PROCESS FOR THE RECOVERY OF SULPHUR DIOXIDE FROM GAS FLOWS '

The present invention relates ^" _' ,to a process for the purifi¬ cation of SO2 containing gases, recovery of SO2 by means of an absorption solution, and regeneration of this, solution. In particular the invention relates to the \ use of aqueous buffer solutions as absorption liquid.

THE STATE OF ART

Two main types of processes in which such aqueous buffer solutions are used, are:

Process type 1. "Regular absorption-stripping processes" where the absorption of SO2 takes place in towers in which gas and liquid flow pass each other countercurrently, and where the liquid is regenerated by stripping with open steam in other towers which,are also operated countercurrently. Little or no change in volume of the liquid stream takes place during the regeneration.

Concentrated SO∑ is obtained after condensation of steam from the steam-S02 mixture leaving the stripping unit, while the flow of stripped liquid is conveyed back to the absorption tower for new SO2 absorption. The stripping may take place at the sa temperature as in the absorption tower or at a higher tempera¬ ture. In this type of process only clear liquid and gas are circulating, which simplifies its operation.

Process type 2. Absorption is carried out as for type 1, but t regeneration takes place by evaporation, in which a steam-SO∑ mixture is driven off, while at the same time the most basic buffer component is precipitated by crystallisation. Concen¬ trated SO2 is obtained as in type 1, while regenerated absorpti liquid ready for new absorption may be obtained after dissoluti of the buffer crystals in water (or condensate) .

FURTHER DISCUSSION OF PROCESS TYPES'! AND -2 BY EXAMPLES u

As buffer in process type 1 Na-citrates may for instance be used (1) . A characteristic feature of said t;_pe of process is that the pH in the absorber must be kept* almost constant and at relatively low level in order to accomplish the regeneration by means of regular steam-stripping. \ This type of process is particularly suitable for gases having a relatively high SO2 content, where the consumption of stripping steam per ton of SO2 produced gets low enough.

The most commonly used buffer in process type 2 is Na- sulphite, Na2 SO3 (2) . The reactions utilized in this "sulphite- process" are the following:

Absorption: SO2 ( a > = SO2 ( 1 )

SO2 ( 1 > + H20< 1 > + Na2 SO3 ( 1 ) = 2NaHSθ3 < 1 >

Regeneration: 2NaHSθ3 < 1 > +n*H20< 1 > = SO2 < g > +(n+l) ^« H2θ< ₃ > +N 2 SO3 < 3 > (n = a large number) .

Due to precipitation of Na2 SO3 < s > during regeneration the p in the residual liquid is stabilized, and this facilitates the evaporation of SO2.

It should be noted that for each mole of SO2 absorbed, two moles of HS03~are produced in the sulphite process.

ADVANTAGES AND DISADVANTAGES OF THE SULPHITE PROCESS.

Today the sulphite process is commercially the most used process for the recovery of SO2 from dilute gases. There are several reasons for this:

1. The consumption of steam in the regeneration step per ton o produced SO2 is at a relatively acceptable level. With double-effect evaporation approximately 7-8 tons are required per ton of SO2 , and this is particularly favourabl for gases of relatively low S02 content. In comparison the said citrate process normally would need about 20 tons of . steam per ton of SO2 for gases with 0,25 mole-% SO2 ,

although this amount can be much reduced with the use of vapour reco pression. .

2. The requirements for a very high purification efficiency with respect to SO2 may normally be satisfied due to a relatively high pH value in the regenerated absorption solution. A regular absorption-stripping process, for instance based on Na-citrate which would have to operate at a substantially lower pH level, cannot so easily satisfy very strict clean-up requirements.

3. The sulphite process is in principle based on very simple chemical reactions. In particular, it is attractive that i generates its own buffer anion from absorbed SO2 , when a Na base is added:

4. The sulphite process was developed for commercial use relatively early. Such a process development is very resource demanding. The success of the sulphite process ma thus have delayed the search for alternative processes.

Although successful, there are features which complicate installation and operation of the sulphite process:

a. When the process is used for gases containing oxygen, Siv- species are lost by oxidation, e.g.

The Na-sulphate formed must be removed from the solution, and the lost Na-sulphite must be replaced. This requires the addition of Na-base. Considerable oxidation losses, corresponing to 3-10% of absorbed SO2 , have been reported. Corresponding losses for the said citrate process have been reported to be less than 1-2%.

b. Selective removal of Na-sulphate from the process liquor, for instance by precipitation, is a complicated operation,

and in addition to the alkali consumption with Na2SO , it will also entail a loss of Na-sulphite/Na-bisulphite due to incomplete selectivity. There is a limited market for Na2SO , and deposit of this water soluble salt is problematic.

c. High concentrations of sulphite - bisulphite in all process stages give rise to undesired disproportioning reactions such as 6 NaHSOs -> 2 Na ₂ S0 ₄ + Na ₂ S2O3 + 2 SO2 + 3 H2O and/or 2 Na2 S0 ₃ + 2 NaHSOs -> Na ₂ S ₂ 0 ₃ + 2 Na ₂ SO4 + H2O. On remedy for reducing the Na2S2θ3 formation is to keep the concentration of said component relatively high, but this again leads to a certain reduction of the solubility of the buffer.

d. The absorption capacity of the sulphite buffer is limited due to limitations in the useful pH-range and the solubilit of Na-sulphite and of Na-pyrosulphite (Na2S2θs). (Na-bi- sulphite is not known in solid form, instead Na∑S2Os _^ is obtained as an anhydrite: 2 NaHSOβ = Na2S2θa + H2O). Belo a pH of about 5.5 and above a pH of about 6.5 the buffering effect is small and rapidly decreasing, so that the acceptable pH-range for commercial operation tends to be from about 5.5 to 6.5. Also higher pH-values than 6.5 have been reported to be detrimental when the gases contain CO2 (CO2 absorption) . This entails that the absorption liquid must be conveyed back to the absorption tower with a considerable content of bisulphite.

e. The evaporation for regenerating the absorption solution is a particularly demanding process step. The solubility of N 2SO3 (and also of Na2 SO4 which will be present in a considerable concentration) drops with increasing tempera¬ ture and therefore the temperature approach in heat exchangers used must be very limited since otherwise deposits will form in the exchangers. This requires very large heat exchangers and very high volumetric circulation rates for the slurry through the heat exchangers. As an

example a slurry circulation of 8.64 m ³ /s = 31100 m ³ /h is reported when the process is used for the production of 8.7 tons of S02/h in a 500 MW coal-fired power plant. Furthermore, in order to meet limited periods when the regenerator system is shut down without disrupting the gas cleaning process, the absorber is often provided with large storage tanks for the feed and product liquor.

DESIRABLE GOALS WITH AN ALTERNATIVE PROCESS.

There would obviously be much to gain if said undesirable properties of the sulphite process could be reduced or eliminate in a new process, while at the same time maintaining the said desirable properties. Such a process could belong to process type 2 above or be a combination of processes in which the majority of the SO2 is removed by regular absorption-stripping (process type 1) while the final removal takes place by absorption-evaporation (process type 2) .

THE INVENTION, GENERAL PRESENTATION

We have now found a new regenerative process for SO2 absorption from gases by means of aqueous buffer solutions combined with recovery of SO2 from such solutions, wherein the following advantages in comparison with known processes are offered:

1. Higher absorption capacity per unit volume of liquid.

2. More complete SO2 removal from the gases.

3. Lower energy consumption for SO2 recovery.

4. Less problematic regeneration of the absorption liquid.

5. Less oxidation losses.

In a comprehensive program a number of buffer systems have i been tested and compared with the sulphite buffer. It was discovered that aqueous Na2HPθ - NaH2PO solutions combine a number of desirable properties to a surprisingly high degree.

The invention utilizes the following reactions:

\ Absorption: SO2 < _g > = SO2 < 1 >

SO2( 1 > + H ₂ 0< i > + Na ₂ HPO4( 1 > = NaHSOs _< 1 > + NaHzP0 ₄ ( 1 _> Regeneration: NaHSOβ ( 1 > +NaH∑PO4 < 1 > + ^« H20 =. SO∑ ^' cg> +(n+l) Η20 _{< g} > + (n — large number) .

As seen from the regeneration reaction equation the phos¬ phate system here plays the role of a buffer within said process type 2 and could be used either alone or in combination with e.g. process type 1 as described above. In process type 1 Na- phosphate was found to be an unsuitable buffer, since it cannot stabilize pH at the relatively low pH level required. In advance there seemed to be good reason also to fear serious problems with Na-phosphate-buffers when these are used in process type 2. In the reaction equations above it is assumed that Na∑HPO4 is the buffer salt which is precipitated during the evaporation. This is not obvious since we are here dealing with very concentrated and nonideal salt solutions. Thus, the concentration of NaH∑PO may get very high already in the S0 ₂ absorption stage, for instance 2.5 M compared with 1.5 M for Na∑HPO4 when a 3.0 M Na2HP04 + 1.0 M NaH∑PO buffer is used to absorb 1.5 M SO ₂ (Example 1 and 3 below) . One may fear that this situation aggravates when the solution undergoes further concentration during evaporation. If NaH2PO4 is precipitated, pH will increase and the SO2 evaporation should then be hampered.

Thus, the chemical conditions encountered when a phosphate buffer is used in process type 2 were at the outset quite unclear. This may have contributed to keeping other research workers from. thoroughly investigating the phosphate system for use as a buffer within process type 2. Another reason may have to do with some apparently negative test results: When measured

SO2 partial pressure of SO2 loaded solutions at normal absorption temperatures is plotted against the SO2-concentration in the solution, the curve for phosphate buffers tends to fall much below the sulphite buffer. This could be taken as an indication that the SO2 evaporation from the phosphate buffer system would require much more energy which would immediately exclude this buffer as a relevant competitor.

The use of Na-phosphate as a buffer for SO2-absorption is known from the literature (3). However, in the phosphate process referred to therein the regeneration procedure is completely different: H2S is added to the SO2-loaded solution, utilizing the reaction 2H2S + SO2 -> 3S _<S > + 2H2O. Thus, the process produces elementary sulphur and not SO2 , as in the present invention. Furthermore, in the said known process relatively dilute sodium phosphate solutions are used (about 10% by weight) , and the SO2 uptake in the absorption unit is typically 4-7 g/1 or about 1/10 or less of the loading according to the present invention.

The starting buffer may according to the invention be a pure Na2HPθ -solution with a concentration of preferably above 1.0 mole/1. However, it has been found advantageous to use a Na2HPθ - NaH∑PO mixture with a molar ratio in the range 30:1 to 1:1, -preferably 12:1 to 2:1. Thereby the buffer curve is flattened out so that undesireable high pH values are avoided for buffers with low SO2 content. In principle all SO2 may then be removed from the solution during regeneration. As mentioned, this possibility does not exist with the sulphite buffer.

Very important findings in our investigations are the comparatively higher absorption capacity (in kmoles per m ³ of buffer solution) and the lower specific steam consumption (in tons of steam per ton of SO2 stripped) that can be reached in comparison with the sulphite buffer. In particular for gases having a relatively high SO2 content very much is to be gained compared with the sulphite process, since the steam consumption can be reduced as far down as to half the amount or less. At the same time considerable savings are obtained in the form of smaller equipment units.

It has also been found that the SO2 content in the purified gas may more readily be reduced toRvalues substantially below th strictest requirements. The phosphate buffer solution is surprisingly easier to deplete of SO2 during evaporation- crystallisation at atmospheric pressure phan the sulphite buffer, despite a substantially lower SO2 partial pressure curve at absorber conditions. .

The oxidation losses have most surprisingly been found to be negligible in comparison with the sulphite process. With the much, lower total concentration level of Siv compounds in the liquid in all process stages the disproportioning reactions are restrained rather effectively in comparison with the sulphite process where such concentrations are high throughout the process.

Unlike Na2 SO3 , Na2HPO4 and NaH2PO4 have positive temperature coefficients for their solubilities, at least as far as up to about 95°C. As long as the regeneration is carried out at temperatures around or below about 95°C there is no risk of deposit formation of the buffer salts in the heat exchangers of • the evaporators as mentioned above for the sulphite system. In addition, since the oxidation losses are negligible the concen¬ tration of Na∑SO4 in the circulating process liquid can be kept very low. Therefore the liquid will not reach the saturation concentration of Na∑SO4 until near the end of the evaporation. The solubility of Na∑SO4 has a negative temperature coefficient, and the final evaporation must therefore presumably be carried out with a small temperature approach in the heat exchanger of the. evaporator. This is a disadvantage which, however, becomes modest since it will only appear near the end of the evaporation. In summary this means that relatively much more compact and cheaper equipment may be used and much less pumping energy will be required.

ILLUSTRATIONS OF THE INVENTION BY RESULTS OF EXPERIMENTS.

A. High buffer and absorption capacities for SO2.

A proper buffer should have a high buffer capacity for SO2 , reflected by a small slope of the buffer curve (small change in pH) over a wide SO2 concentration range.

Figure 1 shows buffer curves at 55°C for (1) 3 M Na2HPO4 , (2) 3 M Na ₂ HP0 ₄ + 0,5 M NaH ₂ P0 ₄ , (3) 3 M Na ₂ HP0 ₄ + 1,0 M NaHzP0 ₄ and (4) 1,70 M Na2SO3 , all charged with SO2. The abscissa is th SO2 concentration in moles per liter, and the ordinate is the pH. We have found that pH is rather difficult to measure correctly i such solutions, and the curves should therefore primarily be use for comparing the different buffers. It is striking how the phosphate buffers keep a relatively high buffer capacity over a wide range of SO2 concentrations. From Figure 1 the following may also be concluded:

* The sulphite system is a poor buffer above a pH of abou 6.6 and below a pH of about 5.5. This is in agreement with the above discussion.

* The three buffer curves from the phosphate system demonstrate the great flexibility of this system. By addin some NaH2PO4 to the Na2HPO4 buffer unfavorably high pH values for low SO2 concentrations are avoided. It has been found that such buffers may be depleted more readily of SO2 during the evaporation.

The concentrations of the buffer used as examples in Figure 1 are believed to be about as high as possible in order to be relevant for commercial installations where precipitation in the absorber due to supersaturation must be avoided.

Figure 2 illustrates experimental gas-liquid equilibrium curves for SO2 in (1) 3.00 M Na ₂ HP04 + 1.00 M NaH ₂ P0 ₄ and in (2) 1.70 M Na∑ SO3 buffers at 55° C. The abscissa is the SO2 concen¬ tration in moles per liter, and the ordinate is the SO2 partial pressure in bar. These measurements were carried out in an

equipment which is described in Section C below. Nitrogen was in these measurements used for the gas circulation circuit which equilibrates with the SO2-loaded solution. The lines are broken at lower partial pressures where the measurements are less accurate. Figure 2 shows the phosphate buffer to have the higher absorption capacity for SO2 , which is further demonstrated by the following data read from Figure 2.

Partial pressure at equilibrium p*U2 (bar) 10- ² 10- ³ 10-

Buffer (moles/1 )

^C S0 ₂

Phosphate 1.93 1 . 23 0 .60

Sulphite 1 . 39 1 . 07 0 . 50

Ratio phosphate/sulphite 1 . 39 1 . 15 1 . 20

Over the highly relevant partial pressure range 10- ² - 10- bar the phosphate buffer has accordingly an ability to bind 1.15 to 1.39 times more SO2 p.er unit volume than the sulphite buffer tested. Together with the observations reported in a following section that the phosphate buffer also is more easily depleted of SO2 than the sulphite buffer upon evaporation, the effective absorption capacity for the phosphate buffer is presumably even higher than indicated by these figures.

The solution fed to the absorber in the present process should suitably have a pH above 5.5, usually above 6.0, prefer¬ ably above 6.2, in particular about 6.5. The solution should suitably contain at least 1.0 mole, preferably about 2 moles and particularly above 2.5 moles per liter of a dialkali hydrogen phosphate, particularly disodium hydrogen phosphate.

B. Low steam consumption and complete absorption.

α Batchwise operation.

Buffer solutions loaded with SO2 to known concentrations were subjected to evaporation by boiling at atmospheric pressure in a 5 liter glass vessel equipped with 4 baffles, a stirrer and a submerged electric heating element in the form of a spiral-.

The escaping vapour was passed through a condenser for condensing the steam and then together with the condensate ^' into a receiving bottle containing an aqueous H2O2 solution in which all SO2 was oxidized to sulphuric acid which was then determined quantita¬ tively by titration with standard NaOH. There were 3 receiving bottles attached in parallell via glass valves to a common tube from the condenser. This allowed the evaporation process to be followed by replacing one receiving bottle with another without disrupting the process about every 8 minutes.

The experiments were stopped when the surface of the suspension had fallen too near the top of the heating element, in order to avoid the latter to get dry (overheating) .

The SO2 concentration in the buffer solution was determined iodometrically. The amount of H20 + SO2 evaporated was deter¬ mined by weighing. The density of the solution was determined by weighing known volumes. After the evaporation was stopped, the crystals formed in the vessel were first dissolved by adding water and then the solution was diluted to the start volume, before the SO2 content, the density and the pH were measured. The following results were obtained.

EXAMPLE 1

Start solution: 3000 ml

3 M Na ₂ HPO« 1 M NaK _z P0 ₄ 1.54 M S0 ₂ p = 1.43 g/ml pH = 5.15

Sample No.

H2O+SO2 evap. (g) 110.6 111.9 114.9 115.1 113.5 115.5 ^" 113.1 115.2 112.0 SO2 evap. <g) 6.39 6.65 6.74 6.93 6.90 7.23 7.24 7.54 7.60

16.3 15.8 16.0 15.6 15.4 15.0 14.6 14.3 13.7

Continued 10 11 12 13 14 15 ;16 17 18 19*

114.4 112.8 112.0 109.9 112.0 109.3 114.3 113.5 120 12.

7.93 8.13 8.35 8 .47 8 .92 9.20 10.61 12.05 14.94 1.6

13.5 12.9 12.4 12.0 11.6 10.9 9. 8 8 .4 7.0 6.3 \

1) Crystallization started 110 in. after start.

No incrustration of the heating element observed.

2) Heating stopped after sample no. 18.

Mass balances:

Total weight including vessel: Before evaporation 8785 g

After evaporation 6721 g

Difference = Evaporated SO∑ + H∑O : 2064 g

Sum of all samples of evaporated H_0 + SO2 : 2051.9 g (The discrepancy is within a reasonable error of analysis) .

S0 ₂ at start 1.54*3-64.1 = 256 .1 g

Sum SO2 all samples (a) ^' 153 .5 g

SO2 remaining after evaporation (b) 141.4 g

Sum (a) + (b) 294.9 g

(The discrepany is within a reasonable error of analysis).

pH in rest solution after dilution as explained 5.83 (55° C) Density of same solution 1.40 g/ml

SO2 concentration in same solution 0.72 M

Comparative example

Start solution 3000 ml

1.70 M Na ₂ S0 ₃ 1.23 M SO2 P = 1.19 pH = 5.40

Sample No.

H2O+SO2 evap. (g) 107.4 109.3 126.1 110.4 122.8 113.5 121.1 115.7

S02 evap. (g) 2.19 2.56 2.96 2.64 2.99 2.80 3.08 3.00 _s ,g H2θ-evap.» 48.0 41.7 41.6 40.8 40.1 39.5 38.4 37.6 g SO2 evap.

Continued 10 ll ¹ 12 13 14 15 16 17

118.5 113.1 118.3 112.7 116.8 113.2 116.7 114.0 119.6 3.16 3.08 3.37 3.58 4.49 5.43 7.13 8.65 11.26 36.5 35.7 34.1 30.5 25.0 19.8 15.4 12.1 9.6

1) Crystallization started 80-85 min. after start. The crystals were observed to form on the heating element.

Mass balances:

Total weight including vessel: Before 8055 g After 6065 g

Difference 1990 g

Weight of all samples 1969.3 g

(The discrepancy is within a reasonable error of analysis) .

SO2 at start 1.23*3*64.1 = 236.5 g

Sum of SO2 in all samples (a) 72.4 g

SO2 remaining after evaporation (b) 155.1 g

Sum (a) + (b) 227.5 g

Loss of SO2 : 236.5 - 7.5 = 9.0 g

Part of this loss can be explained by SO2 escaping as gas when the vessel was opened after the evaporation. pH in rest solution after dilution as explained 5.94 (55° C)

Density of same p = 1.18 g/ml

S02 concentration in same 0.80 M

Conclusions,

The results of the batch evaporation experiments allow the following conclusions, referring to the conditions of the experiments:

- The energy (steam) consumption for driving off SO2 from clear solutions of phosphate buffer is about 1/3 the consumption for clear sulphite solutions. This is sur¬ prising, since from the SO2 gas-liquid equilibrium data at hand referring to 55°C, one would rather expect the phos¬ phate buffer to require distinctly more energy than the sulphite buffer (Figure 2) .

- When the crystallization starts the steam consumption drops for both buffers.

- Evaporation with crystallization is problematic with the sulphite buffer because of its tendency to form deposit on the heating element. This problem was not seen with the phosphate buffer.

β Semicontinuous operation.

An SO2 containing buffer solution was by means of a metering pump conveyed to a 500 ml glass flask serving as a boiler and heated from the outside for evaporating a SO2-steam mixture at atmospheric pressure, with simultaneous crystallization of the basic buffer component Na2HPO4. The experiments were carried out semicontinuously by allowing the crystals- with mother liquor to accumulate in the boiler. The vapour was first passed through a condenser for condensation of steam and then together with the condensate down into an aqueous H2O2 solution, in which all SO2 was oxidized to sulphuric acid which was then determined quanti¬ tatively by titration with standard NaOH. The evaporation procedure was followed by very quickly replacing the H2O2 bottle every 30 minutes.

Before starting the pump the boiler was charged with a SO2- free buffer which was subjected to evaporation. The experiments were stopped when the stirrer could not function properly because of too much and too thick slurry.

The SO2 concentrations in the liquid feed were determined iodo etrically, and the total liquid volume charged was deter¬ mined by weighing the container with feed liquid before and afte the experiment and by density measurements.

The SO2 content in the slurry remaining in the boiler after the experiment had been terminated, was determined iodometrically after addition of water and dissolution of the crystal mass to a known volume.

The supply from the pump was controlled with a rota eter calibrated with pure water. Due to deviations in specific density and viscosity between water and the buffer solutions the stated supply rates should only be taken as indications.

EXAMPLE 2

Start solution: 250 ml

To the start solution 20 ml of water was added, and the resulting solution was evaporated until crystals started to form. Then 15 g of Na∑HPO4 •2H20 crystals were added, and the pump was started.

Feed solution 3 M Na2HPO< 1.50 M SO2 pH = 5.10 P = 1.428 g/ml

Sample No, 3 ¹ 4 72 8

Feed rate, approx.

(ml /min.) 5.0 2.2 3.0 6.5 5.0 3.0 3.0 5.0 5.0

H2O+SO2 evap. (g) 125.8 123.6 115.6 101.6 122.9 120.0 118.9 106.6 33.4

S0 ₂ evap. (g) 4.2 11.45 13.3 8.3 10.3 11.94 12.70 10.94 3.61

28.9 9.78 7.70 11.2 10.91 9.05 8.36 8.95 8.2

1) The feed rate was increased from 3 to 7 ml/min. the last 10 min. of this sampling period in order to reduce the density of the slurr

2) The feed rate was increased from 3.0 to 5.0 ml/min. the last 15 min.

3) The feed was stopped after 5 min. and the sampling was terminated.

Comments to Example 2. .*•.

Material balance: /

Total feed of S0 ₂ buffer: 1.151 1

Total amount of S0 ₂ in feed:' 1.151 • 1.50 ^• 64.1 = 110.67 Total amount of SO2 evaporated . (Samples 1-9): 86.74 g Calculated amount of SO2 remaining in evaporator: 23.76 g Amount of SO2 remaining in evaporator found by analysis (residue) : 25.3 g

(The discrepancy is within a reasonable error of analysis) . Residue of SO2 found as equivalent concentration in the feed liquid = "regenerated" buffer solution: 25.3/1.151 • 64.1 = 0.34

Corresponding pH according to Fig. 1 at 55° C: about 6.35

The steam consumption S = 8-9 tons per ton of SO2 may with doubl effect evaporation presumably be reduced to S = 5-6 tons per ton

SO2. This compares with the 7-8 tons per ton SO2 referred to above for the sulphite process.

Interpolation of the gas-liquid-equilibrium data in Figure gives the following SO2 partial pressures at equilibrium, p*U2 :

^c sθ2 ^{moles 1} P ^* so ₂ ^{(bar )} ' ⁵⁵ ° ^c

0.34 < 0.00004 1.50 0.0024

The absorption unit will often be operated at about 55° C. From these equilibrium pressures it may be concluded that 1.50 M SO2 in the buffer solution may be obtained with relatively low partial pressures of SO2 of the feed gas, and that even with as much as 0.34 M SO2 in the feed solution to the absorber, it may be possible to obtain a surprisingly complete SO2 absorption.

The remaining content of SO2 in the liquid after the evaporation can presumably be brought down to a considerably

lower value than the value of 0.34 M found herein, when a regular continuous crystallizer is used. This will make it possible to obtain an even more complete absorption.

Comparative example.

Start solution in evaporator

0.25 1

1.33 M Na ₂ S0 ₃

74.02 g of H20 was evaporated, and the feed pump was started before any crystals had precipitated.

Feed solution: 1.70 M Na∑ SO3

1.25 M S0 ₂ ^• P = 1.21 g/ml pH = 5.38

Sample No. I 2 3 4 5 6 ² 7 8 9 ³

Feed rate, approx.

(ml/min.) 6.5 7.0 7.0 6.5 6.5 6.5 6.5 6.5 6.

SO2+H2O evap. (g) 129.5 139.7 139.9 141.0 145.4 144.8 145.2 147.6 132.

SO2 evap. (g) 0.22 1.71 4.49 7.07 9.33 10.48 11.05 11.57 10.8

Slg HsO/g S0 ₂ ) 588 80.7 30.2 18.9 14.6 12.8 12.1 11.8 11.

1) Crystallization starts after 10 min.

2) Strongly yellow colour of the suspension

3) The feed pump stopped after 15 min. Vigorous boiling. Heating element turned off after 25 min.

At the end the suspension was extraordinary viscous and had a total volume of about 0.5 1. The crystals were dissolved by addition of water to a volume of 2 1. pH measured in the resulting solution = 6.45 (55°C) .

Further data for the experiment were as follows: Evaporated SO2 according to samples 1-9:

66.76 g = 1.04 moles

Total weight of feed 1730 g

Volume of feed 1.730/1.21 = 1.430 1

SO _z added with feed 1.43*1.25 = 1.79 moles

Na2 SO3 added with feed 1.43*1.70 = 2.43 moles

Na2 SO3 in the start solution 1.33*0.25 = 0.33 moles

SO2 + Na2 SO3 + NaHSθ3 in evaporator at the termination:

3.43 moles

SO2 remaining in evaporator as equivalent concentration in regenerated buffer (3.43 - (2.43 + 0.33) )/l.430 = 0.47 M

S-balance:

Calculated from S-containing material added (SO2 and

Na ₂ SO3 ) :

1.79 + 2.43 + 0.33 = 4.55 moles

Calculated on the basis of evaporated SO2 and residue of S- coήtaining material in the evaporator;

1.04 + 3.43 = 4.47 moles

The discrepancy is within reasonable analysis error.

The high values for the initial specific steam consumption is due to the high pH in the start solution. Presumably close t stable conditions with S = 11 - 12 g/g were not attained until sample no. 7.

The pH values of the start and regenerated solutions are in good conformity with the information given above for commercial plants. This also applies to the concentrations of Na ₂ SO3 and S0 ₂ .

Interpolation of the gas-liquid equilibrium data for SO2 at 55° C in Figure 2 gives the following SO2 equilibrium partial pressures,

^P S02 ^:

'SO2 mol/1.. (bar)

^P S02 55 ^β C

0.47 0.0001

1.25 . 0.002^

The absorption unit will often operate at about 55° C. From thes data it is then derived that 1.25 M SO2 in the loaded buffer requires a rather .high SO2 partial pressure in the feed gas. Also the partial pressure of 0.0001 bar in the regenerated buffe is well above the German norm for clean-up of 0.00014 bar for large boilers.

Conclusions.

The results show substantially lower S-values for the phosphate than for the sulphite buffer, 8-9 compared to 11- 12 tons of steam/ton of SO2. -This is in contradiction to the higher SO2 partial pressures of the sulphite buffer at 55° C. Thus, these experiments indicate that SO2 is surprisingly easie to liberate from the phosphate buffer than from the sulphite buffer upon boiling at atmospheric pressure.

At 55° C the SO2 partial pressure in the regenerated phos¬ phate buffer is only about 1/3 of the partial pressure of SO2 i regenerated sulphite buffer. This would make the phosphate system better suited than the sulphite system when extremely lo residual content of SO2 in the purified gas is required.

C. Low oxidation losses.

A gas-circulation circuit illustrated in Figure 3 comprise a gas-scrubbing bottle A, a teflon membrane pump B, a gas absorption bottle C, and a glass tube D. The circulation circu was connected to an oxygen-filled gas sylinder E, through a reduction valve F, for constant supply pressure. Before the experiment the circulation circuit was filled with oxygen, and the gas scrubbing bottle A was charged with a known volume of buffer solution charged with SO2.

By adjusting the three way valyes H the gas flow could be directed through alternatively (a) the tube D and the gas scrubbing bottle A or (b) the tube D and the absorption bottle C.

The oxidiation experiment was started by starting the pump B to circulate the oxygen gas through the buffer in bottle A.

The experiment was stopped by stopping the pump. The SO2 content of the absorption bottle A Was determined iodometrically before and after the experiment. The SO2 content in the gas phase in the gas circulation circuit of known volume, was determined after the experiment by standard base titration after absorption in an aqueous H2O2 solution in the bottle C. The whole apparatus was kept at a constant temperature.

100 ml of buffer solution was used for each experiment. All experiments were carried out at 60°C and 1 atm. total pressure.

The oxidation losses were calculated from the following formula:

_% oxidized = ⁿ s_ttaarrtt ^"'{n eenndd ⁺ , ⁿ σgaass ^{• }} _{m 1QQ}

SO2 , start where n° start = total no. of moles of S -compounds

(Na2 SO3 , NaHSθ3 and SO2 ) in buffer at start ⁿ end ^{= do} * ^{af ter} oxidation n = no . of moles of SO2 present in the gas volume after oxidation ⁿ S_ ^O2 , st.art. = n , o . of moles of SO2 added to buffer bef _.ore st , art , .

The following results were obtained :

Oxidation time , min . 3_0_ 60_

Oxidation loss %:

Phosphate buffer 0.9 0.9

Sulphite buffer 8.7 23.1

The buffer composition at the start was : Phosphate buffer 3.00M Na ₂ HP0 ₄ +1.00M NaH ₂ P0 ₄ +0 . 50M

Na ₂ SO ₄ +1. 50M SO2 pH=5 . 20

Sulphite buffer 1-. 70M Na ₂ SO ₃ +0 . 50M Na ₂ S0 ₄ +1 . 22M SO2 pH=5 . 30

The oxidation in the sulphite system is seen to be very substantial, whereas the oxidation in the phosphate system is so small that it is considered to be insignificant. This is supported by the following observations:

By closing the valve F during the oxidation experiment the oxygen uptake by the buffer in bottle A due to oxidation could be followed visually by observing the movement of the liquid level in the manometer G. For the sulphite system the level moved rather rapidly. For the phosphate system there was no observable change of the level even after 5 min. This indicates that only the sulphite system exhibited significant oxidation.

Litterature:

(1) Erga, 0.

SO2 recovery by sodium citrate solution scrubbing Che .Eng.Sci. 35. 162-169 (1980)

(2) Nato CCMS Study.

Phase 1.2. Status report on the sodium sulfite scrubbing flue gas desulfurization processs. March 1978. Pedco Environmental, Cincinati, Ohio. Contract No. 68-01-4147.

(3) Eldring, A.K.

The Stauffer SO2 abatement system.

In: I. Chem.E. (London) Symposium No. 57

"The Control of Sulphur and other Gaseous Emissions",

Salford, April 1979, p J1-J20.