Document B8e4eGgZa6qnNp6wDNpyVRXGw
Interoffice Communication
To P. L. Fetzer
From
H. L. Hackett
Date March 14, 1973
Subject
Progress Report, February 1973
Summary
Tars Vacuum Flash Distillation--Flash distillation temperature-pressure-composition data were obtained for mixed plant tars containing 14.2 percent EDC and 35.4 per cent 1,1,2 trichloroethane. A fifty-five weight percent overhead fraction was obtained with a combined EDC and 1,1,2 trichloroethane content of seventy-six percent. The bottoms were calculated to contain two percent EDC and sixteen per cent 1,1,2 trichloroethane. Some decomposition and plugging occurred in the pre heater at 170C. It is planned to get the same type data for predominately EDC derived tars in March.
Reduction of Benzene and Trichloroethylene in the Vapor Feed EDC--Benzene and trichloroethylene removal by chlorination was investigated in a series of experiments with recycle EDC. This information was sought to help determine the feasibility of rerouting R-102 effluent (recycle chlorinator) into R-101 (direct chlorinator). Reductions of less than ten percent each were obtained at conditions close to those in R-101. High dissolved free chlorine and ferric chloride aid benzene removal but ferric chloride suppresses the reaction with trichloroethylene. There is some con current loss of EDC to 1,1,2 trichloroethane.
Oxygenates in EDC Streams--Chlorine treatment of R-102 effluent under conditions similar to those in R-101 gave only a moderate increase in the oxygenate level. The result of 2.8 is well below those found when the recycle EDC was sent directly to R-101.
Homer L. Hackett bs CC: LNV-RDG-RHG-GTH-JDM-RWC-DCS-JCL
CCR 000034029
Tars Vacuum Flash Distillation
The flash chamber was built from 12-inch by 3-inch diameter carbon steel schedule 40 pipe with pipe caps on top and bottom. The chamber was mounted with its axis in the vertical position. It was necessary to filter the tars to get consistant operation of the Milton Roy feed pump. Tars were pumped from a calibrated reser voir at 500-700 milliliters per hour into %-inch stainless steel tubing 22-feet long coiled and immersed in an oil bath. The bath was usually maintained at 170C. It was necessary to wrap the 3-foot section of tubing leading to the chamber and the chamber itself with heating tape to prevent excessive heat losses. Feed temperature was measured by a thermocouple placed in a well projecting into the central cavity of a pipe "T" at the chamber inlet. The chamber inlet was two thirds of the way up.. The vapor takeoff was from the top cap. The overhead tem perature was measured in a thermocouple well projecting into the cavity of a pipe "T". A concentric tube condenser, 3/8-inch inside diameter, sloped downward from the "T". The overhead product was collected in calibrated glassware. Unvaporized bottoms product was withdrawn through the bottom cap, through a valve, then into a calibrated glass receiver at system pressure. The valve was manually operated to maintain a 2-inch liquid level in the chamber as observed by a sight glass. The bottoms level was chosen to give sufficient thermocouple well (installed through the top cap) immersion for accurate temperature measurement. The bottoms drawoff rate was held as near constant as possible. Pressure was measured by a "U" tube manometer and controlled by a vacuum regulator.
The experimental data are summarized in Table I. Certain compromises were neces sary in these experiments. These will be discussed as an aid to evaluation of the results. The general procedure was to line-out the distillation at the de sired feed temperature and pressure with the overhead and bottoms products going to slop receivers. The products were then switched into the run receivers and collected for a period of 0.5 to 1.0 hour. If no significant changes occurred during this period the run and its products were accepted. The amount of over head produced could be accurately measured but the amount of bottoms collected during a run was somewhat arbitrary because it depended on accurate liquid level judgements. The combined measurements of the two are considered more accurate than the feed entry since the feed reservoir calibration was only approximate. It is also important to note that the bottoms sample collected and analyzed is not necessarily representative of the non-volatized material. In most cases the equipment was not operated long enough at run conditions to provide enough dis placements of the retained bottoms. It is probably more accurate to calculate the bottoms composition from the feed stock and the overhead. This was done for part of the runs and entered in parentheses below the analyzed value.
Reduction of Benzene and Trichloroethylene in Vapor Feed EDC
Following the October 1972 installation of the separate recycle chlorinator, with its greatly reduced chlorine contact time, benzene and trichloroethylene levels gradually increased in the circulating EDC. The two compounds are not readily removed by the purification train because they boil so close to EDC. Some pro duction time was lost in January and February due to problems in the HC1 column and severe coking in the furnaces possibly related to higher levels of these im purities. A brief study was made of some of the factors involved in the chlori nation of the two compounds and in the feasibility of routing R-102 effluent through R-101 instead of directly to T-101.
Chlorine diluted with some nitrogen sometimes accompanied with HC1 and ethylene, was passed through either R-102 effluent or vinyl column bottoms. The EDC was maintained at 50C. Samples were withdrawn periodically, usually every 10 minutes for dissolved free chlorine and for gas chromatography analyses. After an initial buildup period, an attempt was made to roughly maintain a desired
, CCR 000034030
dissolved free chlorine for the rest of the run. The results are summarized in Table II. The first three runs were made in glassware where the ferric chloride level could be fixed instead of gradually increasing from near zero as in sub sequent runs. Runs 4 through 7 were made in the apparatus previously described as the chamber for the flash distillations. Except for residence time and higher chlorine concentration, Runs 2 and 3 most closely approximate condition in R-101. The data shows benzene removal dependent on both chlorine and ferric chloride concentration. Increasing chlorobenzene was found in the product. The reaction of trichloroethylene is inhibited by ferric chloride. It was pointed out by Dr. Charles Starks, Research and Development, Ponca City, that trichloroethylene might react by alkylation instead of chlorine addition. The results of Runs 6A and 6B lend support to the suggested reaction pathway. Oxygenates in EDC Streams The effect of taking R-102 through R-101 on the oxygenate level was checked by comparison of the charge and reaction products from Run 7 above after distil lation at standardized conditions. These results and some other checks on oxy genates in EDC streams are given in Table III. Comparison of Distillation Numbers 3 and 4 showed much higher oxygenate level in R-101 with no recycle EDC. There is no obvious explanation for this apparent discreptancy. Com parison of Distillation Number 5 with Number 3 shows an oxygenate reduction of 67 percent across the wash system.
CCR 000034031
TABlV AOUl/Y^^^ASM D/Sf/Lt-A Tsbi) op MtXFO 7ARS
o//44S-ri
/AC- Fti.Tr/Aeo TAX* , r- yee- 3-A/- 73 \
><1 /y.s T\
/,/, 3.V% yoTAt. V9 6 %
OUT.
c.ur
C K'DfT/Ofi'S
<3 oA a/7/r /
FP> VAP ftC c
ertA PRFss. Am.
c i/f. rfj
F0^ Aft Ml. Ml
ems. ML.
orfXftSAo
0,4 3-
stms total
3- ` At-
VOL
n
*T.
55
coMPoSf r/o t/ ey gas cz/soMA to&-ra phy
Ove# MPA>
fforroMS to)
p/asm pj.
7?C wm
/./A TOTAl Wt 91 Wt Yt
eot
wth
TOTAL VltJa 3LL2*.
o/=
3 /AS ft2
/Y.f* Yas So J?o 39.YA 77.7 SJ9.1 7.3 ; 7.3 sa.S YP.S /00 ~ /X.o 39.3 sa.3
m
y /Ab
1-1-13 * ill it 3 ill
7-1-73 / /YA
ii 3 *Y3 1 f
00 89 JV./A 3 70
t
/ta / o/ /A.9A /fo /*Y /OA JO. fA --
ni /** 10.9A /SS /I A /to /0`9A ns /&0 n /3. 7s iff
7/ loo 93. S3 Syo.o 633. S /fa/ /.r 39.6 S?. S ?? /l.o 33,0 j 1
Zl.S /OS ii/.O ma.
YY.0
30.3
S3.6 73.9 9.7 A9.0
so. o 0Ti
JL It I.-1 !
37.7 f-
1 I1 '!
90.0 n
92 9X -
33 7s \
/ao,99 97. of s/t. 6 SS.b sr.l
fi-ka nt.ka 3/s.s 3.9 : YY.9
3.07 9SS0 /3S.S 00.6 ' 30.9
33.73 /VO. <,2 SX3.Y
1 37,0
:1
AY.Y S/,1.
26,0 i st.ff 3/.S ' ss.y
76.o
98.8 90.7
6,9 S. 7 (/.?) Os. 7)
i.o av, o
Y/.C (t7.s)
30.0
v t .
J 'Jl /
i 1
,J
1 Sr' i j
/2&
(.1) (a/.S) (*&*)
/o.y 39, 9.9 (6.6) (a 6.8) (33.Y)
J a
'
A9.A
so.s 79.7 -
ft 3 (s.v)
A?. A 35*. r (ao.k) (3a. o)
IZo T~ l"
(1) Not accurately measured n (2) Bottoms samples may not be representative. Some calculated compositions are shown below in parentheses.
n
Jo
o o o uo * uo>
m
ft O A/ /
R-JOO.
rfW T/W Mtv/ tiT
T^FIr T IT
AVJERA&JT PO 30
HCA
--
Caf/#
--
4NP TSS--4V0"0ST//YlP/jS whtf CfUO
excess tii*)
V*
./
A* ^ G-c /MAUV5V5
PPfl\
TAZ/V/AA. % r/AJU.%
/'0
&Z
. 32"?
, 3$?
TRt /.(.?(, t}i.2. ,a/L,
/, /7 .vj?
4 Jr0vc r/ou %
0
P? --
A /1*/02. /OS'
/-3z-'?3
/o-r <33- --
---
,J3
so.o Bz .327
>39t>
tz.L
,
TAJ 1>(.U :
1,37
*S,U ...
0
*
Pft tblO>
I . *690
--
"{?1
3 A-tOT. no /-3/-?3
VC 0TMS fRS
/-*'?
-
& VC &THS t% 0 i-3i-?3
/06 ao --
--
,/2 ' so
Bz ,W
: 7.9.
-- -
- -r 7/U I.U90 - ./,S9? . 1 . ' 67*~
>*
................. 1
, 1 ---' - '
^rr 1[x-\
?o 30 --
-- 1
AS ' 3S1
Bz ,H2
, 634
?2.S
r. L'J *T*
'
IJ
TfU t>SM i ' AVio
6~o
nu
1
f,t,2 ,t> of
.072 .
i .^1 r'-f
H^ <j-li
?0 $0 . 3 --
.23 . tPO
Bz ,,3&2 J. .002 ' . ??.? , ^ 0
: -
1 m /,xyv .... : /,f?r
: 3'*
SU
6A t/C 077tfi ?0 ...
. so -1
3 0* y
1
3C . to
'Bz ,
_ .39/ *
.1 - 1 - 1 !,, 1 . Tflr t.SV*/ . i /.s-vo .
i
s,r 0.3
66 I'C /3 7MS SO
1 -3/-73
s~0 .. ^ 3o
-- ,0~
< a?
: j -
ISZ j
^ - r --
A
Oz .39/ L ( ,m .
TfU /,sYo 1 ; /,
Ii
:
/s,$ S.s- ,
7 /?V %
S-0 3a -- --
.32 .
! Bz ,3P?
.too
77. 3
/ - 3/- 72
Tftt /.CVi, ;
/.Sto ;
a.
(1) Nitrogen and chlorine flows were varied to help control the excess chlorine level.
(2) This is only approximate since the levels varied considerably during a run.
(3) The first three runs were in Glass ware and FeCl3 level was constant, but all other runs were in steel and level
gradually built from near zero up to this value.
0034033
TABLE III - OXYGENATES IN EDC STREAMS (DISTILLED AT STANDARDIZED CONDITIONS TO CONCENTRATE OXYGENATES)
CONDITIONS: 400 ml EDC sample distilled at atmospheric pressure to
5 ml of Bottoms. Bottoms were diluted back to lOcc with a portion of the overhead and the carbonyl absorbance determined. The absorbance was divided by the cell path to obtain the oxygenate number.
Distillation 1. 2. 3. 4*
5.
6.
EDC Samnles
Remarks
Orveenate No.
R-102 R-102
1-31-73 1-31-73
No treatment Chlorine Treated (R-7 TABLE"II)
1.4" 2.8
R-101
2-20-73 With Recycle EDC (R-102 not in use)
2.9
R-101 T-102
1^25-73 2-20-73
No Recycle EDC (R-102 in service)
6656 Oxygenate reduction in wash systems
6.5 1.0
C-203 BTMS 2-20-73
nil
CCR 000034034