strainwere known, there is no one set of tests that is suitable for all devices. However, a keymechanical test is the tensile test, which measures the force required to stretch a devicethrough a range of extensions.

The most widely used instrument for stress–strain measurement is the Instron TensileTester. This instrument is essentially a device in which a sample is clamped between gripsand jaws, which are pulled apart at constant stress rates.17 A variety of parameters aredetermined such as elongation, elongation at break, breaking strength, and tensile modu-lus of elasticity. The tensile modulus is defined as the ratio of stress to strain, and is deter-mined from the initial slope of the stress–strain curve. The modulus of a material is ameasure of the ability of a specimen to resist deformation. Tensile modulus is also referredto as Young’s modulus. Some plastic materials can be weakened by the process of sterili-zation or disinfection that leads to a decrease in the tensile strength of the material. Tensilestrength data collected before and after cleaning, decontamination, or sterilization can helppredict any changes in the strength of the material.16

Hardness testing is another mechanical test useful to determine the functional charac-teristics of material used in a medical device or instrument. Hardness is generally definedas an indication of the resistance to indentation, scratch resistance, and/or reboundresilience. International Standards Organization (ISO) standards report three methods formeasuring hardness: Shore hardness, a ball indentation method, and Rockwell hard-ness.34,35 It is important to emphasize that hardness values obtained from one method ingeneral cannot be compared with those derived from another, although data can be empir-ically compared. Generally, hardness is used as an end-performance property of materialused in a device. The instrument used to obtain the measurement is called a durometer.

Thermal analysis, the response of a polymer to controlled heating processes, is a fam-ily of techniques widely used in the development and characterization of materials, includ-ing plastics and elastomers. Characterizing the glass transition temperature, melting point,and extent of crystallinity of a polymer is important and often used to produce materialswhose properties are tailored to the ultimate application of the product.36 The primary ther-mal analysis techniques for certifying product quality are differential scanning calorimetry(DSC) and thermogravimetric analysis (TGA). Thermal analysis tests for certifying prod-uct quality are listed in Table 3.


Differential scanning calorimetry results can give rapid measures of thermal charac-teristics, such as the polymer melting point and glass transition temperature. The nature ofthese transitions, in addition to identifying characteristics unique to each polymer, can alsoprovide information about its phase structure, thermal history, and purity. The loss of poly-mer additives can have a dramatic effect on these thermal properties.

Chromatographic Analysis

The various chromatographic methods, such as gas, liquid, paper, and thin-layer, havebecome indispensable aids in the isolation, separation, and determination of chemicals.Gas and liquid chromatography, especially high-performance liquid chromatography(HPLC), have become powerful analytical tools to characterize additives in polymericmaterials used in medical devices, as well as disinfectants and liquid sterilants adsorbedonto surfaces.16

HPLC is undeniably one of the fastest growing and most useful of all the analyticalseparation techniques. The reason for this growth is attributable to the sensitivity of themethod, its ready adaptability to accurate quantitative determinations, and its suitabilityfor separating nonvolatile species or thermally fragile ones. Liquid chromatography,because of its great flexibility and wide-spread applicability, can be used for the analysis ofmore than 80% of all known organic compounds.17

Spectrophotometric Analysis

Infrared

Infrared (IR) instruments measure the vibrational spectrum of a sample by passing IRradiation through it and recording which wavelengths have been absorbed and to whatextent. Since the amount of energy absorbed is a function of the number of molecules pres-ent, the IR instrument provides both qualitative and quantitative information. Since the IRspectrum of a chemical compound is perhaps its most characteristic physical property, IRfinds extensive application in identifying substances and their respective concentrations.Many cleaning and disinfecting agents have distinctive IR spectra that can be used to iden-tify residuals on the surface of devices and instruments.37

If residual glutaraldehyde is present on a medical device, then it can be extracted in asolvent and the solvent analyzed for glutaraldehyde. Most detergents and enzyme (pro-tein) solutions can be identified by this analytical method.

Ultraviolet/Visible Spectroscopy

Much like IR, the ultraviolet (UV) and visible spectrum of a chemical can be used toidentify and determine the concentration of analytes in an extract. Enzymes, which are pro-teins, are very easily detected and their concentration determined by their distinct UVspectrum.20 Most proteins absorb maximally at 280 nm.

CONCLUSION

Successful cleaning, decontamination, and sterilization/disinfection of medicaldevices and instruments requires both the careful selection of materials and consistentmonitoring of the procedures used to process them. Chemical characterization of residues


that may be adsorbed onto surfaces and mechanical testing to ensure functionality shouldprovide sufficient information to evaluate the potential success of cleaned and disinfec-ted/sterilized medical devices. By using a combination of chemical and mechanical analy-sis techniques, both manufacturing and decontamination processes can be optimized toensure a safe and effective product. It is important to note that no test, however foolproofits design, can ever be considered a definitive predictor of clinical performance.

REFERENCES

1. Hart, P.D., Decontamination compliance of porous surfaces—is it attainable? Infect. Control Steril.Technol., 5(3), 43–44, 1999.

2. Caballes, N., de Guzman, D.R., and Voorhis, J.V., An infection control perspective of antibioticresistance, Infect. Control Steril. Technol., 3(10), 18–26, 1997.

3. Avalos-Bock, S., Small hospitals face special challenges in infection control, Infect. Control Steril.Technol., 4(1), 24–27, 1998.

4. Designing, Testing, and Labeling Reusable Medical Devices for Reprocessing in Health CareFacilities: A Guide for Device Manufacturers, AAMI TIR 12-1994, Association for theAdvancement of Medical Instrumentation, Arlington, VA, 1995.

5. U.S. Department of Labor, Occupational Safety and Health Administration. OccupationalExposure to Blood Borne Pathogens; Final Rule,. 29 C.F.R. Part 1910, 1030, December 6, 1991.

6. Spaulding, E.M., Chemical Disinfection of Medical and Surgical Materials. Disinfection,Sterilization and Preservation, 1986, 517–531.

7. Alvarado, C.J., Revisiting the Spaulding classification scheme, in Chemical Germicides in HealthCare, Proc. Int. Symp., Rutala, W.A., Ed., Cincinnati, May, 1994, , 203–209.

8. DeSchutter, C. and Ritz, S., Processing minimally invasive instruments, Infect. Control Steril.Technol., 2(12), 26–28, 1996.

9. Block, S.S., Disinfectant, Sterilization, and Preservation, 4th ed., Lea & Febiger, Philadelphia, 1991.10. Jendresky, L., Garcia, R., Clarke, A., and McDonagh, W., Performance improvement of disinfec-

tion and sterilization systems: a case study, Part II, Infect. Control Steril. Technol., 3(4), 44–53; 1997.11. Rutula, W. et al., FDA labeling requirements for disinfection of endoscopes: a counterpoint, Infect.

Control Hosp. Epidemiol., 231–235, 1995.12. Jawetz, E., Disinfectants and Antiseptics in Basic and Clinical Pharmacology., 4th ed., Appleton &

Lange, Norwalk, CT, 1989.13. Miner, N., Reuse of single-use only devices, Biomed. Instrum. and Technol., 33(1), 5–6, 1999.14. Weller, I.V.D., Williams, C.B., Jeffries, D.J., Leicester, R.J., Gazzard, R.G., Axon, A.T.R., Hanson,

P.J.C., Ayliffe, G., Barrison, I., and Neumann, C. Cleaning and disinfection of equipment for gas-trointestinal flexible endoscopy: interim recommendations of a working party of the BritishSociety of Gastroenterology, Gut, 29, 1134–1151, 1988.

15. Vaughn, S., Disinfectants and their use, in Proceedings of MARE Seminars, 1993.16. Albert, D.E. and Wallin, R.F., A practical guide to ISO 10993–14: materials characterization, Med.

Dev. Diag. Ind., 20, 96–99, 1998.17. Albert, D.E., Materials characterization as an integral part of global biocompatibility, Med. Plast.

Biomater., 4, 16–23, 1997.18. Feldman, L. and Hui, H., Method of Protecting and Sterilizing Aluminum Surfaces on Medical

Instruments, U.S. patent pending, 1997.19. Sloan, F., Chemical attack of graphite/epoxy by hydrogen peroxide, Appl. Spectros., 46, 524–528,

1992.20. Albert, D.E., and Hintz, M.A., in Material Properties and Bioreactivity in Spinal Drug Delivery, T.

Yaksh, Ed., Elsevier, New York, 1999, 395–406.21. Prescott, Harley, and Klien, Microbiology, 2nd ed., Wm. C. Brow Publishers, Dubuque, IA, 1993.22. Basett, R.C., Formaldehdye in Biological Monitoring Methods for Industrial Chemicals, 1980,

144–147.23. Casey, R.L., Glutaraldehyde: concern for use, Infect. Control Steril. Technol, 4(2), 12–16, 1998.


24. Mbithi, J.N., Springthorpe, V.S., and Sattar, S.A., Bactericidal, mycobactericidal and virucidalactivities of reused alkaline glutaraldehyde in an endoscopy unit, J. Clin. Microbiol., 31,2988–2995, 1993.

25. Frobisher, M., Hinsdill, R.D., Crabtree, K.T., and Goodheart, C.R., Sterilization and disinfection:practical applications, in Fundamentals of Microbiology, 1974, 302–318.

26. U.S. Pharmacopeia, Vol. 23, NF18, U.S. Pharmacopeia Convention, Washington, D.C., March 8–10,1990, published by the Board of Trustees, January 1, 1995.

27. Holton, J., Nye, P., and McDonald, V., Efficacy of selected disinfectants against mycobacteria andcryptosporidia, J. Hosp. Infect., 27, 105–115, 1994.

28. Ossia-Ongagna, Y. and Sabatier, R., Comparison of in vitro activity of six disinfectants on bacte-ria from contamination in haemodialysis water, J. Pharm. Belg., 48, 341–345, 1993.

29. Tennant, J.M., Lyon, B.R., Gillespie, M.T., May, J.W., and Skurray R.A., Cloning and expression ofStaphylococcus aureus plasmid-mediated quaternary ammonium resistance in Escherichia coli,Antimicrob. Agents Chemother., 27, 79–83, 1985.

30. Brecx, M. et al., Efficacy of Listerine, Meridol and chorohexidine mouth rinses on plaque, gin-givitis and plaque bacteria vitality, J. Clin. Periodontol., 17, 292–296, 1990.

31. Vigeant, P., Loo, V.G., Bertrand, C., et al., An outbreak of Serratia marcescens infections related tocontaminated chlorohexidine, Infect. Control Hosp. Epidemiol., 19, 791–794, 1998.

32. Best, M., Sattar, S.A., Springthorpe, V.S., and Kennedy, M.E., Efficacies of selected disinfectantsagainst Mycobacterium tuberculosis, J. Clin. Microbial., 28, 2234–2239, 1990.

33. U.S. Pharmacopeia XXIII.34. ASTM D2238, D2583, in Annual Book of ASTM Standards, 08.02—Plastics (II), American Society for

Testing and Materials, Philadelphia, 1991.35. ISO, International Standard 868, 1985.36. DiVito, M.P., Fielder, K.J., Curran, G.H., and Feder, M.S., Recent advance in routine thermal

analysis instrumentation, Am. Lab., 24(1), 30–36, 1992.37. Palley, I. and Signorelli, A.J., Physical testing, in A Guide to Materials Characterization and Chemical

Analysis, Sibilia, J.P., Ed., VCH Publishers, New York, 1988, 273–284.


Documents

1655_05.pdf