Kinetics of the Reaction Between Methane and Sulfur Vapor

304 47 3MB

English Pages 164

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Kinetics of the Reaction Between Methane and Sulfur Vapor

Citation preview

PURDUE UNIVERSITY

TH IS IS TO CERTIFY THAT THE T H E SIS PR E PA R E D U N D E R MY SU PE R V ISIO N

BY

Robert A. Plsher

ENTITLED

Kinetics oT the Reaction Between Methane

and Sulfar Vapor

COM PLIES WITH TH E UNIVERSITY REG ULA TIO NS O N GRADUATION T H E SE S

AND IS APPROVED BY ME A S FULFILLIN G TH IS PART O F TH E REQ UIREM ENTS

F O R TH E D EG R EE OF

Doctor of Philosophy

P

H

r o f e s s o r in

ead o f

S

Chabge

ohool or

TO THE LIBRARIAN:

m TH IS T H ESIS IS NO T TO B E REG ARDED A S CONFIDENTIAL.

GHAD. SCHOOL F O R M 9— 3.49—IM

D

of

Th

e s is

epa rtm en t

TO BE USE0 IN LIBEABY ONLY

KINETICS OF THE BSACTION BETWEEN METHANE AND SULFÜB VAPOR A Thesis Submitted to the Faculty of Purdue University by Robert A. Fisher In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy February, 1950

ProQuest Number: 27712225

All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is d e p e n d e n t upon the quality of the copy subm itted. In the unlikely e v e n t that the a u thor did not send a c o m p le te m anuscript and there are missing pages, these will be noted. Also, if m aterial had to be rem oved, a n o te will ind ica te the deletion.

uest ProQuest 27712225 Published by ProQuest LLO (2019). C opyright of the Dissertation is held by the Author. All rights reserved. This work is protected against unauthorized copying under Title 17, United States C o d e M icroform Edition © ProQuest LLO. ProQuest LLO. 789 East Eisenhower Parkway P.Q. Box 1346 Ann Arbor, Ml 4 8 1 0 6 - 1346

ACKNOWLEDGMMT The author has had many discussions of the project with Dr# J.M. Smith and is indebted for the encouragement and stimulation gained from them# The difficulty of making a practical choice of action was always made easier by the confidence inspired by Dr. Smith and his understanding of the problem# The author is also indebted to George Palmer for his aid in formulating techniques and carrying out analyses, and to Robert Forney and Martha Hull for their interest and help in developing the manuscript.

KINETICS OF THE REACTION BETWEEN METHANE AND SULFUR VAPOR

ABSTRACT The kinetics of the homogeneous reaction between methane and sulfur vapor were studied in a stainless steel reactor at temperatures from 550 to 625°C. and varying reactants ratios and space velocities.

The

data

were analyzed on the basis of the assumption that the rate of dissociation of Sg and Sg to Sg is fast with respect to the rate of the reaction with methane.

A second order reaction between methane and the S2 species

of

sulfur offers a satisfactory kinetic interpretation of the experimental results.

The reaction velocity constant

Qui/(g.mole)(hr.)3

is given by

the equation

k

-

A .9 X 10*^^

e

Kinetic equations postulating that

RT or Sg is the reactive species

of sulfur vapor do not agree with the observed data.

KINETICS OF THE REACTION BETWEEN METHANE

M D SULFUR VAPOR

R. A. Fisher and J. M. Smith Purdue

University

The availability and low cost of natural gas and sulfur has stimu­ lated the investigation of the production of carbon disulfide by the re­ action of methane and sulfur.

The developnent of a satisfactory commer­

cial process has been underway for some years in the petroleum industry. The first published results were those of de Simo (6 ) who found that the reaction was possible in the vapor phase at elevated temperatures (800900®C) and was catalyzed by metallic sulfides.

Later Thacker and Miller (7)

showed that a reasonable rate could be obtained at lower temperatures (500-700°C) with catalysts of the clay type.

Bacon and Boe (1) have

recently corroborated these results and also determined approximately the extent of the homogeneous reaction.

The construction of the first com­

mercial plant for carbon disulfide by this process was reported in 19A9 . While the previous work uncovered effective catalysts for the re­ action and determined the general conversion levels, no attempt had been made to investigate the mechanism of the process.

This problem is of

particular significance because of the partial dissociation of sulfur in the vapor phase.

In the neighborhood of the normal boiling point sulfur

vapor consists of approximately equal amounts of 8^ and Sg at equilibrium, but these species dissociate to a significant extent with an increase in temperature or with dilution to 8^ and 82 • At atmospheric pressure the 8 ^ form does not exist in appreciable concentration below lOOO^C. On the

other hand, 8^ and 82 are the predominant species in the temperature range

2. where the carbon disulfide reaction is commercially feasible.

The purpose

of this investigation was to determine, if possible, the species of sulfur vapor which takes the predominant part in the homogeneous combination with methane and also to study the kinetics of the reaction as a whole. the homogeneous reaction rate is significant, any investigation

Since

of

thé

mechanism of the catalytic process must be preceded by a knowledge of the kinetics in the gas phase.

Accordingly, this preliminary work is concerned

only with the homogeneous gas phase reaction.

EQUIPMENT AND EXPERIMENTAL PROCEDURE

llte problems in constructing and operating the experimental equip­ ment arose primarily from the behavior and properties of sulfur at temperatures.

high

The vapors are extraordinarily corrosive and will attack

platinum thermocouple leads at temperatures near 1000°C.

In the neighbor­

hood of 600°C stainless steel is corroded, but at rates which permit reasonably long life for a reactor made of this material.

Quartz

a

has

some advantages, but silica is known to be an effective catalyst.

For

these reasons the reactor was constructed of stainless steel. Preheating of the methane and sulfur separately is difficult because of the heat absorption wiien ihe sulfur vapor is diluted.

Unless good heat

transfer is available where the gases are mixed the temperature gasos will fall considerably. dissociates to

of

The heat absorbed when one mole of

the Sg

and 82 is approximately 10,000 and 95,000 calories,

respectively. The most serious experimental difficulty was the control and measure­ ment of small sulfur rates.

The unusual viscosity effects hinder the con­

3. ventional liquid rate measurements.

A method based upon passing the sulfur

vapor throu^ a heated capillary tube was found to offer the best solution. The continuous flow system used in this investigation is shown Figure 1.

in

The reactor consisted of standard one inch stainless steel pipe,

six inches long, capped at both ends and with leads to the caps at either end.

The reactor was maintained at the desired temperature level by placing

it in the vertical hollow core (2.5 inches I.D. 24. inches long) of an elec­ trical furnace.

A second furnace directly below the first (Figure 1) ser­

ved to heat the sulfur capillary and preheat the reactants mixture.

The

temperature measurements were made with three chromel-alumel thermocouples, two located on the outside surface of the reactor and one inside the re­ actor at the midpoint.

One of the surface thermocouples was placed at the

same height as the inside thermocouple to permit measurement of the radial temperature gradient, \diile the other was placed near the bottom of the re­ actor in order to check axial temperature gradients. The inside thermocouple was contained in a quartz probe about one-eighth inch in diameter.

To

insure maximum heat transfer capacity, the reactor was packed with approxi­ mately seven mesh rock salt.

Some of the preheating of the charged gases

extends into a shallow layer of the packing at "Wie bottom, as found by pre­ liminary tests. The most successful solution of the problem of joining the quartz tubing to the reactor consisted of a packing gland which held the quartz leads in a packing of asbestos cord.

The method did not permit fine- adjustments of

position in making the connections, but this difficulty was avoided by intro­ ducing a ground joint in the quartz leads at a short distance from the gland and making the adjustments at this position.

The use of these joints also

FLOW

SYSTEM FIG. I

METHANE LJ

SULFUR CONDENSER DRY ICE ACETONE BATH

CAPILLARY i-V W 'T —

REACTOR

HEA TER S

DRIERITE CU FOIL

CAPILLARY AND VAPORIZER AIR PRESSURE

LIQ U ID

SU LFU R

THERMOCOUPLE JUNCTIONS

5. facilitated the introduction of the quartz thermocouple probe.

The glands

are leakproof if they are not subjected to extreme heating and cooling cycles. The sulfur metering system is illustrated in Figure 2.

Liquid sulfur is

maintained at a constant pressure in a reservoir by application of controlled air pressure.

Liquid sulfur is forced upward from the reservoir throu^ a

tube by the regulated air pressure.

The reservoir and tube are of pyrex and

make connection with the quartz capillary section through a standard taper joint in the delivery tube.

The quartz tube is wound closely with nichrome

wire for a length of about one-half inch as indicated in the Figure. sulfur vaporizes at this point at a constant pressure.

The

At a short distance

above the heater, the entrance end of the capillary hangs free in the vapor section.

The capillary itself is suspended from a ringseal at the exit end.

With a constant pressure on the sulfur column rising into the vaporizer, the liquid-vapor interface tends to be stabilized.

Variations in the interface

level aro automatically reduced by the differences in rates of heat transfer to liquid and gaseous sulfur.

If the liquid level tends to rise, the addi­

tional surface covered by the liquid increases the heat transfer, and the sul­ fur is evaporated more rapidly.

Similarly, if the level falls, more of the

surface in the vaporizing section is exposed to sulfur vapor, and the

heat

transfer rate goes down permitting the liquid level to return to its original position.

The vapor temperature is measured vildi the aid of a thermocouple

probe protruding into the vapor space below the capillary.

Hence small

changes in the temperatures of the vapors, which may result from changes in the vaporizer heat input, are detectable.

With proper dimensions of

the

capillary, an essentially linear relation between pressure drop and rate

cf

flow is obtained, and reproducibility of flow rates within a half percent of the average value is possible.

SULFUR

VAPOR

GENERATOR FIG. 2

u METHANE

leads

ARM ELEC TR IC

CAPILLARY

FURNACE

T.C. PROBE BOILER

HEATER

SLEEVE

RESERVOIR BATH

a

' AIR PRESSURE

7. After passing through the reactor, the raixture of gases goes throng the sulfur condenser, where unreacted sulfur is removed.

Finely divided sulfur

is then removed by glass wool filters before the mixture passes into the car­ bon disulfide condenser. In view of the results shown by Thacker and Miller (7), who found that side reactions are not significant, the product gases were analyzed for car­ bon disulfide only.

Carbon disulfide is continuously extracted from the pro­

duct gases in the condenser by cooling them to essentially all of the disulfide condenses.

-60^C, at which temperature

The remainder in the gas was

estimated frcan the vapor pressure of CSg at -60°C. ture minimizes the condensation of HgS.

This choice of tempera­

The condensate was analyzed for car­

bon disulfide according to the xanthic acid oxidation method of Bell Agruss (2) which permits the selective deteimiination of GSg in the

and

presence

of all the possible sulfides. The reactor system and sulfur metering system were maintained within ^ 1 ° C during operation.

In the final set of runs (53 to 79) the radial tem­

perature difference between the inside and outside of the reactor was never larger than 2°C, while the axial change in temperature from bottom to top of the reactor was less than 5®C. In view of the sensitivity of the quartz-steel glands to leakage, it was desirable to minimize the gauge pressure in the reactor. made at reactor pressures of the order of 0.5 inches of water.

All runs were Leakage was

tested before each run by noting the change in pressure of the system as a whole with time. nitrogen gas.

When not in use the system was flushed and filled with

3• RESULTS The conversion to carbon disulfide was measured at four temperatures, 550, 575, 600, and 625°C, witli a methane to sulfur (as S^) reactants molal ratio of 1 to 2 and for space velocities from 27 to 373 reciprocal hours. At 600°C additional data were obtained at ree.ctants ratios of 2 to 1 and 1 to 1.

The methane mass flow rates ranged from 0.03 to 0.42 gram moles per

hour.

Most of the results reported in Table I were measured in the reactor

completely packed (bed depth 6-1/4") vdth seven mesh rock salt.

To test the

homogeneity of the reaction, some preliminary data were obtained i^th the reactor packed to a depth of only 1-1/4".

These results are also included

in Table I. TABLE I For the following runs the reactor \fas empty except for a preheater bed of 10-14 mesh salt, 1-1/4" deep.

The temperature recorded is at a point just

below the top surface end at the axis of the bed. designated as packing C.

Spc-ce Velocity’ '^

The reactor condition is

The void volume of the reactor,

Reactants Molal Ratio xCH^:y(8g)

CH^

was 67.0 ml.

Run No.

Temp, C.

Gr. moles/hr. (8g) OSg

Conversion

25

600

213.

1:2

0.238

0.476

0.0251

0.105

26

600

373.

1:2

0.417

0.834

0.0314

0.075

27

600

106.7

1:2

0.119

0.238

0.0214

0.180

28

550

106.7

1:2

0.119

0,238

0.0048

0.040

29

550

213.

1:2

0.233

0.476

0.0053

0.024

30

550

373.

1:2

0.417

0.834

0.0053

0.013

Run No.

Temp. C.

Space Velocity*

Reactants Molal Ratio xCH^:y(Sp)

31

575

373.

1;2

0.417

0.834

0.0115

0 .028

32

575

213.

1 :2

0.238

0.476

0.0097

0.041

33

575

106.7

1 :2

0.119

0.238

0.0114

0 .096

34

625

373.

1:2

0.417

0.834

0.0531

0,127

35

625

213.

1:2

0.238

0.476

0.0391

0.164

36

625

106.7

1:2

0.119

0.238

0.0312

0.262





®4

Sr. moles/hr. (Sg) CSg

Convers::

The reactor was disassembled and packing C was replaced with packing D, a bed of approximately seven mesh rock salt completely filling the reactor (6-1/4" deep), and of void volume equal to 35.2

ml.

The temperature refers

to the middle of the packing. 37

625

373.

1:2

0.417

0.834

0.0178

0.043

38

625

213.

1:2

0.238

0.476

0.0179

0.075

39

625

106.7

1:2

0.119

0.238

0.0204

0.171

40

625

160 .

1 :2

0.1785

0.357

0.0220

0.123

Packing D was replaced by packing E., a reproduction of the conditions in D.

The temperature is again measured at the axis, ■

at the mid-depth c

the bed. 41

625

373.

1 :2

0.417

0.334

0.0188

0.045

42

625

213

1 :2

0.238

0.476

0.0192

0.081

43

625

160

1:2

0.1785

0.357

0.0203

0.114

44

625

106.7

1 :2

0.119

0 .238

0,0212

0.173

45

625

106.7

1:2

0.119

0 .238

0.0224

0.188

46

625

106,7

1:2

0.119

0.238

0.0213

0.179

10. The flow rate range of the sulfur metering system was altered to permit lower rates.

The temperature at the capillary was also depressed.

The packing

condition remained unchanged.

Run No,

Temp. C.

Space Velocity*

Reactants Molal Ratio xCH^îy(Sg)

53

575

106.7

1:2

0.119

0.233

0.0026

0.021

54

575

213 .

1: 2

0.238

0,476

0,0029

0.012

55

600

1:2

0.02975

0.0595

0.0079

0.268

56

600

213 .

1:2

0.238

0.476

0.0059

0.025

57

600

106.7

1:2

0.119

0.238

0.0078

0.066

58

600

53.3

1:2

0.0595

0.119

0,0086

0 .1 4 4

59

600

106.7

1:2

0.119

0.238

0.0072

0.060

60

625

106.7

1:2

0.119

0.238

0.0160

0.135

61

625

213 .

1:2

0.238

0.476

0.0161

0.068

62

625

53.3

1:2

0.0595

0,119

0.0166

0.279

63

625

26.6

1:2

0.0295

0^0595

0.0136

0.457

64

550

213 .

1:2

0.238

0.476

0.0015

0.006

65

550

106.7

1:2

0.119

0.238

0,0018

0.015

66

550

53.3

1:2

0.0595

0.119

0.0024

0 .0 4 0

67

550

26.6

1:2

0.02975

0.0595

0.0025

0.084

68

575

106.7

1:2

0.119

0.238

0.0032

0.027

69

575

213 ,

1:2

0.238

0.476

0,0030

0.012

70

575

53.3

1:2

0.0595

0.119

0.0036

0.061

VI

575

26.6

1:2

0.02975

0.0595

0.0048

0.161

26.6

CH4

Ch*. moles/hr, (Sg) OSg

Convers: cX

11.

Reactants Molal Gr. moles/hr. CSg (S2 )

Conversion o
erature of the vapor and of the capillary (or furnace) are manually controlled through a Variac and a slide wire resistor* The method proves quite satisfactory when equilibrium is established 3 it is not hard to restrict the temperature to

2

degrees C*, and generally control is even better* Purification* Although sulfur is available in sublimed condition, the few impurities present are a problem sdien continuous vaporization in the boiler deposits a layer on the quartz surface* Small particles flake off and may be thrown into the capillary* Precaution against this case (above) was observed in mounting the capillary, but it was thought necessary to remove most of the impurities beforehand. Filtration of the molten sulfur yielded a black deposit which was not identified: a distillation of the sulfur removed per­ haps BOfo of the impurity, but still permitted some to be entrained* Vacuum distillation was found unsatisfactory because of the strength

16 APPARATUS ARD mOCEDURE

limitations of glass vessels at boiling temperatures of sulfur. A processing of two successive distillations at atmospheric pressure was chosen finally. Calibration. The calibration presents a tedious difficulty. Since the nature of the balance among such factors as heat input rate to the boiling sulfur, the superheating of the vapor to 54(f C. and the main­ tenance of enough pressure in the vapor region to balance the effec­ tive pressure on the liquid sulfur, prevents tiie instantaneous achieve­ ment of equilibrium, it is necessary to make two test runs under iden­ tical conditions but for different lengths of time in order to estab­ lish one rate. Also, they must be developed reproducibly. Thus the determination of a rate may be illustrated by Fig. 6 . The area under each curve is the total amount of sulfur collected by the trap during the run. The difference in the amounts collected is equal to an amount collected at operating rate running for an interval of

The conditions to be maintained continuously during calibration tests are as follows: Furnace temperature (thermocouple)•••••••540**C. / C, Auxiliary heater for sulfur trap.. .47^ C. ^ C. (portable thermocouple) Sulfur delivery tube .................... ,12(f C. ^ C, (portable thermocouple) Sulfur reservoir bath (thermometer)••.♦..120^C. £ 6**C. Methane arm heater: sufficient heat to maintain vapor condition of sulfur in furnace wall section.

.

17

CALIBRATION CYCLE FIG.

6

SULFUR RATE

PRESSURE O FF; BYPASS- OPENED

VACUUM ON

FULL PRESSURE ON

TIME

18 APPARATUS AHD PROCEDURE Remaining test run conditions are developed as follows: Preadjustment»

The trap is weighed (W^) and replaced in position.

The exposed section of the "methane" arm is checked: and free of any sulfur.

it should be cold

Valve (d). Fig. 5, is closed, (a) is opened,

and the desired test pressure set up by adjusting (b) and (c).

How

(a) is partially closed until the pressure falls to about two inches of mercury.

Bypass and vacuum valves (h) and (e) are checked to be sure

that they are closed, and (d) is then opened.

Pressure builds up in

the reservoir and the sulfur rises slowly in the delivery tube.

Mean­

while the boiler element is warmed, and when the liquid sulfur is within about one inch of it, full operating voltage plus about 20^ is applied.

The element becomes a dull red. Test run:

The cold section of the arm is now observed :

when a film of condensed sulfur appears, the stopwatch is started and valve (a) is opened.

The gas lamp is lit and the condensed sulfur

re-evaporates and proceeds to the trap.

The vapor temperature is

checked. At 1*00* valve (h) is opened, and the pressure and vapor temperatures are checked. When the vapor temperature reaches the operating valve, the voltage impressed on the boiler element is reduced to the running valve.

Response of vapor temperature to the volbage setting increases

with the boiling rate. The reservoir level is noted.

19 APPARATUS AHD HiOCEDURB The run is ended by closing valve (a) at termination time, waiting one or two seconds for the reservoir pressure to exhaust, clos­ ing valve (d), and opening valve (e) to the vacuum line*

About ten

inches of mercury vacuum brings the liquid sulfur column down quickly. o The auxiliary heater is switched off, allowed to cool to about 250 C., and the trap is removed to be weighed. A second run of different length is then made.

Experience

with a seven-minute and a twenty-two minute run to give the net weight collected in fifteen minutes has proved satisfactory. Pressure development is handled in two steps to permit simultaneous development of the operating temperature gradient in the liquid.

Premature application of full air pressure cannot be matched

by rapid buildup of pressure in the vapor section, because only the very top of the liquid sulfur column is near the boiling temperature. Flooding of the vapor section would result.

It has been found that

the vapor temperature does not reach the operating valVe, despite the tenq>orary excess in boiler voltage, until two or three minutes after the first appearance of sulfur in the exit arm.

The liquid column extends

into the furnace about one and one-half inches, and the development of the temperature gradient here probably takes place over a similar period of time. The application of the vacuum to terminate the run with­ draws the sulfur column abruptly and sweeps the arm clear of vapor. Data from the calibration runs are reduced to a plot of log (sulfur rate) versus log (pressure drop), as shown in Pig.

20

SULFUR

CALIBRATION

RATE ' GM. M O L E S / H R . HE A D«IN CHES

MERCURY

FIG. 7

LOG (RATE) -

2.0-2

-1 .5 -2

-

1.0 LOG (HEAD)

21

APPJffiATÏÏS ABD PROCEDURE

through the following series of calculations: The weight of sulfur collected is corrected "by a ratio of the average absolute pressure in the capillary when discharged to one atmosphere to the average absolute pressure when discharged to existing barometric pressure* The applied air pressure is reduced by an amount corresponding to the liquid sulfur head reaching from reservoir level to boiler section* A sample calculation of a test run is shown here: Test Test # 82 Vap* 82 Capill* Air Variac 82 LeTemp.(mv) Temp.(mv) Press* Voltage vel

Initial Final Time (min.) Weight Weight Elapsed

38 (a) 22.50

22.50

58*40 22*60

53.55

22.6 A -.470/gm* 7.00 4.3cm. 101.25CTl01.720/

(b) 22.50

22*50

58.40 22.60

55/

22.6 A = 2*990 gm 4.3cm. 101.720 104.710/ 42.00

AH-35.60 in* H2O AH—23.8 cm.82 A W —2*52 gm. AO -35.0 min* Liquid head j- 23.8 cm. 83 1*254 in. % . Applied head 35.60 in. HgO- 2.622 in. Hg* Effective bead ^ 1.368 in. rfg. 3.5 cm. Hg. Log (eff. head)^ .136 ”” Weight of suIfuF collected in 35 min. - 2.52 gm. Corrected to std. pressure condition, ?*52 x/76 / 1.7] - 2,54 gm.

les.2/1.7/" Rate 2

54/64.12 % (60 min./hr/35 min.^ 2 .0681 gm.mole/W.

Log rate - .632-2 Each pair of values is plotted as a point on logarithm coordinates, and the trend yields a line of slight curvature of a

22

APPARATUS AND PROCEDURE slope close to unity.

Reproducibility of calibration points is

estimated to be within about one percent variation on the average. Small variations are noted in the heat input to the boiler section : deposition of impurity on the walls of this section requires higher temperature gradients to sustain the heat transfer rate. Pressure regulation is subject to small drifts over a period of fifteen to twenty minutes corresponding to line pressure changes ; these do not exceed one percent generally.

With periodic inspection

and correction they can be kept to within one-half percent or less.

Methane Rate Measurement The rate measurement of the methane presented no special prohibas. Two grades, 96^ and 99^, were available.

Since no great

precision was expected from the work and since the price for the 99^ grade was about three or four times the amount for the 9 ^ grade, the latter was ordered. Measurement of the gas through a capillary at a constant temperature appeared to be the sis^lest way to secure small rates of flow of gas.

The desired range was 0.025 gm. mole to 0.26 gm. mole

per hour. The capillary and constant temperature bath system is shown in Fig. 8. The capillary is replaceable, being supported only by the standard taper ground joint by which it is connected to the downstream line.

The desired sise was obtained by alternately

METHANE RATE

MEASUREMENT

CAPILLARY AND FIG. 8

COTTON FILTER

cp

CAPILLARY —

-TH E R M O S TA T

HEATER

BATH

PRESSURE — TAP

24

APPARATUS AND HIOCEDDRE drawing out a capillary in a flame and testing it for capacity and for conformity to viscous type flow* The reservoir of gas surrounding the capillary amounts to approximately eighty ml., which may not be sufficient to bring the gas completely in equilibrium with the bath temperature but does provide reproducibility.

The bath temperature is maintained in a

/ 0 £ *3 C. range by a mercury switch and relay arrangement controlling the heater* Pressure regulation is achieved with an ordinary domestic gas regulator by a little modification*

The leather diaphragm is

ordinarily subjected to compression from an adjustable spring*

The

force is distributed over the diaphragm by a steel plate udiich over­ lies most of the area of the leather.

Thus the discharge pressure

of the regulated stream may be built up to a maximum of ten inches of water pressure*

This range is too low for purposes here but may

be increased by impressing air pressure on the spring side of the diaphragm.

In fact, the spring was removed, and the air pressure was

made the reference pressure which would control the discharge pressure of the regulator*

The discharge pressure may be raised to any

reasonable value, while the excellent sensitivity of this type of regulator is preserved.

Reference pressure is available from the air

regulator system described in the section on sulfur* In Fig. 9 manometers 1, m, n, o, p afford a check on pressure at veurious parts of the system.

Manometer (l) reading is

25

METHANE RA"E MEASUREMENT PRESSURE

CONTROL SYSTEM FIG. 9 AIR BLEED

REGULATOR____

,15 PSI SUPPLY LINE

TO VACUUM

TO VACUUM

o

DEOX a DRIER ,

04—

250 .M L ,

REGULATOR CAPILLARY

METHANE

METHANE CA! IBRATION FIG. 10

LOG (RATE)

- 1.5-2

-

10-2 .

RATE = GM. MOLES/HR. HEAD « INCHES WATER

LOG (HEAD)

27 AFPARATCrS m > PROCEDURE maintained necessarily higher than manometer (m) reading, which shows the gauge pressure of the gas entering the capillary*

Manometer

(o) shows the pressure drop across the capillary and is a water manometer approximately eighty inches high.

&bnometer (p) is a combination

relief valve and inlet pressure manometer for the reactor system. A 1000 ml bulb (t) serves as a gas reservoir for maintaining a slight pressure in the system when not in use. Between the capillary and the reactor a drier section of Drierite and a deoa^geimtor section of heated copper foil are included in the gas line. The pressure regulation and metering system may be modified by proper valve settings for filling with methane. Alternate evacu­ ations and fillings are used rather than a thorough flushing with methane, because the void spaces are too inaccessible and large to permit flushing to be efficient.

The leather diaphragm is protected from

overstress by temporarily including the void space above it in the system to be filled.

This is done by means of the bypass valve (c).

Manometers (m) and (o) are protected by shutoff valves.

Reactor The foremost problem in designing the reactor was the choice of materials necessary to withstand the corrosiveness of the sulfur vapors and the complicating effects which arise from this trouble.

Quarts may be used easily in constructing loads and probes, but considerations of size and workability are against it in

28 APPARATUS AMD PROCEDURE fabricating a chamber of any size. corroded rapidly.

Stainless steel is workable but is

Also the use of stainless steel requires a joint

between it and quartz to be maintedned above 50cf C.

Rendering such

a joint gas tight is a task. The choice of the volume of the reactor was based on an arbitrarily selected size of catalyst bed intended for eventual study. This bed was intended to be ten ml. in volume and was used in esti­ mating the desired range of gas rates.

Additional space should be

allowed for varying the size of the catalyst bed and for accommodating an inert packed section, where mixing and preheating could take place. Ik>wever, the entire packed volume, amounting to seventy-five ml., became the effective volume in the calculations.

(See page ^ ).

In view of the dissociation heat effects of the sulfur vapors it is desirable to know the extent of the temperature gradients in the preheating and reaction region.

The measurement of temperature

inside the reactor requires complete protection of the thermocouple wires, and nothii^ less than a quartz probe appears practical for this duty.

Apparatus. The apparatus constructed to meet the problems outlined above is shown diagrammed in Pig. 11 and Fig. 12. Some improvisations were necessary, or time saving when equipment of approximate speci­ fications was already at hand.

29

REACTOR SYTEM FIG. II GRADED SEAL ELECTRIC HEATER SULFUR TRAP

V.

F IL T E R -C E L RADIATION BAFFLE

THERMOCOUPLES

INSULATION

DETAIL OF

JOINTS

FIG. 12

SILICA

TRANSLUCENT

SILICA

SILICA

JOINTS

LEADS

STAIN LESS

STEEL

ASBESTOS

PACKING

THERMOCOUPLE

PROBE

31

APPARATUS AHD PROCEDURE The reactor furnace warms up rapidly when the voltage is applied# because of the rather thick insulation space around the core*

An energy input of 500 watts is enough to bring the furnace up

to 700-800* C«

Current is controlled

by a

stepwise resistor in

parallel with a slide wire resistor and is limited by a fifteen ampere fuse* The preheater furnace is situated directly below the reactor furnace and is aligned with it. inch separates them is running*

A space of about one-half

but is filled with glass wool vdien the system

On the top of the steel cover of the furnace a yoke of

stainless steel strap straddles the core opening and may be adjusted to provide vertical and radisJ. positioning of the reactor# which is supported from it*

The cover is laid over with a thick layer of

glass wool# engulfing the yoke and the auxiliary heater# The reactor is shown in detail and in assembly in Fig* 11 and Fig* 12*

It is 18-8 stainless steel and is connected to the

quarts leads by means of packing gland connections*

The packing

material is asbestos and is compressed around the end of the quarts tube by the packing nut and washer arrangement*

The walls of the

reactor are penetrated by a number of holes in this region to admit air*

the possibility of small leakages of sulfur resulting in freezing

the threads together is thus prevented by the immediate oxidation of any sulfur vapor seeping throgj^h* gas tight seals*

Threaded joints are brazed to assure

32 APPARATUS AND PROCEDURE The cap screws on the reactor until it bears on the edges of the barrel. The surfaces involved are smooth and in practice are sepeurated by a gasket of asbestos.Ventilating ports are out through the cap to permit oxidation of any sulfur vapor which may leak through.

(See Fig. 12.)

The barrel is completely filled with seven mesh rock salt.

Temperature measurements are made possible by three thermo­

couples# two of which aresituated

on the outer surface of the

barrel while the other is at either of two positions along the axis of the salt bed.

(Fig. 12)

In one position the junction is approx­

imately an inch from the entrance end of the reactor; in the other# the junction is at the midpoint of the salt bed.

The surface thermo­

couples are located at corresponding heights for determining radial gradients and are positioned by stainless steel capsules brazed to the reactor wall.

The bed thermocouple is supported in position by the

quartz-to-quarts joint admitting the probe and is entirely protected from the reactant gases by a quartz sheath of about one-eighth inch diameter.

TNhen it is desired to move the junction downward to point

(b) an additional quartz section is substituted at the ”Y*# thus supporting the probe at a lower level.

The standard taper joints may

be assembled dry with a compressive twist severe enough to sustain considerable tension.

They are almost perfectly gas tight.

To cut down excessive axial temperature gradient in the core space of the reactor furnace, a number of thin discs of stainless

33

APPARATUS AHD PROCEDURE steel were fixed to the entrance and exit tubes as baffles as shown in Pig. 11.

Characteristios. The characteristics of the system require some study and measurement. The volinme of voids must be determined indirectly on a re­ production of the packed volume condition in the reactor. was developed as follows*

This method

a glass tube of the approximate diameter

of the reactor was filled a number of times with a fixed wèight of rock salt.

The method of introduction was changed until the volume

occupied was reproducible# idiich condition is made possible by using a suitable funnel and pouring rather slowly.

The determination of

the voids is carried out in the vessel shown in Pig. 13# by repro­ ducing the packed condition of the reactor and then filling the inter­ stices with a measurable volume of liquid.

A liquid which easily wets

the surfaces is introduced at the bottom of the vessel to prevent the inclusion of any air bubbles.

The percentage of void space is

found to be forty-seven percent# and a void space of 35.2 ml. is computed for the bed.

An additional void space exists in the adjacent

part of the exit line for perhaps an inch or two at the temperature level of the reactor; but the inside diameter of this line is someidiat less than 0.5 cm. because of rapid early corrosion.

It is doubtful

that this space should be added to the void space of the reactor in view of the somewhat depressed temperatures at the entrance of the

34

VOIDS DETERMINATION FIG. 13

SALT

TUBE

PACKING

d ia m e t e r :

ABOUT ONE INCH

35

APPARATUS AHD PROCEDURE proheater section.

An error of perhaps as much as three to four percent

is recognized for the estimation of the effective void space# because of uncertain temperature gradients at the entrance and exit. The process must be studied with the chainber fully packed* It had been planned to use only a shallow preheater bed in conjunction with the catalyst bed# but with the decision to investigate the homo­ geneous reaction carefully it was seen that the chamber would have to be fully packed to preserve uniformity of gas temperature.

While the

shallow bed was still in place# however# a number of preheating tests were run with the probe junction in the preheater at point (b)# just below the surface of the packing.

The salt used in this ease was ten

to fourteen mesh and a heat transfer estimation indicates that a layer of one-twentieth inch thickness would suffice to bring the gases to within five degrees of bed temperature for a space velocity of 633. The limiting factor is the ability to transfer heat radially through the bed# as corroborated by tests.

When .the reactor was dismantled

for repacking completely with rock salt it was decided to use a coarser grade, of seven mesh,to keep the pressure drop through the system low fdiile still preserving good equilibrium between packing temperature and gas temperature.

As far as axial gradient is concerned space

velocities of 100 or less produce little change in temperature above the first inch in the preheater section. Temperature regulation is satisfactorily handled by manual control.

The temperature of the bed probe is taken as the t%*perature

36

APPARATUS AHD PROCEDURE of the reaction#

The wall temperature serves to guide control adjust­

ments as well as to check radial gradient, since its response to furnace temperature is faster than that of the probe.

After approxi­

mate equilibrium has been established, the range of temperatures can be generally kept to one degree on either side of the running temperature with little adjustment.

Response to adjustment is relatively slow be­

cause of the thick bank of insulation surrounding the core. In a series of runs between which the apparatus is maintained at running temperature there is no opportunity for checking the thermocouples directly, although there is a roughly indirect check in that the temperature readings in the region of the reactor remain in a constant, close pattern.

Before assembly the

thermocouples are checked during immersion to a depth of six inches in an atmosphere of sulfur vapor.

A large glass tube, two inches

by fifteen inches, is used for refluxing sulfur, with about two inches of boiling liquid in the bottom.

The thermocouples check each other

within a range of one degree, but this temperature is generally about four degrees lower than the corrected boiling point of the sulfur. The only explanation advanced is that heat losses by radiation from the probe take place.

One of these thermocouples was checked against

the melting point of lead satisfactorily, so it is assumed that the discrepancy is caused by the radiation loss. Before all runs a leak test is performed on the apparatus and is effective for the system between the methane capillary and the

37 APPARATUS AHD PROCEDURE carbon disulfide condensation coil*

Leakage is most likely to occur

in the packing gland joints between the stainless steel and the silica tubes.

A simple test was devised* An additional volume of 250 ml. is connected to the otherwise

isolated system and nitrogen gas is bled in until the pressure builds up to three inches of oil as shown by the manometer (p) in Pig* 9. The feed is shut off and the rate of fall of the pressure is timed# A period of twenty seconds or more to fall from two inches to one inch oil pressure is regarded as satisfactory.

The minimum time is

estimated to correspond to a leakage of about eighty ml. per hour at the highest pressure developed in the system; namely, one-half inch of water ahead of the reactor bed. The need for vertical positioning of the reactor by adjusting the yoke support continues into the warming-up period when the length of the steel system increases through thermal expansion.

The position

of the methane arm must be maintained free within the wall port, and the length change is absorbed by raising the yoke. O The temperature level in the exit line should be below 525 C. but above the boiling point of sulfur.

There is s CH^ + /SSg + « Sg + II Sg and Thus the totalnumberof moles of gas present n =

ocCSg -K 2o%Hg8

atany time is

(1 - o() + / 3 + Î T + I I 1 +

=

2 (X

+

/3

+ IT

+

+ 2 CSg + 2 HgS

Hg

=

-25,290 at 700®K 42,700 at 29f K 17,410 cal.

The standard free energy change for reaction (3), (3)

CH4 +

1/2

Sg

GSg + 2 HgS

o

AF

may be similarly calculated. For the reaction 19)

1/4 Sg —

@2

the equilibrium constant for the system at one atmosphere is

a F,

104

APPmDIX

6g

Pg nm. X 1/760 mm.

®

( ■ ^ m r î " 7 7 6 5 ‘iïô*'^



5s- X (1/760 mm#)^ , % (1/760 mm.)

=

^

where Lg is the "L" of Preuner and Schupp defined as "L” —

(Pc mm.)





(pg^mm. )

LogB = =

(l/4)logI^ - (3/4)log760 (l/4)logI^ - 2.100

Values of logl^are drawn from Fig. 25 by interpolation. Reaction (1) may be added to reaction (9) in the form: 1/2 Sg - ►

(1)

C%

(3)

Sg

2 APg

+ 2Sg

GSg + 2 HgS

AP^

GH4 + 1/2 Sg

+ 2 HgS

AfJ

o

The steps in the con$>utation of AF^ are given in the table below: T •k

Log%

InB

lAgB

(l/4)logI%

RT

AF?

-RTlnB

700

2.48

.620

800

1.575

900

6.30 9.27

1000

11.61

2.320 2.905

—1.480 - .525 .220 .805

-3.415 -1.210 .508 1.855

4

2 AFg

AFi

700 800

9,600 3,850 -1,820 -7,380

-28,020 -28,510 —28,860 -29,390

900

1000

1393 1590

1789 1988

AF3 -18,420 -24,660 -30,680 -36,770

4,755 1,924 - 909 -3,690

105 APPENDIX

Approximate heat of reaction: (1)

(3)

CH^

CH^

+

+

2 Sg

CSg

1/2 Sg

2 Sg

1/2 Sg

CSg



6 Hi ==

2 HgS

+

2 AHg

=

AH3

=

2 HgS

-25,290 at ?0(fK 48,500 at

29^K

23,210 cal.

The standard free energy change for reaction (4), (4)

CH|^ +

Sg

CSg

+

2 Hg

may be calculated with the aid of the data concerning the dissociation of HgS. Reaction (1) may be added to reaction (10): 0

f 2 Sg

(1) (10)

2

GSg +

A F

i q

A F i

/T

AFlO* -RTlnK

HlnK *

ÛF^O

CSg + 2 Hg 0

0

T ®K

H2S

—► Sg + 2 Hg

HgS

CHji^ + Sg —

(4)

2

700

38.13

800 900

30.49 24.53

26,700 24,400 22,050

1000

19.74

19,740

A F i

A F ® 0

-28,020 -28,510 -28,860 -29,390

0 AF^

-1,320 -4,110 -6,810 -9,650

The standard free energy change for reaction (5), CB^ +

(5)

1/3

8(,

CSg

2 Hg

AF 5

may now be calculated by adding reaction (4) to reaction (8): (4)

CH4 + S2 — ^

(8)

1/3 @6

(5) * Kelley

0%

+ 1/3 @ 6

CSg + 2 Hg

AF 3

W " Sg

—►

CSg ♦

AF4

2H2

0 A F 5

^

106 APPENDIX

0

0

.T

AF^

. Fj

700 800 900 1000

4,880 2,460 30 -2,390

AFg

-1,320 -4,110 -6,810 -9,650

3,560 -1,650 -6,780 -12,040

The standard free energy change for reaction (6), (6)

+

0

1/4 Sg —

CSg

+

2 Hg

AF6

Sg -

GSg

+

2 Hg

AF^

may be similarly calculated: CHji^ +

(4)

1 /4 Sg

(9) (6)

C%

1/4 Sg

+

AFç

®2 —► CSg

+

0

4

AF 6

AF4

700 800 900 1000

4,760 1,920 - 910 -3,690

AF6

2 Hg

-1,320 -4,110 —6,810 -9,650

3,440 -2,190 -7,720 -13,340

The standard free energy change for reaction (7),

0 CHji^ +

(7)

2 HgS

— ^ CSg

+

4 Hg

AFy

may be calculated by adding reaction (1) to reaction (10) in the form

(1)

CH^ + 2 Sg -► CSg + 2 HgS

AF^

4 HgS

2 Sg + 4 Hg

2 AF^ q

2 2 HpS HgS

CSg + 4 H2 Hg + 4

AF,

o CH. CH^ f

(7)

0 2 AFio

700

53,400

800

48,800 44,100 39,480

900

1000

0

AFi

-28,020 -28,510 -28,860 -29,390

0

AFy

25,380 20,290 15,240 10,090

107

FIG. 2 6

-0 CS 2 + 2 H;

— 10

r— o

— 30

700

1000

T °K

108 APPENDIX Sançtle Calculations The calculations for the following run conditions will he presented: Temperature:

60(f C

1CH^:2(S2> no excess methane, A

Charging ratio:

(1) -

=

0

The first step is to develop the data necessary to construct

the conversion map showing the variation of each constituent during the reaction. The expression used to determine the relation between if and ti is as follows:

-

n)(l/3J

3 -

-

To make use of this expression A and B must be found: they may be drawn from Fig. 25 by interpolation of values of log^and logljp

From page 101,

LogA

^ =

Similarly

Than

LogHg =.

5*231

LogL^ =

8.508

l/3(logH^) 1/3(5*231)

A

.669

B =

.930

|S| =

3.72

-

2/3(log760) 2/3(2.880)

The number of moles of pseudo gas, that is, moles of CH^, (Sg), GSg, and H^S, remains constant at three throughout the

109 APPENDIX

process. The number of moles of actual gas, n, is somewhat less, because of the association of some of the Sg into

and Sg, The

sequence of calculations is as follows. For an assumed value of n, the quantities (A

•“

•ttare

|!f"i(A

+ 3

-

+

3

-

n).

n)(l/3) are calculated; values of

then assumed in order to balance the left and right hand sides

of the equation, IX and

may then be evaluated from the simple

expressions:

II =

tff"i Now choose a value of n such that it is somewhat less than three moles; asuume n

=

n*'* = (A

2,300 moles,

1.320 +

3 -

n)

==

,700

[(!)*."] -

(1 !)^ ^ ^

«

+ 3 -

[ }(2/3)«

n )(V 3 )j =

I.w

I } - ( ]u /3 )«

«

•300

,981

.167

.202

,290

*948

,200

,192

.292

.955_________zl92__________

no APPENDIX II = % n

.0395

=

,663

=

1.742

'/5 % An^n^ =

1.104

=

Hg,

,

.773

_

3n - 4 H

X3 -

.773

-

31Î0 - 4II0

-

.876 -

.158

2.000

.904

= =

_

.096

^

^ 82' *^2^

The values of

H may now

be plotted as increments on the ordinate of the conversion map at the abscissa value of c% ==

.096, as shown in Fig. 27. The sum of

the increments should equal n =

2.300 moles.

As the reaction proceeds the value of n

win

approach

three. The lower limit of n is not known, but assume now a lower value than in the previous trial in order to approach assume n =

of ^

0:

2.200 moles 1.300

n^ = (A

+

3 “ n)

" .800

4.840 = |((a )

^



3

-

,292 1#;

n)(l/3)j =

It w i n be noted that the term (A

+

3

-

n) represents the

number of moles which disappear through association of Sg to Sg and Sg.

Ill

APPENDIX 3 .175 .197 .193 .191 .192 =

.188

2.000

.763

II

-

.980

Assume now that n =

(A

-

.020

=

n^ =

.782

.593 .668 .655 .647 .652 .0219

.170 .098 .095 .115 .108 .112 .116 .111 jO n ___________ a n

112

APPENDIX (/y €1^ =

.577

=. 1.866

i /3

A« %

= =

.720

.934 ^

1

A =

(X =

- 3n - 411 - 3% - 4II0

_

.720

- .577 - .088 2.000

.692

.308

Since the conversions for the present run conditions do not exceed thirty percent, the rates of change in the sulfur vapor compositions require hardly more than three sets of points to construct the initial part of the conversion map. Fig. 27 represents the plot of these points and may be used to read the amounts of all of the gas constituents in the calculation of their concentrations. (2) -

The second step is to evaluate the integral developed

1 I

from the second order reaction rate equation, -

I

i

,

Frcm Fig. 27 it is possible to compile Table 7 to show the mole fractions of methane and of the three species of sulfur vapor existing at a number of different degrees of conversion. The fraction — -—

is then calculable and may be pitted against

available as