Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
l ¦ The invention relates to a process which converts
2 ¦ carbon monoxide (C0) and water (ll20) into carbon dioxide (C02)
~ and hydrogen (H2) by homogenous catalysis and which on suitable
4 selection of the reaction parameters ~ives yields o 85 to 95
in short reaction times.
6 The conversion oE C0 according to
7 C0 -~ H20 _ C02 + ~l2 (1)
8 serves for the technical production of hydro~en from C0-rich
C 9 gases produeed by gasifying coal or hydrocarbons. secause of
the unfavorable equilibrium at hic3h temperatures reaction (I)
ll has to be performed below 500 C. ~owever, the equilibrium
12 is e~tablished very slowly in this temperature re~ion, so that
13 hi~hly aetive catalysts are requirecl. For many years solid
l~ catalysts based.on iron or copper have been utilized at 350 to
500 C (high temperature process, iron catalysts) anci at 210
lG to 270 C (low temperature process, copper contacts). DurincJ
~7 the last years several authors attemp-tcd to develop homocJellous
la catalyst systems which would avoid the hi~h eneryy level of the
19 hi~h temperature process and the sensitivity to catalyst poison
of the low temperature process in systems with minimal technical
21 difficulties. (Bibliography in C. ~ngermann, V. Landis,
22 S.A. Moya, H. Cohen, H. Walker, R.G. Pearson, R.G. Rinker and
23 P.C. Ford, J. Amer. Chem. Soc. 10l, 5922 (1979)).
2~ The main component of the hitherto known systems
usually consists of a transltion metal carbonyl compound which
26 has been tested in ~ater containin~ solvents at C0-pressures
27 of 0.25 to 300 bar and temperatures of 80 to ca. 150 C.
28 The interestiny catalyst: systern deri~ed from iron pentacarbonyl
- 2 - ~
~Z2~
:
~ '
l ~ (Fe(C0)5) and alkalines was already mentioned by W. Reppe in
2 ¦ 1953 (W. Reppe and coworkers~ Liebigs. Ann. Chem. 582, 121 (1953~))
3 ¦ and has been investigated by R.B. King (C.C. Frazier, R.M. Hanes~,
4 A.D. Kiny, Jr. and R.B~ King in Inorganic Compounds with Unusual
5 ¦ Properties (R.~ ing, editor), Vol. II, Chapter 9, pages 94 ff,
6 ¦ Aclvances in Chemistry Series 173, Am. Chem. Soc., Washington, .
7 ¦ D.C. and R.B. King et al, J. Amer. Chem. Soc. 100, 2925 (1g78))
8 and R. Pettit (R. Pettit, K. Cann, T. Cole, C.~. Mauldin and
9 ! w. Slegeir in Inorganic Copounds with Unusual Properties (R.B.
lO ¦¦ King, editor) Vol. II, Chapter 11, pages 121 ff, Advances in
ll Chemistry Series 173, Am. Chem. Soc. Washington, D.C. 1979).
12 The conversion rates so far observed are, however, too small
13 for technical application. Furthermore, the main disadvantage
l~ ,j is caused by the fact that under these conditions even aEter
15 ll longer reaction times the reaction (1) does not establish the
16 ¦ thermodynamic equilibrium of more than 95 % hydrogen yield,
17 1l but stops below 40 % hydrogen yield depending on the experimental
18 1I parameters in an unclear manner.
19 1 These low hydrogen yields do not permit an economic
application of the Fe(C0)5-base-system.
21 C.C. Frazier et al show on page '~7 of the above mentio-
22 I ned reference in Table II the effect of pressure and temperature
23 l on the reactivity of the iron pentacarbonyl-catalized water
2~ 1l gas shift reaction. The iron pentacarbonyl catalyst is used in
25 ~I butanol as solvent which contains the necessary amount of water.
2~ According to this publication temperatures between 140 and 160C
27 ~ are preferred. Runs 4 and 5 of Table II show temperatures of
28 ~l 181 C and 183 C respectively. The turnover numbers (mole
1~ 3 _ i
2~
~ .,
l hydrogen/mole me-tal per 6 hours) are given in the las-t column
2 ¦ of Table II. It is stated on page 96, lines 26 - 28 -that the
3 ¦ turnover numbers represent only the turnover in an early stage
4 ¦ of the reaction. In a further reference (A.D. King, Jr., R.s. I
5 ¦ King and D.B. Yang, J. Am. Chem. Soc 102, p. 1028 e-t seq. (1980)l)
6 ¦ Kinq et al are discussing the homogeneous catalysis of -the water
7 ¦ ~as shiE-t reaction using iron pentacarbonyl in a mainly methano
8 lic solution. From Fig. 4 on page 1003 follows that the -turn-
9 ! over number is not constan-t over the whole reaction time (the
lO ¦ turnover number is propo~ional to the slope of the partial
ll ¦ pressure of hydrogen versus -time-curve). Therefore, -the -turn-
12 over numbers of Table II of Frazier e-t al canno-t be used to
13 , estimate the expected hydrogell yields. Run 6 (161~ and run 9
l~ l (163 C) of Table II of -the Frazier et al reference are des-
15 I cribed in more detail and evalua-ted in form oE Fig. I and
16 ~¦ Fi~. II on pages 98 and 99 respectively. From the last
17 ll measured points of these Figures one can calculate that in run
18 !1 6 after 25 hours less than 40 ~ conversion and ln run 9 after
19 1 6 hours less than 2~ ~ conversion are measured. On page 100,
20 ~ second paragraph, it is stated that the rate of hydrogen
21 1 production increases as the reaction tempera~ure is raised. I
22 1 In thereafter following Figure 3, same page, it is only demon-
23 strated that the rate of the very early stage of the reaction
2~ is raised by increasing temperature when the partial pressure
25 ~l of the carbon dioxide formed during the reaction is neglectable.
2~ 1l However, this does not permit to deduct tha-t the ob-tainable
27 1I hydrogen yields at the end of the reaction are also increased
28 l by raising temperature.
I -- 3a -
l ~ Frazier et al and King et al are using alcoholic so-
2 lutions of the catalysts which solutions contain water in a small
~ amount. King et al state in their publication on page 1030,
4 left column, last paragraph, that the use oE water should be
5 1 avoided because it is stated in line 5 "pure water is seen to
6 be a very poor choice of solvent.". In this connection, Fig. 3
7 ¦ in the SAme column, lowest curve, has to be considered from
8 which follows that at 140 C after 70 hours less than 10 %
9 ll conversion is obtained. As explanation for these results -the
lO ~ low solubility of carbon monoxide in water and -the decomposition
ll ¦ of the catalyst system which is connec-ted therewith is given
12 on page 1032, left column, last paragraph by King et al..
13 It follows from the above discussion tha-t the state of
l~ l¦ the art until immediately before -the priority date of -the
15 ll instant application has dissuaded the ar-tisan -to use high
16 ~ temperatures and water as solvent :Eor the water gas shift
17 ~; reaction of carbon monoxide and water to carbon dioxi.de and
l~ h~drogen.
19 1 The process presented here is based on the surprisi.ng,
20 1 not foreseeable observation that at temperatures above 220 C
21 ! wea~ly basic aqueous solutions of Fe(C0)5 do convert C0 and E-l20'
22 l! into C02 and H2 with yields up to 95 ~. Reaction rates high
23 enough for an application are achieved only above 180 C.
24 1I Completely unexpected is the temperature behaviour of
25 1l the system: Between 180 and 250 C a~ in some cases between 180
2~ and 220 C the reaction (1) does not give complete corlversion,
27 1l
l~ ~ 3b -
28 1
I
1~ ~
but depending on the temperature a cer-tain, no-t to be exceeded,
maximal yield is found which is around 40 % at 180C and can
be raised to 95 % by increasing the temperature. This surprising
behaviour will be demonstrated by means of the attached drawings
in which
Figure 1 is a pressure-temperature-diagram; and
Figure 2 is a pressure-reactiontime-diagram.
In figure 1 curve a d~scribes the theoretical pressure-temperature-
dependence of an inactive system (no conversion, hydrogen yield
0 %) calculated by addition of the ideal CO-pressure (gas law)
and the respective steam pressure of the excessive aqueous phase.
Curve b is obtained in a similar manner for complete conversion
(hydrogen yield 100 ~ corresponds to a twofold amount of gas).
The experimentally measured points of table 1 marked by (o) show
that in a typical run (initial CO pressure ~8 bar) heating
the autoclave rapidly and continuous~y to 222C the reac-tion
starts noticably above ca. 180C and proceeds to high conversions
on approaching the final temperature. Figure 2 shows the
coxrespondiny pressure-time-dependence, whereby the continuous
line between the experimental points represents the isothermal
pressure development when the final temperature of 222C was
reached. ~hen the system was only heated to 181C (experimental
points (x) of table 2) the conversion reached a final reaction
pressure of 115 bar asymptotically after 5 - 7 hours (figure 2)
which according to figure 1 corresponds to a yield of about 40 ~.
-- 4
? ~ 2 ~
On further heating to 198~C the reaction started again until
after 3 to 4 hours another limiting value of about 70% yield
was obtained. Raising the temperature once more to 217C
finally gave the pressure corresponding to a practically
complete conversion. This observed temperature behaviour
clearly indicates that despite of a low catalytic activi-ty of
the Fe(CO)5-alkali water-system higher yields than about 40~ are
in principle not to be attained below 180C. Temperatures of
only 220C and more lead to hydrogen yields which are interest-
ing with regard to a technical application.
Mixtures o-E water, alkali and carbonyl compounds of
the metals Cr, ~o, W, Co, Ni, and Mn give results similar to
the Fe(CO)5 system.
In the conversion using metal carbonyl catalysts
hydrido-metal carbonyl complexes, among others, are apparently
involved, ie. metal compounds having hydrogen bonded directly
on the metal. Therefore, a catalytically active phase can be
prepared using a metal carbonyl compound itself or precursors
which give the metal carbonyl or the hydrido-metal carbonyl
in a preceeding process or under the reaction conditions
themselves. For instance Fe(CO)5 is formed directly from
metallic iron, iron(II) compounds or polynuclear iron carbonyls
(or instance Fe2(CO)g, Fe3(CO)12)by reaction with CO.
Accordingly Fe(CO)5 might be replaced by these substances.
Similarly polynuclear anionic iron complexes react to the
hydrido-tetracarbonyl ferrat anion (HFe(CO)4 ) with CO, as for
instance Fe2(CO)8 and HFe3(CO)13 . Therefore, the application
of such compounds - with or without an additional base -
- 5
~ 7~
1 ¦ similarly leads to catalytically active mixtures. Tllese
2 ¦ examples shall illustrate that each compound or mixtures of
3 ¦ those compounds are sui-table to prepare the catalyst solution
4 ¦ which either form metal carbonyls or hydrido-metal carbonyl
5 ¦ compounds and higher analogs, having hydrogen bonded directly
6 ¦ on the metal or establish e~uilibria with those species.
7 ¦ A11 compounds which form hydroxyl ions, under circum-
8 stances only under the reaction conditions, are suitable bases,
C ¦ preferably the alkali and earth alkali metal hydroxides,
1~ ¦ carbonates, and hydrogencarbonates, especially sodium carbonate
11 ¦ and sodium hydrogencarbonate.
12 ¦ Generally the process is carried out by intimately
13 ¦ mixiny the aqueous phase containing the catalyst component(s)
14 ¦ and the gas phase of C0 or C0-rich gases in a pressure vessel
15 ¦ at temperatures of 190 to 300 C, preferably 200 to 270 C,
16 ¦ and at carbon monoxide partial pressures of 1 - 500 bar,
17 ¦ preEerably 20 to 100 bar. The concentration oE the metal
la ¦ component and the base which has to be applied is dependent
1~ ¦ on the kind of components and the reaction parameters pressure
20 ¦ and temperature. The carbon monoxide partial pressure can be
21 ¦ varied in broad limits, but should not fall below the C0-
22 decomposition pressure of the metal carbonyls. Addition of
23 ¦ organic solvents like ether or alcohols is sometimes advan-
2~ ¦ tageous.
25 1 Definition of some terms used hereafter:
26 1 1.) Composition of the product gas
27 XH = molar fraction of hydrogen in the product gas
28 XcO and XcO are defined in analogy
, , . -~
~ 27g~
l ¦ 2.) hydrogen yield - XH (X~l2 + Xco)
2 ¦ 3 ~ ZM~ ~B = number of catalytic cycles relative to metal
3 ¦ and base, respectively, expressed by mole hydrogen
4 ¦ per mole metal or base
5 ¦ ZM = yleldH mole (C0) mole (metal) 1
6 ¦ Z~ = yieldH '.mole (C0) . mole (base) 1
7 1
81 . "
f 9 ¦ E.Yample 1
I
lO ¦ A 500 ml shaking autoclave with fittings, both made
ll ¦ from stainless steel (DIN 1.4571), was charged with 1;96 g
12 ¦ (0.01 mole) Fe(C0)5, 1.06 g (0.01 mole~ water-free sodium
l3 ¦ carbonate (Na2C03) and 100 g (5.56 mole) wa-ter, pressurized
l4 ¦ initially with 48 bar (0.8 mole) carbon monoxide under exclusion
15 ¦ of oxygen and heated to 222 C with shaking. During the whole
16 experiment the internal temperature and the pressure were
17 measured by an iron constan-tan thermocouple and a pressure
l8 gauge, respectively. Table 1 shows the experimental data of
19 pressure,`temperature and reaction time.
Composition of the product gas: Xl~ 0~493, XcO 0.414,
21 XcO 0 093~ ZM 67~ ZB 67, yield~l 84.1 ~-
22
23 Table 1
24 Reaction Time Temperature Pressure
2S (h) (C) (bar)
26 I 0 20 48
,"l ~.3 130 72
:'ll 0,7 .1f~5 ~1
I . ~ .
z~
Reaction Time TemperaturePressure
(h) . (c) (bar)
1.0 200 109
1.3 210 125
1.5 217 1~2
1.9 219 15~
2.1 220 162
2.8 221 172
3.3 222 178
3.4 222 180
4.1 222 18~
4.4 222 184
Example 2
The autoclave was charged in the same manner as in
Example 1 and heated stepwise first to 181C, then to 198C and
finally to 217 C. The experimental data are given in table 2
XH0 494~ XcO 0.427, XCoO-079, YieldH 86.2~, ZM 69, ZB69
Table 2
Reaction Time TemperaturePressure
(h) (C) ~bar)
O 20 48
0.5 176 87
0.7 183 90
0.8 178 90
llffZZ79
l ~Reaction Time Temperature Pressure
2 ¦ (h) (C) (~ar)
3 ¦ 1.0 178 92
4 ¦ 1.5 ~ 181 9B
5 ¦ 2.0 -. 185 104
6 ¦ 3.0 181 .107
7 ¦ 4.0 181 111
8 ¦ 5.0 181 113
5.5 181 114
lO ¦ 6.0 181 115
ll ¦ 6.4 185 117
12 ¦ 6.9 190 122
13 ¦ 7.3 - 193 124
l~ ¦ 7.8 . 196 129
15 ¦ 8.0 198 137
16 I 8.9 198 1~1
17 10.0 198 145
18 11.0 198 146
19 ¦ 1~.0 198 147
20 ¦ 12.2 205 150
21 ¦ 12.9 215 168
2Z ¦ 13.4 217 - 174
23 ¦ 13.8 . . 217 177
24 ¦ 14.7 217 179
Z6 ¦ Example 3
27 4.9 g (0.025 mole) Fe~C035 were dissolved in 100 (3 1 m po~as-
28 sium hydroxide by stirrin~3 the mixture 6 h at 60 C with
l exclusion of oxygen ~aryon atmosphcre). The brown-yellow
2 solution of potassi,um hydrido-tetracarbonyl ferrat (K~ e(C0~9)
3 and 46 bar initial pressure of carbon monoxide (0.77 mole)
4 was heated to 260 C in.the 500 ml shaking autoclave. AEter
5.5 h (1.5 h heating time, 4 h isothermally a-t 260 C final
6 pressure 210 bar) a gas sample had -the followiny compositi.on:
7 ~I2 ' C02 0-452, XcO 0.060, yield~I 89.1 %, Z 27
8 ZB 9-
10 1
ll ¦ Examples4 to 14
12 ¦ The examples 4 to 14 were carried ou-t similarly to
l3 ¦ Exampl 1. The parameters and res~llts are given in Table 3.
lli
18
19
21
221
23 ~
241
25 i
26
27
28
I';
1 ~ 0~ 0 01` ~ o~
I ~ Ln ~n Ln Ln ~ r~l r~i Ln ~o Ln ~D
l O rd r~l r`i r~J r~l r`~ r~l r~l r~ r~l r`i
21 ~a)û IIIII'I Ii'IiI
U ~1~0 O O O O O O O ~n Ln Ln o
~ I rd ; ~ r~ r~ r~i ~ o o ~ ,~, In ~ ~r.~
v I r~) r N r~i r `i r~ r~i r~i r~l r~ r~ r~J r~i tn
4 1 ~ . u
l tn , U
5 1 Ln `. Ln Ln o ~ ~ o Ln ~ ~ . ) .
I rL) _ O ~ L~ I I~ L~ rJ~ r~ r-- O r~
¦ Pl rd r.~i ~ ~ ~ ~ ~ ~ ~ ~ r~
6 ~:1, . . ~ ~ ~ r~l ~ Ln Ln ~ rt) ~o a~
7 1 ~j .~ ~ t~ r~ CJ) Ln r~J r~l r~ r~ r~l r~l r~
l . ' . _)
E3 1 v . ' Lo
I ~ ~1 . , . ~1 ~ ~ rr~ r Ln L~ ~ r~ r- r~
a I r~ tn s~ ~D o ~D Ot) o ~3~ r ~ D c) t~l 1- r~ r-- Ln L~ L~ Ln u~ Ln ~n
0 I .~1 t rd ~r ~ ~r a~ r ~r ~r ~r ~r .a
P~ . . .
. I f~ . . .
2 I rv . r~
I r~ rc) : . ~r: o ~ O Ln ~r r~ r a~ a) r~i ~
13 I r~ O r~ r ~D rr~ r~ r O ~ ~ ~ ~ ~d ~ a~ L~ ~ a~ à~ c~n ~ ~ a~ '
1~ o f- O O O ~ O O O O O O o ~t ,
~.1 . ,
15 r~ O . . ,
fV ~D ~D rX) ~ D r~l r~i r~i r~i I ~n r~ l Col~ ~ ra~ ra~ ~D r~ ~r rL~ Lt~
Q ~ Ln Ln r~ Ln Ln 'Ln Ln r,'~l f,~l r~l r,~J I .~-1 O O ~ O O O O O O O C)
rd ~I V ~ tn O ~
f,~ .1) ~ Ln Ln r.~l Ln Ln Ln Ln r~i r~ r~l r~ :>~ J O O O O O O O O O O O
7 ~ ~r-~ r~ ~.
_~r~1 0 0 ~ r~ Ln ~r c~ cn o~ Ln O r-
~: O r ) r ) r- In In n c~ cr r l ~D ~n ~n ~o
18 . ~ ~ ::C O r ) r~ ~ ~ m r~ ~r ~ ~r ~ r ~r ~;r ~r ~r ~r ~r
O O O rd r~ fll rd O O O O fd O
_~K X ~ Z X Z ~; X X X X r~ r) O O O O O O O O O O O
19 ~V . Ln l ~
rd f- O f~ O In Ln O O O O O u .n r~ r- r.~l r~ r~J r~ a~ ~ r~ O
m ~- ~ ~ ~ ~ o ~ a) ~ n r~, r~r .n r ~n ~
l . . P- ~C G O O O O O O O O O O
21 l . ro
~n n n m n n In ~D ~ _
_ _ _ ~ LO
2~ I ~ ooooo o ooo---_)
I r ~ ) v ~ o-- I
I o ~ r~J j .
I va ~ ~ rv rl) rv rJ rv O ~ O~~ O I
23 I ~)O ~t ~t ~t ~t tst ~t ~t t) ~ ~ r_) r' _~.
24 ~ o o ~i o n n O n ul n u~ u rl) ~ ~r n r r~ In n ~ r~ ~r
~: . ~t
26 I r~ . r
I ~ .~ n ~ a~O ~ r~ ) ~ ~ O ~r n ~D r- rx) o~ G ~- ~ ra ~r2 7 x z ~ ~ ~ x ~ r-- ~ ~ r
28 I
Examples 15 - 28- Continuous Apparatus
J
The apparatus is consisting of a vertical arranged
reaction tube of 1,3 litre content (2000 mm length, 28 mm
inner diameter) which is heated by oil circulating around
the tube, a fur-ther tube mounted thereon which is heated
in the same manner and has a capacity of 1.7 1 (670 mm
length, 56 mm inner diameter) which served for the sepa-
ration of gaseous phase and liquid, and a reflux condenser
(620 mm length, 56 mm inner diameter) which was water-
cooled. An aqueous sodium carbonate solution (o.l~, 1%
and 10~ respectively) was pumped through a metering pump
into the lower end of the reaction tube at the beginning
of the runs. The filling height was adjusted with an
overflow valve mounted at the upper end of the reaction
tube. The water consumed was supplemented through the
pump when necessary. Carbon monoxide was pressed into the
lower end of the reactor by means of a compressor operating
isobaric at the pressure side through a pressure vessel
(700 ml content, 250 ml iron penta carbonyl) filled with
iron pentacarbonyl and through an oil-heated preheater.
The thermostatisation of the pressure vessel and the down-
stream capilar leading to the pre-heater permitted the
adjustment of a defined partial pressure of iron penta-
carbonyl. The withdrawal of gas was controlled by means of
a release valve on the top of the cooler and recorded by
a flowmeter. The gas analysis was performed with the aid
of a gas analyzing apparatus (Orsat) in defined time inter-
vals. -The temperature was measured in the upper third of
the reaction tube.
The process parameters, results of measuring and the there-
from calculated data of examples 15 - 28 are given in
Table 4 and defined as follows:
(Na ) concentration of sodium ions in mol/l in the aqueous
phase
P total pressure in bax measured at the top of the reactor
T reactor temperature in C
T
e( )5 temperature of the iron penta carbonyl in C
- 12
,~t~,
27~
Fe(CO)5 partial pressure of iron carbonyl in bar
output gas gas withdrawn in l/h (related to 20C
and atmospherlc pressure)
input CO carbon monoxide calculated from output gas and
hydroyen yield - input in l/h (related to 20C
and atmospheric pressure)
input Fe(CO)5 iron pentacarbonyl calculated from input CO
(related to the actual process parameter TF (CO
and P) and Fe(cO)5 P
C/Fe(C)5 calculated molar carbon monoxide iron
pentacarb ratio at the reactor entrance
output H2 productlon rate of hydrogen calculated from
output gas and hydrogen yield in l/h (related
to 20C and atmospheric pressure).
Temperature, pressure, partial pressure of iron penta-
carbonyl and throughput o~ carbon monoxide varied in examples
15 - 23 whereby conversions (=hydrogen yield) up to 70%
were obtained when carbon monoxide was passed once through
the liquid phase. As it can be seen from examples 21 and 23
respectively on one side and 18 and 22 respectively on the
other side, the temperature is the decisive process para-
meter which allows high conversions. I~ the reaction tube
is charged with fi:Llers (cylindric wire nets Erom V2A,
6 mm diameter, examples 24 - 28) thi.s results in higher
throughputs at the same yields. Pumping of a 10% instead
oE a 1~ aqueous sodium carbonate solution has no signi
icant influence, however the activity o~ a 0.1-percent
solution decreases remarkably (examples 27 and 28).
7~
OO~COcO~r,~ D~ iDo~
~D~`U~CO~ ~rOCO
o _
~, o C~ o o o o o o o o o o o o
~ U~ ~ ~ ~ 00 ~D O ~ 00 ~ 1~ CO
u~ In
,, ~
o ~ . o o o o o o o o oo o o o o
S~ _ co r~ ~ o o o o o o o o o o o
.
U~
o~ ~~ 8
~ O O ~ ~ ~ ~ ~ ~ ~1 a~
O h ~_,
C)
O ~ CO O O t` 1
~r o a~ ~ ~ 0~ 1` ~ ~D ~D ~ ~D LO Ul Ln ~r ~ u) u~
U~ t~ H U~ ~1 ~1 ~1 " ~-1 ~1 ~ ~_1 r-l ~ ~1 ~1
U~
~1.~ ~ ~
~1,~.~ o
E~ ~ 3 ~ ~ 1 o ~o i~ O ~r
G)
~ r~ o o u~ o ~ O o O o OO O
ulR ,1 Ql t~ ~ ~ ;r o o ul o
o
~ o ~
~u) ~ ~rgggogogggoggo
1:1 O ` C~ ~ O N t~l ~`1 N N N N N N N N ~ ~
N ~ R o o o o o o o o o o o o O o
L~
P~ ~ o I U~ o o o o o o o o o o o o o
EF4 1 '
O o o o o o o o ~r ~ o ~
O ~ N N N N N N N N N N N N (~1
L~n~ooooooooooo
+^ ~ ~
0 ~ N N N ~ N N N N N N (~1 N O O
Z r~ O O O O O O O O O O O O N O
U~
E ~ ~1 ~1 ~ ~1 ~ t~ N N N N ~ ~`J ~`
