i... . ..
9! a;
L

 

. .12....unshnwahaw5naua. jlvfidfié... «mguuwfiuru
12: . 53“.»... z rt : uni“! . {Infing
. .. . . .4.

            

. A. . . . 4. . . . .

. . ..4.......u............fia.z...J....Swnnn .
. . in s; .2. “mm... .. .Wfinmx... a.

.J. . . .1 0.... ._ .. : . . . .. . . 2.41). . 3. .v . .

u r. mm... . .2 «garb... . .uuwmfmr. . . . . :3. “am“. 4:... at... .2. mm“... .EQ. x.

1:.» (Li?! . L. x . 2... s: 3%. ..f%».. . . . . . . x mm... 4 ~ .. 2. . T. A. n . . 9.3:».i.
v2.fix...¥?zr c, . I. «I. n .G... (5. . . .. . . .éx ~33... :. 1 . a... in. .. . .. push. (J 3 . itifihfifiitit‘?
MARE... . 361...... .3..qu 2 . _ .. .. 1.. 8... Y}... I

..
. at. . .ro... LEM. .. . s a. .. .. . a”... .1111... ., .,...
.a . . . .. . 4.. L . .. .4 s. v . x .
Zihflbmfi :w. ham. L... . 9...... u if»... M». n Mafia... 9...: p.$ti..£:2t.
\ .. . . .. 1..
. . i.

                      
         

 

       

        

2.35....

 

...
x 1:!

 

   

  

.... an... mfmmfiwaiww
..m~b....5. manuaaaxm... 5.4mm...
Q

            

      
     

 

é o .7. S. L. a. .315... thfiud...
3 .. .5 . .. _, ..... . .9. iii... ......... .s -,..i
..., 2.38. an _ . .fiafl. .r z. .mmnflrh» 1. a... . «.Qmmfiflmfi .......s. :Pqu. .. . s. u Mnmuhwmmmnmuhfiwh
. ‘1. . . .. . l . . it. 7 . .. .1... v A . .o‘ : l

.. . . . s ".24“... «hézfifiwmfl? .. . d.................. .3. . . z... .. A . z. 3.... 3.. i. . .ethmuwhfifixfivuua
. . , . ,8flfiff1xw . «u... ..

. I.
3. n... no... {I‘LL . .xttx‘tttnl
. 43. . c A '1 *«31'7»\\’H.2tt1

   

..
.
5...... .8... w.
«X. .01.”... "a... . v. z.) ....
3. HT; r. 1.39.... . 1x...
fiahafl...§.§%ms $3.53 hm.
.... v.5 .. an»... x. . .9...
. . ”4: :mi. .. tatkkonnt. .
. . . ark... new: 2.... r. _.
Aid. QtaanflwfiJ! :5 . ~ , . . T 33".. EQWV 25L ..... .. x 5.50 mki$
kgng... . .. , J ,. ..-...:........§_ XJKW _...u...nu...h..-.z.z..¥..-..wv.m.ng .. .. :2... on... .63....
IL§.11\.§§Y\ . . ~ 5.3%“... Law.“ n59“... ‘17....sz . $.I‘f‘? “c.1132.
:2 . m. e .. t... x

 

    

     

        

 

    

     

                         

fl

    

         
     

       

    

 

      
             

   

 

           
       

   

  

         

                 
                          

 

   

   

           

               

   

                    

      
          

  

 

  
 

.. £3.3514...
. . .. . . . .. . . .X...T.{...L.i . x . «7.; i
.8: .. . . . .... . . . . 1.43.31. . c a .. (9.... . cg JIK 2L1...»
In.“ A... .gkuu... 5315.. 2...... .., ... 1:93....umWflh, an? .. ...§..~.2.i..... 1.3.5.1{33 a: a. . (stag.
9. 1.11 «NA. “$1.2: rs . a. valuiiizxasut... . 5‘1... tag“). . u [txllL
mus}. .65... $2.03. . .51.. ..........n...... . 2w... .1 . . . > .s......u..x.fi...
:. . . .4 . K t l...
, 1.22.. 1 2.... 2...... {mm . . .. _ .. a... N! 1..
.. . ilk .yJ . .. . . . . ..
$5.... . 0...: 1.... Runab...«1..}§sz.:: Inn: .2 z .. . . . . . thWPP§UAtS ,1.
.. . amnfibua . :s : Hahn... 51...} <5. I . {:33 . . éttbizti
d . 2...... $2.)... .h.2.$x....d.z...5 . 4... . . . . 5...... 2.31.... .
2,. i3...fluflw.fibfl..fbnfijwmu§rhfl 2...? . a....l.9.,....1.3 .. . . igvfitkdhsit
y s 1 . 11A. 1. . . . z t Z .
Zuznwgfiiéinnmn...mfi.m.uuxzj.2n}.q§ .95....»«ésgp . . .g t .u. Luann... i
......:...§........2.§l.i.... :Qinwuflu. aha..- u......m..i............._.vmb 49......“ .53.} . .. m3... . IL... -19} .3. .
.C (33:47.7. . 21.4992“. x1....1.(~H..:.L..\... .. z . .1?....R<u€..i.§3 3:4. 11...; 341.121.... ‘12.}; . 5.1.1.0... fist...
’13 . .. . x 38 5.551212). ..... . itc 3: .. 2...}.3731. .5} II 1‘; .. is!) rt‘r. 555.3... to. 73‘... . :11-
. . 31.33}... 1.5 .x 2‘ 1.... .17 .1. 3.... .5. 5.. a...)i1\.i-3 2. . . 3 4 st. by»: I .
3.5.4.5? E: xix“... -... . .122“. 1 . 3»... 1...... 1+ .. n. Tart? an“)... 5 “filming; (adviivrfivfi..-
.. . . . 2 £52.. 1...
$.99. 2mm... fiwfilv m... . . . .5? -...!.t...-....... a
. . 1.2.... x. 3 19...... . .. an)...” x V 3.8.5.1 b1 .Y .35.? A » Mk; .5
,. . . .. . . .xwvfian....flé 3:0»...3... hfllsiu .. .51.. 975.5%: : anthr¢1mmmfitwv
... \xx‘s 1v... . ’1’..qu IJ $0.6: )3), <1)! . . V . . 4. . #1.?» EH: hut! Nun
vggfiké .3. . as 2 fig"... .13.}... 9.37pm.... .. 2. k It‘titsnz. .§.Ef..{fluxéfii.uu¢flinfl
>311.s!$fi\{1\§\.xy1x1vy~flui . . v.53 a. .15.“? l .t. 5 a A... 6.... . .fuoifs! A .. 15”.“: {£330
3.M.§HQ§312§§§ 5%.... , fiSgsovfiéf 2? J a r! .s t . . . ‘AafififihflttififSHsLLK. .0;
. . . . .. ) . . . . [A x;
fxé.:<.§.i.a§..w.m}:3.. . .. .omrxfiauhfifl .3... z. _ .uammmumfl . .. z): n. . z t... .48...
1.113312: 1 ifllifl)? k L. .1 V . irvr\z . . :1 r!¥l«\.}x 519.15... LVJHU
. . .5... in“... 2.131. an... 5.51%. 32...... .1 E. .. 31.12%“... .. r .3 ct... , 2.. IILYJ. 21.2.1... ». x...
. 11.3»... . grfld‘nxditufi. 331?... 1x. 5.1!... . £17.. a $ 1.5 I: are... {it .4...)
?.a.$.flu..?3.3 "nth. $35....) .ufimfialith 1.5).. 5:... a z. .413...)

. . .
.111... \ .4 Sngkhufifiat. «4.33... 3.65:0?! «.3... \Lnrtigl?!
I... 24...ri:42{....&§ fi..2........m..i. $31.1. nflwua;z¢1u wfifitfi. . 1...... ..............su.amm._. .
. . H. )3... 2.13%.?! {1.72.}... 7 3»): :N x. .5 2. a .
. ~ 73;»..5‘2 it. . {55... 4... .11 . . .5. ..
.21.»... Inuzfifi.\1..3iiah381. 33......» .. “.5... 3». 3...... 30......

   

       

fin .. . U I. . u... t. . .n ..
~ . l.‘A.§w§x\I-& \f‘ .I .33.. n . . .
.. w“)! J‘hnafl. .. ‘Etflfbruu’vg‘L-A . 4.. 4»... ‘i.6...§ fig
a . . xx . y rnkatvrué.§rzéd.\t.§t will at}... fifin.§£. .1.
.l .1 LIL_~\M...2L..V( LL11. . 1...... “is .L .i .l..\LL.uvflnI...ln{Ot
fif‘ ‘X’ :33, . u I I . :n.‘ llll
2g... .5 2...... H;.d..\1(tti?ftazm§3.et 21.5.5.2... 34?... $9.... a. 33...... . 51%; Auriga...
I... f... .15. 2.. .Ififl... i... .. . k. +3. . .3. .02.. ...h
. .. 533.1554. ks... . i. la. .1 .1. 93...: “with. 2.. .Hix 1 witnurwwi. so. 131 mu. 3. ....\...P.+l..w......‘.ut~.:tt..9u.«u.ndszi.fl.tsh.u.xl.tuknfln
. 1.. 7:... (0.1.24. 1. '1. “’3. 31.13.11 .. \ter; 3.. , ’11....14. . . \ 3. . . . s it); . ‘O‘Ftélt
.I .91.}! gawk“... filinggt: ?§.::xr§¢. Ida"...1.tu§i: 1m. .5:3.$.M6..~. 1 3%.... ik..nul.\..&.xs\.‘ . u. i 3!. ) a
. 51.3335: Slifo‘l. ”Rx . 3V 3...»; .t ~ \
; . , . 1 i i
>

.5 K (viii;
5...... z. 1. £34.. .mns .. An... .. fr}...
. \. §1\.§ )2)“: L. . . .- . A
. .‘K v \ 3.x 1.1.3}; iixxrlxizkufl. . . . I)... I .
f. 2L... gwnfinz..%. h.........:... ........3... .2. a i RA 2.. .

.
-z...

         

        

                

   

  

                          

                   
          

 

          

      

  

.37... . .

        

      

 

      

1..
:1! .17 x. I. 17.91....41 $\..
714’. A .3145 I)... gfixi?!7:1 4.0.11.5: \ 1 31517,. {vb .mtuflttturnt ( 6L
. .51.»... I... 3... .§:.§P\ 31.1% ..fim~.3_;.x..\\ 2.1 3...”! .I -3 .w. 37.5111 $6.3“; $17.75.... nu.
_ . . . 3- 2w, . : . 1) x... . h . . . «.1. .9???
(Syuvwvmfirfia.fiiruflax§.4....s.u.a\h.v.m¢fi1uhv~uu..w_96.3%»... as. . . . 1.”... an»... . fift....\.fl$ .t;fl2~nflnl is
. . i‘f .| { Tl! .t .'\\ '- bzlt 1‘ \Ap pl
.(........7).. ”Hidi1 x OM... 4., Ekxtxwannvw : .11! 11rnflitltutif‘.
I; an“ wigmri. . .. . . .
‘4.- On ii} ~ i}—
l: ‘ gihiixifa 25:}

     

g; 5... 1..
k1$4 9% kudtfig
1.12.. .535... . .s 3.51.... $38.3;
. W! $.13; k . I!
. 11 t\. 1324,1111); .
5. a3‘ ~“Wr.(.t¢1;‘~l ii” .
ail§gilaxmlabnuiiixkfluhv w"; v. i .
333... 35113.»)? r £3.13? . ) I
31?. . .. 1.53.51“... .\
3“»?3! 1....
t 1 .tfifixn. ,2...
...f.$!!§ . L
If?» .

.- ulg a

     

. nL-chILE 11.3.
f... 1...: Hr§§fi211k§3mfiqi «itflufidvptii l...
. .z“ ‘1‘“? i If. L28‘.1tlvit\dlihln.’rl\ll. :.
:Jugfihnu. 11...»...11. . .Lu...s....r\....s.!..hn....xz.. .
3...... It £75.... 1.5...- . .: .2.-t.-t3.k
it}... is... 1 . .t..?...;-..1.53n......- 1P?
. 3." fits». .\t&..ittltxzi .. (lttslxvl?
. 55:. tutti: . ¢..Ln...l\.u.7.a»x5..$t:.l. Erixbx
. if. giftixtiu 4,
. . tit}? 1133!...1.‘
... sz.......d...ttkfl2 tfluu inauig .- 2..11..R... .....
. ~£LL£AI n... u.-5..3vn~-T.ltnntfi...§.nr..vfiL:Ll.V\)lt~.rtb;-k$’tkab 7.5.5!
. Jaufiéfimu fit... izéci..It...ctt.u...z............u\.vfi
. at... A3“: w. . £LA.«.NI1|KA\L. i .. . u
z \ .l{- {L I Lity.‘35€¢t\f5{1\.£¥
magi: Fit}... 5. tiétzlrfivfivi:
ilé.§r $L....L.¢..(v4\.;... v
. . y ‘ . .. K. wt 1 . .
énufifiti.2.........n.......xu.t..n.ti1\t.i..u. nature..- 1.
n. s. 311.11.. t.n.fl.flrfflii\utx.L4t£§xff_c1tc&lL
Ln...

        

7
.
.1 L .2

                       

        

L Eire.
.--...X...
..........4.:!..
........... tittiri. .. i.
fiitfiiiufluiz......1...t.§t.i.eto.2.. .3 .1
. . . L. .
..

     

 

                        
      

              

 
       
   

 

 

  

    

. . . u I
in... v Lt) \. #r.
"II'ELLAAKUHHHDIL .2 b «4......7... 51.59
p... {LixLibiszlsbvrrfi 2.11.30. kl.
. waters}. .12.?! 3-11»... x...:..s:..l.2 r . Adm If!
.a :{ii‘llltt .llll‘l l.” t
. t0}... . xtrlst ..... ALEI§£IL¥:KI.-i(¥x I: 1?;
I???“ u i i. \xT‘ébbi‘-..ll. .5.AL.«A..\9.I.?..I! J‘.‘l¢9i... . .
. 5...... . ix?..:rt.!..2$nv\.i2 1.4-... ....cu............ . 2......
i‘r is... ti: .9. . f1 {1321\4ttt <11!» titt. l .fiiu 1L4“ ..'T.V u~L1LVvAJII
. :1 2:52.... 3%. ELSE... ti... l!!!)lio‘E t... ‘15... .. \tflt}! . ..l§..u7.\tL ‘ . \\\\\ I}: I
. .l 1 .1f...‘ :91... . . 83v.l.xr?1,l A}... .33.... ......\k.1.1\.rt 4...... {:51}: If t. .? uttlo’
. x1...1.4L....Ldet.vY.0. . 3.... 7....1L‘liis4fifx‘lfiwgis Ix.tu?\$x¥.fizhg-tx.hu§1ikfit . J.» ...
. . . . . \u.....l$6t\v.r7..ix.l..‘ .I.I.Jn7.f....i....1l\.( .01 L... . .IIA‘ §LSI|LL :
T; l. .. . . §?.{. £1,113.52 {ttx.tx}}.zz.7\hL1‘l \( 5.1....972E‘1
1 .1... .15.. . {fillieixrn 1.12.3.4... 1a.41.:1..fr.tix.i..stt -1... $33., “an... .
1...- as . x. .. $9132.19 t.{.uM\:.£2..3kli5i1§$-£ lfl»1i....:}i;x:.:n CL}!!! .lirll. . . .
u! :1 .2: Km; .3h..~u...1xu.léz§t 2.1.7.2.}: A!....£..5..11.\1slbr¢...}. .. . . .2
. . .nfl . £32... 5.... ..s.....t...t.. . 1.11.2.1}...t..3«:i..3!...... .. . .
. . 1.... fit} {#:3.‘ ..Y\..IXLL.YI..A4AII‘~K ltl...£E..£L1IV<&IL.I.CII5.0.1.9.}...Af s ..
1 A: A . . .K..J.l.r? ..¢..l{l...|x 1.1.1.}... v ...|.:al.l:.. t.l....§...l 1.4.711...~.¥.~fis .
3. .Bw..i..i.r¢1. ii. .53.».nfiuwfii . {9...}: ii.......!.r..m.--ii.+. .
. I a 1 3.11 gift; .1 $1132.}..IAA f... £Lx\1’£f.e.th.1. .
:7 .5. Iliztlv‘oi... 1.1.. .9..:7.5!il§.lf7§£..~f\.ilxl..\3 1.24.1! .i‘z.§l..|;,.:lxm.kx .......l...9«..\..!4...\~.«
. . . . .. .
.3 1 .. . .

    

  

            

3...

.3...us

  

. .. .-
. K|LLLlKLl k\.$tL|.1.
f‘l?.~.\.i~r7L-.£ .7?
E.
L»:

THESIS

  
 
 
 

EIBRARY’
M'shiganSme i

umflfity f

 

This is to certify that the

thesis entitled

DESIGN AND OPERATION OF A DEMONSTRATION SCALE
LIQUID-ION EXCHANGE ALUM RECOVERY PLANT

presented by

Gary C. Cline

has been accepted towards fuifillment
of the requirements for

MS degree in Sanitary Engineering

 

NM,W

Major professor

Date May 16, 1980

0-7 639

 

 

<-.» i . w... A Ae—uevn—r-«b— 4——-g«—.~<-,-‘_~€'—."
t
1

\-

":

I " lvk“"’/ff” I 51'
W

 

OVERDUE FINES:
25¢ per dew per item
RETURNING LIBRARY MATERIALS:

Piace in book return to rem0\
charge from circulation recon

(mm L.

 

 

 

 

DESIGN AND OPERATION OF A DEMONSTRATION SCALE
LIQUID-ION EXCHANGE ALUM RECOVERY PLANT

BY

Gary C. Cline

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Civil and Sanitary Engineering

1980

ABSTRACT

DESIGN AND OPERATION OF A DEMONSTRATION SCALE
LIQUID-ION EXCHANGE ALUM RECOVERY PLANT

BY
Gary C. Cline

Based on data developed in continuous flow, bench-scale
studies of alum recovery from raw alum sludge with the .
liquid-ion exchange process, a lO-gpm demonstration-scale
pilot plant was designed and constructed at the City of
Tampa, Florida Hillsborough River Water Treatment Plant.

. The pilot plant was operated a total of 10 months in
1979. Operational difficulties which were initially en-
countered included the emulsification of the insoluble
Sludge solids during aluminum extraction and the subsequent
recovery of solvent from that emulsion. These problems were
overcome by the replacement of the mixer-settler extractor
with an RTL Contactor. The RTL Contractor did not emulsify
the insoluble solids and precluded the need for solvent
recovery.

The results of the pilot-plant operation confirmed
bench«scale results and process scale-up methods. The
liquid-ion exchange process recovered 90 percent of the
aluminum in the sludge and effected a substantial reduction

in the dry weight solids of the aqueous waste stream.

 

ACKNOWLEDGEMENTS

I would like to recognize the following people —-

Mr. Elroy Spitzer and the AWWA Research Foundation
for their generous support,

Professors Mackenzie Davis and John Eastman for their
encouragement in the early days of this study,

My parents for their support in all shapes and forms,
Dave Cornwell for his friendship and hard work,

And last, but not least, my wife Lise -- who makes it
all worthwhile.

ii

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

CHAPTER
1. INTRODUCTION
1.1 Description of the Problem
1.2 Rationale for Research
2. LIQUID-LIQUID EXTRACTION EQUIPMENT
2.1 Principles
2.2 Contactor Classification
2.3 Column Contactors
2.3.1 Unagitated Column Contactors
2.3.2 Mechanically Agitated Columns
2.4 Centrifugal Extractors
2.5 Mixer-Settlers
2.6 Other Contactors
2.6.1 RTL Contactor
2.6.2 Morris Contactor
2.6.3 Motionless Mixers
2.7 Summary

SELECTION OF CONTACTING EQUIPMENT

PILOT-PLANT DESIGN

4.1
4.2

Introduction

Pilot-Scale Design Data Development

4.2.1 Introduction
Extractant Strength
Stripping Acid Strength
Number of Extraction Stages
Number of Stripping Stages
Extractor Dispersion Conditions
Stripper Dispersion Conditions
Extractor Phase Ratio
Stripper Phase Ratio

0 Extractor and Stripper Mixer

Residence Times

0

HOGDQO‘WIPLON

NNNNNNNNN

O

O O O O

vbthrPI-PtPtP-éthsh

iii

Page
vii

viii

 

 

4.2.11 Extractor and Stripper Mixer

Impeller Speeds

4.2.12 Extractor and Stripper Settler

I-PI-b
430.)

4.4.1

Loading Rates

Pilot Program Objectives and Timetable
Proposed Operation and Pilot Equipment Design

Proposed Operation
4.4.1.1 Introduction

Extraction Circuit

Stripping Circuit

Solvent Recovery

Personnel Requirements
quipment Design
Extractor Mixer
Extractor Impeller
Extractor Settler
Extractor Settler

Weir System
Solvent Feed Pump
Sludge Feed Pump
Extractor Organic Recycle
Stripper Mixers
Stripper Impellers
Stripper Settlers
Stripper Settler

Weir System
Stripping Acid Pump

and Accessories
.2.13 Stripper Aqueous Recycle
.2.14 Recovered Alum Treatment
.2.15 Solvent Recovery
.2.16 Tank Design

H-0 0

ecbb p ppmeppe Laceyepbfi

N NNNNNNN NNNNdHHHH

O O

O

0 O O

Hl-‘LOGDQOWU'I IpLONl-‘FIWQUDN

l—‘O

thFnP-bp vb whip-##9##} ##IP-thI-FIP-I-Pth
3—:
N

5. RESULTS OF MIXER-SETTLER OPERATION
5.1 Introduction
5.2 Extraction Circuit Performance

5.2.1

Extractor Mixer
5.2.1.1 Hydraulic Performance
5.2.1.2 Aluminum Extraction
5.2.1.2.1 Introduction
5.2.1.2.2 Aluminum
Extraction
Efficiency
Extractor Settler

iv

Page

79
82
82
82
84
84

85
90

 

 

,__— n-. M ...¢..r__..

5.3 Entrainment and Stripping Circuit Performance 91
5.3.1 Introduction 91
5.3.2 Entrainment Symptoms 92
5.3 3 Causes of Entrainment 94
5.3 4 Other Aspects of Stripping Circuit
Performance 99
5.3.4.1 Introduction 99
5.3.4.2 Stripper Mixers 99
5.3.4.3 Stripper Settlers 99
5.3.4.4 Entrainment 100
5.4 Solvent Stability 101
5.4.1 Introduction 101
5.4.2 Pilot Studies 101
5.4.3 Tampa Bench-Scale Studies 102
5.5 Personnel Requirements 105
RESULTS OF RTL CONTACTOR OPERATION
6.1 Introduction 106
6.2 Equipment 108
6.2.1 RTL Contactor 108
6.2.2 Pumps and Accessories 108
6.3 Plant Layout 109
6.4 Operations 111
6.5 Extraction 112
COST ANALYSIS OF ALUM RECOVERY
7.1 Introduction 118
7.2 Tampa Facility Design 118
7.2.1 Sludge Pretreatment 118
7.2.1.1 Sludge Collection 118
7.2.1.2 Thickener 119
7.2.1.3 Thickened Sludge
Holding Tank 119
7.2.1.4 Sludge Pumping to
Recovery Site 119
7.2.2 Acid Storage 119
7.2.3 Mixer-Settler Extractor 120
7.2.4 RTL Contactor Extractor 120
7.2.5 Centrifuge - Alternative 1 120
7.2.6 Centrifuge - Alternative 2 121
7.2.7 Strippers 121
7.2.8 Solvent Reservoir 122
7.2.9 Recovered Alum Storage 122
7.2.10 Granulated Activated Carbon Columns 122
7.2.11 Waste Stream Neutralization 122
7.2.12 Building 123
7.2.13 Sitework 123
v

\J\l
45w

8. CONCLUSIONS AND RECOMMENDATIONS

8.1
8.2

APPENDICES
A.
B.

C.

BIBLIOGRAPHY

Capital Cost

Operating and Maintenance Cost

.4.1 Acid

0 O

mm®¢¢ bbebpe

Power

0
0 Analysis

\10\)\l\1 \1\)\I\I\J\l\l

2
3
4
5
.6 Labor
7
8
9
1
1

Conclusions

Solvent
Alum Credit
Lime Demand
Carbon Demand

Sludge Conditioning and
Hauling Savings

Downtime

Maintenance

Capital Amortization
7.5.2 Capital Pay Back - Alternative 2

Recommendations

Mixer-Settler Operating Data
RTL Contactor Operating Data

Calculation of Aqueous Entrainment
in Loaded Organic

Vi

Page

123
123
123
125
125
125
125
125

126
126
126
127
127
129
129

131
133

135

148

149

153

 

LIST OF TABLES

Table
2-1 Classification of Industrial Contactors
2-2 Contactor Summary
7-1 Estimated Capital Costs of Tampa Facility

7-2 Estimated Operating and Maintenance Costs
for Tampa Facility

Page
11
41

124

128

Figure
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
2-10
2-11
2-12
4-1
4-2
4-3

4-4

LIST OF FIGURES

Unagitated Column Contactors

Pulsed Column Contactor

Rotary Agitated Column Contactors
Asymmetrical Rotating-Disc Contactor
Treybal Liquid-Liquid Contactor
Draught Tube Mixer

General Mills Mixer-Settler
Davy-Powergas Mixer-Settler

IMI Mixer-Settler

Kemira Mixer-Settler

RTL Contactor

Morris Contactor

Extractor Mixer Impeller

Extractor Mixer-Settler

Stripper Mixer Impeller

Stripper Mixer-Settler

Schematic of Stripping Acid Feed System
Schematic of Tampa Pilot Plant
Pilot-Plant Operating Data

Extraction Coefficient as a Function of
Raffinate pH

viii

Page
13
16
18
20
22
27
29
30
32
33
36
38
58
62
67
69
73
77
8O

88

Figure

5-3

Stripping Acid Aluminum Concentrations as a
Function of Operating Time

Effect of Extractor Mixer Phase Ratio on Recovered
Alum Aluminum Concentration

Aqueous Entrainment as a Function of Extractor
Mixer Phase Ratio

Extraction Coefficient as a Function of
Organic Loading

Schematic of Tampa Pilot Plant with RTL
Contactor

Percent Aluminum Extracted as a Function
of Sludge Detention Time

Percent Aluminum Extracted as a Function
of Phase Ratio

Raffinate pH as a Function of Sludge
Detention Time

ix

Page

93

97

98

104

110

113

115

116

CHAPTER 1

INTRODUCTION

1.1 Description of the Problem

Aluminum sulfate or alum is used as a coagulant for the
removal of turbidity and color in potable water treatment.
The removal of these impurities produces a sludge which is
of low density and contains large amounts of water of hydra-
tion. The production of alum sludges in the United States
exceeds 14 million wet weight tons each year.(l)

Traditionally, alum sludges have been disposed of by
discharge to the water treatment plant raw water source.
However, the enactment of PL 92-500 in 1972 has forced the
water production industry to abandon traditional sludge
disposal methods and replace them with environmentally sound
sludge treatment and disposal techniques.

In order to comply with the terms of PL 92-500, water
treatment facilities must employ the Best Practical Treatment
Economically Available (BPTEA) by July 1, 1982. BPTEA is
generally accepted to be a technology that has been demon—
strated on an advanced laboratory or pilot-plant scale to be
technically and economically feasible.

Currently, the most economical method of treatment and
disposing of alum sludges is dewatering following by land-
filling of the residual solids. To be amenable to landfill-

ing with other waste material, the sludge must have no free

water, which implies a solids concentration greater than 20
percent. Furthermore, it is generally felt that the solids
concentration must be increased to 40 percent in order to
landfill an alum sludge alone.(l) Because of the hydrous
nature of alum sludges, obtaining a solids concentration of
20—40 percent is difficult and expensive. While many sludge
dewatering methods are currently available, each has problems
associated with it.

Co-disposal of the sludge with wastewater sludge can be
practiced if a sanitary sewer is accessible and the wastewater
treatment plant will accept the sludge. The disadvantages
of co-disposal are that the acceptance costs of the wastewater
plant are often high and these costs are subject to inflation.

Gravity methods of sludge dewatering include drying on
sand beds, freeze-thawing, and lagooning. Sand bed drying
and freeze-thawing can be effective if the proper climate
and necessary land coexist. Lagooning of sludge is only a
temporary solution since lagoons eventually fill with solids
and these solids are generally unacceptable for landfilling.

Mechanical dewatering devices such as vacuum filters,
belt filters, centrifuges, and pressure filters are capable
of producing a solids cake suitable for landfilling, but
these devices tend to be capital and operating intensive,
putting the costs of mechanical dewatering methods out of

the reach of smaller plants.

Landfilling costs can also be high. The aluminum and
heavy metal content of the dewatered sludge has prompted
many states to classify alum sludge as a hazardous waste.
This classification requires the sludge to be disposed of in
secure landfills whose user costs are higher than those of
conventional sanitary landfills.

Clearly, other sludge treatment alternatives must be
developed if water treatment plants are to meet discharge
requirements and hold down the cost of water production.
The groundwork for one such alternative was laid down in
1975 by Cornwell.(2) Using liquid-ion exchange, a com—
merically successful operation in the field of extractive
metallurgy, Cornwell proposed a system where aluminum could
be recovered from alum sludges generated by wastewater
treatment plants in the removal of phosphates. In conjunc-
tion with the recovery of aluminum, a substantial reduction
in dry weight suspended solids was also realized. Later,
this idea was expanded to include the recovery of alum from
alum sludges generated by water treatment plants.

Using liquid-ion exchange to recover alum from water
treatment plant sludges is a novel idea, but the recovery of
alum through sulfuric acid treatment of the sludge is not.

(3)

Acidification recovery systems, reviewed elsewhere, could
recovery 50-90 percent of the aluminum in the sludge and

effect a significant reduction in the weight and volume of

 

 

solids requiring disposal. These systems, however, were
plagued by the accumulation of contaminants in the recovered
alum. Also, the recovered alum was often dilute and incon-
sistent in its aluminum content, making accurate dosing
difficult.

Alum recovery through liquid-ion exchange can overcome
the problems associated with acid recovery by selectively
extracting aluminum from the sludge, leaving potential
contaminants in the process waste stream. Liquid-ion
exchange recovery can also produce alum which is consistent
in nature and comparable to commercial alum in strength.

In 1976, Cornwell was awarded a research grant by the
American Water Works Association Research Foundation to
investigate and develop the liquid-ion exchange alum
recovery process. The study was divided into three areas,
each covering one year. The subject of each study area is
listed below:

0 lst year — Batch Optimization and Sludge
Characterization

0 2nd year - Bench-Scale Continuous Flow Studies

0 3rd year - Pilot-Scale Continuous Flow Studies

In the first year, viability of the process was demon-
strated on a batch basis using synthetic aluminum solutions
as feed stock. Extractants, diluents, and stripping agents
were screened and selected. Extraction and stripping phase

ratios were optimized and isotherms developed.

In the second year of research, the previous results
were used to design a continuous flow bench-scale recovery
unit. During the first months of this study area, synthetic
solutions were used as feed stock. Hydraulic and chemical
characteristics of the system were evaluated. When it was
felt that key operating parameters had been defined, the
feed solution was changed to raw alum sludge. This sludge
was taken from the City of Tampa, Florida, Water Treatment
Plant. Tampa sludge was chosen for the following reasons:

1. Tampa's use of 22 tons of alum per day made coagu—
lant recovery very attractive.

2. The sludge generated at Tampa contained large
amounts of organic matter and it was felt that the
processing of this sludge would be a severe test
for the liquid-ion exchange operation.

Key operating parameters were again evaluated with sludge as
the feed solution. The final months of the second year were

spent generating design data necessary for process scale-up.

1.2 Rationale for Research
After two years of laboratory study and development,
the viability of the liquid-ion exchange alum recovery
process had been demonstrated at both batch and continuous
bench scales. Using the sludge generated at the City of
Tampa, Florida, Hillsborough River Water Treatment Plant as
the feed stock, operating parameters and design data were

developed. Based on the proven laboratory feasibility of

the alum recovery process, the City of Tampa, Michigan State
University, and the AWWA Research Foundation undertook a
joint venture to have the first liquid-ion exchange alum
recovery pilot plant erected at the Hillsborough River Water
Treatment Plant. The subsequent design, operation, and
evaluation of that pilot plant make up the third study area

of the research program and are subject of this thesis.*

*The third study area was divided into two parts, the first
being the subject of this thesis. The second deals with
the quality and treatment of the process4exit streams and
is the subject of a thesis by Pryzbyla.

 

 

CHAPTER 2

LIQUID—LIQUID EXTRACTION EQUIPMENT

2.1 Principles

Efficient transfer of solute from one phase to another
and the subsequent separation of those phases is the task of
liquid—liquid extraction equipment. The rate of solute
transfer is dependent on, among other factors, the inter—
facial area and the deviation of the actual solute concen-
trations in the two phases from the equilibrium position.(5)
Thus, without greatly degrading phase separation performance,
an efficient design seeks to maximize interfacial area and
solute concentration gradients.

In liquid-liquid extraction equipment, one phase is
dispersed in the form of drops into the continuous phase.
The interfacial area is a function of phase ratio and mean
drop size. The drop size is dependent on the phase ratio,
degree of agitation, and physical and chemical properties of
the phases. Physical properties would include phase densities,
viscosities, and interfacial tension. Up to a point then, a
decrease in mean drop size yields an increase in interfacial
area which increases mass transfer rate.

The solute transfer mechanism involves both molecular

and eddy diffusion. Molecular diffusion, the random movement

of molecules in the direction of the concentration gradient,

 

is a relatively slow process. Eddy diffusion, originating
in the bulk movement of the fluid is, under turbulent condi-
tions, much greater than molecular diffusion and for the
purposes of liquid-liquid extraction to be encouraged.

Given turbulent conditions in the continuous phase,
eddy diffusion will usually dominate that phase. Turbulence
in the dispersed phase is a function of mean drop size, the
velocity of the drops with respect to the continuous phase,
and drop interaction. A high relative velocity is desirable
because as a drop is moving through the continuous phase,
drag at the interface sets up internal circulation in that
drop. This action promotes mixing with the drop and enhances
the mass transfer. Droplet interaction, the coalescence and
redispersion of two drops, produces mixing within the drops.
This too, promotes solute transfer in the dispersed phase.
Internal circulation and droplet interaction, however, are
functions of drop size. Very small drops behave as rigid
spheres. So, as agitation increases, the mean drop size
decreases and a point is reached at which the drops assume
this rigid sphere behavior. It is here where internal
circulation and droplet interaction cease and the slower
molecular diffusion mechanisms becomes the primary means of
mass transfer in the dispersed phase. So it can be seen
that adequate mixing is necessary to effect efficient mass
transfer. Excessive mixing, however, will have a negative

effect on mass transfer rates.

 

 

As stated earlier, solute transfer rates are also a
function of the difference in solute concentrations in the
two phases with respect to the equilibrium position. This
is optimized when the mixing pattern for the two phases
corresponds to perfect countercurrent plug-flow. In stage-
wise contactors such as mixer-settlers, the concentration
profile is very nearly ideal. Differential contactors,
those which provide continuous conditions for mass transfer
throughout their length, usually deviate greatly from their
ideal concentration profiles.

Transferring solute from one phase into another is only
one part of the liquid-liquid extraction process. Once the
mass transfer has taken place, the two phases must be
separated. The rate of coalescence of the dispersion is
primarily a function of the drop size distribution of the
dispersion, the phase ratio, and the nature of the disper-

(6'7) Also, if solute

sion (organic or aqueous continuous).
transfer in the mixer is incomplete, the transfer will
continue in the settler and this can have either an enhancing
or inhibiting effect on the rate of coalescence.(8) Often,
conditions which are desirable for mass transfer are
deleterious to phase separation. Increased agitation can
increase mass transfer, but the decrease in drop size of the
dispersion reduces the rate of coalescence.(6) Also, mass

transfer from the dispersed to continuous phase has been

shown to promote coalescence.(5)

 

 

 

 

10

Given the interdependence of the many variables involved
in the mixing and settling operations of liquid-liquid

extraction, process optimization can be a difficult task.

2.2 Contactor Classification
Many types of contacting equipment are currently avail-
able ranging in sophistication from simple spray columns to
centrifugal extractors. The classification and description
of the various contacting devices has been carried out by

several investigators.(5' 9' 10)

A brief survey of liquid—
liquid extraction equipment will be the subject of this
section.

Industrial contacting equipment can be divided into two
main categories, differential and stagewise. Differential
contactors provide conditions for mass transfer throughout
their length. The flow of the two phases is countercurrent.
Because the concentration gradient changes along the length
of the contactor, equilibrium is never reached at any point
in the contactor. Stagewise contactors, on the other hand,
provide a number of discrete stages in which the two phases
are equilibrated and then separated before being passed
countercurrent to each other.

Further divisions can be made according to the type of
force which produces phase dispersion. This force may be
gravity, pulsation, mechanical agitation, or centrifugal
force. A list of the major types of contactors is shown in

Table 2-1.

11

 

AOHOSO

 

 

 

 

 

Hwauummluwxfiz :owmummmflcmu
Houomuumm Gad :ESHOO lwocmommamoov
Houomucoo awakens mumamlm>wflm cesuoo aomucoo
cesdoo Hwnwwsom comasm wuwam mmfi3mmmum
Houomuuxo
Hm>mq on
Houomuuxm casaoo
mwammumwz :EsHoo :ESHOO vmxomm unmucoo
Houomuucoo counmsmumszmcao cmxomm cfisaoo Adequcmummwflcv
xmficawfincom Dam comasm hmwmm _ .msoscfiucou
OOHOM cofiumufimm «an couscoum
ammSHMHucmu HMOficmnomz :owummasm mufi>muu t|_:OflmHommfic
mmmnm
mouom
Hmmzmfluucmu >ua>muo

 

 

 

«mm couscoum 30Hm acmuusoumusnou

 

 

 

8.. $035.28. ufimamsoza no 20552535

\ lilfllll!1l.

MIN mam<B

 

 

 

12

2.3 Column Contactors

2.3.1 Unagitated Column Contactors

The simplest type of column contactor, shown on Figure 2-1,
is the spray column. In this contactor, the dispersed phase
is sprayed either up or down through the continuous phase.
Spray columns are simple to construct and are capable cf
providing high throughputs per unit cross-sectional area.
They are very inefficient, however, since there are no
internal obstructions to check axial mixing in the continuous
phase. There is so much axial mixing in these devices that
the continuous phase usually has a uniform composition
throughout the column. Spray columns are usually used for
simple separations or washing processes requiring few stages.

The problem of axial mixing in spray columns is substan-
tially reduced by the introduction of packing material into
the column, resulting in the simple packed column, depicted
on Figure 2-1. The packing material chosen must be preferen—
tially wetted by the continuous phase to prevent premature
coalescence of the dispersed phase and thereby reduce the
interfacial area. The packing will enhance solute transfer
by reducing axial mixing, setting up internal circulation
within the dispersed phase, and promoting droplet interaction.
The use of packing, however, reduces the cross-sectional
area available for flow and the diameter of a column required

for a given throughput will be greater for a packed column

 

 

13

AmvmmOBoflazoo ZZDAOU omaflfiHU¢ZD

Hum mmDon
oumcwwmmm u m ouaaom I m
poem I m uomnuxm I m

2238 9::
ouhgoummm 2:38 omxo: 2238 ><~Em

m» \SEEE/ m»

 

 

 

 

  
  

 

 

 

 

a
000.0.Wan. c.0oaao
a v Q k 0 o k
a . a o a
o
.0120 c
Q Q
o o O c
o . .
r . .1. a
e s . o
n . 05.— Q
/ .. .
/ 5:sz o 2 .
/lomv_o<n_ . u
a o o p .
a
9 .
o o es s
w P Q. v o
o . o a
n .
II II .k 19.!
TIE? I‘ll ‘II
x m m

 

 

 

 

 

 

14

than for a spray column. Packed column performance is
hindered if there are solids present in either feed stream
which would foul the packing.

The third important type of unagitated column is the
perforated plate column, shown on Figure 2—1. Operating in
a stagewise manner, the dispersed phase goes through a
coalescence-redispersion cycle within each stage. These
columns can be operated with either light or heavy phase
dispersed. The dispersed phase moves either up or down the
length of the column, as the case may be, collecting under
or over the perforated plates until enough has coalescenced
to produce an adequate head to force the liquid through the
perforations and disperse it again into the continuous
phase. Interstage flow of the continuous phase is handled
by downcomers or upcomers, depending on which phase is
continuous.

In general, gravity differential contactors are
inexpensive to build and operate. They cannot handle wide
ranges of flow ratios, require large amounts of headroom,
and because of axial mixing are inefficient and difficult to

scale-up.

2.3.2 Mechanically Agitated Columns
If mechanical agitation is provided, the efficiency of

simple packed or perforated plate columns can be greatly

 

 

15

increased. One way of supplying this agitation is to apply
an oscillating pulse to the contents of the column, resulting
in the pulsed column shown on Figure 2-2. It has been
suggested that a gauze packing be used in place of random
packing since the pulsing tends to orient the random packing,
causing channeling and lowering efficiency. The action of a
pulsed column involves a coalescence-redispersion cycle.

The dispersed phase will coalesce above or below the gauze
or plate and will be prevented from flowing through the
perforations until the upward or downward pulse pushes that
phase through the perforations, dispersing it into the next
stage. The continuous phase is drawn through on the other
cycle of the pulse. The application of the pulse improves
the solute transfer efficiency over that of an unpulsed
column by increasing turbulence and interfacial area.

The mechanical problems associated with providing the
pulse limits the size of pulsed columns. In order to produce
larger units there has been developmental work with the use
of compressed air to pulse large columns.(ll)

Because of the considerable amount of energy required
to pulse a liquid column, the reciprocating-plate column has
been developed as an alternative. Research on this type of
column and its industrial use have been reported in recent
years.

While pulsed columns have had many applications in the

nuclear and petrochemical industry, the most important type

l6

HEAVY LIGHT
LIQUID T LIQUID

______ PERFORATED
______ ///F— PLATE

 

 

A J
i LEAST,
LIQUID Q

FIGURE 2-2 (11)
PULSED COLUMN CONTACTOR

 

 

 

 

17

of mechanically-agitated columns are those which use a
rotary device for mixing.
The best known and most commonly used rotary agitated

columns, shown on Figure 2-3, are the Scheibel column,(12)

(13)

the Rotating Disc Contactor (RDC), and the Oldshue-Rushton

column.(l4)

Coalescence occurs only at the ends of the RDC and
Oldhue-Rushton columns. In the Scheibel column, some settling
takes place between stages and operation of this device is
neither stagewise or differential, but a combination of the
two.

In rotary agitated columns, the rotary mixing produces
a more efficient dispersion of the phases. This type of
agitation, however, may promote axial mixing. To minimize
axial mixing, some form of baffling must be used.

The Scheibel column was the first mechanically agitated
column to be introduced commercially and was quickly utilized
by the petroleum industry. The original version of the
device had mixing impellers located between layers of packing
where coalescence occurred. Later designs used baffle
arrangements to replace the packing.

In the RDC, the impellers, which are rotating discs,
are located midway between static ring baffles attached to

the shell of the column. These rings deflect the streams of

liquid into the center of the column producing circulation

18

 

 

fl?

 

 

 

 

 

 

 

 

 

OLDSHUE¥RUSHTON

 

 

VERTICAL

BAFFLE _\\\

 

‘1:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.T T T T. .,.
r; I I I‘-
IIIIIIIIIIIIII
. lllillll-lllli.
rlllllllllllil"
IIIIIIIIIIII I

 

 

RDC

SCHEIBEL'

(5)

FIGURE 2-3 '
ROTARY AGITATED COLUMN CONTACTORS

19

within each stage and reducing axial mixing. Phase dispersion
is accomplished by the shearing action of the rapidly rotating
discs. RDCs have been used throughout the world in the
petrochemical industry to separate aromatics from aliphatics
in the sulfolane process.

The Oldshue-Rusthon column is similar to the RDC in the
use of ring baffles to inhibit axial mixing. Instead of
using rotating discs, the Oldshue-Rushton device utilizes
turbine-type impellers for agitation. Also, vertical baffles
are placed along the shell wall within each stage to improve
mixing efficiency.

In an effort to reduce axial mixing, columns have been
developed which combine the efficiency of mixer-settlers
with ground-space compactness of columns. Two designs were
developed which have settling zones between each mixing
chamber. This arrangement isolates the mixing chambers and
greatly reduces axial mixing.

The Asymmetric Rotating-Disk extractor (ARD), shown
schematically on Figure 2-4, is a semi—compartmentalized
device with phase dispersion accomplished by means of rotating
disks mounted on a shaft that is off center in the column.
Settling zones, located at the side of the column, provide a
path for interstage flow. It is claimed that the amount of
axial mixing in the ARD is much less than that in other
mechanically agitated devices. Axial mixing is not entirely

eliminated, as some does occur through the settling zones.

20

on
4“

 

h
FL

 

 

 

 

 

 

 

1. Rotating disc rotor
2. Mixing zone
3. Settling zone

FIGURE 2-4
ASYMETRICAL ROTATING-DISC CONTACTOR

(9)

 

 

21

A similar type of contactor is the Treybal Liquid-Liquid

Contactor,(15)

shown on Figure 2-5. Originally developed as
a vertical mixer-settler, it has been modified to operate in
column form. This device is composed of two cylinders, the
smaller set off center within the larger. Horizontal baffles
divide the column into compartments. Mixing takes place in
the smaller cylinder, with agitation being provided by a
turbine impeller. The dispersion flows out of the mixing
chamber into the outer annual chamber where coalescence
takes place. The separated phases then flow countercurrently
to their respective mixing chambers. The Treybal unit can
produce high transfer efficiencies and flow capacity is
comparable to that of the ARD extractor.

A design which combines the mixing provided in
mechanically-agitated columns and the simplicity and
flexibility of perforated plate columns is the Kuhni column.

Featured in this unit are the use of double-shrouded turbine

impellers and perforated stator plates.

2.4 Centrifugal Extractors
The contactors discussed to this point make use of
gravity to effect both countercurrent flow and phase separation.
The rate of countercurrent flow and phase separation can be
increased by the application of a centrifugal force. This

principle has led to the development of some very compact

 

22

 

 

 

 

 

Impeller 1?
L. _ -- Dispersion
”E.
L.
”L ' -\
A I ’ - ”’E ~- . 4 -

 

 

 

 

 

Ihvadcummmwmhp
nmm0.1hqmb pmmwmn mu.
shown. LL - ”gift liquid, HI. =-
Inavy liquid. (0.3. Stoneware,
hm) .

FIGURE 2-5
TREYBAL LIQUID-LIQUID CONTACTOR

 

23

contacting units offering short contact times and the ability
to separate dispersions which exhibit poor separation charac-
teristics under normal conditions. Centrifugal extractors
have been used extensively in the pharmaceutical industry.

The Podbielniak extractor(l6)

was the first centrifugal
unit to be used on a wide-spread basis. It is mounted on a
horizontal axis and consists of a rotating drum inside which
are a series of concentric perforated cylinders. The light
phase enters the drum under pressure through the horizontal
shaft and flows via a channel to the periphery of the drum.
The heavy phase is fed through the shaft to the center of
the drum. The centrifugal force, acting on the density
difference between the phases, causes the two phases to flow
countercurrent to each other through the perforated cylinders.
Being differential in nature, one unit can provide more than
one theoretical stage.

The DeLaval extractor is a vertically mounted centrifuge
bowl fitted with vertical baffles to create a complex spiral
of channels from the center to the outside of the bowl. In
the bowl, the heavy phase flows radially outward, while the
light phase is displaced countercurrently toward the center.
Orifices located alternately at the top and bottom of adjacent
baffles allow the liquids to pass between concentric zones.

As the phases pass through the orifices, shear forces create

 

 

 

 

——«~ _ 1

 

24

a dispersion which then separates under centrifugal force
before the next orifice is reached. It has been claimed
that this arrangement can provide up to 20 theoretical
stages in one unit.

The Westfalia Extractor consists of a vertical centri-
fugal bowl which can be divided into several separate chambers,
one located above the next. Each chamber representing a
separate stage. One advantage to this arrangement is that
the light phase does not have to be fed under pressure. The
maximum capacity depends upon the number of stages employed.

The Robatel Extractor utilizes a series of "mixer-
settlers" stacked on their sides with mixing in each stage
being carried out by a stationary disc on the shaft with the
mixing chamber rotating about it. The two phases pass into
the settling chamber before moving on to adjacent stages in
countercurrent flow via a channel system.

The Quadronic design is a differential contactor which
utilizes the concept of gradient mixing energy.

Centrifugal extractors have the advantages of being
able to handle two phases with low density differences,
having low hold-up, and short contact times. Also, they
require little floor space and headroom. The disadvantages
of these devices include high capital and operating costs

and a limited number of stages in a single unit.

25

2.5 Mixer-Settlers

Mixer-settlers are the simplest liquid-liquid contacting
equipment that are completely stagewise in operation. Some
of the earliest commercial liquid-liquid extraction plants
employed mixer-settlers. Early models used separate mixing
and settling units with the dispersion flowing from mixer to
settler by gravity. Interstage flow was carried out by
pumps. A design which required only half as many pumps had
the stages staggered in height so that the light phase could
flow by gravity while the heavy phase was pumped from stage
to stage. Both of these early models were reliable, flexible,
and easy to design. The major drawbacks were the large
ground-space required, the large inventory of chemicals
necessary for operation, and the pumps and piping involved.

Because of the density differences between phases and
the average density of the dispersion in the mixer, all
interstage flow can be gravity driven. The Windscale design
employed this principle and eliminated the need for interstage
pumps. This design also eliminated the need for interstage
piping because it was built as one box with partitions.

Although a major improvement over previous designs, the
Windscale design also required a large chemical inventory.
The driving force for interstage flow was a function of

liquid depth and this inhibited chemical inventory reduction.

26

The flow limitations of the Windscale unit were overcome
by interstage pumping like that provided in the Pump-Mix

extractor.(17)

In this design the mixing impeller also
served as an interstage pump. This allowed an increase in
throughputs without increasing the depth of the unit. The
Pump-Mix extractor was an integral-box design, similar to
the Windscale design.

The Pump-Mix design had the disadvantage that the
pumping action of the impeller could apply a suction to
adjacent stages and destroy the hydraulic independence of
the stages. Also, use of this design reduced the flexibility
of the units. Any change in phase ratio had to be coupled
with an adjustment in the pumping units to avoid one phase
being completely displaced from the settlers, with massive
back mixing of the other phase being the result.

The hydraulic problems associated with the Pump-Mix
extractor were overcome by the use of a draught tube mixer,
shown on Figure 2-6. This design could handle a wide range
of phase ratios and flow rates within the pumping capacity
of the impeller. In the draught tube mixer, the two phases
flow over their respective weirs and are mixed as they flow
up the draught tube after which the dispersion flows by
gravity to the settler. Changes in throughput or phase
ratio are compensated for in the mixing chamber by the

rising or falling of the liquid level in that vessel.

27

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

(I
BAFFLE \ \
w
‘R.
CO
SETTLER MIXER SETTLER
FIGURE 2-6

DRAUGHT TUBE MIXER ( 9)

28

Demand in the last decade for units capable of handling
very high flowrates has results in much research and develop-
ment work on mixer-settlers. \Many different designs have
evolved and each will be briefly discussed. A more complete
discussion of mixer-settler design, performance, and operation
is presented in Chapters 4 and 5.

The General Mills design,(18)

shown on Figure 2-7,
utilizes a round mixer fitted with vertical battles to
minimize vortexing. The pumping impeller is a top-shrouded
turbine. Following mixing, the dispersion flows by gravity
through a trough to a settler equipped with a picket fence
to reduce turbulence and distribute flow evenly across the
width of the settler. A shallow settler is employed to
reduce chemical inventory. This design is in commercial use
for copper extraction.

The Davy-Powergas mixer-settler,(19)

as shown on

Figure 2-8, also uses a pumping impeller. In its basic
form, the Davy-Powergas design includes a square-sectioned
mixer and a shallow, rectangular settler. The feed streams
enter the mixer through a draft tube where they are mixed by
a double-shrouded turbine impeller. A horizontal plate with
a central circular port is fitted in the top of the mixer as

a vortex breaker. The mixed phases flow off the top of the

mixer into the settler where a dispersion introduction

 

 

 

 

 

 

 

29

PICKET FENCE

 

 

 

 

 

 

 

 

 

 

GENERAL MILLS MIXER-SETTLER (5)

/,
I/
-1 LIGHT PHASE rx
I
I A -——-~ //’————_-—~
I r" // '
I I / HEAVY PHASE
I ' I
I I
I I
I I ELEVATION
l I
L—_\ /<————L—I
__...)K _.
——-/ \-——
/»—PICKET FENCE
/
Ii /
III’ I
I"
II I
I?
hi I
III I
'II I
III I
III I
I"
III '
III I
IHXER
SETTLER
PLAN VIEW
FIGURE 2-7

 

 

30

mmnfiafimlmmxHS mdumfi30mlfi><o

Amy

 

 

 

 

 

 

 

 

 

 

 

w4um<m hmdz_ xuappmm

mIN flMDOHh
J N
. «
x7 :5; :5;
8:: 2323:”; I/ _ __ _
l,
W
If 3:: :5: IIIHV WI
\
K
\
I\

 

I

31

baffle leads the dispersion to the proper depth in the
settler. This design has been employed on a large scale for
copper extraction.

The IMI mixer-settler,(20)

as shown on Figure 2-9, was
developed by Israeli Mining Industries, Ltd., and is currently
being used for uranium extraction and phosphoric acid purifica-
tion. Pumping and mixing are carried out by separate impellers
mounted on the same vertical shaft in a cylindrical mixing
compartment. The phases are first mixed and then drawn up a
draft tube before being sent to the settler. The advantage

of separating the functions of mixing and pumping is that

each can be optimized individually. The dispersion enters

the center of a cylindrical settler where the phases disengage
radially outward. The separated phases are removed by weirs

at the periphery of the settler. The settler is also fitted
with turburlence reducing baffles.

(21) shown on Figure 2-10,

The Kemira mixer-settler,
also uses pumping and mixing devices that are separate but
mounted on the same shaft. The dispersion is drawn from the
mixer at a point in line with the mixing impeller. The
cylindrical settler has a conical bottom from which the
heavy phase is removed. The light phase is removed from the

t0p of the settler where it flows by gravity to the next

mixer. It is claimed that this design has the advantages of

32

‘\

MMABBMmIMMxHZ HSH

Amy mIN mmDGHh

m4mm<m mozmnamxnp
\fil I_hz<

 

 

 

 

 

 

/////////</

 

 

 

 

 

\\\\\\\ \\
I

 

 

 

 

 

 

 

 

 

 

Lam. calm mmaon

It a L3: I I
.+.2 _I.2
L3; :2: C a m3: 2
was: :3: Kb) was: :5...
Lm<=a >><L=

mmdfififlmlmmtz flmHZm—M

 

 

-. DID I

 

/(\

33

 

 

:. /

.m¢<bm ch
mm<=m h=o_4 mm<=m h=o_4

I -, //

 

 

 

 

 

.I:
wc<hm 2cm;
um<zm h=c_4

 

 

 

34

easy shutdown, easy start-up, and the elimination of unstable
phase conditions caused by the contact of the phases prior
to entering the mixing chamber.

Lurgi, of West Germany, has two designs which are
currently being marketed.(22) One design is a stack of
mixer-settlers forming a column. Settling compartments and
interstage channels are within the column. All interstage
flow and mixing is accomplished by axial-flow pumps on the
outside of the column. The size of the settlers in this
unit are smaller than those of other designs for a given
flowrate and where applicable electrostatic forces have been
used to promote phase separation. The other Lurgi design is
a horizontal—type mixer—settler which employs an axial flow
impeller to mix and pump. The settler is rectangular and
fitted with multiple horizontal trays to allow higher settler
throughputs.

Holmes and Narver, Inc., has revealed a mixer-settler
design that includes multi-compartment mixers. Little
information, however, is currently available on this design.

It is important to note that any mixer-settler design
which has pumping impellers has the ability to recycle a
phase from the settler to the mixer of a stage. This
recycling ability can be a powerful tool for optimizing

operating conditions within a single stage.

 

35

2.6 Other Contactors
There are several contactors which do not fall into any
of the traditional categories. These are the RTL Contactor,
the Morris Contactor, and the motionless mixer. Each will

be briefly discussed below.

2.6.1 RTL Contactor

The RTL Contactor (formerly the Graesser Raining Bucket
Contactor) shown on Figure 2-11, combines mixing and settling
in the same compartment. Also, it is unique in that each
phase is both continuous and dispersed. Laid out horizontally,
this device has a series of plates mounted on a central
shaft. Located between the plates are C-shaped buckets.
This rotor assembly is fitted into a shell with a narrow gap
between the shell and rotor discs. It is through this gap
that lengthwise flow occurs. In operation, half the contactor
is filled with heavy phase, the other half with light phase.
As the rotor is slowly turned, the heavy phase is picked up
and dropped through the light phase while the light phase is
pulled down and bubbled up through the heavy phase. The
mixing in the RTL Contactor is very gentle, making it attrac—
tive for systems with poor separation characteristics.

Operation of a RTL Contactor will be discussed in Chapter 6.

 

36

 
 
  
  

SOLVENT
INFLUENT

SOLVENT
EFFLUENT

SLUDGE

RAFFI NATE
EFFLUENT

F IGURE 2-1 1
RTL CONTACTOR

37

2.6.2 Morris Contactor

(23) shown on Figure 2-12, is a

The Morris Contactor,
continuously agitated, horizontal unit comprised of a
rectangular vessel divided into a series of compartments by
alternate vertical weirs and baffles. The weirs and baffles
are spaced such that narrow transfer sections are formed
between two consecutive mixing sections. Each mixing section
contains a vertically mounted impeller. Flow is countercurrent
with each phase being introduced into the end mixing chambers
at opposite ends of the contactor. If the light phase is
dispersed, droplets are swept under the dividing baffle into
the adjacent transfer section where they ascend. The heavy
phase is swept over the weir in the opposite direction.

Under normal operating conditions, an interface exists
only at the dispersed phase outlet. This helps avoid the
phase separation problems which sometimes accompany mixer-
settlers.

The Morris Contactor has been used in the manufacture
of trinitro-toluene and the washing and lye treatment of
soap curds.

Morris contactors are relatively cheap to build and are
simple and flexible in operation. Throughput is said to be

greater than that for comparable mixer-settlers because of

38

hzo
mm<=m
h=o_4

Aw.

.mOBUdBZOU mHmmOE
NHIN EDOHE

 

 

 

 

 

 

mm<=m
>><w=

. I

 

 

 

 

 

 

 

 

:—
mm<=m
>><w=

 

 

 

 

 

wm<zm
h=c_4

39

the absence of settler. It is the absence of these settlers,
however, that causes back mixing of the continuous phase in

Morris Contactors.

2.6.3 Motionless Mixer

A motionless mixer is a shell or pipe fitted with
internal static elements which force multiple splittings of
the feed streams as they pass through the mixer and effect
intimate contacting. The geometry of the internal static
elements is generally a proprietary feature of the many
brands of motionless mixers on the market.

Motionless mixers usually require additional pumping
head for feed streams. The additional pumping costs, however,
can be offset by savings in operating and maintenance costs
since the motionless mixers do not have any moving parts.
Also, the static elements provide a predictable amount of
mixing.

Copper is currently produced in Zaire at a liquid—ion
exchange plant using motionless mixers as contactors. Other
motionless mixer applications include the separation of
cresylic acids and mercaptans from catalytic-cracker gasoline,
the polymerizing of caprolactain in the production of nylon 6,
and the hydrolysis of chlorinated organics. Units up to

5.25 feet in diameter have been reported.(24)

4O

Motionless mixers may find applications in systems
prone to the formation of emulsions because the mixing can

be controlled much more closely than in mechanical mixers.

2.7 Summary

Many types of contacting equipment have been developed,
each with their own advantages and disadvantages. They all,
however, utilize differential or stagewise contacting
principles. Variations in equipment that has evolved are
generally specialized applications of one or both of the
basic contacting principles. Table 2-2 lists the better
known contactors, their advantages, disadvantages, and

applications.

41

Hosm Hmoaosc
ucmsuwouu
HouMSmwmmz
Havausmomahmsm
mmusaamuofiowcma
Hflo magnummw>

Houflawuuom

HMOfiuDoomauwfim
ucoeumwuu

Hmum3oumm3

Hosm wmoaosc

mmwsaamuoeoupmn

HMOflEmnoowuom

:0flmuo>coo
Houm3 ocflamm
amusaamuoEOHCSn

 

mcofiumoflammd
HownuwsccH

pans

oamcwm ca mommvm
mo wonfisc pmuwsfla

mumoo
mocmcoucfimfi swag
muwoo mcflpmummo swan
mumoo Hapwmmo zmfln

oflumn 30am
30H: macaw: uoccmo
mfiwumhm
mGHMMHmHSEo
macaw: poacmo
moccnmmmac
mufimcop HHmEm\3
unmamsouan emuasnfi

QSIonom UHDOHMMAU

moamHOHmmo 30H

Eoowcmon nmfln
owumu 30am

ocflz macaw: poccmo
oocouommaw
muflmcmc HHmEm\3

usmamuouau emuflsna

 

mommucm>cmmwo

.v

.m
.N
.H

Shoucw>cw HHmEm
ucmfi

Imnwoqou oommm 30H

oEHp mdnaon uuonm

oEDHo> mspaon 30H
oozeuommac

mufl>wum 30H macaw;

mfiImHmom
Sumo ham>flumamu

mommum mama
umoo wHQMGOmmmH
coflmnommflp poom

.m coauooupmcoo onEHm

.H

mumoo
unflumummo 30H
mumoo Hmuwmmo 30H

 

mommucm>c¢

r-INM'Q‘

cacoupmsa

Hepomnuxo Hmumnom
Houomuuxo maammumoz
Houomuuxo Hm>mqma
Novomnuxo xmflcawflnpom

cesaoo Hansm
Houomucou
eflsanquwaswflq Hmnsmua
Houomwuxm
omHQImchmuom oaMumEESwa
Gazaoo cousmsmIosnmoHO
Houomucoo Oman mcflumuom
:EDHOO Hmnflogom
cesaoo oumamImcHumoonmfiooH
sasaoo pomadm

CEDHOO oumam cmumuomumm
cesaoo ooxomm
Gesaoo mmumm

 

mama

 

wm¢22Dm MOBO¢BZOU

NIN mqmfia

Houomupxm HomDMHHuch

Houomucou afisaou

conspflmfl maamoflcmnooz

Mononucou
GEDHOU pmumuammcn

 

Suommumo

7.
A.

mcooaouonuflcwnp
Emuomaoummo

.HMOHusoomenmsm

.mamofiawno Hmoo

wcmfiumonp
Houm3oumm3
Hosm Hmoaonc
mmHsHHmuofiouwmz
HMUHEopoouuom

 

mGOHHMOAHmm¢
HMHHumnwcH

mcflxflfixomn oEOm

moaflu
coaucwuoo mcoa

oommm Hooam mmHmH
mumoo Hmuflmmo saws

mumoo uozom amen
mofluoucm>cfl memH

 

mommpcm>owmfim

>wousw>cfl 30H
Ewummm
maammamadso

macaw: coo

demHmom oaflmaaou .

mommum mama

soawflonmmm swan

soon coon 30H
oaumw 3oam

oUH3 weapon: .

mcflwomucoo ooom .

I
MVLOKQ

I—IN

 

mommucc>w¢

 

wm¢ZSDm mOBU¢BZOO

AU.#COUV NIN mqm<9

Houomucou mfluuoz

uouomucoo gem

Ho>umz w mofiaom
fimhsq

meSoM

HSH
mmmuw3omlh>mm
mHHHz Hmwmcow
onsu usmsmnw
xflzlmesm
oamompcwz

 

mama

mHOUOM#GOO HOfiMO

HoauuomIHotz

 

muommumu

CHAPTER 3

SELECTION OF CONTACTING EQUIPMENT

The batch optimization studies proved the feasibility
of selectively extracting and recovering aluminum with
liquid—ion exchange. It was then necessary to decide which
type of contacting equipment would be best suited to carry
out the recovery of aluminum.

The contactors had to be flexible, easy to operate,
reliable, efficient, and relatively easy to scale-up.
Flexibility was needed because of the many uncertainties
which remained after the batch studies were completed. The
batch tests were performed on synthetic aluminum solutions
and the continuous flow studies would eventually evaluate
raw alum sludges. It was suspected that the two feed stocks
would have different operational characteristics and the
contacting equipment had to allow a wide range of flowrates
and phase ratios to be evaluated. Contactor performance is
often a function not only of Operating parameters but also
of the size of the contactor. To ensure that full-scale
performance could be reasonably predicted on the bench and
pilot scale, an easily scaled—up contactor was desirable.

A review of industrial contactors revealed that mixer—
settlers satisfied the requirements of the alum recovery

process. Also, the successful use of mixer-settlers in

43

44

other hydrometallurgical applications, particularly those
producing uranium and copper, confirmed the use of mixer-
settlers for aluminum recovery.

The major drawbacks of mixer-settlers, mentioned earlier,
were outweighed by the advantages. The corrections of the
deficiencies of mixer-settlers were seen as refinements to
equipment which had already been optimized.

The type of mixer—settlers chosen for the bench-scale
study were horizontal, box-type units using pump-mix turbine
impellers. The design is similar to the Davy-Powergas unit
shown on Figure 2-7. The decision to use this type of
mixer-settler was based on their simplicity and flexibility.

Later in the research it was decided to use an RTL

Contactor (Figure 2-9) for reasons to become apparent.

CHAPTER 4

PILOT-PLANT DESIGN

4.1 Introduction
The design of the Tampa pilot plant was divided into

three areas:

0 Pilot-scale design data development
0 Pilot program objectives and timetable

0 Proposed operation and equipment design

Following the continuous flow, bench scale studies in which
pilot-scale design data was generated, it was necessary to
establish objectives of the Tampa pilot program. Once these
objectives were defined, an operations program could be
developed to obtain those objectives. When it was known how
the plant was to be operated, the equipment to carry out

those operations could be designed.

4.2 Pilot-Scale Design Data Development
4.2.1 Introduction
The design of the Tampa pilot plant was based on nine
months of continuous flow studies on bench-scale mixer-settler

units. It was felt that data on the following parameters

45

46

would be necessary to have in order to properly scale-up the
process and the bench-scale program concentrated on optimizing

these parameters:

Extractant Strength
Stripping Acid Strength
Number of Extraction Stages

Number of Stripping Stages

UllnP-UJNH

Extractor Dispersion Conditions
Stripper Dispersion Conditions
Extractor Phase Ratio

Stripper Phase Ratio

\OGDQO

Extractor and Stripper Mixer Residence Times
10. Extractor and Stripper Mixer Impeller Speeds

11. Extractor and Stripper Settler Loading Rates

4.2.2 Extractant Strength

The extractant used, mono-di (2-ethylhexyl) phosphoric
acid (MDEHPA), has been shown to be effective in selectively
extracting aluminum from aqueous solutions.(3) That effective-
ness was again demonstrated in the bench-scale testing.

The concentration of extractant needed is a function of
the concentration of aluminum in the sludge to be treated.

3+

Based on an aluminum concentration of 1500 mg/l A1 in the

Tampa sludge sample treated in the bench-scale studies, an

47

extraction concentration of 0.77 M was found to be necessary

to effect the desired extraction.

4.2.3 Stripping Acid Strength

Batch experiments showed 6N H SO to be the most

2 4
effective stripping agent.(3) This solution was used success-

fully throughout the bench-scale program. On this basis,

6N H2804 was used as the stripping agent at the pilot scale.

4.2.4 Number of Extraction Stages

Two extraction stages were originally designed for the
bench-scale unit. This was based on a McCabe-Thiele analysis
of batch extraction isotherms.(3) The number of stages
necessary was eventually reduced to one. There are two
reasons why this reduction in stages was possible and

necessary:

1. The synthetic feed stock used in the batch and
bench-scale studies were fed to extraction circuit
at pH 2.0. It has been shown that the efficiency
of a stage can be increased by raising the pH of
the feed stock.(3,25) Midway in the bench-scale
program, the feed stock was switched to raw alum
sludge from Tampa. The pH of the sludge was 6.9.
This caused the stage efficiency to increase to
the point where only one stage was necessary.

2. It was necessary to reduce the number of extraction
stages to one because of a unique aspect of this
application of the liquid-ion exchange process.
Many extraction operations, i.e., copper extrac—
tion,(26) go to great lengths to eliminate solids
from the solute feed stream. The economics of the
alum recovery process are greatly enhanced if the
raw sludge can be fed directly to the extraction

48

circuit. After extraction of the aluminum from
the raw sludge, the dispersion flowing to the
settler contains loaded extractant, raffinate, and
solids which remain insoluble at the pH of the
aqueous phase in the mixer. The solids collect at
the settler interface and it would be a difficult,
if not impossible task to efficiently move these
solids on to the next extraction stage if more
than one stage were being employed.

4.2.5 Number of Stripping Stages

A McCabe-Thiele analysis of batch stripping isotherms
predicted that two stages would be necessary.(3) This was
confirmed by bench-scale operations and two strip stages

were designed into the pilot plant.

4.2.6 Extractor Dispersion Conditions

It was necessary to disperse the aqueous phase in the
extractor regardless of any consequences which may have
resulted from such a dispersion. When the aqueous phase was
continuous, a stable emulsion formed which would not dis—
engage by gravity means. This phenomena was observed in
batch testing when contacting the Tampa sludge with the
solvent.(3) Based on the problems associated with aqueous
continuous operation of the bench-scale extractor, the pilot

extractor was to be operated organic continous.

49

4.2.7 Stripper Dispersion Conditions

Minimizing organic contamination of the recovered alum
was the overriding consideration in deciding which phase to
disperse in the strippers. It has been found that organic
entrainment is lowest under organic continuous conditions.
It is believed that under organic continuous conditions the
entrainment of organic in the aqueous phase is a function of
phase disengagement and settler operation, rather than mixer

(28) Based on the confirmation of this in the

operation.
bench-scale studies, the mode of operation in the strippers

was to be organic continuous.

4.2.8 Extractor Phase Ratio

In order to minimize mixer volume and solvent pumping
rates, a low solvent to aqueous ratio was sought. A phase
ratio of 2:1 was chosen as the design value. While it may
have been advantageous to operate at a phase ratio lower
than 2:1 with respect to aqueous entrainment,(27) it was
found that operation at phase ratios less than 2:1 invited
phase inversions which created the stable emulsion. The 2:1
phase ratio was not necessary from.an extraction efficiency
standpoint. Therefore, to minimize pumping rates, the
stripped organic feed rate would be equal to raw sludge feed
rate. The phase ratio of the mixer would then be raised to
2:1 by recycling loading organic from the settler to the

mixer.

50

4.2.9 Stripper Phase Ratio

Batch tests indicated that a phase ratio of 3:1 was the
most efficient for stripping.(3) A 2:1 phase ratio was most
effective at the bench scale and this was the phase ratio on
which the design of the pilot strippers was based. Since
the organic flow entering the bench-scale stripping circuit
was typically 30 times higher than the acid flow, it was
necessary to recycle a portion of the settler aqueous

effluent to the mixer to obtain a 2:1 phase ratio.

4.2.10 Extractor and Stripper Mixer Residence Times

A mixer residence time of 15 minutes was found to be
optimum for both extracting and stripping.(3) This residence
time was successfully used throughout the bench program and

was chosen as the design residence time on the basis of that

success .

4.2.11 Extractor and Stripper Mixer Impeller Speeds

An impeller tip speed of 800 ft/minute was found to be
optimal in batch studies in both the extracting and stripping
(3)

processes. Used throughout the bench program without

adverse affects, this tip speed was used for the design of

the pilot mixers.

51

4.2.12 Extractor and Stripper Settler Loading Rates
Developing a basis for settler scale-up was most diffi-
cult. Settlers are normally designed on the relationship
between the throughput of dispersion per unit surface area
of the settler and the dispersion band thickness. However,
the presence of the solids emulsion at the settler interface,
prevented any measurement of dispersion band thickness.
Lacking traditional design parameters, it was decided to
design the pilot settler using the loading rates of other
settler designs. In particular, the scale-up of a successful
copper extraction bench unit used a settler loading rate of
2.0 gal/min-ft2.(6) On the basis of the success of this
plant, 2.0 gal/min-ft2 was used as the loading rate for the
design of the pilot extractor settler. This loading rate

was also used for the design of the stripper settlers.

4.3 Pilot Program Objectives and Timetable

The three main objectives of the Tampa facility were:

0 Process optimization
0 Demonstration of economic feasibility

0 Generation of full—scale design data

Initially, system performance would be evaluated at design
operating conditions. Once performance at these conditions

was established, any modifications needed to raise the level

 

52

of performance to design specifications would be made. The
number and severity of modifications needed would be a
measure of the success of the scale-up. If the important
parameters were identified and optimized at the bench scale
and sound methods of design were used, there should have
been little difference between the performance of the bench
and pilot units. Thus, an important part of process optimi-
zation would include scale-up evaluation and the identifi-
cation of any operating parameters important to scale-up
which were overlooked at the bench scale.

Once the system had been optimized, the recovery plant
would be operated to generate data related to the Operating
costs of the system. As operating cost data were developed,
the operation of the pilot plant would be refined so that
the most cost-effective method of operation could be found.

With the system economically optimized, data necessary
for the design of full-scale equipment would be generated.
with this data, the full-scale equipment could be sized and
capital costs determined.

The pilot plant was to operate for six months; the
first three months were to be devoted to making any modifica-
tions necessary to bring the process to design specifications
and to optimize operating parameters. During the last three
months, the system was to be operated at optimal conditions

in order to develop operating and full-scale design data.

53

4.4 Proposed Operation and Pilot Equipment Design
4.4.1 Proposed Operation
4.4.1.1 Introduction

The Tampa alum recovery facility was designed on the
assumption that the extraction and stripping circuits would
operate continously, treating 10 gpm of raw alum sludge.
The characteristics of the sludge as determined at the bench
scale included a solids concentration of 0.8 percent and an

aluminum concentration of 1500 mg/l a13+

at pH 2.0. While
variations in sludge quality were expected, these values
were chosen as representative for design purposes. How
these basic assumptions apply to the liquid-ion exchange

process will be detailed below.

4.4.1.2 Extraction Circuit

The single extraction stage would operate organic
continuous at a phase ratio of 2:1 with a detention time of
15 minutes. The feed flow would be 1:1, the additional
organic being recycled. The treatment of 10 gpm of sludge
would produce approximately 2 gpm of bleed solids. These
would collect at the extractor settler interface where a
siphoning manifold would remove and transport the solids to
the solvent recovery operation.

The raffinate would flow by gravity into the Tampa

Water Treatment Plant sludge handling system.

 

 

54

Feed sludge would be pumped from a mixed sludge holding
basin offering a representative sampling of sludge. Stripped
organic was to be pumped to the extraction circuit from an
organic reservoir.

Process control would be achieved by monitoring the

following parameters:

0 Solute and organic feed flows

0 Recycle flow

0 Mixer phase ratio

0 Raw sludge characteristics

0 Stripped and loaded solvent characteristics

0 Raffinate characteristics

4.4.1.3 Stripping Circuit

The two strip stages would operate organic continuous
at a phase ratio of 2:1. The detention time would be 15
minutes. The solvent to acid feed flow ratio would be 29:1,
requiring recycle of the aqueous phase within each stage.

The 6N H2504 used as the stripping agent would be mixed
in-line prior to being pumped into the stripping circuit.

The recovered alum would be passed through a granulated
activated carbon (GAC) column to remove any color or entrained

organic which may be present. The alum would then flow by

 

 

I
I

55

gravity to a day tank from which it would be pumped into the
water treatment plant's existing alum storage tanks.

The stripped solvent would flow by gravity from the
stripping circuit to the solvent reservoir.

Control of stripping would be maintained by monitoring:

0 Stripping agent flow

0 Stripping agent strength

0 Recycle flows

0 Mixer phase ratios

0 Solvent characteristics before and after each stage

0 Recovered alum characteristics

4.4.1.4 Solvent Recovery

Because of the large amount of solvent contained in the
bleed solids siphoned out of the extractor settler, economics
does not allow bleed solids disposal without recovery of the
solvent. The bench-scale studies had a great deal of success
recovering solvent with a solid bowl centrifuge. For this
reason a centrifuge would also be used for solvent recovery
at the pilot plant. To reduce maintenance and operating
costs the centrifuge would Operate only 8 hours per day.
During the other 16 hours the bleed solids would be stored

in a holding tank. The recovered solvent would be returned

56

to the extraction circuit via the organic recycle line,
while the residual waste solids would be disposed of with

the raffinate.

4.4.1.5 Personnel Requirements

It was felt that after any modifications were made and
the system was brought to steady-state, operation of the
system would require only one full—time person to operate
the centrifuge. During the other two shifts, an hourly or
bi-hourly inspection of the alum recovery system by water
treatment plant personnel would be sufficient for proper
operation.

In addition to the operation of the equipment, detailed
operations and chemical inventory logs would need to be

maintained.

4.4.2 Pilot Equipment Design
4.4.2.1 Extractor Mixer

Using a residence time of 15 minutes, a solute feed
flow of 10 gpm, and an organic to aqueous ratio of 2:1 (the
organic feed to be equal to the solute feed, a 2:1 phase
ratio to be achieved by recycling loaded organic), the pilot
mixer was designed by maintaining geometric similitude with
the bench mixer. The volume of the mixer is found by:

v = [10 + 2(10)]gpm(15 min) = 450 gals = 60.2 ft3.

57

Making the mixer cubic in shape yields the dimensions,

47" x 47" x 47", plus 7” freeboard.

4.4.2.2 Extractor Impeller

The turbine impeller chosen for the extractor mixer

(19)

consisted of six, flat, top-shrouded, radial vanes. The

length of the vanes was equal to the shroud radius. Entrain-
3D2

ment levels are lowest for turbine impellers when N 520,
where,

N = rotational speed Of impeller, rps

D = diameter of turbine, feet

Maintaining a tank width to impeller diameter of approximately
2:1, yields an impeller diameter equal to 28 inches. Solving

for the rotational speed of the impeller,

N3 = 20/1)2
N = 20/(2.33)2 1/3
N = 1.54 rps = 93 rpm

However, the tip speed for this impeller at 93 rpm is 680
ft/min. It was felt at this time that a trade-off could be
made between impeller speed and entrainment, so the extractor
impeller was designed at 680 ft/min instead of 800 ft/min.

Shown on Figure 4-1 is the impeller used in the extractor.

4.4.2.3 Extractor Settler
The function of a settler is to efficiently separate a

dispersion of two or more phases. In the alum recovery

58

28“

 

 

 

 

 

PLAN

STAINLESS STEEL

; COUPLING

 

 

 

3"
'H
==I

 

 

 

 

ELEVATION

FIGURE 4-1
EXTRACTOR MIXER IMPELLER

59

process the extractor settler must handle three phases: the
loaded extractant, raffinate, and solids which remain insoluble
during extraction. Due to the intensity of mixing developed
during extraction, a large amount of solvent and raffinate
become attached to the solids. This emulsion, termed bleed
solids, collects at the top of the raffinate in the settler,
but does not cross the interface into the solvent. The
Tampa sludge sample tested in the bench-scale unit exhibited
this type of behavior for the duration of the evaluation
period. The presence of the bleed solids undoubtedly had a
detrimental effect on phase separation but it did not hinder
settler operation to the point where coalescence failed to
occur. since the solids accumulated downward into the
raffinate, the volume of solvent in the settler never
decreased, and hence, the mean velocity of the solvent
through the settler never increased for a given solvent
flow. An increase in the mean velocity of the solvent may
have led to a washing out of dispersed raffinate and solids
into the stripping circuit. However, no contamination of
the recovered alum with raffinate or solids was observed.
While the bench settler was oversized to accommodate several
hours storage Of bleed solids, it was not over-designed to
the point where secondary dispersions of raffinate and

solids would have settled out. Based on the bench-scale

60

findings, the bleed solids could be tolerated if they were
removed periodically. The pilot settler was designed
accordingly.

Given the design sludge flow of 10 gpm and a design
mixer phase ratio of 2:1, the total flow of dispersion to

the settler is 30 gpm. The surface area of the settler is

then:
30 gpm = 2
2 m 15 ft
ft2
Using a length to width ratio of 2:1 yields settler area

dimensions of 32 3/4" x 65 3/4".(19)

The depth of the
settler was set at 40 inches plus 7 inches freeboard. This
was based on the maintenance of about one-foot layers of
raffinate, organic, and bleed solids. While much effort has
been directed at minimizing the size of settlers, there were
no restraints on the size of the units to be used at Tampa.
Thus, priority was given to ensuring efficient phase separa-
tion, not minimizing settler size.

A dispersion introduction baffle was installed to

(20) A proven aid in imprOVing

assist in phase separation.
settler performance, the introduction baffle covers the full
width of settler and extends downward to the level where
phase disengagement is taking place. Located five inches
from the mixer outlet, the baffle allows the dispersion to

be introduced to the dispersion band by providing a route

61

down through the bulk organic phase after leaving the mixer.
The dispersion introduction baffle eliminates turbulence and
re-entrainment of the phases in the Vicinity of the settler
inlet.

Another settling aid employed was a picket fence.
Effective in breaking up turbulence and distributing flow
evenly across the width of the settler, picket fences have
reduced entrainment by 50 percent in some installations.(19’20)
The "fence" was made of a row of vertical bars stretching
across the width of the settler. The space between bars
being 1/2 inch. This row was followed by another row of
bars located in such a way that the spaces of the first row
are covered by the bars of the second row. The gap between
the two rows of bars being 1/4 inch. The picket fence
extended to the bottom of the settler and was located five
inches downstream of the dispersion introduction baffle.
Shown on Figure 4-2 is the extractor mixer-settler used at

Tampa.

4.4.2.4 Extractor Settler Weir System

The weir system used to remove the separated phases is
shown on Figure 4-2. The fixed organic weir was located at
a level which allowed seven inches of freeboard in the
settler. The aqueous weir was movable, allowing the level

of the interface to be controlled. The use of full width

62

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ADJUSTABLE AQUEOUS WEIR
T. [ID‘
é— fie:
- l
MIXER SETTLER g _
' L
:Q‘ I!) q.
K)
II
IL ,rgz
RECYCLE SQKWATE
FLOWMETER a I
I-I/2" I.D. ORGANIC RECYCLE:
ELEVATION
47' g
j
I 65" 3/4" 4" 4n 5"
I-I/2" OUTLET TO
STRIPPERS\__..O :
1‘ 5L1 5“ C) 5g
s I-I/z" OUTLET To .
I: RECYCLE~\__,,,O 9|,
\ ‘3 4 x

 

 

 

 

 

 

 

 

 

 

KPICKET FENCE
ORGANIC WEIR

DISPERSION INTRODUCTION
BAFFLE AQUEOUS

WEIR
PLAN

FIGURE 4-2
EXTRACTOR MIXER-SETTLER

63

weirs is recommended because discharge to a pipe or reduced
width weir would cause the velocity of flow to that point to
increase and this can reduce the effectiveness of the

settler.(19’20)

To determine the size of pipe into which the weirs
discharged, the following equation was used:

H = 1.1 Qz/Azg

where H = depth of liquid in weir, ft
A = area of pipe, ft2
Q = flow, cfs

Different pipe areas were inserted into the equation until a
suitable depth was found. It should be noted that the use
of this equation assumes that the only force acting on the
liquid is gravity. In the case of this mixer-settler, the
pumping action of the impeller of the mixer to which the
orifice in question is connected to will also act on the
liquid. Hence, the result of this equation represents a
"worst case" situation and was used because it offers a
conservative answer.

In the case of the organic weir, it would have two
orifices, one leading to the strip circuit and the other to
the organic recycle line. Each orifice must pass 10 gpm.
Using a pipe with a l—l/Z-inch diameter resulted in an

acceptable liquid depth in the weir of 1.1 inches.

 

64

1.1(.02)2

_ = 0.09 ft = 1.1 inches
32.2(.Ol)

The orifice of the organic recycle line was flush with
the bottom of organic outlet weir, while the orifice of the
pipe leading to the strip circuit was raised three inches
above the bottom of the weir. This arrangement assured that
the organic recycle line would always be flooded so no air
could be entrained into the recycled organic. This entrained
air could have a detrimental effect on coalescence.

The velocity of flow in the organic recycle pipe would
be 1.6 ft/sec. The head loss through the recycle line would
be approximately 0.25 feet or three inches. While this loss
is equal to the head on the recycle line developed in the
weir, it was considered acceptable in light of the pumping
capacity of impeller.

A l-l/2-inch pipe was used in the aqueous weir; however,
its sizing is not as critical as that in the organic weir

because of the large amount of head available.

4.4.2.5 Solvent Feed Pump

The head requirements of the solvent pump were
negligible. Also, it was felt the pumping action of the
impeller would not have a detrimental effect on the pump. A
Cole-Parmer Instrument Company No. 7087-40 high capacity

centrifugal pump was chosen to feed the solvent. Flow would

65

be controlled by means of a valved recycle around the pump
and would be measured by a Fisher & Porter 10A2227A Series

all metal rotameter.

4.4.2.6 Sludge Feed Pump

The sludge feed pump would have to overcome 15 feet in
losses due to 300 feet of 1-1/2-inch plastic hose which
would serve as the sludge feed line and a five-foot suction
lift. To meet these demands, a Multi-Duti Manufacturing
Company No. 1131.31 centrifugal pump was selected. This
pump had a rating of 10 gpm at 80 feet of head. Sludge flow
was controlled by a valved recycle and measured by a Badger

Meter Recordall Model 15 water meter.

4.4.2.7 Extractor Organic Recycle
Organic recycle flow would be controlled by a gate
valve and measured with a Blue—White, Inc., 0-20 gpm

flowmeter.

4.4.2.7 Stripper Mixers

Two strip stages were required at Tampa, the dimensions
of each being identical. Using a residence time of 15
minutes, an organic flow of 10 gpm, and a phase ratio of
1:1, the pilot mixers were designed by maintaining geometric

similitude with the bench-scale mixers. Since the acid flow

66

to the strippers would be only 0.34 gpm, aqueous would have
to be recycled to achieve a 2:1 phase ratio. While the
strippers would be operated at 2:1, they were designed to

accommodate a 1:1 phase ratio. The mixer volume is found

by:
v = [10 + l(lO)]gpm(15 min) = 300 gals = 40.1 ft3
A cubic mixer would have the dimensions of 41" x 41” x 41",

plus 6" of freeboard.

4.4.2.9 Stripper Impellers

The type of impeller chosen for the strippers was the
same as for the extractor. The impeller diameter was set at
approximately one-half the tank width or 24 inches. The

rotational speed would found by keeping N3D2 $20.
2)1/3

Rotational speed of impeller, N (20/2

N 1.71 rps = 103 rpm

A rotational speed of 103 rpm yields a tip speed of 647
ft/min. As with the extractor, it was felt that some mixing
could be sacrificed in order to lower entrainment, so the
pilot mixers were designed at 647 ft/min instead of 800
ft/min. Shown on Figure 4-3 is the impeller installed in

the strippers.

67

24"

 

 

 

 

 

PLAN

 

 

STAINLESS STEEL
_ j COUPLING
'N
h
\»

 

ELEVATION

FIGURE 4-3
STRIPPER MIXER IMPELLER

A 4 fir—w H... Afl—_‘..——~“‘_,

68

4.4.2.10 Stripper Settlers

The stripper mixer-settlers used at Tampa are shown on
Figure 4-4. Unlike the extractor settler, the stripper
settlers had to separate only two phases, acid and solvent.
Some crud did accumulate at the Strip I settler interface in
the bench-scale unit. These solids consisted mainly of
Ca2504 and regular cleaning of the interface kept this crud
from interfering with coalescence. The interface of the
Strip II settler remained crud—free throughout bench-scale
evaluation of the Tampa sludge. It was not expected that
crud accumulation would cause any great problems at the
pilot scale and was not given any special consideration in
the design of the settlers.

Given the design solvent flow of 10 gpm and a phase
ratio of 1:1, the total flow of dispersion to the settler

was 20 gpm. The surface area of the settler is found by:

.35LJHEH = 10ft2
2 SEE
ft2
A length to width ratio of 2:1 yields settler area dimensions

of 27" x 53". The depth of the settlers were set at 24
inches plus 6 inches of freeboard. This was based on
one-foot layers of aqueous and organic phases.

The same settling aids used in the extractor settler
were used in the strip settlers. The dispersion introduction

was located four inches downstream of the mixer outlet. The

69

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4 “’I . .12- 3
K)
MIXER SETTLER I ‘0
V0
gt»
¢
9'.
I_.__. L
.IL.
.. l/2" I.D.-ACID
I-I/2 I.D.- TO NEXT
ORGANIC TO STAGE OR GAC
NEXT STAGE COLUMN
RECYCLE OR RESERVOIR
FLOWMETER
I- I/2" I.D.- AQUEOUS RECYCLE
ELEVATION
4|"
w 53" v 8" 4" ‘4? 5"
1 I
. O
= 4" 4" g
3 ° '5
I‘ \ O
K \ \ 5 I
KI=>ICI<ET FENCE
4’— SURFACE AREA
DISPERSION INTRODUCTION BAFFLE ADJUSTMENT BAFFLE
ORGANIC WEIR
ADJUSTABLE AQUEOUS WEIR
PLAN
FIGURE 4-4

STRIPPER MIXER-SETTLER

70

picket fence was located four inches downstream of the
introduction baffle. The space between vertical bars was
equal to 1/2 inch and the gap between the two rows of bar
was 1/4 inch.

Also added to each of the strip settlers was a baffle
which intercepted the dispersion band near the effluent end
of the settler. When the baffle was lowered, a 8.5 ft2
surface area could be evaluated. When raised, the original

10 ft2

surface area could be investigated.
4.4.2.11 Stripper Settler Weir System

The design of the stripper weirs followed that of the
extractor weirs. Since no organic is recycled in the
strippers, only one outlet is needed in the organic weirs
and 10 gpm must flow through that outlet. The organic weirs
were dimensioned as shown on Figure 4-4.

The aqueous weirs differed from those in the extractor
in that a portion of the aqueous flow would be recycled to
the mixer. The recycle flow is a function of feed flow
ratio and mixer phase ratio. Given an organic feed flow of
10 gpm, an aqueous feed flow of 0.34 gpm, and seeking a
mixer phase ratio of 2:1, it can be seen that 4.7 gpm of
aqueous must be recycled. Using the same analysis as in
4.4.2.3, a l-l/2-inch diameter orifice and pipe was found to

be suitable for transmitting the recycle flow. The velocity

71

in the recycle line would be 0.81 ft/sec with a head loss of
one inch. This loss is well within the pumping capacity of
the impeller. The other outlet in the aqueous weir is that
which passes the stripping agent to the next stage or out of
the stripping circuit. A 1/2 inch diameter outlet is more
than sufficient for this purpose. The aqueous weir arrange-
ment, shown on Figure 4-4, serves two purposes; first, it
means that at steady state, the recycle lines will always be
flooded, and second, it allows a large amount of head to
build up on the recycle line ensuring proper recycle

performance.

4.4.2.12 Stripping Acid Pump and Accessories

To introduce the stripping agent into the stripping
circuit, it would be necessary to meter, feed, and mix
concentrated sulfuric acid and water. The rates of feed of
the acid and water would be such that the mixture was a 6N
solution. The 6N acid demand is a function of the volume of
sludge entering the extraction circuit and the amount of
aluminum in that sludge. At the stated design conditions,
the acid demand for the production of alum containing 44,000

mg/l A13+

would be 0.34 gpm or 1.29 l/min.
The acid metering pumps chosen were Fluid Metering,
Inc., Models RPID-CP and RPIG400-CP. It was known that the

characteristics of the sludge would vary and these pumps

72

provided a wide range of acid feed rates. The concentrated
acid demand at design specifications would be one-sixth of
1.29 l/min or 0.22 l/min. The source of the concentrated
acid would be a 10,000 gallon elevated storage tank located
at the recovery plant site.

The potable water feed would originate from a faucet
adjacent to the acid storage tank. The design potable water
demand would be five times the acid demand or 1.10 l/min. A
gate valve would be used to control water flow and a Badger
Meter Recordall Model 25 water meter would measure the flow.

Figure 4-5 shows the acid-water feed system. The two
liquids would be fed into a one inch diameter Pyrex standpipe
connected to the Strip II inlet structure. The standpipe
rose above the liquid level in the mixing tank to prevent
spillage in the event of mixer failure. It was felt that
enough turbulence would be created by the liquids as they
hit the liquid surface in the standpipe to provide adequate
mixing.

The mixing of acid and water demanded a Pyrex standpipe
because it could withstand the heat generated mixing and
then dissipate some of that heat before the diluted acid

reached the PVC portion of the acid feed system.

4.4.2.13 Stripper Aqueous Recycle
Aqueous recycle flows would be adjusted by gate valves

and measured with Blue White, Inc., 0-20 gpm flowmeters.

73

H “=th
20K“.

0.25:5

2MBmMm Qmmm DHUQ DZHmmHmBm m0 UHBdEmmUm

w40>owm
mDOwDO<

_-

mlv mmDOHm

 

 

_:

 

mmx...‘ fl Emhm

 

IIOII

_ $52 52;

II

UH

tuba; w4m<hom

 

 

r0

 

 

 

 

Dun

 

 

 

wo<m0hm o_o<

74

4.4.2.14 Recovered Alum Treatment

Bench-scale studies indicate some of the color in the
sludge is carried into the stripping circuit by the loaded
organic and stripped out along with the aluminum. Also,
organic is entrained in the recovered alum during stripping.
To remove these contaminants, a granulated activated carbon
column was placed in line between Strip I and the recovered
alum day tank. Laboratory studies showed a large reduction
in recovered alum TOC after contact with Filtersorb 400, a
product of the Calgon Corporation.

The column was a simple design consisting of a
four-foot section of 12-inch diameter PVC pipe and filled
three feet of GAC. The recovered alum flowed by gravity
from Strip I to the GAC column. The alum left the column by
means of a screened manifold located at the bottom of the
column. After GAC treatment, the alum would then flow by
gravity to the day tank. Under design conditions, flow
through the GAC column would be 490 gal/day.

4.4.2.15 Solvent Recovery

The solvent recovery operation was designed on the
assumptions that 2 gpm of bleed solids would be produced by
the extraction circuit and that the pilot centrifuge would

perform as efficiently as the bench centrifuge. With the

75

extraction circuit operating continuously and the centrifuge
operating 8 hr/day, the pilot centrifuge would have to
process 6 gpm of bleed solids.

The unit chosen was a Sharples P660 Scroll centrifuge.
This solid bowl device would be mounted on a platform under
which a wheelbarrow could be parked into which the solids
cake would fall. These solids would then be removed and
studied to determine their drying characteristics. The
recovered solvent and raffinate would be discharged at the
liquid end of the bowl and would be returned to the extrac—
tion circuit via the organic recycle line. It was felt that
the return of these liquids to the extraction circuit would
not hinder extractor performance.

A siphoning manifold, situated appropriately in the
extractor settler, would withdraw the bleed solids. With-
drawn continuously and discharged to a small holding tank,
the bleed solids would be pumped to a larger holding tank
when the centrifuge was not in operation or directly to the
centrifuge when it was operating.

The pump chosen to perform the transfer of the bleed
solids was a Robbins and Meyer Moyno pump. This progressing
cavity pump, equipped with a variable speed motor was piped
in such a way that bleed solids could be pumped from the
small storage tank to the large storage tank, from the small
tank to the centrifuge, or from the large tank to the

centrifuge.

76

4.4.2.16 Tank Design

The design of the various holding tanks and reservoirs
mentioned in the previous sections will now be discussed.
These include an overnight solids holding tank, a short-term
solids holding tank, a solvent reservoir, a recovered alum
holding tank, and a spill and utility tank. Shown on
Figure 4-6 is a layout of the alum recovery plant at Tampa.

The overnight solids holding tank was designed to store
24 hours worth of bleed solids or 2880 gallons. The tank
chosen was a vertical cylinder with an eight-foot diameter
and a 3000 gallon capacity.

The short-term solids holding tank was arbitrarily
sized at 300 gallons. Its major purpose was to provide a
mixing chamber where future solids neutralization studies
could take place.

The solvent reservoir was a 1000 gallon horizontal
tank. This sizing is based on the assumption that the bleed
solids would contain 20 percent solvent by volume. Thus,
when the large solids tank was full, it would contain 600
gallons of solvent. This would leave 400 gallons of solvent
in the reservoir, which was enough to compensate for any
changes in phase ratios or interface levels within the
extracting or stripping circuits.

The recovered alum day tank was to have a volume of 750

gallons, providing 1-1/2 days of alum storage. To transfer

77

BZflQm BOQHQ 45min. m0 UHEgmmUm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mus mmeHm
Inns. mozzom
wh<ziu<z v.24? >40
E53 “:2; I--- H.826 34: .222 8568.1
hzw>40m ouzw>oomz .lllflu
838 83a -II
23¢ 3568:
SE; 32%“ III
a ..... I. . .
0.33:6 Dwaizpm III! I m4<w COD 0.
22(95 awe}: IIII ”
won—Dam 35.. IIIII " wwdmobrm
2238
.35 coo. 98
£953: pzwiom .......... . ......... c ............... gig
. I i. l
.IIIII.I.I.I.I.I.I.ImEo<Exm 52:5 _
I flql+lod -
— w !
9135:2525 QT “
= .
- a - . _ .55 II
= .I

  

.mIEG DOOM

ww<ZOhw
mo..5m waI—m

    

AQI
I.

 
    

 

 

 

 

 

 

 

 

IIIVIIL

 

4150de w mowm
02.55 oz<m Oh

78

the alum to the water treatment plant's storage tanks, a
Corcoran Model 2000E centrifugal was specified. This pump
had a rating of 20 gpm at 60 feet of head. The Corcoran
pump would also be used for utility purposes.

The spill and utility tank would have a capacity of 300
gallons and be used in clean-up and maintenance operations.

All of the above tanks were made of fiberglass.

CHAPTER 5

RESULTS OF MIXER-SETTLER OPERATION

5.1 Introduction
Construction of the pilot plant was completed on
February 21, 1979. Operation began on February 27, 1979,
and was scheduled to continue through August, 1979. However,
equipment breakdowns and other delays caused operation of
the mixer-settlers to continue through September, 1979. The
pilot plant was in operation for a total of 479 hours and
processed 141,900 gallons of sludge. Sludge flows varied
from 1 to 6 gpm and sludge aluminum concentrations ranged

from 80 to 1180 mg/l A13+

at pH 2.0. The low capacity of
the centrifuge prevented the 10 gpm design sludge flow from
being evaluated. Figure 5-1 presents sludge flow, aluminum
extracted, and raffinate pH as a function of time.

Upon start-up of the pilot unit, it was observed that
the solids in the extractor settler were behaving differently
than those in bench study. The pilot-plant solids situated
themselves on the organic side of the interface, and as they
accumulated their mass extended up into the organic portion
of the settler. Initially, it was felt that the impeller
speed might be responsible for this unexpected behavior.

Various impeller speeds were evaluated, but the solids

continued to accumulate in the organic phase.

79

val-BOW ‘lo SI
'IDb RSI

vdI-Iaow °/.I I
105 022
val-{sow °/.S'L
155 229
VdHEOW °/. 91
105 OII
Veil-{30W 7.97.
'I06 OvI

Veil-BOW °/oOI
'IDb 9v

vdH30W°/
VdI-BOW °/. 65 ‘

 
 

SOLVENT MAKE

80

 

11F

 

ISO 200 250 300 350 400 450 500

TIME OF OPERATION-(HOURS)

I00

 

441441 |o

 

BIVN IddVEd

(JLI /'5CII)
QELDVELXB
WFINIINI‘I'IV

O

0 SD" NC)

(mdb)
033:! 390018

FIGURE 5-l
PILOT PLANT OPERATING DATA

81

The first week of operation was thus spent varying
operating parameters in an effort to control the bleed
solids in the extractor settler. However, it appeared that
the unusual solids behavior was not related to any operating
parameter. To determine if some aspect of the sludge had
changed with respect to the alum recovery process, a sample
was shipped to the bench-scale laboratory to be treated
under the conditions determined optimum at the bench scale.
The results of the bench test showed the solids to behave
exactly as they had at the pilot scale. So, it appeared
that some characteristic of the sludge had changed and the
new behavior of the solids had to be dealt with.

Another problem which quickly arose was the performance
of the centrifuge. When first operated, it could only
process a fraction of the 6 gpm design value. Also, solvent
recovery efficiency was poor. Since efficient solvent
recovery was the key to cost-effective operation of the
plant, optimization of the solvent recovery operation was
given priority over other operations. Because of the low
centrifuge throughput, the extraction and stripping circuits
could not be operated continuously. Bleed solids production
far exceeded the capacity of the centrifuge and when the
solids storage capacity of the solvent recovery system was
reached, the extractor and strippers had to be shut down

until the stored solids were processed. The lengths of the

82

mixer-settler runs were very short, varying from 4 to 15
hours, depending on the raw sludge feed rate and the solids
concentration of the sludge. This discontinuous type of
operation meant that steady-state conditions in the mixer-
settlers could not be attained for any useful length of time
and without steady-state conditions, the evaluation of the
mixer-settlers as proposed would be difficult if not
impossible. It was under these conditions that operating
data were collected. The presentation and discussion of

operating results were divided into the following areas:

0 Extraction circuit performance
0 Entrainment and stripping circuit performance
0 Solvent stability

0 Personnel requirements

5.2 Extraction Circuit Performance
5.2.1 Extractor Mixer
5.2.1.1 Hydraulic Performance
The hydraulic performance of the extractor mixer was
inconsistent. Phase inversions occurred at phase ratios
ranging from 5:4:1 to 1:9:1. Mixer phase continuity was
generally stable at these phase ratios during bench-scale

tests. Also, variations in the pilot mixer phase ratio were

83

observed under constant mixer conditions. Both of these
phenomena indicate a lack of agitation in the mixer.
Variations in local phase ratios, a result of poor mixing,
can lead to local phase inversions. These local inversions
could then spread to the rest of the mixer. Therefore, the
inversion of the whole mixer was a function of impeller
rotational speed and not overall mixer phase ratio.

A phase inversion to aqueous continuous operation was a
major upset to both extracting and stripping circuits. The
stable emulsion which formed during aqueous continuous
operation quickly filled the extractor settler and if allowed
to pass into the stripping circuit would cause the stripper
mixers to invert to aqueous continuous and foul the interfaces
of the stripper settlers. To remedy the situation, the
whole system had to be shut down to clean the extractor and
stripper settlers and invert all mixers to the desired
continuous phase. Because aqueous continuous operation in
the extractor mixer upset the entire system, the extractor
was generally run at high phase ratios to avoid phase
inversions.

A simple remedy to the mixing problem would have been
to increase the impeller speed. However, a mechanical flaw
in the mixer drive train prevented the impeller from being

rotated at the design speed. Because sufficient agitation

84

could not be provided in the mixer to ensure phase continuity
stability at phase ratios lower than 3:1, the design phase
ratio could not be evaluated.

Since the design mixer speed could not be attained, it
was not possible to evaluate the method in which the mixer
was scaled-up. While the successful use of N3D2:520 as a
basis for mixer scale-up has been reported,(19) others
suggest that the safest method of scale-up is based on
maintaining geometric similitude and constant power input

per unit volume.(28)

5.2.1.2 Aluminum Extraction
5.2.1.2.1 Introduction

It was originally planned to operate the extraction
circuit in such a manner that flow rates and phase ratios
could be systematically evaluated to determine the optimum
values of these parameters. However, the poor performance
of the centrifuge and the limitations imposed by poor mixing
conditions did not allow the proposed evaluation to be
conducted. The steady-state conditions necessary to evaluate
these parameters never existed and it was in this context
that the data from mixer-settler operation, presented in

Appendix A, were analyzed.

85

5.2.1.2.2 Aluminum Extraction Efficiency

The extraction efficiency of the extractor was generally
high, 85 percent or more, which is surprising in light of
the poor mixing conditions which existed.

It is widely accepted that the bulk mechanism of mass
transfer in liquid-liquid extraction mixers involves two
zones. The first is in the immediate vicinity of the
impeller. Here, dispersed drops are sheared into smaller
droplets, passing through film and filament stages. The
best conditions for mass transfer exist in this zone. The
residence time of the droplets in this zone, however, is
very short and equilibrium is usually not reached in one
pass. The other mass transfer zone includes the rest of the
mixer where, if it is being operated properly, mildly
turbulent conditions exist. In this zone, a slower mass
transfer rate is generated by droplet coalescence/
redispersion.(21)

Inadequate mixing can affect mass transfer in three

ways:

1. By reducing the volume of the zone of high mass
transfer rates in the vicinity of the impeller.

2. By reducing the number of times a given drop will
pass through the zone with higher mass transfer
rates.

3. By reducing the amount of coalescence/redispersion

in the more tranquil zones of the mixer.
It is likely that all of these were taking place in the

pilot mixer.

86

The effect of the poor mixing on mass transfer rates,
however, was balanced by mixer residence times which were
usually far in excess of the necessary 15 minutes and by
high phase ratios. Thus, the extended mixing periods and
increased concentration gradients allowed the reduced mass
transfer rates sufficient time to approach equilibrium.

The effect that the recycling of organic had on extrac-
tion cannot be determined, but other research in this area
suggests that the recycling of the continuous phase has both
beneficial and detrimental effects. It is beneficial because
the recycling of either phase will enhance extraction. It
is detrimental because the addition of more of the continuous
phase will dilute the dispersion and decrease the amount of
coalescence/redispersion. Experimental results suggest that
these effects balance each other and have no net effect on

extraction.(28)

If it is assumed that these findings are
applicable to the alum recovery system, then it is probable
that the recycled organic had little effect on extraction
efficiency.

Assuming that the organic recycle has little effect on
efficiency, it is possible that there exists an optimum feed
phase ratio within the limits imposed by mixing conditions.
If such a phase ratio does exist, it is not evident from the

operating data. The apparent disorder of the data can be

attributed to:

1. The lack of steady-state conditions.

2. The extending of residence times far past that
required to reach equilibrium.

3. Varying residence times which did not allow
isolation of the feed flow ratio.

Given a mixer that is operating with an optimum impeller
speed, a constant residence time, and organic continuous
conditions, an optimum feed flow ratio would be one which
was as low as possible and still be stable with respect to

the desired dispersion conditions.(28)

As the phase ratio
increases, the dispersed phase is diluted, decreasing
droplet interaction and decreasing mass transfer rates.

The pH of the raffinate was used as a relative measure
of extraction. This was based on bench-scale results which
indicated that changes in raffinate pH closely reflected
changes in extraction efficiency. This relationship can be
explained by the mechanism by which aluminum is extracted.
The extraction reaction involves the exchange of aluminum
from the aqueous to organic for protons from the organic to
aqueous. Thus, as more aluminum is extracted, extraction
efficiency increases, and the raffinate pH is depressed. A
plot of extraction coefficients as a function of raffinate
pH is shown on Figure 5-2.

The extraction coefficient, E2, is defined as the ratio

of the metal concentration in the organic extract to the

metal concentration in the aqueous raffinate and is a measure

mm MBQZHmm/mm m0 ZOHBUZDm 4 mmm BZm—HUHmMMOU ZOHBUIRMEXH

 

mum mmDon
In
3 3 ca ..~ 3 . 3 RN
C I I. I d d d
m. G B
. a a
O a

88

525.225
Guzmofizoo wz< mm}; zogzwhmo

<2. 8.... on:
to... 8... 2L
.\s=oo~-omo
<9: com- com
to... 26. or...
.xszoo¢-oo~
:9: 65-8

wees—4m >3: 2. 232.2244

8860068

I

 

O.

ON

On

0?

On

3

89

of the amount of metal extracted for a given solute metal
concentration. As can be seen on Figure 5-2, there is no
evident relationship between extraction coefficients for a
given group of like solute concentrations and raffinate pH.
The apparent lack of order of the data plotted on Figure 5-2
can be explained by the following:

1. The lack of steady-state conditions did not allow
the isolation of the parameters in question.

2. Most of the data points on Figure 5-2 lie between
pH 2.3 and 1.8. Batch studies under controlled
conditions found that a drop in pH from 2.3 to 1.8
of the Tampa sludge represented a 4 percent increase
in the aluminum available for extraction. Given
the conditions under which the pilot plant operated,
it is unlikely that such a small change could have
been identified.

3. While MDEHPA is highly selective for aluminum, the
presence of extractant in concentrations beyond
that necessary to extract a given amount of aluminum
will encourage the extraction of other cations.
Since the aluminum concentration in the Tampa
sludge fluctuated greatly from day to day, it was
necessary to maintain a level of extractant in the
solvent which could accommodate a wide range of
sludge aluminum concentrations. This meant that
when sludge aluminum concentrations were low,
there was excess extractant present. Once
equilibrium was reached with respect to aluminum
exchange, another cation, calcium, which was often
present in significant concentrations, would be
extracted. This would cause the pH of the
raffinate to be lowered without an accompanying
increase in aluminum extraction and could yield
misleading results.

The raffinate pH can be a useful tool in monitoring
extraction efficiency, especially from an operational point

of view, if the factors which alter this relationship are

90

understood. The most reliable measure of extraction, however,
is to directly measure the aluminum in the aqueous phase
before it enters the extractor and after it leaves the
extractor.

Other factors which may have effected extraction were
solvent stability, to be discussed later in this chapter,
and alkalinity and solids in the sludge.

The presence of alkalinity in the sludge could affect
extraction by buffering the aluminum exchange reaction.
Extraction coefficients could be lowered by exchanged protons
being consumed by alkalinity instead of depressing sludge
pH.

Investigating the effect that the solids in the sludge
had on extraction was beyond the scope of this study; however,
it is a variable which cannot be ignored. The solids can be
divided into two groups, those which are solubilized during
extraction such as, the alum floc and color, and those which
remain insoluble, such as silt and turbidity. Little litera-
ture is available on this subject as most operations find it
cost-effective to remove all solids from the solute feed

stream.

5.2.2 Extractor Settler

The performance of the extractor settler was very good
when the adverse conditions under which it operated are
considered and the scale-up of this unit must be considered

a SUCCESS .

91

The two settling aids employed, the dispersion introduc-
tion baffle and the picket fence were most useful. The
settler was operated for a short time without them and there
was a great amount of turbulence in the settler during that
period. It was felt before start—up that the solids might
clog the picket fence, but operation of the settler proved
this fear unfounded as the solids passed easily through the
slots of the fence.

It has been suggested that the use of horizontal plates,
meshes, or packing can increase the efficiency of a settler.
It is unlikely that any of these could be used in alum
recovery because the bleed solids would quickly foul this
type of settler aid. However, the use of a packing such as
Knit Mesh,(6) made of a material wetted by the raffinate and
solids, and placed directly in front of the organic weir
might be able to reduce the amount of impurities going into

the stripping circuit.

5.3 Entrainment and Stripping Circuit Performance
5.3.1 Introduction

Entrainment between stages can result in reduced effi-
ciency, while entrainment into effluent streams can mean
solvent losses and product contamination. Thus, it is
imperative to minimize entrainment to ensure efficient.

mixer-settler operation.

92

Bench-scale studies did not reveal any entrainment
problems in the extraction circuit and none were expected at
the pilot scale. Organic entrainment losses into the
raffinate were assumed to be small when compared to losses
due to bleed solid processing and were given no consideration
during the pilot study. Raffinate entrainment into the
extract, however, turned out to be a major problem which

affected stripper performance and recovered alum quality.

5.3.2 Entrainment Symptoms

Two problems which arose during mixer-settler operation
were the solids which collected at the Strip I settler
interface and the recovered alum aluminum concentration.

After start-up, it was quickly evident that a large
amount of solids, identical to the bleed solids, were
accumulating at the Strip I settler interface. Two to five
gallons of these solids were scooped out of the settler
daily with a strainer. Occasionally, some of the solids
would be washed out with the organic into Strip II. More
importantly, it is probable that the presence of the solids
interfered with coalescence in the settler. Obviously, the
solids were being carried from the extractor to the stripping
circuit by the loaded organic.

Figure 5-3 presents a plot of stripping acid concentra-

tions as a function of operating time. Throughout.1nost of

93

OOm

MZHB UZHBflmmmO m0 ZOHBUZDm
4 mfl mZOHfiflmBZmOZOU SDZHZDA< QH04 UZHmmHmBm
mlm WMDOHM

AmmDOIV I MS; OZF<KMQO

 

 

g,

Om e OO? Omm OOn Om N OON Om. OO. Om O
I — d - G I I C I I d
.H .H
o
“m m
r. .2
a. .m. \
u n 55.2 3568: \I\
_ . 93 .— EEBIII \(
u n \
_ . \I\
1 _ .
_ \
_ x
. . \
_ . x
x _ _ \
a _ . \a \
. _ \ <2
f. ., . \I\ ‘
/_ 2 n .x. \
.1 z 27 \ .
l g u /\

WONIWI‘I'IV OIOV 9N IddItIlS

(1,. Iv I/fiw)‘- SNOIlvalNBONOO

 

94

the study, the Strip II acid contained more aluminum than

the recovered alum. Normally, the alum could contain several
thousand more milligrams per liter of aluminum than the

Strip II acid. Even if the stripping efficiency of Strip I
was zero, the alum would still have the same aluminum concen-
tration as the Strip II acid. The alum was being diluted
and the only influent stream to Strip I which could supply
the aqueous solution for dilution was the loaded organic

from the extractor.

5.3.3 Causes of Entrainment

Initially, it was felt that the solids and aqueous
carry over into the stripping circuit were caused by the
high scour velocities developed in the organic portion of
the extractor settler when bleed solids were allowed to
build up in the settler. A visual inspection of the settler
with approximately six inches of free organic revealed
locally high velocities and the movement of fine solids
particles within the bulk organic phase. Also, the loaded
solvent was cloudy, indicating aqueous entrainment.

In an attempt to eliminate the solids and raffinate
carry over into the stripping circuit, a deeper organic
layer was maintained in the extractor settler. This reduced

the high local velocities and bottom scour, but the loaded

95

organic remained cloudy. The deeper organic layer was
maintained from hour 310 to hour 380. While the amount of
solids that collected in Strip I decreased, it would appear
from Figure 5-3 that it only aggrevated the alum dilution
problem.

Entrainment in a system with a well designed settler is
a function of phase continuity, mixer speed, and phase
ratio. The phase continuity and mixer speed of the pilot-
plant extractor were considered to be fixed and it was the
phase ratio which was investigated with respect to the
aqueous entrainment.

Other research indicates that for extractors operating
organic continuous, the level of aqueous entrainment increases

with increasing phase ratios.(28)

The proposed mechanism
for this relationship states that for a given set of mixer
conditions, there is a "preferred" phase ratio at which
entrainment will be minimal. In the case of organic
continuous operation at phase ratios above the "preferred"
phase ratio, there is an excess of organic over that required
to form the "preferred" ratio. This excess organic "flashes
off" when leaving the mixer and when entering the settler
immediately joins the bulk continuous phase, carrying with
it small dispersed aqueous droplets.

It can be seen from the stripping circuit operating

data, presented in Appendix A, that from hour 220 to hour

330, the extractor was operated at generally high phase

96

ratios. As a result, the difference in stripping acid
aluminum concentrations increased because of the increase in
aqueous entrainment.

From hours 370 to 435, the extractor was operated at an
average phase ratio of 3:1. The response of the aluminum
concentration in the alum was both immediate and positive.
Within 30 hours the aluminum concentration rose from 27,000

3+ 3+

mg/l A1 to 39,000 mg/l A1 By hour 426, the alum strength

had surpassed that of the Strip II acid and by hour 432 had

reached 48,000 mg/l A13+.

Also, slightly fewer solids
collected at the Strip I settler interface.

The portion of Figure 5-3 following hour 435 represents
a period of operation in which priority was given to the
extraction circuit and stripping circuit parameters were not
monitored closely.

The effect that the extractor mixer phase had on the
recovered alum aluminum concentration is shown on Figure 5-4.
The higher phase ratios, producing higher levels of entrain-
ment and dilution caused the alum aluminum concentration to
decrease. At phase ratios between 6:1 and 7:1, the amount
of dilution was equal to the amount of aluminum coming from
Strip II. At phase ratios below 6:1, the alum generally
increased in aluminum concentration.

Figure 5-5 presents a plot of aqueous entrainment as a

function of extractor mixer phase ratios. Two of the data

2
:3
E
:E1:‘
Di
in
U’
25
as
<9
F.
3%
5,”-
<2
mm
0:0
U2
20
-0
21‘
373::
—3
Es
.1‘.’
<2
>28
:3
4E
(E
ZUJ
"2’
E30
(U
LIJ
Q:
0
Lu
c:

400

300

200

I00

-I00

-200

U

V

 

 

 

    
 

m_______________________.______

 

 

P 0
d
(9
0
XTRACTOR MIXER PHASE RATIO
II 2‘4 I til 54 07H 8'=l %" IOTI

 

 

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
3;
|
I

___ao .‘

PHASE RATIOS AT
WHICH PHASE
INVERSIONS OCCURRED

FIGURE 5-4
EFFECT OF EXTRACTOR MIXER PHASE RATIO ON
RECOVERED ALUM ALUMINUM CONCENTRATION

98

OHB4m mm4mm mmxHE mOBU4mem m0 ZOHBUZDm 4 m4 BZMEZH4mBZm mDOMDO4

 

mIm mmDOHm
OF<¢ wm<Id mute—2 mOhO<mhxw
z . .o. .5 E. E. .6 B. Z. .G .a Izo
q u q . u a q . q C .

OOO.

_l

4 OOON

L

OOOm

OOOV

J

 

owmaw<w2 O
Dmh<4304<o O

. OOOm

 

(wdd)-.LNEIWNIva1N3 SOOBI‘IOV

99

points on this plot were measured, while the remainder were
calculated. Again, the relationship between phase ratio and
aqueous entrainment is clearly evident. The method of
entrainment measurement and a sample entrainment calculation

are provided in Appendix C.

5.3.4 Other Aspects of Stripping Circuit Performance
5.3.4.1 Introduction

Time and personnel limitations did not allow the
strippers to be operated and evaluated as originally planned,
but several general comments can be made with respect to

their performance.

5.3.4.2 Stripper Mixers

As with the extractor mixer, the design mixer speed did
not provide enough agitation and the validity of the use of
N3D2:520 as mixer design criteria, at least in this case,
should be questioned. A mixer speed of 140 rpm, equivalent
to a tip speed of 850 feet per minute, was necessary to

provide stable phase continuity conditions at phase ratios

less than 2:1.

5.3.4.3 Stripper Settlers
In contrast to the mixers, the stripper settlers were

properly designed as dispersions bands varied in thickness

100

from one to two inches for the flow rates and phase ratios
at which the settlers Operated.

In addition to the bleed solids which washed out of the
extractor settler, the Strip I settler also had to deal with
the CaZSO4 which precipitated in the Strip I mixer. The
source of the calcium was the raw sludge. The calcium was
extracted along with the aluminum and was also carried into
the stripping circuit by the entrained raffinate. What
effect the calcium had on stripping efficiency is unknown,
but the presence of CaZSO4 at the settler interface
definitely interfered with coalescence.

The CaZSO4 problem can be eliminated by the use of HCl
combined with H2804 as the stripping agent. This would
reduce the SO4 concentration and thus, reduce Ca2504 produc-
tion. Hydrochloric acid costs more than sulfuric, but the
savings in maintenance costs may well offset this expense.
Just as at the bench scale, all Ca SO4 precipitated in

2
Strip I; Strip II remained free of this contaminant.

5.3.4.4 Entrainment

No entrainment measurements were made, but a study on
entrainment in strippers operating organic continuous
indicates that in the range of phase ratios at which the

pilot strippers operated, organic entrainment is very low

101

and independent of phase ratio. However, aqueous entrain-
ment under these conditions was high.(28) Precisely how
much acid was entrained cannot be determined, but its
presence was confirmed by the cloudiness of the stripped
organic. High levels of stripping acid entrainment can
reduce aluminum extraction efficiency by lowering the pH of
the feed sludge and increasing the aluminum content of the

sludge.

5.4 Solvent Stability
5.4.1 Introduction
Bench-scale studies with the mixer-settlers indicated
that under constant operating conditions the extracting
ability of a given solvent inventory increased when fresh
extractant was added to the system. One of the goals of the
pilot program was to investigate and quantify the degradation

of extractant quality.

5.4.2 Pilot Studies

The extractant at Tampa behaved much like it did at the
bench scale. As can be seen on Figure 5-1, after an addition
of fresh extractant to the system, the pH of the raffinate
dropped. However, it was not possible to evaluate extractant
performance over a long period of time because make-up

solvent additions were made so frequently that the extractant

102

was never allowed to stabilize. As shown on Figure 5-1,
make-up solvent had to be added approximately every 50
hours. This frequency reflects the poor performance of the

solvent recovery operation.

5.4.3 Tampa Bench-Scale Studies

Since it was not possible to evaluate solvent stability
with the pilot-scale mixer-settlers, a 4-inch diameter by
36-inch long RTL Contactor was set up to investigate the
behavior of the extractant. Starting with a five-gallon
inventory of fresh 7.5 percent (V/V) MDEHPA in Kermac 627,
the solvent was contacted 23 times with Tampa sludge
containing approximately 1000 mg/l A13+. Each time the
entire inventory had been contacted with the sludge, it was
stripped with 6N H

SO4 in a batch stripping operation.

2
Throughout this evaluation operating conditions were held

constant at the values listed below:

Rotor speed = 6 rpm
Organic feed = 20 ml/min
Sludge feed 2 20 ml/min

After the solvent was stripped, it was contacted with a

synthetic solution containing 1000 mg/l A13+. The aluminum

concentration of the synthetic raffinate was then measured

103

and the results are plotted on Figure 5-6. It is evident
from this plot that there is a dramatic decrease in extract-
ing ability after only three contacts with the sludge. By
the time the solvent has been loaded 13 times, the extrac-
tion coefficient decreased from 18.7 to 0.25, a 99 percent
decrease.

Recent work by Cornwell,(25) has provided some insight
into the problem of solvent degradation. An analysis of
extractant that had been contacted with sludge many times
revealed that the "mono" portion of the extractant was gone.
Titrations of the extractant with a basic solution confirm
these findings and explain the observed reduction in extract-
ing ability. Fresh extractant contains approximately equal
molar quantities of mono(2-ethylhexyl) phosphoric acid and
di(2—ethylhexyl) phosphoric acid. The mechanism by which
the "mono" portion of the extractant is removed or neutralized

is not yet known. Three possible explanations are listed

below:
1. The "mono" portion of the extractant is soluble in
the raffinate.
2. The "mono" portion of the extractant is neutralized
by a reaction with impurities in the sludge.
3. The "mono" portion of the sludge precipitates out

of solution with the calcium in the sludge.

104

OZHD4OQ UHZ4UmO m0 ZOHBUZDm 4 m4 BZWHUHMMMOU ZOHEU4MBXM
mlm MMDOHW

02540.. 22.335 .10 mmZ; “.0 mmmzsz
mu 8; m. .. o. m

 

 

 

 

o 3
1 D O
. 4.
.1 O
. H
.6
. s
. v
. N
. O
.o. W
. O
1
L V
. m
.0. W
. M
. fl
. W
Lo~S
I O
I q...
. fl
. .......
Rm
.4

 

105

5.5 Personnel Requirements
The operation of the mixer-settlers required very
little attention. The operation of the centrifuge, however,
required a great deal of attention to operate it at peak
efficiency. For this reason it would be necessary to man a

full-scale alum recovery plant with one person at all times.

CHAPTER 6

RESULTS OF RTL CONTACTOR OPERATION

6.1 Introduction

As the mixer-settler pilot program progressed, it was
realized that the Sharples centrifuge was not capable of
performing in a manner which would make the alum recovery
process cost-effective. Centrifuge throughput was low and,
more importantly, solvent recovery was extremely poor. To
continue operation with the Sharples centrifuge would have
been fruitless and a decision had to be made as to what
course the remainder of the pilot program would take. Two
alternatives were available. The first was to install and
evaluate another centrifuge which was more suited to the
separation problem posed by the bleed solids. The second
was to replace the mixer-settler extractor with a contacting
unit which did not produce a waste stream from which solvent
had to be recovered.

Fortunately, it was possible to explore both alternatives.
A DeLaval BRPX - 207 Solids-Ejecting Centrifuge was leased
for a two-week period and the evaluation of its performance
is the subject of another paper.(4) The mixer-settler
extractor was replaced by a RTL Contactor, the evaluation of

which is the subject of this chapter.

106

107

The problems associated with the treatment of the bleed
solids as outlined in Chapter 5 were the reasons for the
unsuccessful economic performance of the mixer-settler/
centrifuge system. Bench-scale studies conducted at Michigan
State University with a four-inch diameter RTL Contactor
showed that in cocurrent operation, aluminum could be
extracted from the Tampa sludge without the production of
bleed solids. The solids which remained insoluble during
extraction in the RTL did not form an emulsion, but remained
in the raffinate. This meant that solvent losses from the
RTL system would be lower than from the mixer-settler system
and these reduced solvent losses could be attained without
the use of a solvent recovery operation.

The formation of bleed solids in the RTL Contactor was
prevented by the relatively gentle mixing provided by the
slow rotation of the buckets on the rotor shaft. While the
dispersion formed in the mixer-settler was very fine, the
rotating RTL buckets formed dispersed droplets many magni-
tudes larger than those in the mixer-settler. These large
droplets coalesced into their bulk phases much more quickly
and easily than the smaller droplets of the mixer-settler.

Based on the success of the bench-scale RTL studies, a
one gpm RTL Contactor was constructed and installed for

evaluation at the Tampa alum recovery pilot plant. The main

108

objectives of this evaluation were to establish the maximum
throughput, solvent losses, and the operator input for

proper process control.

6.2 Equipment

6.2.1 RTL Contactor

The RTL Contactor was supplied by Riotinto Til Holdings
S.A. of London, England and was constructed of stainless
steel throughout. A glass sight-port was installed at the
discharge end of the contactor to monitor the interface
level and observe solids behavior. The rotor contained 76
stages, each stage containing eight buckets. The chain
driven rotor was powered by a gear reduced 3/4 HP electric
motor. Sprockets were supplied which allowed the rotor
speed to be varied from a maximum of ten rpm, to a minimum
of two rpm. The contactor shell was supported by two
saddles, these saddles were supported by a rectangular base

frame.

6.2.2 Pumps and Accessories

Organic was pumped into and out of the RTL. Two Cole-
Palmer Model 7314 Variable Flow pumps were used to handle
the organic. These pumps had a capacity of 0-5 gpm. Sludge
was fed to the RTL with a Wallace and Tiernan Diaphragm

Metering Pump. Organic flow was measured with two Fisher

109

and Porter Rotameters. Sludge flow was monitored by manual
measurements with a stopwatch and graduated cylinder.
Organic and sludge transfer lines consisted of 3/8 inch ID

Tygon tubing.

6.3 Plant Layout

The RTL Contactor was tied into the existing stripping
circuit as shown on Figure 6-1. Stripped organic was pumped
from the effluent end of the settler of Strip II to the
organic inlet of the RTL. Loaded organic was pumped from
the organic outlet of the RTL to the Strip I mixer. Sludge
was drawn from a five gallon sludge reservoir which was fed
with sludge pumped by the mixer-settler sludge feed pump.
Sludge discharged into the reservoir caused enough turbulence
to keep the reservoir well mixed.

While the strippers were oversized with respect to the
organic flow generated by extraction with the RTL, their use
was suitable for the short-term testing of the RTL.

Raffinate was not pumped out of the RTL. The raffinate
discharge line contained an air break which allowed the
interface level to be controlled by the relative flow rates
of the stripped and loaded organic. If the influent organic
flow rate was greater than the effluent, the interface would

drop and the raffinate flow would be greater than the sludge

mOBU4BZOU ABm mEHB .524an BOQHm 49.449 m0 UHB42mmUm
Hlm mmDOHm

 

woo: m 34:

womDOm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

23< 3569.... ..........

wh<zEu<m l.l..

o_0< I..I

22407.5 ammo—Ebb .II
0.25.010 0w0<04lll

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“.83..“ SEI
WQ<COPm ............. jib. IIIIIIIIIIIII IIIII— I
2344 TILFII _ -
3368... 2.238
93 a 4...; @1i. m
r? u .I
l u o
n O
o . m.
u _ _ V
34.55 . w
29. a
H 4.5 _
L I _
PIA! ................... rl.||..'l.ll_
L
203%.“... ”.83... 35. ‘

.750de .0 momm
02.5.0 oz<m 0b

 

111

feed flow. If the influent organic flow rate was less than
the effluent, the interface would rise and the raffinate

flow would be less than the sludge feed flow.

6.4 Operations

Between September 18, 1979, and September 27, 1979, the
RTL Contactor was operated for 208 hours. The data from
that run are presented in Appendix B.

The extractant used for the RTL run was that which
remained in the mixer-settler system. The extractant
strength was approximately 10 percent (V/V) MDEHPA.

Operational parameters which were monitored are listed

below:

0 Organic flow
0 Sludge flow
0 Raw sludge characteristics

0 Raffinate characteristics

The RTL was operated by plant personnel and readings were
taken hourly. This frequent attention was for data collection
purposes only and under actual operating conditions the RTL

would require much less attention.

112

Equipment performance was generally good. The RTL
Contactor was operated at two rpm for the duration of the
run and operated flawlessly. Low operating and maintenance
costs are a major advantage of the RTL Contactor. Sludge
flows held steady at the desired levels, as would be expected
with a metering pump. Maintaining the organic flow rates at
equal values, however, proved at times to be a problem.
Several times the effluent pump speed drifted such that it
was slightly greatly than the influent pump. This caused an
air gap to form at the top of the RTL. In turn, this caused
the effluent pump to draw air. With no liquid to pump, the
effluent pump overheated and shut down. With the effluent
pump out-of-service and only the influent pump running, the
contactor interface quickly dropped. However, no major
upsets occurred due to the effluent pump shutdown because
the problem was noticed and remedied in the course of the

hourly system monitoring.

6.5 Extraction
Sludge flow rates from 0.40 gpm to 0.88 gpm were evaluated
at various phase ratios. While the RTL run was short and
the amount of data collected limited, it was possible to
discern several relationships. As shown on Figure 6-2, for
a given sludge aluminum concentration, phase ratio, and

rotor speed, the amount of aluminum extracted is proportional

113

MSHB ZOHBZmBmQ MOQDQm m0 ZOHBUZDW 4 m4 QMBU4Mme SDZHEDA4 Ezmommm

 

mlo MMDOHm
AEEV I 52.... zo_._.zm._.mo man—:40
OD. . 00. On
A
I
L
4
U
_
4
L
IA
U

.55 ... Rana .36:
3.. u 2:... was...
”.25.... 35. z. A... :9: o... .

00

N0

em

00

00

00

N0

v0

00

mm

00;

 

°/o

GELDVHIXB IN 0 N I WIT-IV

114

to detention time. The RTL approached equilibrium at a
sludge detention time of 88 minutes or a sludge flow rate of
0.40 gpm. Extraction of 90 percent of the aluminum in the
sludge under the conditions listed on Figure 6-2 required a
sludge detention time of 65 minutes or a sludge flow rate no
greater than 0.54 gpm.

The phase ratio at which the RTL was operated also had
an effect on the amount of aluminum extracted. As shown on
Figure 6-3, for a given sludge aluminum concentration,
sludge flow rate, and rotor speed, the percent of aluminum
extracted is proportional to the phase ratio. Higher phase
ratios would increase the amount of aluminum extracted by
increasing the aluminum concentration gradient throughout
the RTL.

Raffinate pH was used as an indicator of aluminum
extraction throughout the RTL run. Figure 6-4 shows raffinate
pH plotted as a function of detention time. Longer detention
times resulting in lower raffinate pH's. When this relation-
ship is considered in conjunction with Figure 6-2, the
validity of the use of raffinate pH as an indicator of
extractor is proven. The fact that steady-state conditions
existed in the RTL and did not exist in the mixer-settler
can explain why this relationship developed during RTL

operation.

115

OHB4m mm4mm m0 ZOHBUZDm 4 m4 QMBU4Mme SDZHEDH4 Bzmummm

mlm mmDOHm

A<\OV .. 0.5.5.. mm<In_

..md ...

70..

..N

Tad

..m

 

I 1

sea .... Rana mos:
EA... and ... 26.... 33.5

33$ 35. z. A... :9: 8..

I

00

mm

00

N0

¢m

 

CIBIOVHIXE WONIWO'IV °/o

pH

RAFFINATE

116

3.0 -

2.9 - 4IO mg/l AI3“ IN RAW SLUDGE
PHASE RATIO = I.7=l

2.8 _ ROTOR SPEED: 2rpm

 

 

22 J J n L 1 L L L 4

 

50 ' 75
SLUDGE DETENTION TIME -Imin)

FIGURE 6-4
RAFFINATE pH AS A FUNCTION OF SLUDGE DETENTION TIME

117

Due to time limitations, only one extractant strength
could be evaluated. Because of the long detention times
involved with the RTL, very large contactors would be needed
to carry out efficient aluminum extraction. To reduce
capital costs, it would be desirable to reduce the necessary
detention time. One way to do this would be to increase the
extractant concentration. While bench or pilot work would
be necessary to evaluate the effect of extractant concentra-
tion on detention time, it is not unreasonable to assume
that, for a given sludge aluminum concentration, higher
extractant concentrations could be used to reduce detention

times and still maintain high extraction efficiencies.

CHAPTER 7

COST ANALYSIS OF ALUM RECOVERY

7.1 Introduction

This analysis is based on the results of pilot-plant
operation and compared mixer-settler and RTL Contactor
options. Three alternatives were analyzed:

1. Mixer-settler with centrifuge for solvent recovery

2. RTL Contactor with centrifuge for solvent recovery

3. RTL Contactor

In order to conduct this analysis, it is necessary to
make assumptions with respect to the sludge flow rate and
average amount of alum used. The following values were used
in this analysis:

1. The average sludge flow rate is 300,000 gal/day.

2. The average alum usage is 22 tons/day.
These values are for the purposes of this analysis only.
Pilot work has shown that the character and volume of sludge
produced at Tampa varies greatly and a final design would

have to take this into consideration.

7.2 Tampa Facility Design
7.2.1 Sludge Pretreatment
7.2.1.1 Sludge Collection
The sludge can be collected via the existing mechanical
removal system and pumped into a thickener with existing

sludge handling pumps.

118

119

7.2.1.2 Thickener
An elevated thickener is currently in use at the sludge
dewatering facility. This unit could be moved and installed

adjacent to the thickened sludge holding tank.

7.2.1.3 Thickened Sludge Holding Tank
The existing mixed holding tank will be used to collect
the thickener underflow and to keep the sludge mixed for

uniform feed to the recovery circuit.

7.2.1.4 Sludge Pumping to Recovery Site

The sludge will be pumped to the recovery site by an
existing submerged pump in the thickened sludge holding tank
via an existing six-inch PVC line which runs directly to the

recovery site.

7.2.2 Acid Storage

The concentrated sulfuric acid demand of the stripping
circuits will be 1750 gallons per day. An existing elevated
acid storage tank with a capacity of 10,000 gallons will be
moved to the recovery site. Another elevated tank of equal
size will be purchased. Together, these tanks will provide
about 11 days of acid storage. The acid will be pumped to

the stripping circuits via metering pumps.

120

7.2.3 Mixer-Settler Extractor

Two stainless steel mixer-settler units are required
for Alternative 1. The 4815 gallon mixers would measure
9' x 9' x 9' and would contain two equally sized mixing
chambers separated by a horizontal baffle.' The settlers

would have a surface area of 162 ft2

and be 5 feet deep.
7.2.4 RTL Contactor Extractor

If Alternative 2 or 3 is chosen, four stainless steel
RTL units measuring 7.5 feet in diameter and 26.2 feet long

are required.

7.2.5 Centrifuge - Alternative 1

Alternative 1 requires two 36 gpm centrifuges (DeLaval
Model PX) to recover solvent from 75,000 gallons of bleed
solids produced each day in the mixer-settlers. The bleed
solids will be siphoned from the settler by three indepen-
dently operated, perforated manifolds. The siphons will
transport the bleed solids to a 500 gallon holding tank from
which two positive displacement pumps with 35 gpm capacities
would feed the centrifuges. The operation of the centrifuges
will require one operator on all shifts. The cost-effective

operation of the mixer-settler option is dependent on the

121

efficient operation of the centrifuges. Pilot results show
the solvent losses of two gallons per 1000 gallons of sludge

can be expected.

7.2.6 Centrifuge - Alternative 2

The raffinate leaving the RTL Contactor contains 0.5 to
1 percent by volume of floating material which contains
recoverable solvent. The loss of solvent in the floating
material corresponds to 1.5 to 3.0 gallons of solvent lost
per 1000 gallons of sludge treated. The use of a 10 gpm
DeLaval Model PX centrifuge to process the floating solids
can reduce solvent losses to 1 gallon per 1000 gallons of
sludge. The centrifuge would operate 40 hours per week.
The floating solids would be collected in a 3000 gallon
float cell fitted with a surface skimmer. The floating
solids would then flow by gravity to a 5000 gallon holding
tank.

7.2.7 Strippers
In all alternatives, four stainless steel mixer-settler
strippers will be used. The 2570 gallon mixers will measure

7' x 7' x 7'. The settlers will have a surface area of 85

2

ft and be 2 feet deep.

122

7.2.8 Solvent Reservoir
The stripped solvent reservoir will have a 5000 gallon
capacity and be located such that it is fed by gravity from

the strippers.

7.2.9 Recovered Alum Storage

A 3500 gallon wet well will be provided for storage of
recovered alum. A 100 gpm transfer pump will transfer the
recovered alum to one of the two existing alum storage
tanks. The other tank will be used to store commercial alum

for make-up and emergency purposes.

7.2.10 Granulated Activated Carbon Columns

The granulated activated carbon column required for
each circuit will be two feet in diameter and six feet high.
It can be operated as a gravity feed column (if head require-
ments exist) or a pressurized column. Backwash provisions
3

will be required. The carbon volume of the column is 18 ft

for each unit.

7.2.11 Waste Stream Neutralization
It will be necessary to neutralize the raffinate and
the centrifuged waste solids prior to disposal. A 1000

gallon (five-minute detention) mixed tank will be provided

 

neutralization tank.

7.2.12 Building
For Alternative 1, a 20' x 40' building will be required
to house the centrifuges, laboratory, and control room.

Alternatives 2 and 3 would require a 20' x 30' building.

7.2.13 Sitework

The Sitework includes preparing the ground for equipment
installation, setting foundations for the mixer-settlers
and/or RTL Contactors, moving the thickener and acid tank,
preparing the site for the solvent reservoir, alum wet well,

float cell, float solids storage tank, and pump house.

7.3 Capital Cost
The capital costs of the alum recovery alternatives aria

shown in Table 7—1.

7.4 Operating and Maintenance Cost:
7.4.1 Acid
The demand is 11 tons Of concentrated sulfuric ac1d per

day. The bid price of acid in Tampa is about $38 per ton.

 

124:

TABLE 7-1
ESTIMATED CAPITAL COSTS OF TAMPA FACILITY

Alternative 1 Alternative 2 Alternative 3

RTL Contactor RTL Contactor

with Centrifuge

Mixer-settler
with Centrifuge

 

 

Mixer-settler

extractors $125,000 - -
RTL Contactors - $800,000 $800,000
Strippers 188,000 188,000 188,000
Additional acid
tank 18,000 18,000 18,000
Solvent reservoir 8,000 8,000 8,000
Neutralization
tank 5,000 5,000 5,000
GAC Filters 13,000 13,000 13,000
Float cell - 19,000 -
Float storage - 25,000 -
Bleed solids
storage 2,000 - -
Centrifuge 650,000 113,000 -
Centrifuge pump
system 20,000 10,000 -
Alum holding tank 6,000 6,000 6,000
Alum transfer pump 6,000 6,000 6,000
Building 96,000 72,000 72,000
Sitework 100,000 100,000 100,000
Piping, valves 120,000 120,000 120,000
Electrical/Instru-
mentation 95,000 95,000 95,000
Organic inventory 37,000 40,000 40,000
SUBTOTAL $1,565,000 $1,638,000 $1,471,000
Technical Fees
and Contingencies
(35 percent) 548,000 574,000 515,000
TOTAL CAPITAL COST $2,212,000 $1,986,000 $1,986,000

 

125

7.4.2 Solvent

The cost of mixed solvent is $2.30 per gallon, based on
$1.00 per gallon for kerosene and $.90 per pound of MDEHPA.
Solvent losses for the three alternatives are listed below:

Solvent Loss, gal per

 

 

Alternative 1000 gal sludge
1 2.0
2 1.0
3 3.0

7.4.3 Alum Credit
Alum will be recovered at a rate of 19.8 tons per day.

The cost of alum is assumed to be $100 per ton.

7.4.4 Lime Demand
Waste stream neutralization will require 1.5 tons of

quicklime per day at $50 per ton.

7.4.5 Carbon Demand
The carbon demand is assumed to be 30 pounds per day at

$1.00 per pound.

7.4.6 Labor
Alternative 1 will require one operator at all times or
a total of four. The City of Tampa has one man available

and only three additional operators are necessary.

 

 

 

 

126

Alternative 2 will require one operator and this person
is already available.

Alternative 3 requires no additional personnel.

7.4.7 Sludge Conditioning and Hauling Savings
Tampa personnel indicate that savings in truck rental
for dewatered sludge hauling and polymer used for sludge

conditioning will amount to $102,000 per year.

7.4.8 Downtime

In Alternative 1, it is assumed that the system will be
down 20 percent of the time due to dependence on the centri—
fuges. This is reflected in reductions in sludge conditioning

and hauling savings and alum credit.

7.4.9 Power

Power costs are based on $0.04 per KWH.

ALTERNATIVE 1

 

4 strippers @ 10 HP each = 40 HP @ $0.04/KWH $11,000
2 extractors @ 15 HP each = 30 HP @ $0.04/KWH 8,000
2 centrifuges @ 100 HP each = 200 HP @ $0.04/KWH 52,000

TOTAL $71,000

127

ALTERNATIVE 2

 

4 strippers @ 10 HP each = 40 HP @ $0
4 contactors @ 5 HP each = 20 HP @ $0
1 centrifuge @ 30 HP each = 30 HP @ $0

ALTERNATIVE 3

 

4 strippers @ 10 HP each

4 contactors @ 5 HP each 20 HP @ $0

7.4.10 Maintenance
The annual maintenance cost will be

percent of the initial equipment costs.

40 HP @ $0.

 

.04/KWH $11,000
.04/KWH 5,000
.04/KWH .91999

TOTAL $19,000
O4/each $11,000
.04/each 5,000

TOTAL $16,000

assumed to be three

A comparison of the

operating and maintenance costs of the three alternatives is

presented in Table 7-2.

7.5 Cost Analysis

It can be seen from Table 7-2 that Alternative 2 is the

least expensive alternative and can be operated with a net

savings in operating and maintenance costs. The costs of

the three alternatives will be examined by two methods:

capital amortization and capital pay back.

ESTIMATED OPERATING AND MAINTENANCE COSTS FOR
TAMPA FACILITY

Acid

Solvent

Alum credit
Lime demand
Carbon demand
Labor

Sludge savings
Power
Maintenance
Patent royalty

TOTAL ANNUAL
OPERATING COST

128

TABLE 7-2

Alternative 1

Mixer-settler

with Centrifuge

$153,000
504,000
-578,000
28,000
11,000
60,000
-82,000
71,000

39,000

50,000

$256,000

Alternative 2

RTL Contactor
With Centrifuge

 

$153,000
252,000
~723,000
28,000

11,000

-102,000
19,000

23,000

M

-$289,000

Alternative 3

RTL Contactor

 

$153,000
756,000
-723,000
28,000

11,000

16,000
18,000

50,000

$207,000

129

7.5.1 Capital Amortization
Listed below are the equivalent annual costs of the

alum recovery alternatives:

 

 

Alternative Equivalent Annual Cost(1)
1 $455,000
2 -$ 80,000
3 $394,000

 

1. Assumes seven percent interest over 20-year
period (P/A = 0.09439)

Analyzing the cost data in this manner shows that the
sludge treatment savings and alum credit in Alternative 2
not only offset operating costs, but also offset the

amortized capital cost.

7.5.2 Capital Pay Back — Alternative 2

The City of Tampa has expressed an interest in deter-
mining how many years would be required to recover their
capital investment. To conduct this analysis, an inflation
rate of 10 percent was assumed. Only Alternative 2 can be

analyzed in this manner. The results are presented below:

130

 

 

Year Savings Capital Outstanding
0 $ 0 $2,212,000

1 289,000 1,923,000

2 318,000 1,605,000

3 350,000 1,255,000

4 385,000 870,000

5 424,000 446,000

5.96 446,000 0

The capital pay back period with 10 percent inflation is

5.96 years if Alternative 2 is implemented.

CHAPTER 8

CONCLUSIONS AND RECOMMENDATIONS

8.1 Conclusions

In spite of the handicaps under which the mixer-settler
pilot plant was Operated, the feasibility of the alum recovery
process was demonstrated at this scale. The cause of the
discontinuous operation which prevented the desired evaluation
of the alum recovery plant was not the fault of the basic
recovery processes, extraction and stripping. Rather, it
was the fault of the poor performance of the original
centrifuge; a machine whose performance has no bearing
whatsoever on the extracting and stripping of aluminum. The
data which were collected showed that 90 percent of the
aluminum in the feed sludge can be recovered as
contaminant-free, functional alum.

The scale-up of the mixers, both extractor and stripper,
was poor, this was caused by the method in which they were
designed. Mixing was inadequate at design mixer speeds and
this greatly reduced the flexibility of the entire system.
Future mixers should be designed, not on the basis of
N3D2 $20, but on the maintenance of geometric similitude and
constant power input per unit mixer volume.

The optimum extractor mixer operating phase ratio was

2:1. This is based on the results of the aqueous entrainment

131

132

analysis. At operating phase ratios greater than 2:1,
excessive raffinate was entrained into the loaded organic.
This entrainment diluted and contaminated the recovered
alum. Operating phase ratios less than 2:1 invited phase
inversions.

The design of the settlers, in both extraction and
stripping circuits, was proven sound by results of the pilot
study. When properly operated, the settlers efficiently
separated the dispersions fed to them. To prevent solids
carry-over from the extraction to stripping circuit, the
bleed solids layer had to be kept as low as possible in the
settler to prevent high scour velocities in the organic
portion of the settler.

The picket fences and dispersion introduction baffles
were effective as settling aids and should be included in
the design of future settlers.

The solvent was found to be unstable and lost extract-
ing ability with use. Bench-scale tests with an RTL Con-
tactor revealed that after being loaded with aluminum 13
times, the extractant could extract only 20 percent of the
aluminum that it could when it was fresh. The mechanism by
which the extracting ability of extractant is lost is not
yet known.

A pilot-scale RTL Contactor was also evaluated at Tampa
and the data produced indicate that aluminum can be extracted

from the Tampa sludge at an efficiency of 90 percent or

133

greater. The major advantage of the RTL Contactor is that
it does not produce bleed solids. Because there are no
bleed solids, the RTL Contactor can make the liquid-ion
exchange process cost-effective without a solvent recovery
operation. To achieve 90 percent aluminum recovery
efficiency, a detention time of 65 minutes was required. It
is believed that the detention time can be reduced by using
higher extractant concentrations and operating at phase
ratios greater than unity.

A cost analysis comparing construction and Operation of
full-scale mixer-settler extractors and RTL Contactors at
Tampa favors the use of RTL Contactors in conjunction with a
centrifuge to recover solvent from the raffinate. If the
cost of the recovered alum is credited against the amortized
capital and operating costs, the City of Tampa could save
$80,000 per year and recover their capital outlay in six

years.

8.2 Recommendations
Based on the findings of this study, the City of Tampa
should proceed with the design and construction of a full-
scale liquid-ion exchange alum recovery facility for sludge
treatment.
Alum recovery with liquid-ion exchange should be given

serious consideration as a sludge processing alternative at

134

other facilities where the economic feasibility of coagulant
recovery exists. Basic researching of the interaction
between the extractant, diluent, and constituents of coagulant
sludges is also warranted. In particular, the mechanism by
which extracting ability of the extractant is lost should be

investigated in detail.

APPENDICES

APPENDIX A

 

 

APPENDIX A

MIXERPSETTLER
OPERATING DATA

1979

Parameter

Time of
Operations, hrs. 6 12 16 21 30
Average sludge

Flow, gpm 2.6 3.0 3.2 3.0 3.0
Average Solvent

Flow, gpm 13.0 13.0 13.0 13.0 13.0
Feed Flow

Ratio, O/A 5:1 4.3:1 4.1:1 .3:l 4.1:1
Mixer Phase

Ratio, O/A - 5.5:1 - - 6.6:1
Mixer Detention

Time, min - 23 - - l9
Al3+ in Sludge

@ pH 2.0, mg/l - 320 — 320 -
A13+ in Raffinate,

mg/l - 19 21 - -
Raffinate pH 1.8 1.8 1.8 1.85 -
Strip I Phase

Ratio, O/A — 6:1 - - 9:
Al3+ in Recovered

Alum, mg/l - 7900 8600 - -
Strip II Phase

Ratio, O/A - 6:1 - - 5.5:1
Al3+ in Strip II

Acid, mg/l - 8200 8600 - -

 

135

Date

 

Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention
Time , min

A13+ in Sludge
@ pH 2.0, mg/l

A13+ in Raffinate,

mg/l
Raffinate pH

Strip I Phase
Ratio, O/A

3+

Al in Recovered

Alum, mg/l

Strip II Phase
Ratio, O/A

+ .
Al3 in Strip II
Acid, mg/l

136

MIXER-SETTLER
OPERATING DATA

 

 

4-1 4-6 4-7
37 45 52
3.2 3.2 3.1
13.0 13.0 13.0
4.1:1 4.1:1 4.1:1
6.7:1 - -
18 - -
6 - 10
- 2.0 -
7850 - 7800
13500 - 13000

4-10

 

58

3.6

13.0

3.6:1

5.0:1

20

8700

15000

4-11

 

65

4.5

13.0

20

80

9600

16000

Date

 

Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A
Mixer Detention
Time, min

+ .
Al3 in Sludge
@ pH 2.0, mg/l

3+

Al in Raffinate,

mg/l
Raffinate pH

Strip I Phase
Ratio , O/ A

3+

A1 in Recovered

Alum, mg/l

Strip II Phase
Ratio, O/A

A13+ in Strip II
Acid, mg/l

70

5.7

12.0

19

80

137

MIXERFSETTLER
OPERATING DATA

 

4-16

75

12.5

2.05

10600

15000

4-17

83

5.2

12.5

2.4:1

3.0:1

21

80

4-20

86

2.8

12.0

4.3:1

5.3:1

25

4-21

 

92

3.3

12.0

3.6

[—1

4.4:1

25

82

13000

19000

Date

 

Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention
Time, min

+ .
Al3 in Sludge
@ pH 2.0, mg/l

Al3+ in Raffinate,

mg/l
Raffinate pH

Strip I Phase
Ratio, O/A

Al3+ in Recovered
Alum, mg/l

Strip II Phase
Ratio, O/A

A13+ in Strip II
Acid, mg/l

98

3.3

12.0

3.6:1

880

22

16000

2000

2138

MIXERrSETTLER
OPERATING DATA

 

4-24

107

12.5

3.3:1

5.7:1

17

400

38

21000

4-25

 

113

4.1

12.5

3.1:1

540

2.25

4-26

122

3.4

12.0

3.5:1

5.7:1

19

18000

4.0:1

21000

4-27

133

12.5

3.0:1

4.0:1

21

510

2.45

3.7:1

Date

 

Parameter

Time of
Operations, hrs.

Average sludge

Flow,

gpm

Average Solvent

Flow,

Feed

gpm

Flow

Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention

Time,

A13+

@ pH

Al3+
mg/l

min

in Sludge
2.0, mg/l

in Raffinate,

Raffinate pH

Strip I Phase
Ratio, O/A

A13+

Adum,

in Recovered
mg/l

Strip II Phase
Ratio, O/A

Al3+

Acid,

in Strip II
mg/l

4-28

142

13.0

2.7:1

2.9:1

139

MIXER-SETTLER
OPERATING DATA

 

4-29

147

13.0

21

34

2.15

3.1:1

21000

27000

5-1

158

3.1

12.0

3.9:1

5.3:1

23

580

 

167

3.4

11.5

22

107

2.1

7.6:1

25000

30000

5-3

176

3.0

12.0

25

1120

2.0

11:1

4.0:1

£140

MIXERFSETTLER
OPERATING DATA

 

Date

5-11 5-12 5-15 5-16 5-17
Parameter
Time of
Operations, hrs. 181 189 199 205 210
Average sludge
Flow, gpm 2.8 2.3 2.3 2.3 2.7
Average Solvent
Flow, gpm - 13.5 11.0 11.5 11.5
Feed Flow
Ratio, O/A - 5.9:1 4.8:1 5.0:1 4.3:1
Mixer Phase
Ratio, O/A 9.0:1 5.9:1 9.0:1 7.6:1 4.3:1
Mixer Detention
Time, min 16 30 19 22 21
A13+ in Sludge
@ pH 2.0, mg/l - 1130 - 510 600
A13+ in Raffinate,
mg/l - 195 38 30 30
Raffinate pH 2.0 2.4 2.3 — 2.2
Strip I Phase
Ratio, O/A 5.7:1 4.0:1 - 3.3:1 3.0:1
Al3+ in Recovered
Alum, mg/l - 31000 31000 30500 31000
Strip II Phase
Ratio, O/A 4.0:1 4.0:1 - 3.2:1 2.9:1
A13+ in Strip II
Acid, mg/l - 37000 35000 36500 36000

Date

Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention
Time, min

A13+ in Sludge
@ pH 2.0, mg/l

+ . .
Al3 in Raffinate,

mg/l
Raffinate pH

Strip I Phase
Ratio, O/A

+ .
A13 in Recovered

Alum, mg/l

Strip II Phase
Ratio, O/A

A13+ in Strip II
Acid, mg/l

215

3.1

12.0

27

570

20

34000

39000

141.

MIXERrSETTLER
OPERATING DATA

 

 

5-21 5-22_‘
227 242
3.6 3.9

12.0 12.0
3.3:1 3.1:1
5.9:1 4.0:1

18 23
280 150
10 10
1.9 2.3

4.4:1 3.9:1

31000 30000

4.8 4.0:1
35000 36000

 

5:23 5-24
250 257
3.3 3.1

12.5 13.0
3.8:1 4.2:1
4.0:1 5.0:1

27 24
350 520
52 67
2.3 2.25

4.0:1 4.6:1

31000 28000

4.0:1 4.2:1

35000 39000

Date

Parame ter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio , O/A

Mixer Detention
Time, min

A13+ in Sludge
@ pH 2.0, mg/l

+ .
Al3 in Raffinate,

mg/l
Raffinate pH

Strip I Phase
Ratio, O/A

+ .
Al3 in Recovered

Alum, mg/l

Strip II Phase
Ratio, O/A

A13+ in Strip II
Acid, mg/l

269

13.0

21

430

2.25

4.0:1

4.0:1

142

MIXERFSETTLER
OPERATING DATA

 

279

13.0

30

285

13.0

31

160

1.9

4.2:1

30000

43000

290

12.5

5.4:1

6.4:1

26

1180

200

30000

4.7:1

42000

298

13.0

20

830

30000

3.6:1

43000

143

MIXERFSETTLER
OPERATING DATA

 

Date 5—31 6-4 6-5 6-6 6-8
Parameter
Time of
Operations, hrs. 304 316 323 335 348
Average sludge
Flow, gpm 1.8 3.0 3.0 3.5 1.0
Average Solvent
Flow, gpm 12.5 12.0 12.5 12.5 8.5
Feed Flow
Ratio, O/A 6.9:1 4.0:1 4.2:1 3.6:1 8.5:1
Mixer Phase
Ratio, O/A 9.0:1 5.7:1 4.6:1 - 9.0:1
Mixer Detention
Time, min 24 19 24 - 44
A13+ in Sludge
@ pH 2.0, mg/l 770 360 - 210 950
A13+ in Raffinate,
mg/l 30 - 17 32 -
Raffinate pH 2.1 2.2 2.1 2.3 1.95
Strip I Phase
Ratio, O/A 5.3:1 - 3.9:1 4.0:1 9.0:1
A13+ in Recovered
Alum, mg/l 28000 — 30000 26000 -
Strip II Phase
Ratio, O/A 4.0:1 - 3.8:1 3.9:1 3.0:1
Al3+ in Strip II
Acid, mg/l 43000 — 44000 47000 -

Date

 

Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention
The,Mm
Al3+ in Sludge
@ pH 2.0, mg/l

+
Al3 in Raffinate,
mg/l

Raffinate pH
Strip I Phase
Ratio, O/A

3+

Al in Recovered
Alum, mg/l

Strip II Phase
Ratio, O/A

A13+ in Strip II
Acid, mg/l

144:

MIXERrSETTLER
OPERATING DATA

 

6-9 6-13
359 369
1.0 2.0
6.0 6.0

6.0:1 3.0:1
11:1 3.0:1
37 68

- 57
2.0 2.2
7.3:1 8.1 1
- 27000
2.7 1 3.0 l
- 44500

372

0:1

64

378

2.0

3.0:1

3.0:1

58

700

31

28000

3.4:1

43000

383

6.0

2.6:1

2.6:1

56

3.3:1

 

 

 

Date

 

Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention
Time , min

Al3+ in Sludge

@ pH 2.0, mg/l
Al3+ in Raffinate,

mg/l
Raffinate pH

Strip I Phase
Ratio, O/A

Al3+ in Recovered
Alum , mg/l

Strip II Phase
Ratio, O/A

Al3+ in Strip II
Acid, mg/l

145

MIXERPSETTLER
OPERATING DATA

 

7-6 7-11
391 399
2.5 2.7
9.0 9.0

3.6:1 3.3 l

3.6:1 5.8 1

46 24
210 560
15 18
- 2.2
13:1 9.0:1

33000 39000
- 3.7:1

43000 47000

405

7.0

38

410

2.9

9.0

3.1:1

3.1:1

38

1.85

4.7:1

417

45

41

2.1

3.6:1

40000

3.6:1

42000

Date

 

Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow, gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention
Time, min

Al3+ in Sludge
@ pH 2.0, mg/l

A13+ in Raffinate,

mg/l
Raffinate pH

Strip I Phase
Ratio, O/A

Al3+ in Recovered
Alum, mg/l

Strip II Phase
Ratio, O/A

A13+ in Strip II
Acid, mg/l

146

MIXERPSETTLER
OPERATING DATA

 

7-20 7-23 9-6
423 426 431
3.1 6.1 3.1
9.0 14.5 -

3.9:1 2.4:1 -
2.9:1 2.7:1 3.2:1
37 19 34
690 180 600
49 10 77
2.2 2.4 1.8
5.7:1 4.9:1 9.0:1
42000 40000 48000
2.1:1 1.9:1 4.8:1
46000 37500 45000

3.1

17

5.4:1

21

3.3:1

450

4.3

22

570

34000

5.7:1

42000

Date
Parameter

Time of
Operations, hrs.

Average sludge
Flow, gpm

Average Solvent
Flow , gpm

Feed Flow
Ratio, O/A

Mixer Phase
Ratio, O/A

Mixer Detention
Time, min
3+

A1 in Sludge
@ pH 2.0, mg/l

Al3+ in Raffinate,
mg/l

Raffinate pH

Strip I Phase
Ratio, O/A

Al3+ in Recovered
Alum, mg/l

Strip II Phase
Ratio , O/ A

Al3+ in Strip II
Acid, mg/l

147

MIXER-SETTLER
OPERATING DATA

 

 

9-13 9-14
461 467
4.7 4.4
- 14
- 3.2:1

3.9:1 4.0:1
20 21
770 680
52 —
2.05 2.0

2.3 l 3.6 1

30500 -

3.3:1 3.8:1

40000 -

474

3.9

14

21

450

29

32000

480

3.3:1

540

99

2.1

36000

34000

APPENDIX B

 

om we owe e.~ Hum.v pv.o omee ms\e~\m
mm av emu o.~ Awm.v mv.e once me\e~\m
me am one s.~ H.~.H em.o emmm as\e~\m
mm mm ohm v.~ Hum.~ oe.o once me\e~\m
me am cam e.~ H"G.H no.6 emmm ms\m~\m
mm as emv m.~ Hum.H «4.0 eons mhxqum
on we emu n.~ H.¢.H 64.6 echo me\n~\m
em as ewv v.~ Hum.H oq.e ommu mexmmxm
64 vm omv m.n Hum.H ov.o ommm me\o~\m
mm mm emv ~.~ ane.~ mm.o em- ma\m~\m
mm em evm m.~ Ane.e ov.e omqo ma\ma\m
no as com 4.~ Hua.e mm.e ommm mh\ma\m
mm as omv ~.~ anm.e ov.e emma ms\oa\m

H\m.= .0.~ mm 0 Haws HES. .0.N mm 0 :0 :50 Scam

ouaefiuuem +m~< oumeammmm +nHa maesflm 3am +maa «presumes 040mm mmmem amenam 3mm mafia muse

 

4849 OZHE4¢HQO mOBU4BZOU ABM

m xHazmmm4

148

APPENDIX C

APPENDIX C

CALCULATION OF AQUEOUS ENTRAINMENT IN LOADED ORGANIC

I. INTRODUCTION

Measurements of aqueous entrainment into the loaded
organic were made by allowing one liter of organic taken
from the extractor settler organic weir to settle for 24
hours. Two measurements were taken. The amount of entrained
aqueous and the phase ratio at which the mixer was operating

at the time of sampling are listed below:

Mixer Phase Ratio Aqueous Entrainment, ppm

 

10:1 4100

2.4:1 100

II. SAMPLE CALCULATION
The measured entrainment values were confirmed by

calculated values of entrainment. The calculated values

were based on operating conditions existing at the time of

entrainment measurements. Presented below is the calculation

of an entrainment value based on the following operation

conditions:

Aluminum in sludge at pH 2.0 = 950 mg/l Al3+

Sludge flow = 1 gpm

149

150

Aluminum in raffinate = 32 mg/l Al3+

Organic flow to stippers = 8 gpm

Aluminum in alum = 26,000 mg/l A13+

Aluminum in Strip II acid = 47,000 mg/l Al3+

These are the conditions that existed when the measured

entrainment equalled 4100 ppm.

The aluminum transferred from Strip II to Strip I was

 

 

equal to:
ml
77 mIn 3+ m Al3+
x 47,000 mg/l A1 = 3600 9 .
ml min
1000 I7

The aluminum transferred to Strip I from the extractor
can be founded by subtracting the amount of aluminum leaving
the extractor in the raffinate from the amount of aluminum
entering the extractor in the raw sludge. This assumes that
the amount of aluminum leaving the extractor with bleed
solids was negligible.

(950 mg/l A13+

3+

X 3.785 1/gal X 1 gpm) -
mg Al3+

(32 mg/l A1 min

x 3.785 1/ga1 x 1 gpm) = 3500

151

Therefore, every liter of solvent entering the strip-

ping circuit contained 120 mg/l Al3+.

--. 3500 mg Al3+
= 120 mg/1 A1

1
3.785951X 8 gpm

3+

If it is assumed that 40 mg/l Al3+ are removed from the

solvent by Strip I, the total amount of aluminum leaving the
stripping circuit in the recovered alum was:

3+
mg Al + (40 mg/l A13+ x 3.78 l/gal x 8 gpm) =

3600 min

mg Al3+

4800 min

If the recovered alum flow was 77 ml/min, the recovered

alum aluminum concentration in the absence of any dilution

should have been:

 

 

mg A13+
4800 min
= 62 mg A13+
77 ml ml
min
62 m A13+ 1000 1 3+
9 x E— = 62,000 mg/l A1

 

ml 1

 

 

 

152

Having calculated what the recovered alum aluminum
concentration should have been and having measured what it
actually was, the amount of raffinate being carried into the
stripping circuit as entrainment can be calculated. The
variable, E, being the amount of entrainment having the

units of ml/min.

3+ 3+

(62,000 mg/l A1 X 77 ml/min) + (32 mg/l Al
3+

x E)

(26,000 mg/l A1 ) (77 ml/min + E)

_ 110 ml _
E - 'HEB’" — 3600 ppm

This calculated value of entrainment is very close to
the measured value of 4100 ppm. The two measured entrainment
values are plotted on Figure 5—5. The other data points on
Figure 5-5 are calculated entrainment values derived with

the above calculations.

BIBLIOGRAPHY

 

 

BIBLIOGRAPHY

Susan, J., "Characteristics of Acid Treatment Alum
Sludges for Alum Recovery in Water Treatment Plants,"
Masters Thesis, Department of Civil and Sanitary
Engineering, Michigan State University, (August 1977).

Cornwell, D.A., "Recycling of Alum Used for Phosphate
Removal in Domestic Wastewater Treatment," J. Water
Pollution Control Federation, 49:4, 600 (1977).

Lemunyon, R.M., "Optimization of Liquid-Liquid Extraction
and Stripping Batch Systems for the Recycling of Aluminum
Coagulants in Water Treatment Plants," Masters Thesis,
Department of Civil and Sanitary Engineering, Michigan
State University, (1977).

Przybyla, J.M., "Quality of Products and Waste Streams
Produced in the Liquid-Ion Exchange Alum Recovery Process,"
Masters Thesis, Department of Civil and Sanitary Engineer-
ing, Michigan State University, (1980).

Bailes, J.C., C. Hanson, M.A. Hughes, "Liquid-liquid
extraction: The process, the equipment,"
Chemical Engineering, 86 (1976).

 

Jackson, I.D., J.B. Scuffham, G.C.I. Warwick, G.A. Davis,
"An Improved Settler Design in Hydrometallurgical Solvent
Extraction Systems," Proceedings International Solvent
Extraction Conference, vol. 1, 567 (1974).

 

Slater, M.J., G.M. Ritcey, R.F. Pilgrim, "Aspects of
Copper Extraction in Mixer-Settler, Proceedings Interna-
tional Solvent Extraction Conference, vol. 1, 107 (1974).

 

Hanson, C.(ed), "Recent Advances in Liquid-Liquid
Extraction," Chap. 14, Pergamon Press, Oxford, 1971.

Hanson, C., "Solvent Extraction — Theory, Equipment, .
Commercial Operations, and Economics," Chemical Engineering,
75:18, 76 (1968).

 

153

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

154

Warwick, G.C.I., "Liquid-liquid contacting equipment."
Chemistry and Industry, 403 (1973).

 

Baird, M.H.I., G.M. Ritcey, "Air Pulsing Techniques for
Extraction Columns," Proceedings International Solvent
Extraction Conference, vol. 2, 1571 (1974).

 

 

Scheibel, E.G., "Fractional Liquid Extraction - Part 1,"
Chem. Eng. Prog., 44, 681 (1948).

Reman, G.H., R.B. Olney, "The Rotating-Disc Contactor -
a new tool for liquid-liquid extraction, Chem. Eng. Prog.,
51, 3, 141 (1955).

 

Oldshue, J.Y., H. Rushton, "Continuous Extraction in a
Multistage Mixer Column," Chem. Eng. Prog., 48. 297,
(1952).

 

Treybal, R.E., "A versatile, new liquid extractor,"
Chem. Eng. Prog., 60, 5, 77 (1964).

 

Barson, N., G.H. Beyer, "Characteristics of Podbielniak
Centrifugal Extractor," Chem. Eng. Prog., 49, 5, 243
(1953).

 

Coplan, B.V., J.K. Davidson, E.L. Zebroski, "The 'pump-mix'
mixer settler, a new liquid-liquid extractor, Chem. Eng.
Prog., 50, 8, 403 (1954).

 

Agers, D.W., E.R. DeMent, "The Evaluation of New LIX
Reagents for the Extractions of Copper and Suggestions
for the Design of Commercial Mixer-Settler Plants,"
General Mills Chemicals Division.

Warwick, G.C.I., J.B. Scuffham, J.B. Lott, "Considerations
in the design of large scale solvent extraction plants

for the recovery of metals," Proceedings International
Solvent Extraction Conference, vol. 2, 1373 (1971).

 

Mizrahi, J., E. Barnea, D. Meyer, "The Development of
Efficient Industrial Mixer—Settlers," Proceedings
International Solvent Extraction Conference, vol. 1,
142 (1974).

 

 

Mattila, T.K., "The Kemira Mixer-Settler Extractor,"
Proceedings International Solvent Extraction Conference,
vol. 1, 169 (1974).

22.

23.

24.

25.

26.

27.

28.

155

Mehner, W., E. Mueller, G. Hoehfeld, "The Lurgi multi-
stage liquid-liquid extractor," Proceedings International
Solvent Extraction Conference, vol. 2, 1265 (1971).

Mumford, C.J., "Liquid extraction is a continuous agitated
horizontal contactor," Proceedings International Solvent
Extraction Conference, vol. 1, 260 (1974).

Rosenzweig, M.D., "Motionless mixers move into new
processing roles," Chemical Engineering, 95, (1977).

Cornwell, D.A., Asst. Professor of Civil and Sanitary
Engineering, Michigan State Univeristy, Personnel
Communication, (Sept. 1979).

Hopkins, W.R., A.J. Lynch, "Anamax oxide plant: a new
U.S. dimension in solvent extraction," Eng. Mining J.,
56 (Feb. 1977).

 

Rowden, G.A., J.B. Scuffman, G.C.I. Warwick, "The Effect
of Changes in Operating Organic/Aqueous Ratio on the
Operation of a Mixer Settler," Proceedings International
Solvent Extraction Conference, vol. 1, 81 (1974).

 

Godfrey, J.C., C. Hanson, M.J. Slater, "The design of
solvent extraction equipment," Chemistry and Industry,
713 (Sept. 1977).

 

   

IMIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII”