Note: Descriptions are shown in the official language in which they were submitted.
~13~706
Today conventional pH-measurement is industrial processes includ-
ing water-conditioning and waste-control, is usually performed with a glass
electrode combined with a reference electrode and sometimes a thermometric
sensor, in a so-called combination electrode or electrode system. This
system is then electrically connected through conduits and junction boxes
to an amplifier system with read-out and/or recording and controlling
functions, in a separate location more easily accessible to controlling per-
sonnel. For the measurements thus obtained to be meaningful, the electrode
system has to be checked and maintained on a frequent schedule, most times
daily, sometimes weekly. Due to the technology of these electrode systems,
this maintenance has to bc performed carefully by specially trained technic-
ians - a cost factor of pH-control which can be higher than the initial cost
of the system itself over the life span of the electrode system, short as
that life span may be. Even more dlsconcerting is the fact that the tech-
nology applied in today's electrode systems may lead to sudden, u~predictable
breakdown of its pH-reporting function even with a frequent maintenance
schedule. Since such breakdown unforeseeably interrupts process control the
result wlll be losses for the manufacturer or, in waste control, pollution
problems for the community.
This invention provides a novel form of glass electrode. As men-
tioned in the book by Bates, mentioned in detail below, on p. 340 and 341;
in 1929, McInnes and Dole published details of manufacture of a glass elect-
rode with a flat pH-sensitive membrane welded on to a glass support tube
where the membrane was made so thin on purpose~ that it shimmered with inter-
ference colors. Cn the inside, a special electrolyte filling the support
tube and an inner electrode were used to transfer the EMF from the membrane
to the measuring instrument outside. Such a membrane was extremely fragile
and could hardly support the hydrostatic pressure of the inner liquid.
- 1 -
~.~
:. ,
`:
1~3~7~
Efforts were made in British Patent 492,936 to reinforce
this flat membrane as well as dispose of the inner electrolyte by backing
up the pH-sensitive membrane with a metal layer, serving both as a support
and an electric contact. For several reasons, disGussed herein, such con-
struction was not satisfactory and never became a commercial success, one
of the reasons being the vulnerability of the extremely thin membrane, which
was also the reason why the construction pioneered by McInnes and Dole was
abandoned in favor of a construction, described by Cary et al, United States
Patent 2,462,843 ~col 9, lines 50-66): a little glob of molten pH-sensitive
glass is picked up at the end of a glass support tube and blown into a
bulb - a technique which is still the present state of the art. Different
types of pH-glasses are used today, and were used by Cary but his standard
of comparison was bulbs blown of the classical glass composition called
7'Corning 015"*, previously used by McInnes and Dole. Although technically
the pH-sensitive bulbs could be blown to any diame~er and any wall thickness,
practice has always been to restrict the wall thickness to about 0.1 mm.
e largest diameter bulb with good mechanical strength with this wall
thickness is about 9 mm diameter ~see Simon, United States Patent 3,025,174
col. 2, lines 32-35, issued March 13, 1962). The reason for this limitation
on wall thickness can be found in the textbook: Determination of pH: Theory
and Practice, 2nd Ed. 1973, by Bates, on p. 351 as follows: "There is
apparently no connection between the resistance of the membrane and the
B pH-response. Pyrex glass, for example, does not develop a satisfactory
pH-function even when the resistance of the membrane is low, and electrodes
of ¢Corning) 015 glass fail to de~elop the theoretical response if the mem-
brane is too thick~ There appears to be a critical thickness for each glass,
above which voltage departures occur. This inoperative threshold thickness,
* Trademark
,~
~13~ 6
which is a function of hygroscopicity, ranges from 54 to 130 /um ~0.054 to
0.130 mm)". A footnote mentioned as the source of this information the
original publications of the National Bureau of Standards of 1951 and '52.
It has therefore, been the object of this invention to reduce
the frequency of scheduled maintenance and to virtually eliminate the known
sources of unpredictable breakdown. Instead, the electrode system may be
expected to work for an extended period without causing losses, until finally
gradual symptoms of aging will lead to replacement at one of the scheduled
check-ups.
In accordance with this invention, a solid state i~l-sensitive
electrode has an unsupported disc of an ion-sensitive material thin enough
to avoid excessively increasing its electrical resistance and appropriately
thick to resist chipping and cracking during electrode fabrication. A porsus
glass backing is attached to the elcctrode disc whereby stresses are absorbed
in the glass backing and not transmitted to the electrode disc. Soft yielding
metal conductive parts are connected to the electrode disc to further min-
imize stresses applied to the electrode disc.
The thickness of the disc is only limited by a resistance which
is not so great that it causes a voltage drop of more than 0.1% when the
operational amplifier ~ITIC) is applied to it, Such a thickness is, for
example, from 0.3 to as much as 2 mm and even more in thickness. A highly
efficient disc is, for example, from about 0.5 to 1.5 mm and particularly
1 mm thick. A particularly effective electrode has an interface between
the disc and the soft yield metal conductive contact parts in which ions of
measurable mobility in the ion-sensitive material are disposed.
The invention will now be described in greater detail and will be
better understood when read in conjunction with the following drawings in
which:
.
:
. :
1131'7~)~
Figure 1 is a schematic cross-sectional view in ele~ation of a com-
bination of electrode system which is the embodiment of this inventlon;
Figure 2 is a three-dlmensional view in cross-section and partly
broken away of a button glass electrode of this inventlon.
The major sources of unpredlctable fallure of present-day glass
electrode systems are:
Breakdown Factor A: Electrlcal leakage of high-impedance clrcuit;
Breakdown Factor B: Fragility of the necessarily thin pH-glass membrane.
On factor A: Condenstatlon of molsture or creepage of conducting
process llquld between external ground and metal parts dlrectly connected
to the output of the glass electrode shorts out the electromotoric force
~=EMF) of the glass electrode, slnce thls EMF ls developed over the hlgh
resistance ~108 to 101 ohms, depending on the process temperature) of the
glass membrane, even though this ls kept as thin as is practical. This high
source resistance requires an electrlcal insulation of 1012 or ~ore ohms of
the conducting clrcuit leadlng this electrical signal to the pH-amplifier
input. Evèn in cases where a miniaturized amplifier has been built into
the electrode system, there ls an exposed connector to enable exchange of
defectlve glass electrodes. By necessity this connector is accessible to
the hand of the maintenance man, hence to fatal traces of dirt and moisture.
Building the amplifier into the electrode system has here only shortened the
length of the hlgh impedance clrcuit, but not eliminated the essential possib-
llity for electrical leakage across gaskets and screw threads.
In order to radlcally cut out all sources of such electrical leak-
age, the inventlon provldes a water- and vapor-proof permanently sealed
enclosure in which both the glass electrode and an impedance-transforming
integrated circuit ~=ITIC) are hermetically sealed as a monollthic unit.
A llnear integrated circuit combining the advantages of very high impedance
1131~7ql6
MOS/FET input, single supply capability, small package in a TO-5 can and
very low price is the CA3130AT ~RCA-numbering) used in the voltage follower
mode. Similar in results is the LF155AH ~National Semiconductor Corp.) used
as a non-inverting unity-gain amplifier, although a moderate amount of gain
can be tolerated without loss of input impedance. The Circuit which connect-
ed the outpu~ of this ITIC with the outside world and process-controlling
equipment is still susceptible to leakage paths formed by condensate, dirt
or creeping solutions as was the electrode output before. But the ITIC has
a low output impedance of around 103 ohms or less, and to corrupt this out-
put, leakage paths would have to be of the order of less than 105 ohm. Con-
ventional conduit- and junction-box technology can casily prevent such leak-
age from occurring, which is a different problem altogether from that of
providing a 1012 ohms insulation for the glass electrode lead against the
vagaries of daily inspections by human h~nds.
The same technique which is used to hermetically seal the glass
electrode and its connection to an ITIC can be doubled, with the second glass
electrode functioning as a reference electrode. By dolng this, a sealed
complete combination electrode system can be achieved. Outside the hermetic
enclosure the reference electrode is in liquid contact with a gelled refer-
ence liquid, i.e. an ionic medium of known pH, which known pH is the refer-
ence.
In conventional glass electrode systems and combination electrodes,
the reference electrode is usually a metal-metalchloride system where the
metal chloride is to a high degree insoluble, built into a reference cell
containing a potassium chloride solution of specific strength. This solution
is in liquid or ionic contact with the process fluid through a so-called
"liquid junction" which may be a fine channel, a slit, a porous medium filled
with the potassium chloride reference liquid, or gelled reference liquid.
.
31706
Reference cells of this type, surrounding the stem of the glass electrode
in a concentric manner, and with junctions encïrcling the pH-sensitlve glass
measurement membrane, are well known. However, the use of the process liquid
itself, brought to its desired endpoint pH, as a reference liquid for a solid-
state glass electrode as reference electrode, used with a similar solid-state
glass electrode as pH-measurement electrode is novel and has several unique
advantages.
If now, during process control the process liquid is brought to
its proper pH endpoint, we will haveJ at this moment of correct endpoint,
at the electrode system the following potential-forming chain:
Process fluid at endpoint pH/measuring glass electrode/lnput of
ITIC/ reference glass electrode/gelled endpoint-pH process fluid.
This chain if fully symmetrical, hence should have zero volts out-
put. More important: it will have this output at any process temperature,
regardless of ~he changes in absolute pH of the fluids inVolVed due to
changes in temperature-dependent ionic dissociation equilibria. At every
process temperature, the manufacturer will be assured that at zero volts
output of the electrode system, the process fluid has the same composition,
pH-wise~ as the ideal endpoint fluid prepared in his lab. Deviations from
zero volts output clearly and unambiguoùsly indicate that the process liquid
is not ~yet~ at its desired endpoint and requires further pH-control.
Gelling of the reference liquid can be achieved by adding fumed
silica or carboxymethyl cellulose or any other practical gelling agent to
the reference liquid. In those installations where it is more desirable to
have an output of zero volts at pH 7, as is common with many electrode sys-
tems today, a gelled buffer of pH 7 can be used as a reference gel. At any
pH-value outside of that of the reference gel, the BMF produced by this com-
bination electrode system will be identical to the EMF produced by a convent-
ional electrode system.
-- 6 --
.
'~
;
1131~7~)6
Every reference llquid has to have unimpeded ionic contact with
the liquid to be measured. ~sually this contact is provided by a potassium
chloride ~=KCQ) solution of specific strength. So-called "sealed reference
electrodes" utili~ing gelled KCQ-solutions and a statlonary junction have
been used, but it will be clear that hera dilution and contamination of the
KCQ-gel by the process liquid is just a matter of time. Wlth this system,
using a gelled process fluid of endpoint pH, such dilution or Gontamination
is impossib]e, since on average the process fluid under pH-control measured
at the electrode system, will have the same composition as the reference
fluid; plus or minus deviations in pH average out. This desirable condition
and the symmetrical potential-forming chain are only possible with this
electrode system, using glass electrodes for both measurement and reference.
This ls a basic difference from the asymmetrical electrode chain convention-
ally used, with a reference cell filled with KCQ-solution. The most desir-
able aspect of this electrode system with two glass electrodes, however, is
that it enables one to provide a hermetically sealed combination electrode
with extended service life, and to eliminate the earlier mentioned Breakdown
Factor A: Electrical leakage.
In order to eliminate Breakdown Factor B; this invention provides
a novel form of glass electrode. It was, accordingly surprisingly found
that when the membrance voltage is measured with an ITIC with such a high
input impedance and such a low bias current that loading the open EMF of
the membrane with the I~IC does not change the membrane voltage more than
0.1% ~as compared to measurement with a true electrometer~, the allowable
thickness of the pH-sensitive glass membrane is only limited by the resis-
tance which causes this voltage drop, and well exceeds the limit of thickness
as cited by Bates. This discovery made it possible to fabricate glass elec-
trodes of a novel structure, whereby the pH-sensi.tive glass membrane is not
-- 7 --
''.
:~ ``J
1~3~L70~i
thin and not supported at the edge by a glass tube for proper strength dur-
ing handling as prior art electrodes, either accordlng to ~cInnes and Dole
or to Cary's design, nor encapsulated in a metal plug as shown by Grauer in
United States Patent 3,700,5~7. The present glass membrane is instead an
unsupported disc ~t least 0.3 mm thick but preferably thicker which can be
handled during production without chipping or cracking, and which, with its
porous glass backing and sandwiched-in metal contact, now can be used as a
component in a combination electrode of considerably more sophistication than
possible with the prior art which ~sed an edge-support of the fragile membrane
early in its fabrication, thus limiting the freedom of design of the eventual
mGasuring electrode combination. -
A disc of pH-sensitive glass thick enough to be handled unsupported
without cracking or chipping has practically to have a thickness of at least
0.3 mm. Even small reductions of this thickness fatally weaken the strength
of the disc since the load-bearing capacity of a membrane diminishes as the
third power of its thickness. If membranes of the limit thickness of 0.13
mm as mentioned by Bates were unsupported discs, they would be more than
twelve times as weak as applicant!s minimum thickness discs of 0.3 mm which
are marginally fragile. Such prior art limited thickness membranes cannot
be handled as unsupported disc, without chipping or cracking.
This present invention utilizes discs about 0.5 to 1 mm thick or
thicker, such as 1.5 mm, 2 mm or even more in unsupported assemblies as
illustrated by Figure 2, and referred to as button elcctrodes. The disc ln
Figure 2 is, for example, 1 mm thick. If the electrical resistance of such
discs made of the classical Corning 015 glass, is considered too high for
application with low-costs ITICs, a low-resistance pH-sensiti~e glass may
be used, as e.g. disclosed by Cary in the abo~e mentioned United States Patent
2,462,843 in col. 7, line 3~-51. Such lithium-lead glass would, for a mem-
* Trademark
-- 8 --
11317V6
brane of the same dimensions, reduce the electrode electrical resistance to
0.004 times the resistance of Corning 015.
In contrast to the earlier mentioned Brit. Pat. 492,936, the con-
tact metal in the button electrode is not used to reinforce a very thin
deformable membrane, but should be made of soft, yielding metal ~hich will
not set up stresses due to differences in thermal expansion, in the stiff,
undeformable glass disc, in an effort to eliminate the mechanical component
of earlier-mentioned Brea~down Factor B. However, such prior art dry or
solid state electrodes contained another factor which led to electrical
rather than mechanical breakdown. This is a second category which has
received little attention so far but is just as unpredictable and pernicious
to an industrial electrode system with a longer projected service life:
erratically shifting assymmetry potentials which add on to the legitimate
pH-potentials. By providing an even and intentionally well defined distri-
bution of mobile ions at the glass-metal interface during manufacture by
steps to be later discussed, this problem has been solved by producing a
glass electrode which is particularly well adapted to the construction of
this combination electrode system. A system, ihcorporating the button elec-
trodes, is shown in Figure 1. Here the measurement glass electrode button
1 and the reference glass electrode button 2 are sealed into shell 3. This
shell may be made of plastic, enameled steel or stainless steel coated with
an insula~ing coating wherever the shell is in contact with gelled reference
solution ~, contained between shell 3 and the mantle S which may be made of
the same or anothar material as the shell as long as the mantle isolates the
reference gel electrically from the surrounding process li~uid 6 except for
seam 7 which forms an ion-permeable junction between reference~and process-
liquids.
Trademark
. -
113~7V~i
The electrode leads of both measurement glass electrode 1 andreference glass electrode 2 are directly and permanently connected to the
highly isolated, high input reslstance floating inputs of ITIC 8. Usually
such an IC has provisions for driving a guard shield. This output may be
connected to a metal wire sleeve ~not shown) surrounding the lead~ire
between measurement electrode 1 and ITIC input, and to screen off the back
of the glass electrode. However, this precaution is not essential in our
construction.
The outputs of ITIC 8 and service connections for electrical supply,
instrument ground etc. designated 12, are connected to the inside of the pins
of sealed-on cable connector 9. A platinum or gold redox electrode 10 for
additional measurement of redox potentials, using the reference glass elec-
trode for reference, may also be sealed into the shell by way of an insulat-
ing seal, and inte m ally connected to a separate pin of cable connector 9.
In order to keep the measuring glass electrode and the junction
seam clean and in unobstructed contact with the process fluid, a piezo-elec-
tric transducer 11 may be mounted inside the shell, with inte m al electrical
conncctions to pins of cable connector 9. I'he transducer may provide ultras-
onic or vibration cleaning of the active areas. Conventional glass electrode
systems never allow close mechanical contact between the transducer and the
glass electrode since the strong vibrations may strip the coating off the
internal reference electrode inside the glass e]ectrode. The solid-state
button glass electrodes, however, are not sensitive to this effect, nor do
they have the built-in stresses in the glass which may be triggered by the
vibration into forming hairline cracks.
Before final sealing of shell 3, it is filled with potting compound,
polymerizing monomer or foam fill 13. Right before local installation, the
space between mantle and shell may be filled with reference gel particular to
- 10 -
;:
1~3~7~
the local application, and closed with 0-ring 18. Installation may be done
with a demountable mounting flange 14, held by a gasket 15, washer 16 and
retainer nut 17. ~t may be advantageous for reducing electrical noise pickup
to fabricate the mantle of non-corroding metal, and to ensure that it has
metallic contact with the metal wall of the vessel on which the system is
mounted, or with earth ground, so that it may act as a s~ield. It is essent-
ial, however, that the inside of the mantle bc coated with an insulating
coating to avoid direct electrical contact between the mantle and the refer-
ence gel.
The construction of the button glass electrode is shown in Figure
2. A disc 21 of pH-sensitive glass is centrally covered by a metallic or
otherwise conducting electrode contact 22. Under the influence of the hydro-
gen ion concentration in the liquid adjacent to the glass surface, ions will
travel through this glass disc and transfer their charge to the electrode
contact 22. This again will give rise to an electron current travelling
through leadwire 23 to the ITIC 8 of Figure 1.
In order to prevent the build-up of spurious polarization or
asymmetry potentials between electrode contact 22 and glass disc 21, the
interface between disc and contact is during manufacture charged with the
specific concentration of ions which have a measurable mobillty in the glass
matrix, li~e silver ~= Ag) ions. In prior art dry or solid state electrodes
some sort of uncontrolled, arbitrary ionlc transition was made between
metallic lead and contact area, and the pH-sensitive glass, by depositing
a layer of silver chloride or other silver salt which often also acted as
a cement to keep the sandwich together. An abrupt, uncontrolled change in
concentration and mobility of ions oxisted between glass alld metal. Under
the influence of time and temperature this transition was unstable, causing
drift and changes in contact potential which made dry or solid state pH-
113~7~i
electrodes unattractive for general use, and only acceptable for specific
purposes as clearly demonstrated by Riseman et al in United States Patent
3,306,837, issued February 28, 196~. This invention made it possible to
maintain contact potentials stable o~er the life of the solid state electrode
by establishing a reproducibly graded transition of mobile ions from the
contact metal through the interface into the pH-sensitive glass at manufact-
ure of the button electrode assembly, e.g. by using a contact made of or
coated ~ith metallic silver, and during manufacture electrolyzi.ng silver
ions into the glass at a higher temperature. This may be done by making
leadwire 23 positlve with regard to a layer of graphite powder on which glass
disc 21 is resting, and passing through this assembly a predetermined amount
of coulombs of electrical charge over a prcdetermined span of time at a
predetermined temperature. This way, always the same number of silver ions
will be driven or diffused to the same depth into the same area of glass to
form a yellow diffused layer 25, resulting in a uniform contact potential
in all electrodes manufactured. The amount of coulombs to be passed through
should be selected such that it exceeds by at least a factor 100 the number
of coulombs expected to pass through the electrode ~worst case) during its
lifetime of pH-measurement or -reerence.
To preven~ excessively increasing the electrical resistance of
the pH-sensitive glass disc, its thickness is held relatively thin, but still
appropriate in view of abrasion or erosion during a reasonable lifetime
which, according to Bates in his above mentioned book on page 349, is limited
to 9 months to 2 years due to the processes of corrosion and leaching. In
industrial use, ove~pressures or shockwaves on an immersed electrode may re-
quire the use of an extra heavy disc. On the other hand the possibility of
thermal shock, ~hich would crack a thick disc, may call for a light disc
of pH-sensitive glass. As regards reinforcements, the prior art represented
by British Patent 495,303 discourages the use of
- 12 -
~.
~1~17~)6
the sintered-on porous layer of glass~po~der because: "the membrane is
deformed in the slntering process on account of the load of the glass grains,
or the shrinkage of the glass grains in this process prever.ts a strengthen-
ing exactly of parts having to sustain particular mechanical stress, vi~.
parts in contact with the thicker glass" ~col. 1, line 31-3~. This is
true for the edge-supported thin membranes of the prior artJ regardless
whether the electrodes are dry or liquid-filled. However, by avoiding edge-
support in the present novel button electrode, and by using a thicker pH-
sensi~ive faceplate, it was found that such rejected technique offered a
new advantage: supporting the pH-sensitive dlsc with a mechanically strong,
heavy bacXing which due to its sponge-like nature prevents or absorbs local
stresses on the solid pH-glass disc, which otherwise might cause hairline
cracks. It is this sintered porous glass backing 24 which also anchors the
button in the electrode holding shell 3 of Figure 1 when using a shell of a
preformed plasticl or of a compound cast around the essential components or
polymerized in place. For sealing the buttons in a metal shell, conventional
glass~to-metal sealing techniques may be used. But in both cases, the porous
structure of the backing will absorb any stresses caused by the sealing pro-
cess rather than passing them along to the pH-sensitive glass disc where they
might cause hairline cracks.
In an effort to consistently remove all causes of unpredictable
electrode failure by hairline cracks, this invention ~tilizes, for electrode
contact, soft, yielding metals like annealed copper or silver, in the form
of thin foils, wire screens or thin, spongy disc. Leadwires should also be
soft and thin so that they will yield rather than set up stresses in the
glass.
For the same reason, the use of different glasses ~ith different
coefficients of expansion for faceplate and for backing should be avoided.
- 13 -
: ' : ' ~`
~13173)6
It may be practical to use fo~ porous backing a glass with a lower softening
point then the p~-sensitive face place.
A practical example will be given for the fabrication of button
electrodes according to the description:
A thin homogeneous paste is mado by ball-milling silver oxide with
a predetermined few percent of silver sulfate or silver sulfide and a sel-
ected amount between lO and 40 % of the fine glass powder used in the next
step. The liquid phase in this paste may be water with a binder like carbox-
yme~hylcellulose, or ben7.ene with polystyrene, or any other suitable vehicle.
~ith silkscreen printing techniques, discs of a uniform diameter of 15 mm
are printed on ashfree filter paper in a widely spaced dot pattern, and dried.
Well centered, on top of the discs, is printed a pattern of larger diameter
discs of a fine powder of pH-sensitive glass suspended a similar vehicle as
used in the previous step. The thickness of the glass powder discs is con-
siderably more than that of the silver compound discs, and should be such
as to yield in the final button a pH-sensitive faceplate of appropriate
thickness, This thickness is to be held as uniform as possible. After dry-
ing, the sheet with the composite printed-dot pattern is clamped between two
flat graphite slabs with spacers of about the thickness of the printed dots.
This assembly is slowly heated in a furnace in air up to about 400 C, until
all the filterpaper has burnt away and the silver oxide is reduced to silver;
and then continued in a neutral or slightly reducing atmosphere ~to prevent
undue oxidation of the graphite) up to a temperature where the glass powder
fuses to a non-porous disc, but below the melting point of the silver or its
compounds. Upon cooling and taking apart a collection of pH-sensitive glass
discs with electrode contacts ~ill have been obtained. It has been found
that the use of this method for making thin discs of 0.3 mm thickness or less,
not only leads to discs ~hich are very fragile, and chip and crack easily
- 14 -
3L131'706
during handling, but that a rapidly growing percentage of discs show through-
holes. Apparently, durlng the process of surface melting which leads to
sintering, the high surface tension of the molten glass draws the adjacent
grains together. If now ~he distribution of glass grains has been slightly
uneven the pull will be uneven across the area and holes are pulled open.
Discs which may be used satisfactorily in the subsequent steps of manufactur-
ing button electrodes should preferably be about 0.5 or 1 mm thick, or thicker.
Now thin silver leadwires are mounted on the electrode contacts with a silver
oxide paste and in a graphite mold a cylinder of coarse glass powder is
filled in on top of each electrode contact. l`his assembly is heated to a
suitable temperature well below the melting point of silver or the softening
point of the pH-sensitive glass, but above the dissociatlon temperature of
silver oxide until the glass powder is fused or sintered into a solid but
open-pored block, as is done in the fabrication of sintered-glass filters.
Such a loosely sintered mass behaves somewhat like foam-rubber. Stresses
exerted upon it are cushioned and absorbed. This is the case with stresses
caused due to differences in thermal expansion as well as due to difference in
water absorption of the glass or stresses caused by a solidifying surrounding
polymer. As a result, the solid pH-sensitive glass faceplate is not subject-
ed to potentially fatal stresses, as it would be if the adjacent glass were
not porous but solid~ This also applies to a faceplate of other ion-specific
material.
An example may also be given of a different method of fabrication
of the same type of button electrodes, which is especially useful for face-
plates of pH-sensitive glasses which tend to devitrify rapidly or to deterior-
ate in non-oxidlzing atmospheres as does e.g. the low-resistance glass descr-
ibed by Cary et al and containing both lithium and lead. A batch of molten
glass is heated until sufficiently bubble-free, and cast into rods of e.g.
- 15 -
1~13170~
12 mm diameter and 50 mm length. After slow cooling to avoid setting-up
stresses, these rods are sliced to discs of e.g. 0.5 or 1 mm thickness or
thicker with a diamond saw blade or by using a wafering machine as especially
developed for slicing thin wafers of hard materials for the semiconductor
industry. Separately, cylinders of coarsely crushed glass, of about the
same thermal expansion coefficient as the faceplate discs, passing 0.5 to
l mm mesh openings in a sieve, are pressed and sintered, or sintered in
graphite forms. Each cylinder with e.g. a height of 3 mm and a diameter of
3 to 10 mm has a thin silver leadwire incorporated through the center, or
has a small central hole like a doughnut through which a leadwire can be
passed after the cyllnder is sintered at a temperature to just consolidate
the glass powder to a mass with open pores and cooled again. It may be
appropriate to have about 30 mm of leadwire extend on one side of the cylind-
er for electrical connection to an ITIC, and about 3 mm on the other side,
bent flush wi*h that side of the cylinder. The surface of that side is now
coated with a sllver paste containing glass powder as described before, pre-
ferably leaving uncoated a rim at the edge of that side. The glass face-
plate is then brought into contact with the coated side of the cylinder and
fused on at the fusion temperature of the glass in the silver paste.
Instead, a round patch of silver paste may be applied to the center of one
side of the pH-glass disc and the cylinder is pressed onto this patch with
its side containing the short, bent piece of leadwire, again taking care that
the silver paint does not extend all the way across the dlameter of the
cylinder, as shown in Figure 2. After drying, the assembly again is fired
to convert the silver paint into a solid contact area cementing the cylinder
with leadwlre to the glass faceplate. In both cases, the firing may be con-
tinued for an accurately reproducible time at an accurately determlned temp-
erature to allow silver ions to thermally diffuse lnto the glass faceplate
- 16 -
`
~3~7~
to a controlled and r~producible depth, which is visible as a yellow layer.
Alternatively, such heating operation may be combined with the
controlled electrolysis of the silver compound as described earlier. This
finishes the fabrication of the button electrodes.
A combination electrode assembly in a sealed shell may now be
fabricated by placing in a mold thc necessary component par~s with the button
electrodes in recessed cavities. The ITIC's are wired to the electrode~
and the cable pins, and so located in the mold that their cans touch the
mold walls or fit into shallow recesses. The mold is now filled with poly-
merizing closed-cell foam or silicone or other potting compound with high
electrical insulation value and excellent water-resistance. Upon silidifi-
cationJ a core is obtained which holds all component parts embedded and which
will be referred to as "the assembly".
This assembly is then placed in a wider mold and a suitable epoxy
compound or other polymerizing or congealing fluid poured around the foam
core to form the hermetically sealing shell. Particular care should be
taken that the button electrode faces remain exposed. Process liquid, or
liquid to be tested, will preferentially creep between the glass of the face-
plate and the surrounding plastic, so speclal precautions should be taken
to prevent this, by the use of bonding agents etc. Also, the polymer skin
may be use become electrically better-conducting than the glass of the face-
plate. In this case, shorting out of the pH-EMF across the faceplate by
the polymer may be prevented by avoiding direct contact between the outer
shell of polymer and the metal contact area on the faceplate or the leadwire.
It is for this reason that a rim, free of contacting silver paint, has been
provided at the edg~ of the area where porous glass backing and faceplate
touch.
The I~IC-cans may have to dissipate a small amount of Joule heat
..
., '` ~ , '' ~
-
':
1~3~l7~
during use. For good thermal contact to the outside world, they should touch,
or preferably have a heatsink embedded in the outer shell without becoming
exposed to the outside of the shell.
It may be clear that by substituting ion-specific glass or ion-
specific compounds instead of pH-sensitive glass, ion-specific button elec-
trodes can be manufactured. In combination with either an ion-specific or
a pH-specific reference, combination electrode assemblies can be made which
are ion-specific and may have automatic compensation for undesired pH-
dependence as may be the case with redox potentials.
Electrically, the built-in ITIC may be biased to show a "live zero"
of predetermined magnitude at the process endpoint to produce a one-way-going
or single-polarity control signal as is often used.
Instead of encasing only one pH-measuring electrode button in the
shell, it has unique and novel advantages to encase three or more separate
and independent pH-measuring electrode buttons, each connected to its own
ITIC input, but sharing the reference. An appropriate external comparator
circuit will continually or intermittently monitor the measured-pH outputs
and decide whether all measuring electrodes produce the same voltage with
regard to the reference. As soon as one electrode starts to deviate within
certain norms from the "majority" of remaining equal-potential indicating
electrodes, an alert will be given so that the complete combination electrode
system may be exchanged for a fresh one at a convenient moment. In the mean-
time, the "majority" will keep the process under proper control until com-
bination-electrode-replacement, without risks or economlc losses due to off-
specification operation. Although more measuring electrodes may be used,
even a triple-electrode combination electrode system will completely elimin-
ate unpredictable breakdown of the system, yet be not any bulkier nor any
costlier than present-day combina~ion electrodes in industrial holders.
- 18 -
.
11317~ti
For the same benefits, three or more reference electrodes may be mounted
in the same combination electrode.
Although it will always be possible to correct externally, in the
control room, for experimentally measured differences in offset-voltages
of ITIC's or electrode buttons, the two-step method of fabrication of this
invention is particularly advantageous since the assembly, obtained after
the first step, can be tested during manufacture for voltage-offset between
duplicating electrode buttons and their ITIC's. Such offsets can usually
be adjusted according to instructions by their manufacturer, with small
potentiometers wired to the ITIC and embedded in the central core in such
a way that adjustment of these potentiometers is still possible. After
adjustment, the whole assembly is encapsulated in the outer shell as already
described. No handling of the sealed unit can now disturb the proper offset
adjustments, and true exchangeability between combination electrode systems
is insured, provided their reference pH is the same.
Internal offset-correction may also be of value for single-measure-
ment-electrode systems, to insure proper potential difference between
reference and measurement electrode with their associated ITIC's and to
eliminate pre-application calibration by the user except for the most exact-
ing applications.
:. - 1 9 -
: '
