Note: Descriptions are shown in the official language in which they were submitted.
5~
~1-57A
~iETHOD FOR RECOVE~ING PLATINUM
IN A NIT~IC ACID PLAI~T
Nitric acid is produced commercially by passing
ammonia and air over an oxidation catalyst which is
usually a gauze woven from platinum-rhodium alloy wire.
Typically, the temperature of gas leaving the gauze
ranges from about 810~C. to about 960C., most often
above 850C. As ammonia is oxidized, platinum is slowly
lost from the gauze, possibly in the form of the more
volatile oxides. Rhodium is also lost, but this is not
so severe a probleM. The rate of loss depends upon the
type of plant. Typically, for each ton of ammonia
converted, a high pressure plant will lose more than one
gram of platinum, while lower pressure plants will lose
less. Even thouyh the rate of catalyst loss is slow when
expressed in terms of weight, the cos-t is usually quite
suhstanti~JO In many operations, the cost of platinum
lost ~uring production has been said to be the secon~
l~r~est. expense of the operation, exceeded only by the
c~st G~ ammonia feedstoc~.
Many approaches have been tried to recover some of
the platinum and rhodium. Filters of various materials
have been placed downstream of the catalyst gauze to
mechanically catch and retain solid particles of platinum
and rhodium. Later, it was discovered that various
palladium alloys had the ability to withdraw platinum-
containing ~apor from the gas stream. The mechanism of
--2--
this withdrawal has been a subject of some controversy,
but it has been theorized that, in the course of -the
reaction, platinum oxide in -the gas phase may revert to
platinum, which either returns to the catalys-t gauze or
is carried away by the stream to possib]y alloy wi-th
palladi.um and catalyze formation of volatile pal.ladium
compounds. (See Holtzmann, Chemie ~Lnleur-I'echnik,
vol. 40, No. 24:1229~37, 1968.) ~ variety of alloying
elements have been selected, mainly for their ability -to
improve the mechanical properties of palladium. Typi.cal
commercial palladium alloys have contained about 80%
palladium and 20~ gold by weight. The recovery alloys
are usually employed ln the form of multiple sheets of
woven gauze but knitted meshes or other foraminous
elements can al.so be used. The recovery gauze is usually
placed as close as possible to the catalyst gauze, often
withili a few millimeters, usually no more than 10 mm.
Since the catalyst gauze in a nitric acid plant is
changed regularly on a schedule of from about every 35 to
every 270 days, depending on plant design, as a practical
matter, the recovery ~auze is usually replaced when the
catalyst g~uze is changed, although it is possible to
replace it less frequently. This technology, which is
currently widely applied, is described in more detail in
U.S. Patent 3,434,820; Platinum Metals Review, Vol. 13
[No. 1]: Pages 2-8 (Jan., 1969); British Patent
1,082,105; and in Chemie- In~enieur-Technik, Vol. 40, No.
24: 1229-37 (1968). As applied, the recovery efficiency
of each sheet of recovery gauze obtained using this
technology has ranged irom about 10% to 60%, primarily
depending upon the type of plant which is usually speci-
fied in terms of the nitrogen loading of the plant.
Using the method of the present invention, it is
possible to increase the recovery efficiency of each
sheet of recovery gauze by several percent resulting in
the annual recovery of more than a hundred additional
--3--
troy ounces of platlnum in a medium pressure plant or -to
obtain equivalent recoveries wi-th fewer sheets of yauze.
These improvements can make a very siynificant difference
in the economic viability of a nitric acid plant. This
savings would be less for lower pressure plants, but for
higher pressure plan-ts, the savinys become even more
substantial. These savinys are now possible because i-t
has been discovered that it is possible to es-timate the
efficiency of platinum recovery of hiyh pal:ladiurn conten-t
gauzes based on the model that -the process is mass
transfer limited, that ls, the rate of withdrawal of
platinum from the stream of gas cominy from the catalys-t
gauze is determined or limited by the rate at which the
platinum species diffuses through the gas to the surface
of the recovery gauze, the rate at which platinum at the
wire surface can be trapped or retained or "alloyed" with
the palladium in the yauze beiny much yreater than the
rate at which the platinum species can diffuse to the
wire surface from the yas stream. On this basis, it is
possible -to rationally design and optimize the configura-
tion of the gauze to obtain improved efficiency without
incurring e~cessive pressure drop.
tJsing yauzes designed according to the present
invention, it is possible to increase the efficiency of
each sheet in -the gauze pack by several percent, depen-
ding on the nitrogen loading of the plant. The method of
the present invention is especially desirable for use in
high pressure plants, since not only is more platinum
lost per ton of ammonia converted, but also the number of
tons of ammonia processed is much greater than in lower
pressure plants. Fur~her, prior art sinyle gauze re~
covery efficiencies for high pressure plants have been
distressfully low, as recovering platinum in these plants
is extremely difficult. Thus, any improvement is par-
ticularly significant for the economics of -these plants.
Recovery gauzes, according -to -the present invention,
are designed and fabricated by the process compris:ing the
steps of
(1) measuring the flow rate, conditions
and composition of the gas- eous stream -to be
treated with the gauze, then
(2) de-termining the physical properties
of t~le stream, either by measurement or
calculation;
(3t determining the mass velocity (G),
dynamic viscosity (~ ), and Schmidt No. (Sc)
for the process stream in which the gauze is
to be employed;
(4) estimating the recovery efficiency
of a selected gauze sheet based on -the
assumption that platinum recovery is essen-
tially mass -transfer limited; and
(5) fabricating and providing a gauze
which will provide a single sheet average
recovery efficiency over the catalyst cycle
within the range of this invention.
For example, recovery efficiencies can be estimated
for gauzes until an optimum configuration is determined,
which will have an average recovery efficiency exceeding
that given in column 2 of Table I.
Brief Description of Drawings
Figure 1 is a chart showing the predicted instan-
taneous recovery efficiency of various recovery gauzes in
a nitric acid plant, having a nitrogen loading of 15 tons
of nitrogen in ammonia per square meter of catalyst gauze
per day, operating at a catalyst temperature of 900C.
and an ammonia concentration in the feed of ~0 m/o tmole
percent)O
? i~
--5--
Figures 2 and 3 are charts which are anaklgous to
Figure 1, except that -the corresponding nitrogen loadings
are 57 and 100 respectively.
Figure 4 is a plot of the recovery function " ~ " as
a function of wire diame~er for a varie-ty of mesh numbers
for a linen weave gauze.
Figure 5 is a chart showing anticipated platinum
loss as a function of nitrogen loading for a typical
nitric acid plant.
Figure 6 is a graph showing the average recovery
efficiencies obtainable with the present invention over
the catalyst cycle as a function of nitrogen loading for
a typical plant.
Figure 7 is a comparison of predicted recovery
efficiencies wi-th a number of experimentally determined
points.
Figure 8 is an isometric view of a typical linen
weave gauze.
Figure 9 is a cross-sectional view of the catalyst
and recovery gauze package in a nitric acid reactor.
Figure 10 is a schematic of the nitric acid reactor.
In many cases, high efficiencies can be obtained by
using a gauze configuration in which -the initial product
of the mesh (in wires per unit length) and wire diameter
exceeds at least about 0.2 for gauzes containing a major
proportion of palladium and a minor proportion of nickel.
Preferably, the initial product of the mesh and the wire
diameter will be in the range of from about 0.2 to about
0.9. For the lower swelling alloys, such as palladium-
gold, the initial product of the mesh and wire diameter
should be at least 0.3 and preferably in the range of
from about 0.3 to about 0.9, more preferably from about
0.35 to about 0.9. For 95% Pd:5% Ni gauzes, it is
preferred -that the mesh N be in the range of from about
10 to about 80, dw is in the range of from about 0.003 to
about 0.090 inches, and their respective values are such
that the initial product of N and dw is greater than at
least about 0.2.
--6--
In practice, fo.r purposes of the present invention
under the conditions encountered .in most commercial
plants, the ini-tial. instantaneous recovery efficiencies
(~ ), that is the percentage of platinum in the stream
that is recovered by a single gauze having hiyh palladium
content, may be e~stimated by use of -the quasi-empirical.
formula
exp ~ 2C . a w _ _
L 2 / 3 Re ~ ~
wherein "~ " is -the volumetric void fraction of the gauze
which is less than 0.76, preferably less than 0.685, but
greater than 0; "Sc" is the Schmidt number for -the
diffusion of oxidized platinum in the effluent from the
catalyst gauze, which is usually between .8 and 1.0; "Re"
is the Reynolds number based on the wire diameter and
average velocity of the process stream just upstream of
the recovery gauæe~ i.e., Re = G dw, where "G" is the
-
mass velocity of the gaseous stream fed to the catalyst;
" ~ " is the dynamic viscosity of the effluent from the
catalyst gauze; typically, the Reynolds number will be
between 10 and 200, most often from 20 to 50; "C" is the
appropriate mass transfer correlation coefficient for the
geometry of the trial gauze which usually falls within
the range of from about 0.4 to 1; "a" is the specific
bulk surface area of the gauæe; that is, the total
surface area of one square inch of gauze divided by its
superficial volume, "a" usually has a value within the
range of from about 53 to 640 reciprocal inches; "m" is
the appropriate mass transfer correlation exponent for
the configuration of the gauze usually having a value of
from about 0. 6-Oo 8; and "d " is the diameter of the wires
in the gauze. Most of the wires used in gauzes according
to the present invention, will have diameters varying
between 0.0015 and 0.02 inches.
For a square linen weave gauze, such as those most
often encountered in practice, the following approxi-
.7_mations are useful:
a = rr[ 1 + N2dW ]~N
where "N" is the mesh or number of wires per i.nch, and
~ NdW ~ N dw )~
Methods of determining the appropriate mass transfer
correlation coefficient "C" and mass transfer correlation
exponent "m" are well-known to those ski.lled in the art.
A notable summary of the literature pertain.inc~ -to the
usual configurations is found in "Estimation of P1atinum
Catalyst Requirement for ~mmoni.a Oxidation" by Roberts
and Gillespie in Advances in Chemistry Series, Number
133, Chemical Reaction Engineering II, 1974 pp. 600-611.
For more unusual configurati.ons, these constants may be
determined experimentally. For the common stacked screen
gauzes, suitable correlations may be found in Satterfield
and Cortez, Ind. Eng. Chem. Fundamentals (1970) 9, 613
and Shah, Ph.D. Thesis, University of Birmingham, England
(1970). For the purposes of this inv~ntion, equation 1
will work adequately with values for "C" of .94 and for
"m" of .7 for the screens, reactors and flow conditions
described in -this application, if the values of Schmidt
number and viscosity given below are used, even though
the diffusing species may not necessarily be platinum
o~ide.
To expedite design of the recovery gauze for a
particular plant, efficiency vs. wire diameter graphs
similar to Figures 1, 2 and 3 can be constructed using
formula 1.
As a practical matter~ the properties of the gas
streams vary only by small amounts over the temperature
ranges encountered in practice of from about 810 to
about 960, so that properties at 900~C~ can be used with
only slight errorO Similarly, the concen-tration of the
feed to the catalyst is normally regulated to be-tween
10.0 to 10.5 /o (mole percent) ammonia and 90.0 to 89.5
m/o air, so the composition of the reaction products from
the catalys-t gauze remains cons-tant, so that physical
properties in tha-t range can be used. In -these ranges,
the Schmidt No. is about .9 .95 for diffusion of
platinum oxide vapors in air and the dynamic viscosity of
the gas is about ~2 x 10-5 poise.
Accordingly, the e~ficiency ~ i5 determined
primarily by the mesh "N", and wire diameter 7~ ~I / for a
given nitrogen loading "L", where the nitroyen loadiny
"L" is the number of short tons of nitrogen (irl ammonia~
passed through each square meter of -the ca-talyst gauze
per day. Thus efficiency can be plot-ted as a function of
wire diameter for a variety oE mesh sizes. Further, void
fraction can be shown parametrically on the same graph,
so that efficiency and the void fraction can be deter-
mined simultaneously for each given combina-tion of wire
diameter and mesh. For a given void fraction and number
of gauze sheets, the pressure drop through the gauze can
be estimated using known correlations. To obtain high
recovery efficiency without excessive pressure drop
across the gauæe, it is pre~erred that the volumetric
void fraction (~) be between about 0O76 and about 0.5~
Volumetric void fractions from about 0.5 down -to about
0.3 can provide even better recovery efficiencies, bu-t
care mus-t be exercised to properly support the recovery
yauze so that it is not damaged or displaced by the force
oE the stream of gas passing through it. In many applica-
tions, volumetric void fractions between about 0.685 and
about 0.5 will provide an excellent combination of
especially high recovery efficiency with acceptable
pressure drop. Void fractions of about 0.3 and lower can
be used to provide extremely high recovery efficiencies,
but many existing plants would require modlfication of
the gauze supports to withstand and properly distribute
the resulting force of the str~am on the gauze~ In some
circumstances, the cost of power due to pressure drop ~a~
also be of some significance. However, in practice, it
is normally sufficient to limit consideration to volu-
,.
metric void frac-tions above about 0.3 and preferably in
the range of from about 0.5 -to about 0.76. The most
preferred range of void fractions is from about 0.5 to
about 0.68~.
The method of fabrica-ting gauzes according to the
present invention is easily accomplished by plot-ting at
least a portion of the appropriate effi.ciency vs. wire
diameter graph for the condition.s, such as temperature,
pressure and nitrogen loading of -the plant under con-
sideration. Then the catalyst cycle length line can be
plotte~ on this graph using the followi.ng proce~ure, such
that if a mesh and wire diameter combination near the
catalyst cycle length line is chosen, the averaye re-
covery efficiency of the gauze over the catalyst cycle
( ~ ) will be within the range of this invention.
The catalyst cycle length line is plotted by deter-
mining which gauz-s will yield efficiencies ~ ) within
the range of this invention by first consulting Figure 6
and drawing a horizontal line corresponding to the
minimum efficiency determined from Figure ~ across the
appropriate efficiency vs. wire diameter graph, such as
Fi.gures 1-3. Thcn the appropriate recovery gauze cycle
lengths "T" for a variety of mesh sizes and wire dia-
meters above this horizontal line are determ1ned using
-the formula
T - W
1.25~ bL
wherein "W" is the weight of each square rneter of the
recovery gauze sheet and "b" is the amount of platinum
lost per ton of ammonia processed. In accordance with
the mode] of the present invention, the rate of platinum
recovery is approximately constant at least until the
recovery gauze cycle length has been reached, but de-
creases rapidly thereafter. "W" in general is ~ Pw ~.
" ~ " for a single linen weave gauze can be determined
from Figure 4. For gauzes of a weave other than linen
weave, the weight may be calculated in a similar fashion
1 0 -
from first principles or if necessary may be dete:rm.ined
empirically. If no hetter data is avai]able from the
plant hi.story or the history of a similar plant, "b" may
be estimated from Figure 5, presenting loss of platinum
per ton of nitrogen processed as a function of nitrogen
loading on -the catalyst gauze. Finally, -the catalyst
cycle length line is drawn connecting the poin-ts where
the recovery gauze cycle leng-th "~ " co:incides with -the
planned catalyst cycle length of the plan-t, "T~ Then a
gauze giving an acceptable efflciency and pressure drop
is chosen near this line. Preferably, to mi.ni~li.ze
interest costs, the minimum weight gauze which will both
yield an efficiency within the range of this invention
and match the planned catalyst cycle lenyth of the plant
should be chosen. It is preferred that the gauze sheets
used have a weight of less than 2.05 Troy ounces per
square foot or more preferably less than 1.9 Troy ounces
per square foot.
Provided that the recovery gauze cycle length of
preceding gauzes has not been exceeded, the recovery
gauze cycle length of the nth gauze is determined by
using the formula
Wn
Tn n~1
1.25~n bL [ ~ i)]
where " ~ i" is the recovery efficiency of the ith re-
covery gauze sheet and Wn is the weight of the nth gauze.
~s a practical matter, gauzes can be added until costs of
lost palladium, interest for the cost of the gauze,
fabrication and installation over the operating and
recovery cycle are not justified by the weight of the
platinum recovered~ Normally, from about .3 to about .5
grams of palladium will be lost from the recovery gauze
for each gram of platinum recovered. In many cases, it
will be advantageous to use gauxes of relatively coarse
mesh and large diameter wires in the initial layers of
the recovery gauze, and to use finer mesh, thinner wires,
or both, in the succeeding gauzes, even though the
efficiency of the initia:L yauzes may not be as high as
could be obtained. ~y appropriately choosing -the mesh
and wire diameter for each gauze, it is possible to
obtain recovery gauze cycle lenyths which are close to
the planned catalyst cycle length for each gauze in the
pack. This result can be obtained since the efficiency
of the downstream gauzes can be rnade yreater than the
efficiency of the upstream yauzes.
If it is desired to design recovery packs so that
there are approximately equal recovery gauze cycle
lengths obtained for each sheet in -the pack, the first
sheet of the pack should be designed as described pre-
viously, so that it will have an average recovery effi-
ciency over the catalyst cycle within the range of this
invention; i.eO, greater than 1 - exp (- 3.45/L 7).
Preferably, the recovery gauze cycle length for this
first gauze sheet will be in the range of from about
nine-tenths -to eleven-tenths of the planned catalyst
gauze cycle length for the plant. The geometric con-
fiyuration of each succeediny yauze sheet may then ~e
chosen, so tha-t the followiny relationship is approxi-
mately satisfied for each gauze sheet:
an(~n)l-m= 2- (G)mln ~ i ¦
where an, dn and ~n are the specific bulk surface area,
wire diameter and void fraction, respectively, for the
nth sheet in the gauze; ~1 is the average recovery
efficiency of the first gauze; ~i and~i are the respec-
tive recovery efficiencies and recovery functions for the
i h gauze sheet; Sc, G, C, m and ~ are as defined
previously, while n is the number of the gauze sheet
~12-
being desiyned in the pack~ For instance, for the second
gauze sheet in the pack, the relationship ~hould be
approximately satisfied wi-th n = 2, the third with n - 3,
and so on. Greatly improved results can be obtained by
insuring that at least one gauze (preferably at leas-t
two) in the pack has an average recovery efEiciency
exceeding 1 - exp (-3.45/L-7~ and that at least one, but
preferably at leas~ two, gauze sh~ets have a recovery
gauze cycle length of from about nine- terlths to about
eleven-tenths of the p:Lanned catalyst gauze cycle len~th.
Using the method of the present invention for a
given plant, it is possible to obtain average single
sheet recovery efficiencies over the catalyst cycle ( ~ ~
which are greater than the values given in column 2 of
Table I. Using preferred configura-tions, it is possible
to obtain average efficiencies greater than those given
in column 3. Figure 6 is a graph illustrating the
average recovery efficie~cies obtainable over the cata-
lyst cycle length with the gauzes of the present inven-
tion as a function of nitrogen loading, as compared to
eEficiencies reported in the prior art.
In practice, recovery gauzes almost always contain a
major proportion of palladium or gold and minor additions
of other alloying elements which improve mechanical
properties. By major proportion of palladium, it is
meallt that the recovery gauze contains at least about 70%
palladium by weight. Preferably, the recovery gauzes
will contain at least about 80% palladium and more
preferably 90%. The most preferred recovery gauzes
contain at least about 95% palladium hy weight. Perhaps
the most widely used alloy has been an alloy containing
80% palladium and 20% gold. While this alloy has found
wide use, alternatives have been sought, since inclusion
of gold greatly increases the cost of the gauze. Other
alloying elements for palladium include other platinum
group metals, nickel, manganese, chromium, carbon, horon,
and the like. Particularly useful palladium alloys
-13~
include palladium/gold, palladium/platinum, palla-
dium/nickel, palladium/copper, palladlum/ruthenium, and
palladi.um/silver. Alternatively, gauzes containing
major proportion of gold and a minor proportion of a
platinum group me~al have been suggested, since it has
been reported tha-t gold does no-t volatilize to the same
extent as palladium. The ability of these gold-rich
alloys -to withdraw platinum seems -to be somewhat less
than the ability of palladium-rich alloys. In the sarne
fashion as the palladium-rich alloys, the mechanical
properties of the gold-rich alloys may be improved by
adding metals which have a greater a-ffinity for platinum
than for oxygen, such as tantalum, niobium, and the like.
Other suitab]e alloying elements include -titanium,
zirconium, chromium, nickel, manganese, and the like~
TABLE I
Plant Loading Ef~iciency
Tons of Ni-trogen of Gauzes of Efficiency
As Ammonia Per the Present of Preferred
m2 Per Day _ Invention Gauzes
10--15 49 52
15-20 41 44
20-25 35 37
25-30 31 33
30-35 28 30
35-~0 25 27
40-45 23 24
45-55 21 22
55-65 19 20
65-75 17 1~
75~85 16 16
85-100 15 15
100 + 13 14
r~
-14-
For purposes of the present inventlon, the preferred
alloys are pal]adium/gold and palladium/nickel alloys,
particularly alloys containing at least about 80% palla-
dium. 95~ palladium and 5% nickel is a particularly
advantageous alloy for the practice of the prescrlt
invention, since it is re]atively inexpensive, is easily
fabricated and upon exposure to the hot platinum-
containing effluent, the wires swell and may double in
diameter before they are to be removed. In some cases,
the diameter of -the wires in the gauze may more than
double, reaching approximately 2~ times thelr ini.tial
diameter~ When properly allowed for, this swelling can
be particularly advantageous, as the efEiciency of the
gauze increases as the wires swell. For example, in a
plant having a nitrogen loading of 57 tons per square
meter per day, a 36 mesh by .0068 in. wire diameter gauze
with an ini-tial efficiency of about ll~ could provide an
efficiency of about 14% after the wires swell to .012
in., and over about 18% if the wires reach 2~ times their
initial diameter, Thus, a gauze which provided an instan-
taneous ef~iciency which was initially outside the range
of the present invention, can swell to provide an average
eEEiciency in the range of the present invention provid-
ing a much higher efficiency than would have been pre-
dicted based on i-ts initial configuration.
Thus, when nickel/palladium gauzes are used, a gauze
may be selected such that its recovery efficiency based
on its initial configuration is less than
1 - exp (-3.45/L-7), but upon swelling, these gauzes
provide an average recovery efficiency over the catalyst
cycle in excess of that glven in Table IA.
--15-
_BLE IA
Efficienc~ Loading
48 10-15
15-20
3~ 20-25
25-30
27 30-35
2~ 35-40
22 ~0-~5
45-55
1~ 55-65
16 65-75
75-85
14 85-100
12 100-~
In the case of 95% Pd:5% Ni, the average recovery
efficlencies over the catalyst cycle ( ~ ~ correlate best
when recovery ls predicted based upon the geometric mean
of the initial and swelled diameters, but adequate
correla-tion for the 80% Pd:20% Au gauzes can be obtained
if recovery is predicted based upon initial diameter,
since the effect of swelling seems to be somewhat less
pronounced. If it is desired to account for the effect
of swelling in a palladium-gold alloy gauze, the
geometric mean wire diameter may be es-timated by
multiplying the initial diamete.r by 1.1. Often for 95~
Pd:5~ Ni, the geometric mean diameter can be estimated
satisfactorily by multiplying the initial diameter ~y a
factor in the range of from about 1.4 to 1.6, depending
on the location of the gauze in the recovery pack with
the higher end of the range being used for the first or
second layers in the pack and the lower end for the fifth
and sixth layers. See Operating Example 11 for more
details. Thus, Equation 1 can al 50 be used to estimate
-16-
average efficiencies if geometric mean wire diameters are
used and the recovery gauze cycle length is not exceeded.
Specific Embodiments
As illustrated in Fiyure 8, the recovery gauzes of
the ~present invention may be employed in the form of
screens 10 having wires 20 and openings 30. ~s ex~
plained, the combination of -the diameter of wi~es 20 and
the mesh or number of wires per lineal inch determininy
the mass transfer parameters (MTP) of the screen accord-
ing to the formula
~TP = a dw
-- 1 --
Then, the number of mass transfer units (MTU)
represented by a single gauze may be determine~ from the
relationship
MTU _ 2C / MTP.
Sc
As shown in Figure 9, prior to a typical run, a
gauze ensemble 20 is placed into reaction chamber 40
(Figure 10) of a combustion vessel 42. This ensemble 20
includes recovery gauze pack 21 and catalyst pack 25
placed adjacent to one another. Catalyst pack 25 con-
tains individual sheets 24 of catalyst in the form of
net-tings or screens stacked one atop the other. In
Figure 9, the catalyst pack is depicted with seven sheets
of catalyst, but it is to be understood that the precise
number of sheets is not critical and they may be
increased or decreased as needed to effect an essentially
complete conversion of ammonia to nitrogen oxides. One
such catalyst consists of 90~ platinum/ 5% rhodium/5~
palladium, but other platinum- containing catalysts may
also be employed with good results. Recovery gauze pack
21 contains two sheets of recovery gauze 22 sandwiched
between separator screens 23. The recovery gauze packs
must be of sufficient mechanical strengt~ to withstand
the force of the process stream at high reaction tempcra-
tures while simultaneous]y enduring the corrosive effects
of -the residual ammonia, oxygen and ni-trogen oxide
proclucts which are formed during the process.
Design Example I
A recovery gauze is to be desi.gned for a nitric acid
plant operating at 900C., 10% NH3 and a loadi.ng of 15
U.S. tons of nitrogen in amrnonia per square meter per
day. The plant operates on a cyc:Le length of 130 days,
at a pressure of 100 p~s.i.g. To begin, a diagram
(Figure 1) is prepared of the single sheet efficiency oE
a recovery gauze as a function of mesh size and wire
diameter. Figure 6 is then consulted~ and i-t is deter-
mined that an efficiency in excess of 40% should be
obtainable. It can be seen from Figure 1 that a 50 mesh
gauze with wires .0095 in. in diameter would provide a
suitable instantaneous efficiency (~). Therefore, to
allow for swelling, a 50 mesh gauze with wires .006 in
diameter is prepared from 95% Pd:5% Ni. Upon use in the
reactor, the gauze swells by a factor of about 2.5 to a
wire diameter of about .015 in., providing an efficiency
(based on -the geome-tric average wire diameter of .0095)
in excess of ~0%. From Figure 5, it can be estimated
that such a plant can be expected to lose about .~ to .9
grams oi platinum for each ton of nitrogen converted.
Thus, about 12.75 grams of platinum per day are presented
to each square meter of gauze which weighs about 916
g/m2. Upon operationl the first gauze sheet can be
expected to remove over 40% of this for a recovery of
about 5.1 grams of Pt per day per square meter of gauze,
or about 665 g/m2 over the ca-taylst cycle. The recovery
gauze cycle length coincides closely with the planned
cycle length of the plant, so this gauze may be used
without a heavier, but less efficient gauze upstream of
it. About .3 to .4 grams of palladium can be expected to
-]8-
be lost for each gram of Pt recovered. Three screens are
used -to achieve an average recovery efficiency of 78%. A
successive finer and lighter screen may about be used
downstream -to recover a portion of -the residual platinum,
if so desired.
Design Example II
A gauze is to be designed for a plant similar to
-that in Design Example I, except that the loading is 57
tons/m~-day, and the cycle leng-th is 60 days. According
to Figure 5, a plant of this type can be expectecl to lose
between about 1.4 and 1.6 yrams of Pt per -ton oE amrnonia
converted. Figure ~ shows that an efficiency of more
than 17~ can be obtained. It can be seen from Figure 2
that this can be obtained with a 60 mesh screen having a
wire diameter of .006 in. An 80% Pd:20% Au screen having
these dimensions is selected. Upon operation, about 85
grams of Pt are presented to each square meter of the
screen and about 14.5 grams are collected each day. Six
screens are used to provide an overall average recovery
efficiency of 67%.
Design Exarnple III
._
A gauze is to be designed for a plant having a
loading of 100 -tons of nitrogen in ammonia per square
meter per day, and a cycle length of 60 days. According
-to Figure 5/ a plant of this size can be expected to lose
between about 1.7 and l.9 grams of platinum for each ton
of nitrogen converted, ~hile an efficiency in excess of
12% can be obtained. However, if an 80% Pd:20% Au gauze
having a mesh of 80 and a wire diameter of .005 in. is
used, even though an efficiency of over 15% is obtained,
the recovery gauze cycle length is shorter than the
catalyst gauze cycle length. Therefore, coarser, heavier
gauzes should be inserted upstream of the finer, lighter
--19~
recovery gauzes after the ca-talyst gauze. Since an 80
mesh by 0.005 in. wire diameter gauze of 80~ Pd-20% ~u
has a recovery gauze cycle length of 60 days with a
platinum recovery of 948 g/m2, the number of grams of Pt
presented to each square meter of the first gauze mus-t be
decreased from about 180 grams -to about 105. Thus, 4
coarse gauzes of 50 mesh by .0085 wire diameter should be
followed by 4 fine gauzes of 68 mesh by .006 wire
diameter to achieve an overall recovery oE 67%.
Design Example IV
A recovery gauze system is to be designed for a
nitric acid plant operating at 4.5 atmospheres pressure
and a nitrogen loading of 13.2 tons of nitrogen in
ammonia per square meter per day over a catalyst gauze
cycle length of 150 days. The catalys-t loss rate is
known to be 0.144 g. of Pt and Rh per ton of nitrogen.
The production rate of the plant is 330 tons of HNO3 per
day, and the effective area of the reactor is 5.8 square
meters.
If two s-tandard 80 mesh by .0031 in wire diameter
recovery gauzes of 80~ Pd:20% Au are used, the predicted
recovery gauze cycle length is only about 130
days, resul-ting in an average recovery efficiency of the
plant cycle length of approximatley 46% per gauze or a
total of 71% for both yauzes.
By following the procedure of Design Example I, it
can be seen that if two 50 mesh by .0064 in wire diameter
recovery gauzes of 95% Pd: 5% Ni are used instead of the
standard gauzes, the predicted recovery gauze cycle
length for -the first gauze slightly exceeds 150 days for
an average recovery efficiency over the cycle length of
72% per gauze for a total of 92%.
Thus, each square meter of the improved gauzes of
the present invention recover over 370 additional grams
of platinum over each cycle length.
~ 3
-20-
In th~ following Operating R~amples, a gaseous
stream of air containing about 10% N~13 by volume was fed
to the reactor at a rate of 680 standard cublc feet per
hour.
Prior to beginning a run, the feed gas was preheated
to a temperature within the range of from about
290-310C.; during the run the gauze exit -temperature was
maintained at a rela-tively constant 930C. In Operating
Examples 1~6, infra, the run was conduc-ted over an
appro~imately 1~6 hour period, and in Operating Examples
7 and 8, the runs were maintained for approxirnately 292
and 483 hours, respec-tively; however, it will be
appreciated that, in practice, the reaction period may be
varied over a wide range. In Operating Examples 9, 10,
11 and 12, the experiments were conducted in opera-ting
nitric acid plants.
'rhis invention will now be illustrated by making
reference to examples which specifically describe the
gauzes of this invention and their use in recovery
processes; however, these examples are illustrative only
~nd this invention should not be construed as being
limited to these examples. In these examples, all
proportions for the metals comprising the oxidation
catalyst and recovery gauzes are in weight percent,
unless otherwise stated. All tonnages in this applica-
tion are in U.S. ( i.e., short) tons.
Compar~ = ~ ti ~
A recovery gauze pack consisting of two 80% Pd:19.~%
Au:.6~ Ru, 80 mesh by .0039 inch wire diameter gauze
sheets was placed between three separa-tor screens, as
shown in Figure 9, and this ensemble was placed into a
reaction chamber below 10 sheets of 90Pt/5Rh/5Pd oxida-
tion catalyst having a weight of ~.6769 g. The recovery
gauzes had a mesh (N~ of 80 and wire diameter (d ) of
0.0031 inches. The sur-Eace area of each recovery screen
B5~
-21-
(bulk surface area of the wires ~er unit volume of
screen) was 263 i.n. 1.
Feed gas consisting of ammoni.a and air was ~orced
through the oxi.dation catalyst and recovery gauze pack as
a mixed gas stream under a pressure of 100 p.s.i.g. for a
loading of 57 tons of nitrogen per square rneter per day.
The yield of nitrogen oxides (NO ) was about 95%.
The average recovery ef:Eiciency (~*) for -the P-l/Au
recovery gauze pack was determined from assay data by
measuring the Pt gain of each recovery gauze and the Pt
loss of the ammonia o~idation catalyst ( l.e., the
oxidation gauze pack) as follow~:
= Pt Gain Per Recovery Gauze
Pt Loss of Oxida- Total Pt Gain For
tion Gauze Pack - Preceding Gauzes
Following the run, the catalyst weighed 4.3951 g.,
and subsequent assay data showed a Pt loss of 0.4203 g.
in the oxidation catalys-t. By comparison, the first
layer of the recovery gauze pack weighed 0.5965 g., with
a P-t gain of 0.0853 g. Based on this data, the average
Pt pick-up efficiency (~') of the first recovery gauze
was found to be 20.3%. In calculating the Pt pick-up
efficiency of the second recovery gauze, the weight of Pt
gained by the first recovery gauze must be taken into
account. The second ]ayer of the recovery gauze pack
weighed 0.57~7 g. with a Pt gain of 0.0592 g., and the Pt
pick-up efficiency of the second recovery gauze was found
to be 17.67%. The average Pt recovery efficiency was
found to be 19.00~.
Similar comparative studies on similar recovery
gauze screens of varying mesh (N) and wire diameter (d )
were conducted; using the procedure described in Example
l to further confirm the applicability of the present
model. In each study, the gauze screens and catalysts
employed were weighed prior to use and immediately
thereafter. Gauze assays were conducted for individual
22-
gau~e sheets -to obtain the average Pt recovery efficiency
(~). Table II summari~es the geometry o~ the recovery
gauzes employed in Operating Examples 1-6. Based upon
these geometries and the flow conditions, the
dimensionless mass transfer unit for a slngle screen was
calculated from
MTU = 2C _ . ad
Sc 2/3 Re ~
Predicted average recovery efficiency (~ ) of a single
sheet could then be estimated using the formula
~ = l-exp ~-MTU). In a].l runs, the conversion of
ni-trogen oxides was in the range of 95-98.9~.
The Pt loss for the catalyst of Operating Examples
1-6, the P-t gains for the respective recovery screens,
and their recovery efficiencies are set forth in Table
III. These results are represented in Figure 7, illus-
trating the correlation of the present invention be-tween
recovery efficiency (~) and MTU.
~
o ~ ~ l~
~ r~
E~ ~ ,~ ~ o
o o o o
u l
~ u -
h~ ~)
~ ~~
r~ O O O O O
O O ~l
r~
S~
a~ -,, ~ i~ O r~
~ ~--
~1 C) ~ ~1 ~ r1
U S~
~) H ~1
_~ O ~ a~ o
J . ~" ~v~ ~) CO
O O O O
~ ~ O O O O
,~ ,~ 3 o o o o
E~
~1 o o o ~r
(1~ N CO CO
O r~ ~
Z
a)
N O
E~
U)
X rl 1~
.5~
I :~ ~,;
~ ~,
~ .~ ,~ .. ,
O ~ ~ o l- o (~ ~r ~r ~, r~l ~'
r~ O\o ~ D o ~r `D If-) ~D r`J r~
rL) r~-- . . . . . . I . . .
o 1--r~ co ~g 1-- ~
~H ~1 ~ ~ r-1 ~1 r~ 1 ~ H
;~
,_
.~ r~) r\l ~g ~r ,~ O
~_ u~ ~ ~ Ln ~g ~r
I r Lr) I ~g ~r r~ r~
t~ o o o o I o o
,~) w , . . .
P~ o o o o o o
u~_~ r~ r~
O ~ o ~r r~
_s ~ ~ r,~
_~r r~l r~
P~ O o O
_
r~ o o o oro a: o r~
5-1 ~ r~ r~ r~ g o ~g r~
~ ~) 0\o . .. . .. . ..
0 ~4-- ~D ~r o 1-- o ro [~ ~ O
r~ ~ ~ r ~ r~
~r
H
H ~ ~ u~ 1-- ~ r~ Ll~ r~
H ~ In ~g ~r r~ ~g ~1 c~ ~ o
~r rd ~ ~ ~ a~ g r--~g r.
~1 ~ r~ cr r~ r~
~a ~ ~
m ~ ~ ~r o o ~ro o ~r o o
E~
~ 0\o o o o
p,,_ ~ I I r~
(~) ~ ~ n r.~l r-- o r.~ ~ o
.~ ,~ _ ~g r~Ln r~ ~ o ~
~ ~ ~ ~ ~r ~r ~ r~ r~) co ci~ r70
`~ ~g u~ g r~ ~
C ~~ . .. . .. . ..
H ~: ~r O O ~r o o ~r o o
~H rJ~
O
O ~ ~ O ~ O
. rD ~ ~1 ~
Z'~ .
~ ~ h
Ln rv r
a ~J ~ ~r rd rd --~ ' ¢
n In O
~ ~ ~ ~~ Q ~ Q;
::~ ~ ~ rn ~ ~ ~ ~n ~ ~) ~ rn
tYi ~ E~ ~1 r.~ ~ r.~ ~ r~
r~ ~ ~ r~
'~
i~
(~ r~
O C ~ n ~ ~ cn ~ ~ ) ~"
;) ,1 c~O cn ~ ~ c~ o ~ ~r o t-
O U-- I . . ,I . . . I. . .
f~ ~ f"l ~D ~t' r~ cn ~o c~ c~
P~W
c,., r~ r~ r~ O ~r
_~ n c~ cn ~D c5~ co
I r~) r~ I rn ~r Ir~) rn
oo oo oo
-- , . .. ..
~ oo oo oo
.~ . .-1
u~
.. u~ ,~ ~, ~ o
.~ O ~ ~
. ~1 ~ t- I I c~ I I u~I I
_
P~ o o o
~n o o o ~ o o o ~ c~
,~ ~ ~ c~;r r~ n ~ 1 cn
~ ~) G\O . , . ~ . . . . .
O ~4-- o ~ ~ o,t ~ ~ ~ n
~ CS~r--l ~ CS~ ~ ~ C~O
~D
H ~ ~ n n o ~ o ~ n~r c~
,L _ COC~ CS~ O ~ r') n c~ o
d ~ ~ h ~ ~ r~ o
~r ~1 .,~o_ .........
~1 ~ h ~ ~r o o ~ o o ';r o o
.,L~ G\O O O O
~~ cn I ~ cn I I cn I I
~1
f~t ~ n ~ cn n o a~ c~
,1,c ,_ n ~ ~ ~ c~7 Ln
h cn cn ci7 Cl~o ,~ cn o ~
rl rl ~ ~D r.~l f.~l ~O r~7 r~7 ~ ~O ~9
H ~2 ~ O O ~r O O ~r O O
4-1 (1~
O
~ O ri r~ O ,~ r~ O ,~ r~
O ~ r-l rl rl
t
~7 ~ 7
~t f~d,~ ~7 fd fd ~ f~ f~t
~; ~ ~ ,~ ~1 ~1
~n ~-,1 fl) .f~ f~ u~ fl)
~t ,~ ~ ~ ~ Q
~t ~ ~ u~ ~ u~ u~
Y; .~ ~ ~ [~
fl~
,~
f~t ~r f7 ~D
a
~erating E ~
A recovery gauze pack, designed accordiny to -the
principles of thc present invention, consistiny of two
recovery gauze sheets (95Pd/5Ni) was placed between three
separator screens, as depictecl in Fiyure 9, and this
ensemble was placed into a reaction chamber of the type
shown in Figure 8 below a 90P-t/5Rh/5Pd oxidation catalyst
(15 sheets). The recovery yauzes were 60 mesh and had
wire diameters (dw) of 0.006 inches. The ammonia
oxidation catalyst weighed 7.1066 g. The separator
screens were in -the form of a wire mesh gauze constructed
from a ferrous alloy.
The ammonia and air feed was forced through the
reaction chamber over a 292 hour period as a mixed gas
stream under a pressure of 100 p.s.i.g. for a nitrogen
loading of 57 tons/m2-day.
The average Pt pick-up efficiency (~) for the Pd/Ni
recovery gauzes was determined by measuring the Pt gain
of each recovery gauze and the Pt loss for the ammonia
oxidation catalyst from assay data.
Fo]lowing the run, the catalyst weighed 6.1783 g.,
and the subsequent assay data showed a Pt loss of 1.0393
g. in the oxidation catalyst. The first layer of the
recovery gauze pack weighed 1.0110 g., and recovered
0.2709 g. of platinum, based on the gauze assay data for
an average platinum pick-up efficiency ~) of 26.07%.
The second layer of the recovery gauze pack weighed
0.9560 g. and recovered of 0.1998 g. of platinum, based
on the gauze assay data for a platinum pick-up efficiency
of 26.0~. The average platinum recovery efficiency was
found to be 26.04%, which is an extremely significant
improvement over known getters operated under similar
reaction conditions.
The gauze screens and catalyst employed in each
study were measured and weighed prior to use and immedi-
ately thcreafter. Assays were conducted on the catalyst
and individual gauze sheets in the manner described in
-26-
the preceding paragraph to determine the average platinum
recovery efficiency (~).
The configurations of the recovery gauzes employed
in this study are set forth in Table VI:
E~ co ~
r-l ~
o o
o l
~l u^
s~ ~
a~ ~D
In
~D
~ o
r~ o o
~o ~ ~
O ~D
~ o f~
s~ ~- ~ ~
o h
U~=
a)-
~D
~_ o a) ~ o ,~
r~ o o
~> rl
H 3 r~ 3 O O
~ ~_
rC (L) ¦
U~ N -
~I) rl ~ O O
Ul rl
æ
a
N a)
~ ~1 r~ r~
t~
4-1
O
O ~ O
,1 h
~ U~ X; ~ p;
-28-
On the basis of this study, -tile geometric average
mass transfer unit for the recovery gauzes was calculated
at 0.260.
Operating Example 8
The procedure of Operating Example 7 was repeated,
except that the recovery gauze was opera-ted over a period
of 483 hours. The results of this study, inclusive of
platinum loss, platinum gain for the recovery gauzes, and
their recovery efficiencies are se-t forth in Table V.
These results are also represented in Figure 7.
~, ~
a) c) rd
~ ~:
O (1) ~ I-- o o r-l ~
rl 0~ I O O O Ir~l r~) ~I
0~ . . . . . .
P~ rl ~ D ~D r~ N G
~) ~H ~I r~lr~ I r~) r~
rl ~ ~ O 1-~
o r,~ r~
r;) I
~ ` , ~I r~
P.l ` O O ~.~ O
I .
t~ ''
n ,-~1 r~) r~
() ~_1 ~ ~0
t~ r~
~1 ~_1 r~l
O O O O O O
_ I~ corJ~ ~D ~ Lr,
~ o\o
r1 Ç4 ~ ~ ~ ~ ~ CO ~ r~) r~
n tn
~1
.rl ~ _S
O O
r~l S O O ~ ,~ r~l 1~)
S ^ r,~ ,r--l ~) U~ r--I L(')
tl~ ~1t-- r~lr1 U) Lr) CO r~
1 -rl 1~ ~-1 rs~o ci~ r.~ r~ o r5
m ~-- ~ . c ~ . .
r~ _ ~ ~ r-l O Lf) ~ r-l O
~ O O
,1 p,,_ a~ I I ~ I I
rl
~ ~ ~ ~ Lr~ r~J r~
rl S ^~D a) O ~ O r~
~: t5~ ~1 O ~' 1.') r~ ~ r~
1--1 rl !~ r-l C~) ro ~ ~ r~
a,)--
~ ~ O O 1~ 0 o
4~ tn
O
a )
C)
O S ~:
ZU~
' ~S-l r~
tn j
tn~ rj(~~ rl iTj r~
J ~ Zt- l t~l S ~Z; t~
n ~ i tn
Q, tn ~ ~ tn
~) ~r-lr~ ~ ~, r-~ (~I
~ P~
r ~ .
I_ r~
X
-
-30-
The improvement in Pt recovery efficiency ~ for the
95Pd/5Ni recovery gauzes of Operating EY~amples 7 and 8 is
illustrated by Table VI. The benef:i.cial effec-ts
attributable to the use of palladium/ni.ckel and the high
average mass transfer unlts fo.r the Pd/Ni yauzes of this
invention makes -them particularly suitable for
platinum/rhodium metal recovery. The data in Table VI
demonstrates the advantages of the Pd/Ni recovery gauzes
of this invention and the improvement in platinum
recovery efficiency for the 95Pd/5Ni recovery yauzes of
Examples 7 ard 8, when compared against an 80Pd/19.4Au/
0.6Ru recovery gauze of Examples 3 and 4 having a similar
initial MTU.
O ^
Lli r-l
.
r~ ,-1 0
P~ d~ oY'
Il~ .-1 .. ~ ~r
cn I ~ ) I I o ~1
.
~ ~ O
~ ~I t
O aJ ~D ~
PJ Id
~) ~ `-rl
u) ~a rl
o a) ~
~ ~ ~ H
H U ~D
~_) ~1
~ ~ .
~) H (U
O ~ r~ ~
E~ C) ,_, ~ o
o~P
. ~ I~
X ~ ~ ~D
:Z CO p~ ~ O
H ~
O S rl
1~.,1 Dl
~1.~
tl~ H
O o--
~1
X
Comparative Operating Example 9
A recovery gauze pack consisting of five 80%Pd:
20% Au, 24 m~sh by 0.008 inch diameter wire gauze sheets
were placed between six separator screens in an
arrangement similar to that shown in E'iyure 9. This
ensemble was placed immediately downstream of a platinum
alloy ammonia o~idation catalys-t pack (90Pt/5Rh/5Pd) in a
nitric acid plant having a nitrogen loading of 7g tons
nitrogen (calculated as ammonia) per square me-ter oE the
effective cross-sectional area of -the recovery gauze per
day ~ i _ , 78T(N)/m2/d). The plant was operated for 77
days, during which the oxidation catalyst los-t 205 troy
ounces in weight, of which 92% of 188 troy ounces were
estimated to be platinum. At the end of the 77 day
operating cycle, the recovery gauze ensemble was removed,
weighed and assayed to determine the amount of platinum
recoveredO Platinum recovery was found to be 42 troy
ounccs or approximately 22~ of the estimated lost
platinum.
The mass transfer unit (MTU] for a single gauze in
the recovery pack was calculated to be 0.05, based on its
mesh (24), wire diameter (0.008 inches~ and nitrogen
loading (78T(N)/m2/d). The total calculated platinum
recovery Eor the five sheets was 24%, a figure which
compares favorably with the observed recovery of 22%.
*o illustrate the effectiveness of this system, a
recovery gau2e pack was constructed by placing five gauze
sheets (manufactured from an alloy of 80Pd/19.4 Au/0.6Ru,
having d mesh of 36 and a wire diameter of 0.0071 inches)
individually between six separator screens. The recovery
gauze pack thus constructed was placed into a reactor
with a nitrogen loading of 78~(N)/m2/d. In this
operation, the single gauze mass transfer unit (MTU) was
calculated at 0.082, and it was predicted that five
sheets of recovery gauze would recover about 34% of the
pla-tinum lost from the oxidation gauze catalyst.
The recovery gauze pack was installed in the plant
immediately downstream of the oxidation qauze pack and
the plant was operated for 78 days, ~uring which the
oxidation gauze lost 213 troy ounces in weiyht, of which
92~ or 196 troy ounces was estimated to be platinum. At
the end of the 7~ day cycle, the recovery c~auze pack was
removed and the quan-tity of platinum recovere~ was foun~
to be 35%, based on the recovery yauze pack weight and
platinum assay. This figure compares favorably with the
predicted recovery of 3~%. These data are represen-ted on
Figure 7.
Comparative Operating Example 10
A recovery gauze pack consisting of six 36 mesh and
0.0071 inch diameter wire recovery gauze sheets were
individually placed between seven separator screens. The
recovery gauze sheets were manufactured from an alloy
composed of 80 weight percent palladium, 19.4 weight
percent gold and 0.6 weigh-t percent ruthenium. The
recovery gauze was placed immediately downstream of a
platinum alloy ammonia oxidation catalyst pack
(90Pt/lORh) in a nitric acid plant having a nitrogen
loading of 65 tons (in ammonia) per square meter of
reactor cross-sectional area per day, ( i.e.,
65T(N!/m2/d). The plant was operated for 61 days, during
which the catalyst pack lost 137 troy ounces in weight,
of which 92% or 126 troy ounces were estimatcd to be
platinum. Based on the wire size and mesh of the
recovery gauze sheet and the nitrogen loading for the
particular plant, the mass transfer unit (MTU~ of a
single gauze was found to be 0.093, and the predicted
total pack recovery for platinum was calculated at 43~.
This predicted recovery figure (43~) compared favorably
with the actual or observed platinum recovery of 52%.
This result is shown on Figure 7.
-34-
Operating Example 11
A platinum recovery gauze pack consisting of 95%
Pd/5% Ni were individually placed between seven separa-tor
screensO This pack contained six sheets of pla-tinum
recovery gauze, the first three having a mesh of 45 and a
wire diame-ter of 0.0083 inches and the last three having
a mesh of 60 and a wire diameter of 0~005 inches. This
pack was placed immediately downstream of a 90Pt/5Rh/5Pd
alloy ammonia oxidation catalyst gauze pack in a nitric
acid plant having a nitrogen throughput of 38 tons (in
ammonia) per square meter effective gauze cross-sectional
area per day ( i.e., 38T(N~/m2/d). The furthest upstream
of the platinum recovery gauzes was gauze sheet
1~ followed by gauze sheets 2, 3, 4, 5 and 6, that is,
gauze sheet 6 was located the furthest downstream of all
of the gauzes. The plant was operated continuously for
71 days during which the ammonia oxidation catalyst pack
lost 443 tro~ ounces in weight, of which 408 troy ounces
(92%) were estimated to be platinum.
At the end of the 71st day in the operating cycle,
the platinum recovery gauze was removed from the plant
and disassembled for inspection. During operation, -the
recovery gauze wires increased in size over
their original diameter, and this increase significantly
affected their mass transfer unit values. The wire
swelling factor (S) for each gauze sheet was determined
accord.ing to the following equation:
S = (Avera~e Final Wire Diameter~-(Initial Wire Diameter)
~Inltial Wire Diameter)
and the results of these determinations are set forth in
Table VII.
~ .
-35-
TAsLE VII
___
Average Swelling
Gauze Sheet Factor (S)
~.
l 1.~5
2 1.30
3 1.08
4 1.0~
0.90
6 0.90
On the basis of the initial wire diameter, mesh size
and nitrogen thxoughpu~, the total platinum recovery for
the gauze pack could be predicted to be 69.8%. On the
basis of the final wire diameters for each platinum
recovery gauze with identical mesh and nitrogen loading
parameters, t.he total platinum recovery could be
predicted to be 83.7~. The recovery, properly based on
geometric mean of the final and initial diameters of the
wires in the recovery gau7.e, is 76.4%.
An as~ay of the platinum recovery gauze pack showed
an actual total pack recovery of 306 troy ounces of
platinum (75.0~). The observed recovery of 75.0%
compares favorably with the predicted recovery of 76.4.
A summary of the parameters for the recovery gauze
packs of Operating Examples 9-11 and their respective
recoveries of platinum metal. are set forth in Table VIII,
infra.
-
~ ~a ~ r~
h ~ ~\oN ~) U~ 1--
a) h _
O U~
~d C~ .4
P~ a) o
-i ~
rd ~ "~i *
ci .y
O ~: ~ ~~ ~rr~) o ~D
r ~ O o\o
~d ~ _
,-i ~
~ ~4 ."
.; j ~; a
~1 ~ ',~
,~ O W ~ ,r~
~r ~ ~a~
~ ' ' ' ' ~d rl
.-i ,i ~d o o O O ,I ra
O O ~i ~
~ ~ ~1
_
~Lr) ~ ~ a~
E-lL~'l r~cn ~
~O o o .~
O O O O
~-- O
a) ~ ~
ri tn :~t--ro ,~ 'a
rJ ~ d~ r~
~ aJ a :~ ~r a
u~--
* o c
,~ rj r~ r
H l
H . ,~ ~ O O O ;Z. ~a ~i h ~
H ~ o\o~ ~\jr,~
;> o o o-- O o a
ri ~.ri ~ ~ a a ~ co ~ .
-i 8 ~
C~ co oo r~ S
-- ~ 1 0 S '~
r~ r-- r--co ~ ~ --i ~ O
~ ~O o oo O -i
a) ~ iO O OO O ~ U.i ~
h id ~,, ,, . o ul -,i O
,i ~,iO O OO O Ql O U~
~ a ~ 8 r h .4
u~~r ~ o c~ O
r~ r~~r ~ ;
~: ,-i .
rd r-i ~i O
h u~ -,i 'i U~ O
O d~ ~ ~D~ r~ æ ~
z~ ... * *
~ u~
~ ~j
~i E~
~ ~ o
a) ~ z
~ ~_co r~ ~ rD
O ~ri1--r~
h O Z
h~
,i~
a
,I h
~d ~ cs~ o -i
~C ~ r-l r1
~i ~
~5~
-37-
~perating Example 12
A recovery gauze pack consisting of two yauze sheets
(80 Pd/19.6 Au/0.4 Ru) were placed be-tween three
separator screens, and this ensemble was placed into a
first reaction chamher below a 90 Pt/5 Rh/5 Pd oxidation
catalyst llO sheets). The Pd/Au gauzes have a mesh of 36
wires per linear inch (N~ and a wire diameter (dw) of
0.0071 inches (N x dw: 0.256~. The oY.idation catalyst
weighecl 4.6963 g., and the recovery gauze pack weighed
5.1737 g. prior to the run. The surface area "a" of each
recovery screen, tha-t is, the surface area of the wires
per unit volume of screen packing, was 117 ina/in3 and
0.555 ftZ per Troy ounce of recovery gauze. The
separator screens were in the form of wide mesh gauze
constructed from a ferrous alloy.
In a second chamber, located immediately behind the
first chamber, there was placed a second gauze pack
consisting of two recovery screens (80% Pd:20% Au),
sandwiched between three separator screens.
The two chambers were preheated to 300C. and
ammonia and air were channeled therethrough as a mixed
gas stream under a pressure of lO0 p.s.i.g. at a total
flow of 680 SCFH. Ammonia constituted 10% of the gaseous
mixture representing a throughput of 57 tons nitrogen per
square meter per day, that is, 57 t(N2)m2d. During this
run, the first chamber was maintained at a temperature of
930C., and the second chamber was maintained at 890C.
The test was run over a 146 hour period, and the yield of
nitrogen oxides ~NOX~ was 98.4%.
The weight recovery efficiency ~') for the Pd/Au
gauzes was determined by measuring the weight gain of
each recovery gauze pack and the weight loss for the
ammonia oxidation catalyst. The difference in weight was
then converted to weight recovery efficiency ( ~ ')
according to the following equation:
~ R)l/n
-38-
where n and ~' are as defined hereinabove, and ~ is the
weight of precious metal recovered by the recovery yauze
pack divided by the weight of precious metal in the
stream presented to the pack.
Following -the run, the catalyst weighed 4.3973 g., a
loss of 0.2989 g. from its starting weight. By
comparison, the recovery gauze pack in the first chaTnber
weighed 5.2350 g., a gain of 0.0623 g. Based on -this
data, the pick-up efficiency ( ~') of the recovery gauze
pack in the first chamber was found to be 14.6%. In
calculating the weight pick-up efficiency of th~ recovery
gauze pack in the second chamber, the weight recovery
efficiency of the first chamber must be taken into
account.
Comparative studies were conducted by repeating this
procedure on recovery gauze screens of varying mesh (N)
and wîre diameter (dw). In each study, the gauze screen
and catalyst employed were weighed prior to use and
immediately thereafter, and the the weight changes were
converted to weight recovery efficiency
( ~'). Bo-th runs were conducted over an identical trial
period. The configurations of the catalyst and recovery
gauzes employed in Examples 12-17 are set forth in Table
IX. Bo-th experiments afforded yields of nitrogen oxides
in the range of 96-98.9%.
In ~
x ~ ~ ~~ ~ ~
~ o o o o o o
N
O u~ o ~ o ~ r~
~ LO ~ ~ a~
O
~ E~ o ,~ o o o o
S~ ~
~. 4~
a) ~: r- r- o ~ r~ r-
S~ ~ ~ Ll~ ~ ~o ~ r~
O ~ ~ ~ ~ ~ ~1
~n .~,
rl
~) ~
~1 ~ N ~ ~ I r~Lr~ cn
~1 ~ O ;~ ~ ~ O ~ ~D
x t) a)
H (L) ID ~1 ~`1 ~i ~I N ~P t~l
a~ ~ O
1 U~
I ~ tnE~
E~
S~ ~
~ a) ,, ~ a~~ ~ o
O -~ ~ r~ ~ ~~ r~ co
a) r~ o oo o o o
o o oo o o
rd ,l
.,~ _ o o oo o o
~ æ ~D O O O O ~r
~~ ~ ~ u~a~In ~
X
E~
. .
~ ~ ~ ~ ~U)~D r~
O ~ ~1
-40-
The weight loss for the catalyst of Operating
Examples 12-17 and the weight gains for the respective
recovery screens and their recovery efficiencies are set
forth in Table X:
~ --
' ~-- I~D Ln I ~ Ln¦ r-l ~ I L~ O ~1
.~ I . . I . .I . . I . . I . . I . .
n ~ ~ ~ I ~ ~I r~ I o ~
~1 ~ ~1
~ `
ra 3 o ~1 ~)o ~r o r-l ~ n O 1~) CO ~
~ ~1) G~ r-l r-l~1~ ~D ~ Ot) In~1~ Ln 1~1 r ~1 o ~D ~
~ ~ D Ln r~ r- ~oLn r~ ~ Ln ~D In
I ~ O O ~I O O ~) O O~I O ON O O ~) O O
~(U OOO OOOOOO OOO OOO OOO
I + + I + +I + + I -1-1 -1- -tI + -1-
~: ~ _~
~ ~J~
o Ln Ln~;rLn ~ ~r
~d .~ ~) ~ O ~ 1 0r~l ~D ~ ~ t~ ~D Lf) ~ t~)
a) r~ ~D LnCD r~o ~ ~~ ~D ~ Lt-) O ~D ~ r-l r-l
n ~ r~ ~7~ Ln Ln~ ,~ r c~ ~ co r~
:~ (~ r~ ... ...
t~ ... ... ... ... ~r ~P Ln
0 h ~ Ln Ll-) ~r ~ ~ ~r ~r ~ ~r ~r Ln
~1 ~\
L~ ~ 11~
O ~ r~ o ~ o~ C~
1~ . t~ N O ~ Ln O r~ r~ Ln o ~ ~D Ln Ln a~ ~ ~ a
tJl ~c~ Ln rn r~D ~ ~ n ~ c~ o ~ o a~ ~r Ln s~
~r~ ~ ~ a~ r~ r ~ ~ o ~ ~ r r~ o u~
~1 ,~ ~ ~D r~ r~ D r~
m ~ ~ ... ... ... ~ n ~ r O
~ 3
1~0
o
O ,L- 0 a)
(I) ~) ~ ~ ~) t~ (~ ~ t~ t~) (~ (r) (~-~ L:
~ a
0~ O X XO X X X X o X X O X X O X X
Z U~ ~ r~ ~ ~ ~ ~ U~ ~
~ ~ ~a
'd ~a~d 'd'C5 'd c~ 0
I:L G ~ ~ P~ ~ ,~
Ln LnLn LnLt`l Ln ~ U~
~
n ~ ~ Ln ~ ~ ~ ,1 a~
Ln ~ ~Ln a) a)Ln ~ a~ Ln ~ ~ Ln a) ~ Ln ~ a~ =
~,~ ~ ~ ~ ~~ :~ ~ ~ ~ ~
~:4 E~ N
. . -~ 0
~:1, X ~ r Ln ~ I~ tJ~
O ~ ~ r~ ~ ~ r~ ~ ~C
-42-
This data confirms the hiyh order of recovery (~'~
attributed to the recovery gauzes of this invention.
Fur-thermore, it demonstrates that -the gauzes of Operating
Examples 15 & 16, having values of -the product of mesh
and wire size above 0.3, provide excellent recovery
efficiency.
Assays were conducted to compare platinum recovery
efficiency (~) against weight recovery efficiency (~'1.
Also, these assays confirm that the gauzes recover hoth
platinum and rhodium. See in this regard Table XI, where
the results of these studies are set forth inclusive of
platinurn and rhodium recovery expressed as a ratio.
TABLE XI
_
Type N x dw ~' (%) ~Pt/~h Recovery
3 .195 11.1 16.5 ----
2 .248 13.2 10.9 46.3
4 .312 15.9 17.5 39.9
On the hasis of these studies, it was determined
that palladium/gold recovery gauzes having a mesh size
(N) in the range of from about 50-80 and a wire diameter
(dw) in the range of from about 0.003 to 0.018 inches,
exhibit particularly suitable precious metal recovery
properties provided they possess an N x dw of at leas-t
about 0.3.
It was also found that the recovery gauzes in the
first reaction chamber and the recovery gauzes of the
same material and configuration in the second reaction
chamber exhibited no significant differnece in weight
recovery efficiency (~ '). Moreover, it appears from
this data that significantly improved results are ob-
tained using recovery gauzes of the type described in
Operating Examples 15-16, which possess the required
~3-
values o~ N x d . Significantly, -these particular
recovery gauæes exhibit an initial N x d parameter of at
least 0.3~
The foregoing data shows that the efficierlcy of an
80% palladium and 20~ gold recovery gauze in ammonia
oxidation process is significantly improved by cons-truc-
ting said gauze to an initial ~ x dw parame-ter of at
least about 0.3. A preferred embodiment of this inven-
tion comprises a recovery gauze ensemble comprised of
several such recovery screens sandwiched between several
separator screens.
The following example illustrates the fmprovement in
weight recovery efficiency which can be realized with
Pd/Ni recovery gauzes.
OPERATING EXAMPLE 18
A recovery gauze pack consisting of two recovery
gauze sheets (Type Ni-B: 95% Pd/5% Ni) was placed between
separator screens and this ensemble was placed into a
first reaction chamber below a 90% Pt/5% Rh/5~ Pd oxida-
-tion catalyst (15 sheets). The recovery gauzes contained
60 wires per linear inch ~N~ and had a wire diameter (dw)
of 0.006 inches (N x dw = 0.36). The ammonia oxidation
catalyst weighed 7O107 g., and the recovery gauze pack
weighed 5.164 g. prior to the run. The separator screens
were in the form of a wide mesh gauze constructed from a
ferrous alloy.
In a second chamber located immediately downstream
Erom the first chamber, there was placed a second gauze
pack also consisting of two recovery screens (Type
Ni-A: 95% Pd/5~ Ni) sandwiched between three separator
screens. I'he recovery gauzes contained 45 wires per
linear inch (N) and had a wire diameter (dw) of 0.006
inches (N ~ dw = 0.27). The recovery gauze pack weighed
4.666 g. prior -to the run.
The -two chambers were preheated to 300C. and
ammoni,a and air were channeled therethrough as a mixed
r j t~
~4-
gas stream under a pressure of 100 p.s.i.g. at a total
flow oE 6~0 SCFII. Duriny the operation, the first
chamber was maintained at a temperature of 930~C. and -the
second chamber was maintained at 890C. Arnmonia
consti-tuted 10% of the gaseous mix-ture, represen-ting a
throughput of 57 tons ni-trogen per square meter per day,
that is, 57 t(N2)/m2d.
The weight recovery efficiency ( ~'~ for the Pd/Ni
recovery gauzes ln each reactor chamber was determined by
measuring the weight gain of each recove~y gauze pack and
the weight loss for the ammonia oxida-tion cataJyst.
These measurements were then converted to weight pick-up
efficiency (~ ') as per -the equa-tion~ (1 - R) /
wherein n, ~l and R are as previously defined. Fol-
lowing the run, the catalyst weighed 6.178 g., a loss of
0.929 g. from its original we ght. By comparison, the
recovery gauze in the first chamber weighed 5.452 g., a
gain of 0.288 g. The recovery gauze in the second
chamber weighed ~.826 g., a gain of 0.160 g. On the
basis of this da-ta, the weight recovery efficiency ( ~'~
of the second chamber recovery gauze was 13.4~.
The used ammonia oxidation catalyst pack and the two
recovery gauze assemblies were assayed to determine their
actual platinum recovery efficiency (~) as well as
confirm that the Pd/Ni alloy recovery gauze recovered Rh
to the same extent as the Pd/Au alloy recovery gauzes.
The results of these assays are shown in Table ~II where
platinurn and rhodium recoveries are expressed as a ratio.
~59..~r;
--45--
TABLE XII
Efficiencies Recove.ry
Chamber Type N d (in) Nxdw ~ ) of Pt/~h
2 Ni-A 45 0.006 0.27 13.4 10.8 27.0
1 Ni-B 60 0.006 0.36 16.9 26.0 50.7
Operating Example 18 illustrates that both nickel-
containing alloys and gold-containing alloys are effec-
tive in recovering platinum and rhodium lost fxom ammonia
oxidation catalysts, and that the characteristically
improved recovery efficiency associated with a high
N x dw produc-t applies equally to the gold and
non-gold-containing alloy recovery gauzes of this
invention.
