|
IV. Cleaning
procedures
The cleaning procedures outlined in this section will be divided into
three broad categories: acid solution cleaning, alkaline solution cleaning,
and dry cleaning. The first two
categories involve liquid cleaning, whereby particle removal may be achieved
at the same time as surface activation.
The third category will have little or no impact on particle removal,
but will generally provide a rapid and simple way of activating surfaces.
The alkaline cleaning generally involves a light etching of the glass
surface. The acid cleaning,
although it may leach components from the glass surface, generally does not
involve an etching of the glass.
The advantages and disadvantages of these cleaning procedures may be
summarized as follows. The
liquid cleaning procedures are simple to set up and rapid to use.
However, they generate considerable waste products and require either
the use of significant operator time, or the use of sophisticated machines.
As such, they are well suited to the cleaning of occasional samples
in research laboratories.
A
strong caveat for these techniques is the requirement of rinsing and drying
the samples. The latter step is
delicate and invariable compromises the obtaining of a uniform, clean and
dry surface on the substrate.
Vapor-degreasing
(Pul84), is not discussed. This
technique relies on the condensation of vapor from a freshly-distilled
solvent. It may remove dust and
organic contaminants. It will
not generally provide a glass surface that is uniformly wettable to water,
due to residual organic molecules that remain adsorbed to the glass surface.
The dry cleaning procedures require the investment in suitable
equipment. They have the
advantage of being able to clean large quantities of samples with little or
no waste products and little operator input.
They also have the distinct advantage of directly generating a clean,
uniformly wettable and dry surface on the substrate.
IV.1 Acid
solution cleaning
This is one of the most commonly used procedures for cleaning glass
and silica surfaces in small research laboratories.
It has the advantage of being easy to set up, requiring the purchase
of basic chemicals and the use of a fume hood.
However, it generates significant toxic waste products and is thus
rarely scaled up for industrial applications.
The basic concept is that of mixing a concentrated acid with a strong
oxidizing agent.
The original formulation was the use of a saturated potassium
dichromate solution in concentrated sulfuric acid. The solution, known as chromerge, was prepared by completely
dissolving 20 grams of potassium dichromate (K2Cr2O7)
in 90 grams of water. To this
was slowly added 900 grams of concentrated sulfuric acid. The mixing reaction is highly exothermic, exposing the
operator to a high risk from spattering.
This solution could be re-used as long as it remained brown, rather
than green, indicating the oxidation state of the dichromate.
The sample was dipped in the chromerge for 20 minutes,
followed by 20 minutes in a 1:1 concentrated hydrochloric acid/water
solution to remove chromium ions from the surface.
A pure water rinse and the drying of the sample surface follow this.
The latter should be practiced as described in the particle removal
section above, taking care not to allow water streaks across the surface.
The use of chromerge has been all but completely banned in most
research and industrial establishments. This is because the chromium ion is highly toxic to the environment
and poses a severe waste disposal problem, even in small quantities. Further, the hexavalent chromium present in the above solutions is
considered by regulatory agencies to be a potent human carcinogen. The addition of chloride or halogens to chromerge solutions can
generate the highly toxic and volatile carcinogen, chromyl chloride. This mandates the use of chromerge under a chemical fume hood.
Chromerge cleaning, as described in the paragraph above, was found to
be adequate for generating a uniformly wettable surface on the native oxide
layer of a silicon crystal. This
cleaned substrate, while still wet, showed uniform deposition of a
charge-adsorbed monolayer of surfactant (Bir94). To obtain a uniform dry surface, this monolayer was burned
off with UV/ozone radiation. This
gave excellent results for the deposition of a monolayer of anionic
surfactant from a thinning film of surfactant solution (Bir95). Further, the re-hydration of this surfactant layer showed a
relatively uniform thickness, indicating uniform surface properties of the
hydrated substrate surface.
As a replacement to the highly successful but toxic chromerge
solution, Godax laboratories (Takoma Park, MD, USA) has developed a product
called Nochromix .
This is reportedly dispensed in a powder form, costing approximately
$2.50 per bag. It is added to
concentrated sulfuric acid. As the solution turns brown, the solution is spent and a new
package may be added to the sulfuric acid.
Due to the hygrosopic nature of sulfuric acid, it is important to
keep concentrated sulfuric acid solutions closed to air while not in use. This avoids their dilution with water following prolonged exposure to
air.
Prior to the commercial existence of Nochromix , an efficient alternative to chromerge was
developed. This consisted of
mixing two parts concentrated sulfuric acid with one part hydrogen peroxide
solution, providing the strong oxidizing agent. The mixing process is strongly exothermic, resulting in an extremely
hot final solution. Cleaning
with this mixture is recommended at 90° C for 20 minutes. This is followed by a rinse in pure water and careful drying
as described above. This
solution removes hydrocarbons from the surface of the substrate with vigor. Care should be taken when using the hot cleaning solution.
It should be used under a fume hood, together with a protection
against spills from the breaking of glass containers.
Verification of a successful cleaning procedure following acid
cleaning is essential. The
sample should be completely wettable to water during the pure water rinse. Wettability of a surface to water in this context implies that a film
of water on the surface does not dewet. Dewetting is observed when dry patches are formed where the liquid
film has receded from an area of surface and they are characterized by the
presence of a liquid film with thickness of order one millimeter adjacent to
a "dry" zone (by visual observation). This may be further verified, if necessary, by halting the water flow
and holding the substrate vertically. The
water film will then thin. As
it thins, light interference fringes should be observed. If dry patches are formed before the interference fringes are seen,
the glass is not completely wetting in these areas and the cleaning has
failed. On completion this
test, the glass should immediately be re-wet with water and dried carefully
as described above.
Finally, these acid cleaning procedures are not effective for the
complete removal of fluorocarbon molecules, such as perfluoro-silanes
grafted to a glass surfaces. Further,
poly-dimethyl siloxane (PDMS, or silicone oil) contaminants are not well
removed using these acid cleaning solutions. PDMS contamination is frequently found in environments where
vacuum pump oils or silicon rubber seals are used.
The PDMS molecules adsorb strongly to glass surfaces and are
generally not removed by surfactant solution cleaning or by acid cleaning.
Care should be taken that no PDMS or fluorocarbon silane molecules
come into contact with the acid cleaning solutions, as they may contaminate
the solutions to a point where they will no longer render glass hydrophilic
and may increase its hydrophobicity.
IV.2
Alkaline
solution cleaning
These solutions, generally made from sodium hydroxide (or potassium
hydroxide) dissolved in water, are designed to lightly etch the glass
surface. Solutions above pH 9
begin to etch glass surfaces. The
manufacture of an aqueous NaOH solution is a relatively simple procedure. Its etching of the glass ensures the removal of surface residues,
resulting in a clean surface. Glass
may cleaned by exposing it to a saturated aqueous sodium hydroxide solution
at room temperature for fifteen minutes. Should the glass be coated with a non-wetting layer, this exposure
may need to be longer for the solution to attack the glass under the
non-wetting layer. Alcoholic
solutions of NaOH or KOH may be used to improve the wetting properties of
the solution.
The cleaning with
alkaline solution should be followed by a five minute dip in 1M hydrochloric
acid to neutralize the alkali attack on the glass surface. This should be followed by a twenty
minute dip in pure water. The
sample should then be checked for surface cleanliness by wettability, as
described in the above section, followed by a careful drying in air, as
described in the particulate removal section.
An efficient alternative to concentrated sodium hydroxide solution is
the use of a surfactant known as Hellmanex.
Developed by Hellma (Müllheim, Germany), this surfactant was
designed to clean quartz spectroscopy cells, without altering their optical
properties and leaving no residue. Its
recommended use is a one hour dip at room temperature or a 30 minute dip at
30° C for a 2% solution by weight in water. In practice, ten to twenty minutes under ultrasonic agitation have
given satisfactory results. Hellmanex
surfactant yields an alkaline solution of small molecule surfactant. The alkaline solution contributes to a light etching of the glass
surface, while the surfactant ensures wetting of the substrate, even if
coated with hydrophobic contaminants. The
use a small molecule surfactant is believed to contribute to the absence of
any surface residue after rinsing. Following
cleaning, it may be useful to soak the substrate in pure water for twenty
minutes to leach out any small molecules absorbed into the glass surface. This is followed by a pure water rinse and a careful drying, as
described in the particle removal section, above. We have found this cleaning solution to exhibit excellent cleaning
properties, coupled with a simple application process. Another similar alkaline small surfactant molecule cleaning powder is
sold under the trade names of Alconox.
Finally, glass surface may be etched by exposure to a solution of
fluoride ions, such as that generated by an aqueous solution of hydrofluoric
acid. These solutions have been
included in this section, despite their being acidic, because they rely on
etching the glass surface to achieve cleaning. The fluoride ions rupture siloxane bonds.
These cleaning solutions are generally dilute hydrofluoric acid, of
the order of a few percent in HF concentration. The concentration and cleaning time should be adjusted to the glass
composition in order to achieve cleaning without excessive etching. The solutions are generally used at room temperature. More concentrated HF solutions are used to dissolve
macroscopic pieces of glass. These
concentrated HF solutions should not be used to clean glass surfaces, since
their rapid etching increases the surface roughness of the glass. Great care should be exercised when handling hydrofluoric acid, even
in dilute solutions. It is
extremely toxic and skin exposure (or inhalation) can be fatal. The acid penetrates the skin and migrates in towards the
bones. During its passage, it
neutralizes nerve endings, causing no pain or feeling. On reaching the bone, a decalcification takes place.
This may lead to a necrosis and a possible amputation.
Exposure of skin over an area approximately larger than the palm of a
hand may result in heart failure. Thus,
the use of hydrofluoric acid is not recommended outside of well-controlled
environments.
IV.3
Dry cleaning
This category encompasses a several processes that clean a glass
surface without touching it or significantly altering its surface
composition. The machines used
and their safety protection require a significantly larger capital
investment than the above chemical cleaning procedures for cleaning small
numbers of glass samples. They
rely on high temperatures or the use of plasma to degrade and desorb
contaminant hydrocarbons from the glass surface.
The
dry cleaning
procedures present the significant advantage of being non-contact cleaning
and resulting directly in a clean and dry substrate surface, ready to
receive a subsequent coating. They
generate few waste products and require minimal operator input, making them
highly attractive to the scaling up required by industrial processes. Their efficiency at cleaning different types of contaminants
from a variety of substrates makes them attractive for use in research
laboratories.
One of the most basic processes for cleaning glass surfaces is the
use of pyrolysis. This involves
heating the glass to a temperature above 300° C, generally 500 ° C for
over 30 minutes. A typical
cycle will consist of a rise in temperature from room temperature to 500° C
over four hours, followed by five hours at 500° C and a return to room
temperature over five hours. The
slow increases and decreases in temperature are designed to avoid glass
breakage from thermal shock. This
process generates few waste products. It
relies on the degradation and desorbtion of organic contaminants from the
surface of the glass. It has
been found to be effective in increasing the wettability of glass surfaces
contaminated with any organic pollution, including such molecules as PDMS or
fluorocarbons. This process may
generate soot residue and is not effective in removing this soot. When using this process, it is important that the residual
contamination on the glass surface be of the order of nanometer thickness or
less. This is in order to avoid
macroscopic residues on the glass surface, such as carbon or other inorganic
matter. This technique does not
remove dust particles from the glass surface. Care should be used when pyrolyzing sodalime glass, where the
presence of moisture may leach alkalis from the glass, generating a sodium
hydroxide residue on the glass surface. This is generally not a problem when using air with ambient humidity,
provided that air circulation around the samples is limited.
A more efficient process is the use of plasma to clean substrates. This process is suitable for industrialization and has been used in
some large industrial processes. The
cleaning may involve an oxygen plasma to directly oxidize hydrocarbons, or
an argon plasma to degrade them and desorb them. Such processes are suitable for cleaning glass surfaces and for
activating polymer surfaces. The
use of a plasma bears the distinct advantage of being able to penetrate
inside complex structures. Generating
plasma requires a specifically designed reactor. Generally, the specific gas to be used is injected to form an
atmosphere at reduced pressure. Electromagnetic
radiation (RF or microwave frequency) is then coupled into the enclosure,
generating the plasma.
Some
plasma systems operate at atmospheric pressure, removing the need for a
vacuum-sealed enclosure. Plasma
cleaning has an interesting feature, whereby its, since its
"effective" temperature (its kbT equivalent thermal
energy) is of several thousand degrees. This is achieved by coupling energy into the gaseous phase without
strongly increasing the substrate temperature. As an example, this process may clean carbon residue from surfaces
without excessive heating.
The plasma cleaning techniques also find application in the
activation of other inorganic substrate surfaces and organic polymer
surfaces.
Other
dry cleaning techniques
One reported technique for cleaning glass surfaces is the use of
laser cleaning (Wür86). This
relies on an infrared high energy beam to irradiate the surface and desorb
contaminants. While no
wettability data are given in the report, the technique is reported to
remove organic contaminants that have diffused into the glass surface, lying
at depths of order 10 angstroms.
Rendering glass surfaces hydrophilic and highly wettable may also be
achieved using ultraviolet (UV) radiation and ozone, as described by John R.
Vig (Vig85, Vig87). This
technique may be considered as an evolution of corona discharge cleaning,
whose use dates back to 1956 (Pul84).
UV/ozone cleaning is most easily implemented using a
low-pressure mercury lamp, emitting radiation at the wavelengths of 184.9
and 253.7 nm. The first of
these generates ozone from oxygen , while the second combines with ozone to
oxidize hydrocarbon contamination at the glass surface. Such a lamp may be obtained from BHK Inc. (Claremont, CA, USA).
It will need to be mounted in an enclosure allowing the sample to be
positioned 1-3 cm from the lamp and protecting the user from UV radiation. The production of ozone in large quantities requires that the lamp be
mounted in an area ventilated to a fume hood.
Ozone in significant quantities is considered an environmental
pollutant and the use of a catalyst at the exit of the fume hood may be
recommended. Alternatively, a
complete UV/ozone cleaning equipment may be purchased from UVOCS (Montgomeryville,
PA, USA). The distance from the
sample to the lamp is critical, since wavelength used to oxidize
hydrocarbons is absorbed by the ozone. One to three centimeters are found to give good results.
A non-optimized cleaning system will result in exposure times of 15
to 30 minutes to obtain a wettable glass surface.
Optimizing the system may reduce this time to below two minutes. This process provides a high degree of versatility and ease of use.
It cleans substrates without contact in relatively short times
without strong heating. The
sample temperature generally does not rise above 60° C. The system may be used to activate the surface of polymer substrates.
Its effectiveness on fluorocarbons is lower than on hydrocarbon
surfaces, presumably due to its oxidation mechanism and the UV resistance of
fluorocarbon molecules.
The
UV/ozone cleaning will effectively remove organic contamination layers with
thickness on the order of one nanometer. For coatings with thickness of order one micron, only the surface of
the coating will likely be rendered hydrophilic. For this reason, it is important to pre-clean the substrate with
organic solvents or surfactant solutions to degrease and remove macroscopic
contaminants. This cleaning
process will not remove dust or other particles from the substrate surface.
V. Influence of Cleaning
The previous section described a variety of procedures for rendering
a glass substrate hydrophilic and exposing its silanol groups for reaction
with the molecules in the sol-gel coating. While all these procedures are designed to obtain a hydrophilic glass
surface, wettable to aqueous and alcoholic solutions, the cleaning processes
impact the glass surface, leaving different states for the glass surface. The glass surface may be altered in its chemical composition, as well
as its roughness. Exposure to
aqueous solutions may leach components from the glass surface, significantly
altering its surface chemical properties. Etching solutions may significantly increase the physical roughness
of glass surfaces, from a few angstroms rms for fire-polished glass, to nm
or even micron roughness.
This
section will briefly discuss the influence of surface cleaning on the glass
surface and its application.
A glass surface is not inert.
The
cleaning procedure used will generally alter the glass surface. A clear example of this is the influence of acid cleaning on sodalime
glass surfaces. The sodalime
glass contains 13 % sodium oxide. This
component, as well as the calcium oxide in the glass, are leached from the
surface region by exposure to the acid solution. An acid-cleaned sodalime glass surface is closer to a silica surface
than the as-formed glass surface. Conversely,
cleaning a sodalime glass surface under a UV/ozone lamp will have little or
no influence on the glass surface composition. If pyrolysis is used, where the sodalime glass is taken to 500° C,
just below its softening point of about 550° C, this may re-generate the
glass surface composition to a state close to that found on the as-formed
surface. Lelah and Marmur
(Lel79) describe the wettability of sodalime glass surfaces, measured by
observing the spreading rate of water drops, following a combination of
liquid cleaning and pyrolysis, covering several cleaning procedures.
Depending on the liquid cleaning procedure used, the wettability of
the sodalime glass is either determined primarily by the final pyrolysis
step, or is influenced by the prior liquid cleaning. Differences in the glass surface behavior following heat cleaning may
be associated with a leaching of soluble components, such as sodium oxide,
where the heat treatment is not sufficiently long or at sufficiently high
temperature to re-generate the original glass surface composition.
We have found that the contamination of glass surfaces following
cleaning may depend strongly on the procedure used to clean the glass.
For this experiment, three typical glass surface compositions were
used. Sodalime glass, in the
form of microscope slides from Erie Scientific; aluminoborosilicate glass,
in the form of Corning code 1737F glass; and silica surfaces, in the form of
the native oxide layer on silicon wafers, supplied by Unisil Corporation
(p-doped or n-doped four-inch wafers from standard production), were
examined.
The three glass
species were cleaned using chromerge and were then passivated in
hydrochloric acid, as described in the "Acid Cleaning" section
above. These samples were
placed in contaminated octane liquid and the contact angle of sessile water
drops measured as a function of immersion time. While the contact angle of each deposited water drop remained
constant, the water drops deposited at later times showed higher contact
angles. This increase was seen
over a period of five to seven minutes, after which the contact angle values
were fairly constant. Following
removal from octane and blow-drying, water sessile drop contact angles on
the substrates were of the order of 30-50°. This indicates a contamination of the substrates.
The three acid-cleaned substrates behaved in a similar manner. However, when the substrates were cleaned by pyrolysis, it was found
that the silica surface was contaminated more strongly (the water contact
angle under octane went up from about 20° to about 60°), while the
sodalime glass remained wettable to water, indicating no strong adsorption
of contamination from the octane. The
pyrolysed aluminoborosilicate glass showed a similar contamination behavior
to the acid-cleaned substrate.
All
of these surfaces were hydrophilic following acid or pyrolysis cleaning.
However, the difference in their susceptibility to contamination
indicates significant differences in their surface composition.
Lelah and Marmur (Lel79) report differences in the wettability of
cleaned glass, following exposure to ambient air, varying with the technique
used to clean the sodalime glass.
For coating applications, we have generally found a trend in quality,
with silicon wafers providing the best substrate, followed a close second by
Corning code 1737 glass, and finally sodalime glass. All three glass species provide high quality substrates for the
adhesion of silane or sol-gel (silanol-based) coatings. As an example, when used as a substrate for silane coatings, the
native oxide layer on silicon surfaces has consistently shown the highest
coating quality results.
The
Corning code 1737 glass comes a close second, and sodalime glass shows
slightly poorer results. The
coatings used for this evaluation were self-assembled monolayers, where
defects generated by substrate inhomogeneities rapidly lead to measurable
wettability changes.
Mechanically polished glass surfaces present very different cleaning
issues from as-formed (or fire-polished) glass surfaces. The polishing process relies on using first coarse, then
progressively finer, abrasive powders to grind the surface to the desired
smoothness. Some of the powder
may remain embedded in the glass surface and latent scratches may appear as
the glass is cleaned and the powder removed.
Further, the polishing process may generate microscopic cracks in the
surface of the glass. These cracks may be amplified during the cleaning process.
Critical parameters for limiting the increase in roughness and
defects in the surface of polished glass include the hardness of the glass
and its chemical durability. The latter is primarily determined by the soluble components
in the glass composition. The
cleaning of such glass surfaces is described in (Gro83).
VI Conclusions
A large variety of cleaning procedures exit that render glass
wettable and free of particles (Lel79, Pul84).
Reviewing the literature, one rapidly learns that the cleaning
processes used vary with the subsequent application of the glass surface. In fact, some coating applications may tolerate a significant level
of organic contamination of the glass surface. In these circumstances, a less aggressive cleaning procedure may
enable the use of less toxic and less aggressive cleaning agents.
Finally a common point in all glass cleaning procedures is that the
surface should be coated or otherwise used as soon as possible after
cleaning (Pul84). Glass
surfaces, and more especially clean and wettable glass surfaces, have a high
surface energy. Thus, they have
a tendency to adsorb particulate and organic contamination from the ambient
environment. Storing the
surfaces in freshly oxidized aluminum containers may reduce the adsorption
of organic contaminant molecules from the ambient air (Pul84). Particulate
contamination will generally create unacceptable defects. Adsorbed organic contaminant molecules, generally less than a full
monolayer coverage, of order nm thickness, will generate a heterogeneous
wettability. This may lead to
non-uniform coatings, in particular if deposited from liquid media. We generally work with ambient contaminants.
The glass surfaces are never perfectly clean in ambient air. It is generally recommended that the glass surface be used within a
few minutes following its cleaning. Our
experiments generally work despite the presence of ambient contaminants on
the glass substrate. It is
desirable to minimize the risk of coating defects caused by random ambient
contamination.
|