Note: Descriptions are shown in the official language in which they were submitted.
~3~l~3
The use of porous glasses as su~strates for the molecu- ~ ;
lar deposition of selected materials has shown great promise in `
the production of materials with selected physico-chemical proper-
ties and selected property variations. This process, called "Mo-
lecular Stuffing" or doping has been described in detail in U.S.
Patent 3,938,97~ and Can. Patent Applications Serial Nos. 247,547
and 247,545 filed March 10, 1976. In addition to such porous glas~
ses, the present invention is applicable to porous glass produced
by such other methods as chemical vapor deposition. (See U.S. ~ ~-
Patent No. 3,859,093~.
By the molecular stuffing process, selected solutions
containing materials which alter the physico-chemical properties
of high silica glasses are diffused into the pores of a high sili-
ca base glass preform to achieve a homogeneous concentration of
the modifying agent or agents (solute~. For step concentration
profiles these modifiers are subsequently precipitated and a
cladding region is formed by their removal from the outer regions
of the preform by a suitable solvent before drying and sintering
` ~ of the preform assembly. For graded concentration profiles, the ~ `
concentration of selected modifiers is altered to a desired varia- -
tion by a second soaklng of the preform in selected solvent solu-
tlons containing selected concentrations of modifiers. This is
` followed by precipitation of the modifiers and subsequent drying
and sintering of the preform assembly. (See Patent Appl. Serial
No. 247,545).
The change in physical property achieved by the addition
of dopants is a function of the dopant concentration. Therefore
i~ the addition of high dopant concentrations generally induces a
~ large change.
1 :
3,~ 30 Several products benefit from large variations in physi-
cal properties and therefore large variatians in dopant concentra-
tion. For example, in fiber optics, a large change in index of
.,, ~ ~
~,~,. . . . . . . .
. ~ .
.
183
refraction between the core and cladding regions of a fiber yields
a high numerical aperture, while in strengthening brittle mate-
rials, a large change in thermal expansion coefficient and/or in
glass transition temperature between the surface and the interior
of an article allows the formation of large surface compressions
(pre-stressing) and thus the achievement of correspondingly in-
creased strengthening.
The numerical aperture, NA, of a light transmitting de-
v~ce is a measure of its acceptance angle. In optical waveguides
the numerical aperture is related to the difference in refractive
index, n, between the axis or center of the waveguides and the off -
axis elements. An increase in numerical aperture is obtained by
increases in the index difference between these elements (for -
example, in waveguides with step index profiles, the difference
is between the refractive index of the core, nl, and the clad, n2, ~-
regions; thus NA = ~nl - n22~.
Since numerical aperture is related to the angle of ac-
ceptance of the incident light beam, high numerical apertures are
desirable since this allows transmission of relatively more energy
from a given light source. High numerical apertures are also de-
sirable from the standpoint of reducing microbending losses in
optical waveguide fibers, and for the preparation of lens elements
and other optical elements.
The process described in Patent No. 3,938,974 and Patent
. .
Application Serial No. 247,545~ demonstrates how molecular stuffing
`j of porous glasses may be conducted using a series of dopants
both individually and in groups to develop integrated optics
components with tailored refractive index clistribution and
.
strengthened articles with tailored thermal expansion coefficient
and glass transition temperature distributions.
.
This invention employs glass compositions and dopants
- 2 -
', '
~ . . .
1083~83
.:
similar to those in Patent Applications Serial Nos. 247,547 and
247,545. ~owever, we have found that certain dopants, such as
lead, induce noticeable scattering in the glass when large dopant
concentrations are used. Though scattering has little effect on
most uses such as integrated optics components and strengthened
members, some uses such as the transmission of very high quality -
images (e.g. in cystoscopes) or long, extremely low loss, (e.g.
below 20 dB/km) optical fibers with a high numerical aperture
(e.g. N.A. greater than 0.35~ may be limited by the amount of
light scattering present. For example, when lead is used as a do-
pant oxide at doping concentrations above 40 grs. Pb (N03)2/100
cm3 of water, scattering is observed in the final giass. This is
an important limitation because lead has a high atomic polariza-
bility and is useful to obtain high index glasses. It would be
of considerable advantage if glasses could be made with large con~
concentrations of Pb as a dopant in those situations where a com~
i bination of high NA and very low scattering loss were desired.
We have discovered that when the dopant in molecular
stuffing is composed of certain combinations of lead and/or bis- ~ ;
muth with cesium, rubidium, and/or potassium, then a remarkable ~ ;
and unexpected decrease in the scattering loss occurs. Further,
we have discovered that certain of these combinations which give
low scattering losses may be used to obtain high numerical aper~
tures and/or high pre-stressing levels as well.
; The desired combination of dopants leads to glass arti-
cles with the following final composition. The composition of
these glasses consists of at least 85 mole percent SiO2 with im-
provements which comprise at least 7 wt. percent of at least one
member selected from Group (I) consisting of PbO and Bi203 and at
least 1.5 mole percent of at least one member selected from Group
(II~ consisting of K20, Rb20, and Cs20.
Even though the maximum dopant concentration is limited
- 3 -
',
''','':'; ' .. ~ ~, .- ; ',
::~
1~3~33
by the concentration of Si02, the broad limi-ts are a maximum of 25
wt. percent for Group (I) and a maximum of 9 mole percent for
Group (II). Our preferred range covers at least 2 but not more
than 9.5 mole percent B203 and at least 7 but not more than 20 wt.
percent of Group (I) and at least 1.5 but not more than 7 mole
perc~nt of Group (II).
In another embodiment of this invention, we have disco-
vered that a method which comprises adding a dopant to a porous
matrix with interconnective pores, immersing the porous matrix in .. -
a~solution of a dopant causing the dopant to precipitate in the ~ :
matrix, removing solvent and where necessary, decomposition pro-
ducts, from the porous matrix and collapsing the porous matrix to
a solid form, can be used to produce high silica glasses. These
glasses can be produced with the following mixed dopant composi~
tions: Group I being Pb and/ox Bi and Group II being K, Rb, and/or .
; Cs in the form of nitrates, carbonates, acetates, borates, phos-
~: phates, arsenates and/or silicates in either hydrated or unhydra~
ted form of mixtures therefrom used to pro~luce a glass having . ~:
` composition 7 to 25 wt.percent (.preferred range 7 to 20 wt. per-
~` 20 cent) of the oxide equivalent of at least one member selected from ;~ .
Group I consisting of Pb and Bi and 1.5 to 9 mole percent (the .: ~ .
preferred range being 1.5 to 7 mole percent) of the oxide equiva-
J lent of at least one member selected from Group II consisting of ..
K, Rb, and Cs.
:. Finally in another embodiment of this invention, letting
I represent [Pb(N03)2~ and CBi(.N03)2~, and II represent the alkali
nitrates of Cs, Rb and K, taken either singly or in combination
the ranges of dopant which yield glasses with high numerical aper- :~
ture and low scattering loss are~
Broad 3 Preferre~
g/100 cm g/100 cm
solution solution .
Group I 45-200 50-150
- 4 - .:~
': . - .: . ~ ', ' .
~: ' ' . ' : , :
1~3~33 :
Group II 40-200 50-150
where the weight represents the weight of ~t least one member of
the group in the form of a nitrate salt. $he solutions may be wa-
ter, optionally with small amounts of low molecular weight alco- -
hols such as methanol. The solvents used in precipitating the do-
pants may be lo~ molecular weight alcohols such as methanol and
ethanol.
In another embodiment of this invention when multiple
dopants are used and precise control in the variation of dopant
concentration near the surface of-the article, e.g., to produce a
composition step profile, is desired, we have found that after the
dopant has filled the pores of the article, it is necessary to im- , -
~
; merse the articles in a solution of multiple solvents which are
selected so that the solubility of each dopant compound is appro-
.
ximately the same. Thus there will be approximately the same
change in concentration of each dopant in the region of the arti-
cle near the surface, preferably the dopant solubilities of 1/2 to
15 grs of dopant material per 100 ml of solvent solution. It :lS ~ ~.
,~ . .. : '
preferred to further wash the glass article in a soluti.on selected
to control the solubility of each dopant compound within a range
of 0 to 2 grs of dopant material per 100 ml of solution.
EXAMPLES
, . ~ .
:
Porous glasses are used as substrates for the deposition
`, I of the selected dopant combinations. Any porous glass preform is
~ ~ satisfactory. In this example we prefer to describe a spec1flc
-~ ~ method for forming the porous glass substrate by phase separation,
j :
although other processes are just as useful. (See for example
$ Schultz, Patent No. 3,859,073).
~n alkali borosilicate glass of composition 57% SiO2, -~
' 30 36% B203, 4~ Na20 and 3% K20 is melted in a Pt crucible in an
electric furnace at temperatures between 1300 and 1450C. The
- .
~1 melt is homogenized by stirring with a Pt stirrer, and is then
. ~ ,
. . ~
.:
,`~ '~ . ~ ' ' ' :
, .
~83~33
pulled in the form of rods 5/16" diameter by 4' length. These ~ -
rods are then cut into rods ~" in leng-th which are hea-t-treated
at 550C for 1 1/2 hours to induce phase separation and subsequen-
tly leached in a 3N HCl acid solution. During the phase separa-
-tion heat-treatment, the homogeneous glass decomposes in-to two
phases, one with high silica content and one with high boron and
alkali content and lower silica content. These phases are inter-
connected sufficiently that exposure to the leaching solution com- j
pletely removes the alkali rich phase leaving behind a high silica
porous glass substrate. The porous glass is washed with deionized
water.
The porous glass substrate is immersed in a solution
containing the desired concentrations of dopants (see Table I) for
3 hours or longer to allow the solution to fill the pores comple-
tely. The dopant compounds are then precipitated from solution. ;~
In following this process we have found it desirable to
achieve precipitation of the dopants by thermal means; that is,
lowering the temperature of preform and solvent to a point where
the solution within the pores becomes supersaturated with the do-
pant causing the dopant to precipitate in the pores. The sampleis then transferred to an unstuffing solution of sol~ent whereby
some of the dopant is allowed to dif~use out of the pores yielding
; a sample with graded dopant concentration. This step is necessary
in both fiber optics to achieve high numerical apertures and in
strengthening to achieve high surface compressive stresses. When
graded properly, the dopant concentration is nil near the surface
of the object thus yielding a low refractive index and/or a low
thermal expansion coefficient in the cladding region. The unstuf- ~ ~
fing step is often conducted sequentially in two different solu- i
tions to insure maximum dopant removal from regions near the sur-
face of the object (see Table I~ steps (a) and (b). The sample
is kept at 0C where it is exposed to vacuum for 24 hours and
: - 6 -
: : . .
' ..
~: , ,
,
~83:~L8~3
then heatecl at 15C~hour up to 625C under vacuum and sintered
between 830 and 850C.
EXAMPLES I, II ~ND III
Table I reports details (concentration of solution, tem-
perature and times) of the preparation and measurements of refrac-
tive index in the core (central) and cladding (outer) regions of
the objects made by using lead nitrate and cesium nitrate as do-
pants. Corresponding numerical apertures are also listed. Table
II lists the compositions at the center of the final glass arti-
cles. A fiber was pulled from Sample II and scattering losses were
found to be less than 20 dB.km in each case.
EXAMPLES IV TO VIII
A porous glass preform prepared as described in Examples
I to III is soaked in a dopant solution as described in Table III
at a temperature specified in Table III for 16 hours. This tempe-
rature is chosen to be at or above the solubility temperature of
~ the designated dopant concentrations. This allows the dopant so- `
- lution to fill the pores of the preform completeIy and uniformly.
The preform is then removed from the dopant solution and is soaked
20 in a single solvent as specified in Table III for three hours in
order to precipitate the dopant within the pores. No dopant remo-
val was attempted since these glasses were made only to observe
index change with concentration. The sample is then kept at 0C
` where it is exposed to vacuum for 24 hours and warmed up at 15C
per hour up to 625C also under vacuum. It is then heated to bet-
~i ween 830 and 850C where sintering occurs. Refractive index mea~
surements conducted on these samples are reported in Table III.
Table IV details the compositions of the final glass articles lis-
- ted for Examples VI and VII.
"
~..
.
1~3~3 -
a) ,. . .
~J h :
O ~S7
Z ~ O O O
.
0 X a~
~1 ~ 0 ~ ~r et'
p; H U
, ,
0 ~
~ ~ O ~ ~ ~ O ~; ~
~ .
~ ~ ~ o `~
O O ~ O O ~ O O ~1
0~ 0~ ~'~ O>1
o o ~ o O ~ a~ o
~ u' e ~ ~ Ei3 ~ '; .'
H X ~1 0 ~-1 O 1-l O h ~1 : ::
o 1~ a) o\ ~- a) o\ ~ o d~
~ ~J O ; .' ',~
. 1 1 H O (d O 0 0 00 11~ 0 ~1
U) 3 ~ 3 ~ ~ 3
D ~ oC~ o o~ o~ o~) oU _
o o o o o o '. ;- ~
O . .
.` ~ ~ .C ~ ` . '
0 Q 0 R ; ~:
: ~ ~ . . ~
O ' ': !
~)~rl ~ 3
0 ~I Q~ O O O
. ~o ~ a~ ~ O O ,~ , ,: .,
'',,' ~ ~ o _ O ~ O ~ , " :'
,~ O t~l ~ ~1 ~ ~ ~rl ~)~,1 U~ :
.~.,, ,_ 0 41 .~J~ O~ ~ ~`1 o q-l ~ .,~ . ..
z o ~ ~ z o ~ ~ æ o ~
O -- ~1 O --~I H~) --~--I ~ : : . .
~.~~ æ u~ ~ 0 H æ u~.-, O H oU~ -1 0 ~ `~ . :. .
- 0 H ~ C,) ~ H-- ~ n H æ r~~ ~ rl
.4 R -- IH .
o u~ O ~ ~ o u~ ) o u
r ~ O ~ ~1 0 ~ 0 ~ ~ .
~: ~ ~ o ~1 0 ~ o ~1 o Q~ O -1 0 ~ ~ ~
~ 0 0 ~ h ~ 0 o ~5~ :~ 0 ~t~ S ` ~
o x ~ ~ X O O t:~ X O a) t~ ~
H + ~ 0 ~ r-l + ~ 4 0 ~1 C0 + 1 ~ 0 K ~:
- .
.. .. . . . .
83
TABLE II
Example SiO2 B203 Cs20 PbO PbO
mole % mole % mole % mole % wt %
I 86 3 5 6 16
II 90 3 3 4 14
III 87 3 6 3 9
TABLE III ~
Dopant solution Refractive -~
(.per 100 ml of Dopant solution Solvent index in
EXAMPLE aqueous solution~ temperature temp. core
IV 50 g KnO3 + o ethOnol
95 g Pb(N03)2 120 C O C 1.506 ;
- 63 g RbN03 + ethanol .
V 100 g Pb(N03)2 120C 0C :1.514
~: 120 g CsN03 +
VI 100 g Bi(N03)3 + ethOnol :~
5H20 110C O C 1.514
.
136 g CsN03 +
: 57.1 g Pb(N03)2 +
;. VII 57.1 g Bi(N03)3 ethOnol
: 20 5H20 110C O C 1.516 - :
TABLE IV
-
` Example SiO2 ~ PbO PbO Bi23 Bi23
mole % mole % mole % mole % wt % mole % wt %
VI 91 3 4 0 0 2 9 ~ :
.~ .
~ VII 89 3 4.5 2.5 7 1 5 :
.': - ~:: ~.
` '~ :
:-,
3Q
_ 9 _ :
.
:, . - . . - :
- ~:., . ~ . -
:: - . :
: ::
~83183
In cases where multi~le dopants are used and precise
control of the profile variation in the cross-section of the arti-
cle is desired, we have found that the following uns-tuffing pro-
cess yields good results.
The porous glass preform is prepared as discussed above
and a combination of dopants are diffused in its pores. Let us
denote these by dopant A and dopant B. Precipitation is achieved
by either cooliny the glass article or immersing it into a solvent
with extremely low solubility for both dopants A and B.
In the ensuing unstuffing process, where good control of
the dopant concentration is necessary, we have found it desirable
to use mixed solvents. It is useful to have the dopants A and B
both with the same solubility in the solvent mixture. This is ac~
complished by selecting a minimum of 2 solvents with the following
properties.
The first set of solvents or single solvent has the de~
sired unstuffing solubility which in the case of step profiles is
0.5 to 15 grs/100 cc of solution for dopant A. The second solvent
or set of solvents has the desired unstuffing solubility of 0.5
`~ 20 to 15 grs/100 cc of solution for dopant B at the same temperature
The solvents are chosen so that when mixed together, the two sets
of solvents have the desired unstuffing solubility of .5 to 15
grs/100 cc of solution for both dopants A and B. Best control of
profile is afforded when the solubilities for dopants A and B are
equal to within + 50%; that is, if the solubility of dopant A in
solution is 5 grs/100 cc of solution, then the solubili-ty of B is ~ ~
between 2.5 grs/100 cc and 7.5 grs/100 cc of solution. ~ -
In the stuffing step, dopant concentration is in the or- -
der of 100 gm per 100 cc of solution. In contrast the first
unstuffing solution solubilities given above are in the order of
a factor of 10 less than during the stuffing step.
Preferably the glass article is further washed in a
:' :,
' - 1 0 - , ,,~
,:~ .
-~: :: . ,
~L~83~33
wash solution in which the dopants are even less soluble, e.g.,
a solubility of 0 to 2 grs of dopant per 100 cc of solution. This
wash step ensures that substantially complete precipi-tation of do-
pant has occurred before drying is commenced.
Solvents can be selected from the following group:
1. Water, in which the alkali me~als are very soluble.
In water, potassium, rubidium and cesium nltrates have a high
temperature dependence oE solubility which is useful in thermal
precipitation.
2. Lower molecular weight aliphatic alcohols (i.e.,
less than 6 carbons/molecule) in which the alkali metal and lead
salts are progressively less soluble with increasing molecular
weight. Alcohols with six or more carbons per molecule have such
low solubilities for the dopants used here that they are not com-
monly used.
3. Acids: the solubility of bismuth is strongly de-
pendent on the pH of the solution. This solubility can be con~
~` trolled by introducing an acid. The acid and any resultant salts
formed by reaction of the acid with the dopants of group I and II
must either evaporate or decompose into an oxide compound before
- sintering occurs. We prefer to use nitric acid.
Ternary dopant systems are handled in the same way with
several solvents in solution with each o-ther.
In both these cases of multidopant stuffing and con-
trolled profiling, it is possible to develop high residual stres-
ses in rods by selecting dopants which alter the thermal expan-
sion coefficient or the glass transition temperature or both pro-
perties of the porous glass. ~
The two examples below describe processes used for two ~;
and three joint dopants with controlled profiles to develop a re-
sidual compressive stress in the outside layer of the consolida- ~ ;
ted glass articles.
- 11 -
-.
: ~,: : . : : :
,., ~ ,. .. : . . . , :
:: ~
- ~ .
lS3
EXAMPLE IX
In this example we describe a procedure for 2 dopants
which are jointl.y diffused into the pores of a porous preform.
Profiling of these dopants is accomplished by using a so].vent .
consisting of a solution of 3 solvents.
Porous glass articles in the shape of cylindrical rods
were prepared as in Example 1 above. The porous preforms were : ~.
immersed in a stuffing solution containing 120 grs of CsN03 and
100 grs of Bi(N03)3 5H20 per 100 cc of solution, at 108C for a -:
period of 24 hours. The dopants within the pores were then pre- ~.
cipitated by transferring to an unstuffing solution consisting of
three solvents at 0C in the following proportion: 6.4% methanol,
10.8~ of 70% nitric acid in water solution and 82.8% water by vo-
lume. The rods remained in this solution for 5 hours at 0C. A
clear region appeared within the outside surface. The samples
were washed for 48 hours in a solution of 82.2% methanol, 1.8% of ~.
70% nitric acid in water solution and 16% water at 0C, and then
were dried at 0C for 7 hours. The samples were then heated in
vacuum to decompose the nitrates and sinter the pores, reaching a . ~ -
20 temperature of 825C. .~ ;~
: The consolidated samples were examined and consisted of
a cladding region relatively free of dopant compounds adjacent to .
the outside surface of the sample, and a core region within the ~;
~ cladding which contained substantially all the dopants. Measure-
ments with an optical microscope revealed a ratio of cladding
.~ thickness to rod radius of 0.342 and a compressive stress in the
cladding region of 14,000 psi. Measurements of light scattering
: loss from these samples showed an intrinsic scattering loss of ;
3.4 dB/km at 1.05 ~m.
EXAMPLE X ~ ,
In this example we describe a procedure for 3 dopants
which are jointly diffused into the pores of a porous preform.
- 12 ~
;.
.: . . .: . . .. . .: .. :
. .; ~ . . : . ,
:: ' ' '' , ' , : ; . '.''.. ; ' ' .
~83~l33
Profiling is accomplished by using a solven-t consisting of a solu-
tion of 3 solvents.
Porous glass preforms were prepared as described above.
The porous preforrns were immersed in A stuEfing solution made by
saturatincJ 50 ml of water with CsN03, then saturating the solution
with KN03 and clissolving Bi(N0313 . 5H20 in the solution at a tem-
perature of 100C. The inal composition of the stuffing solution
consisted of 50 ml H20, 94 grs CsN03, 142 grs KN03 and 234 grs
Bi(N03)3 . 5H20. The rods were left immersed in this solu-tion for
48 hours at 100C. The dopants within the rods were then precipi-
tated by transferring the rods from the hot stuffing solution to
methanol at 0C. The rods were kept in methanol for 10 minutes
; and then were transferred to an unstuffing solution consisting of
three solvents at 0C in the following proportion: 6.5% methanol,
10.8% of 70~ nitric acid and 82.7% water by volume. The rods re-
mained in this solution for 1.5 hours at 0C. A clear region ap-
peared over the external surface of the samples. The samples were ;~
then washed in methanol at 0C for 24 hours and were dried in va- ~;
cuum at 4C for 24 hours. The rods were then heated in vacuum to
decompose the nitrates and sinter the pores at 850C.
The consolidated rods were examined and consisted of a
. .
cladding region relatively free of dopant compounds adjacent to
- the outside surface of the sample and a core region within the
:
cladding which contained substantially all the dopants. Measure~
i ments with a microscope revealed a ratio of cladding thickness to
; rod radius of 0.16 and a compressive stress in the cladding region
of 25,000 psi. Measurements of light scattering loss from these
samples showed an intrinsic scattering loss of 1.6 dB/km at 1.05 ~;;
~m.
'~ 30
:'
.
- - . : .
' ' ': , ' , ' , ''
.
