,t-,\.!4,

Proceedings of EMC 2a'

o

Recycling of Mg-sludge: an Investigation to

realise a Closed Loop Recycling of Magnesium

Dipl.-lng. Alf Niederle, Prof. Dr.-lng. Bernd Friedrich

IME, Process Metallurgy and Metal Recycling, RWTH-Aachen

Intzestral]e 3

D-52056 Aachen, Germany

Abstract

Magnesium and magnesium alloys belong to the lightest engineering metals with densities about

1.8 g/cm3. Further properties like good heat conductivify, high specific strength and excellent damp-

ing behaviour explain the increasing use ofmagnesium as a basic material for light weight construc-

tions like in planes or automotives. Parallel to the increase of Mg production and use of Mg prod-

ucts the quantity of Mg scrap rises. Throughout all melting processes (primary magnestum

production as well as remelting of magnesium scrap) the formation of dross or sludge cannot be

avoided. Dumping or disposal ofthe dross or sludge rn the ocean is not a f inal solut ion in matters of

saving the resources and our environment.

IME, Process Metallurgy and Metal Recycling, Department of RWTH-Aachen started a three years

government funded BMBF collaboration research project in April 2002 to develop recycling con-

cepts. The main target is the recycling of magnesium scrap with a low amount of Mg-metal, e.g.

dross and sludge. The intention of this project is to separate the waste into three parts: Mg-metal,

oxides and a cleaned reusable salt.

This paper gives an overview ofthe experimental part in this project. It starts with the calibration of

the gas fired crucible fumace and shows the results of the tests in which different salt volumes were

used. Subsequently sludge is characterised, which has been used once, twice or three times as melt-

ing salt. Furthermore a new resistance heated Mg fumace with a 35 I crucible is presented, espe-

cially designed and constructed for this project. Finally the perspective offurther work is given.

Proceedings ofEMC 2003 395

Niederle. Friedrich

i Motivation

Increasing research and development activities in magnesium technology during the last decadeopen new applications for magnesium and its alloys. But in comparison to aluminium the develop-ment has just started. Parallel to the optimisation and the rising application possibilities in automo-tive and aeronautics, a significant rise in the consumer area is fbrecasted. During the production andwith the end ofuse ofthe consumer goods, scrap occurs which has to be recycled to save resourcesand the environment. Figure 1 gives an overview of the mass flow in magnesium production andrecycling.

The main target of lME, RWTH-Aachen and the BMBF is to close the loop of magnesium produc-tion scrap without downgrading the magnesium quality. During magnesium processing and magne-sium production scrap emerges. Actually this magnesium containirrg waste is used tbr alloying.desulphurisation of iron, metal lothermy or i t is dumped [,2], especial ly when highly oxidised orwith a high surfäce. ln the firture dumping must be avoided. Through recycling of scrap like dross.

sludge or turnings contaminated with oil, emulsion or water, the loop of magnesium production

scrap can be closed.

iÄr!v,/

Figure l: Mass flow in magnesium production and recycling

3 9 6 Proceedings oJ EMC )t

il''\v/

a rt Expenmenta l sct I

In bench scale tests indusrn mi . : _iron crucible fumace wrth a 5 i rr f : , : .Mg plates al lowed to f ix paramer::. . .f luence ofthe l iquid phase during . - . ,--

the charging sequence of materra:. . , . :--from Rheinkalk HDW, Gennan\ rr . : - . :

nal l i t with components of magnc.t. . :- .amount of 12 Yo calcium fluon,tc2.28 g/cm3. Some further päräl1rr,rc,:. . .

fixed parameters:

temperature; approx. 700 .Catmosphere. argon, 5 -101 r.

3 Results

3.1 Pre-tr ials using i i i t l ! : r :

An average metal yield of 9_< :!- ,- .reached. I f the amount ol salr u.c.: . - . .

r a t i o n o f m e t a l a n d s a l t o r s l u _ _ . _ . . .

20 Yo of salt rs charged. If thc :.r : .,-f lux (Figure 3). The best resulr\ j rr . : , . , .

Proceedings ctf EMC 2a03

l ight dross,

Figure 2: gas fired crucible firmai:

'^.Y

-: :.';hnoiogY during the last decade

:-::rison to aluminium the develoP-

- = ::plrcation possibilities

in automo-

r -,.,.rt.d. During the production and

. -. :.ls to be recycled to save resources

.' ".,* in magnesium production and

:. - ' :e the looP of magnesium Produc-

r. ::ragnesium processing and magne-

- l- , .r ,ning waste is used fbr al loying'

i . tspecial ly when highly oxidised or

-::ough recycl ing ofscrap l ike dross'

i:. :he loop of magnesium Productton

Proceedings oJ EMC 20t

Ai',v/Rea,cling of |vlg-s ludge

2 Experimental Set Up and Process Parameters

ln bench scale tests industry melting conditions were to be simulated. At the beginning a gas firediron crucible furnace with a 5 I iron crucible (Figure 2) was used for calibration tests wtth compactMg plates allowed to fix parameters like composition and amount of flux and temperature. The in-

fluence ofthe liquid phase during charging, the use ofa mechanical stiner (60 mm $; 400 rpm) and

the charging sequence of materials were tested. The best results were achieved with flux 12, a saltfrom Rheinkalk HDW, Germany in a range of 5 tested industrial salts [3]. Flux l2 consists of Car-nallit with components of magnesium-, sodium- and potassium chloride . Number l2 stands for theamount of 12 oÄ calcilum fluoride. The salt has a melting point of 424"C and a density of2.28 glcm3. Some furlher parameters of the melting process are shown in Figure 2.

3 Results

3. I Pre-trials using different salt volumes

An average metal yield of 95 Y" regarding to the weight of charged compact Mg plates can bereached. If the amount of salt used is between 20 and 50 o/o of the charged metal mass a good sepa-ration of metal and salt or sludge resulted. Only one mix phase is formed of salt appear if less than20 % of salt is charged. If the salt amount is more than 50 o%, magnesium is finely dispersed in theflux (Figure 3). The best results are achieved ifmagnesium is charged into liquid salt at 660 'C. The

Proceedings ofEMC 2003

!crap,

argon inlet ?gitator

t id

light dross, sal refractory

Mg-melt cast iron crucible

heavy dross, saltgas burner

fixed parameters:

temperature: approx. 700 "C

atmosphere. argon, 5 - '10l /mincrucible volume: 5 IMg content, e.g. briquettes: 3 kg

Figure 2: gas fired crucible fumace [4]

Niederle, Friedrich

possibility of an overflow of the melting salt had to be observed. Regarding the average metal yieldof 95 oÄ of the calibration tests with magnesium plates, the process in a gas fired furnace with opti-mised parameters is one possibility for melting magnesium. During melting of compact magnesiumplates the flux used prevents magnesium from burning and leads to a good separation ofmagnesiumadd salt if the amount of salt lies in the above mentioned boundaries [5]. Contrary to these results,the industry uses only 3 % salt. This could be explained by the crucible size. The industry used big-ger furnaces e.g. 500 I crucibles, so the ratio of surface to volume is different.

Figure 3: Magnesium castings; molten with 12 %o salt (left), 25 % salt (middle) and 60 %salt (right) [5]

3.2 Melting pure plates and contaminated Briquettes withconventional Salt (virgin and used)

While using flux l2 twice or three times its viscosity raised slowly when pure and massive compact

magnesium plates were molten. ln contrast to this the viscosity increased signifrcantly when melt-

ing oil contaminated Mg briquettes until a crumbly sludge appeared. This sludge seemed to be moresolid than liquid due to increasing content ofimpurities. Because ofthe higher viscosity ofthe salt,more magnesium metal was dispersed in the sludge, so the metal yield decreased. Also the separa-

tion between metal and salt was worse. Significantly more inclusions were in the taped metal. Fur-

thermore it was noticed that the amount of sludge depended strongly on the tlrye of charged mate-

rial (Figure 4). Contaminated material produced more sludge than e.g. pure Mg plates [5].

Proceedings ofEMC 2003

/^\\Y./

/r-\\vj

398

Figure 4: Amount of sludge formed dq

The protection against oxidation oftbOnly small amounts of fresh salt on top Iing salt foamed very quickly and coverqbetween 20 and 50 %o of the metal werr Iity and metal dispersion ofsludge rise irshown that the magnesium losses duriqusing sludge than new salt. Also no hyminimised (Figure 5). In addition lost rncsludge two times[S]. This procedure opadisposal.

Proceedings ofEMC 2003

X Mg briquettes with 10 o/o oil

,ir'Yr

:e metal Yield

:ce with oPti-

:t magnesittm

..i magnestum

: these results.

::stry used big-

: ; le) and 60

I

: : lassive cof iP; ' '

, i t l v when n r c '

. : : rned to be m' - -

. .Lrs i fy of the s i

, .-\lso the sePa:'

i : . iped metal . F"

. .ri charged m'':'

. , :es [5] .

tings oJ EMC '

6000

of sludge [g]

5000

4000

3000

2000

1000

0

Temoerature: 700 oC

AtmosDhere: arqon --z

Salt: virgin flux I 2

ra.@i-tsa-/

1@0 1500 2000 2500

Xf ,4gbr iquet teswi th l0%oi l ö [ ,4gp la tes A iFc-Dross OCleanWZLbr iquet tes Id ry tumings

Ä:'\vl Recycl i ng of llg-.s I tr ctgc

Figure 4: Amount ofsludge formed depending on the type ofcharged material and salt added[5]

The protection against oxidation of the metal is nearly the same like in case of using fresh flux.

Only small amounts of fresh salt on top of the melt are necessary to stop small bumings. The melt-ing salt foamed very quickly and covered the surface ofthe melt. For good isolation salt quantities

between 20 and 50 oÄ of the metal were needed and could also be used sludge. Even though viscos-ity and metal dispersion of sludge rise it makes sense to use flux l2 twice or three times. It can beshown that the magnesium losses during melting due to saturation and metal dispersion is befter

using sludge than new salt. Also no hydrate water is present so the oxidation of magnesium isminimised (Figure 5). In addition lost metal droplets are brought back into the process when usingsludge two times[5]. This procedure opens a potential to reduce costs for new salt and for sludgedisposal.

Proceedings of EMC 2003 399

Niederle, Friedrich

Figure 5: Magnesium losses depending on the on the type of"salt" used [5]

One advantage ofthe process substitutidg salt by sludge is that the metal yield rises, especially ifrelated to the charged magnesium briquettes or plates. The second is that sludge which had nor-

mally to be dumped can be reused. The metal yield, calculated by the following formula is also de-pending on the chaxged material.

Metal yield (%) =[tapped magnesium metal (g) / charg_ed metal (g) (e.g. Mg briquettes)] * 100

With a mix of sludge used once and oil contaminated briquettes a metal yield of 85 % is possible.

Further results are shown in Fisure 6,

Proceedings of EMC

,^\v /a\\v)

4 Characterisation olIn order to investigate the change ofsrlsalt used once, twice or three times (cdl

4.1 Characterisation ofqFIux 12 powder with a grain size of 5 uTable l: ICpAnalysisofFlux 12

Figure 7 shows a micro_graph pictne oflare 4 marked points. The resutb of ücircan be assumed to be water or OHgoq

Proceedings ofEMC 2003400

Figure 6: Metal yield subjected to tbe c

' - j : ' : . ol "salt" used [5]

: :- - : : .1t the metal yield r ises, especial l-v i :

' -. .:cond is that sludge which had nor-

: - - .:::J. by the following formula is also de-

rrr:rd metal (g) (e.e. Mg briquettes)l * 100

- - ' - - : : : j i a metal yield of 85 % is possiblc

Proceedings oJ EMC 2Ür

i l '\v/Recycling of Mg-sludge

4 Characterisation of non metallic input/ output Materials

In order to investigate the change ofsalt composition during the melting process new salt as well assalt used once, twice or three times (called sludge) were analysed.

4.1 Characterisation of commercial Salt (virgin Material)

Flux I 2 powder with a grain size of 5 mm resulted to the following composition (Table l ).

Table I : ICP Analysis of Flux I 2

Ms t%j A r t % l Na [%] K l % l C a l % 1 cr t%t F t%l Balance [%]v .ö 0.2 1 0 . 3 10.4 4.4 50.4 3 .8 10.7

Figure 7 shows a micro-graph picture of Flux l2 made by scanning electron microscopy. Noticeableare 4 marked points. The results of their EDX-analysis are given in Table 3. Most of the balancecan be assumed to be water or OH-groups.

Proceedings of EMC 2003 401

Metal yldd depending on charg€d matslal

0

E

g

0 20 40 60 80 100

SMaximum EA€rage Elue 8l \4inimum I Melal y ield lo/ol

Figure 6: Metal yield subjected to the charged material [5]

Niederle, Frietlric'h

Figure 7: micrograph of flux l2 by scanning electron microscopy left (enlarged l:50) and right(enlarged l :500)

Aluminium, fluoride and oxygen rich zones (Table 2) are located in Figure 7 left.

Table 2: EDX Analysis of Flux l2 (Figure 7 left)

Art%l K I%I Ca l%1 c r t% l F lo/"1 s i t%t o t%tPo in t ' l 23.9 1 . 2 0.9 5 .6 27 .8 3ö. /

Figure 7 (right) and the chemical analysis (Table 3) prove different phases existing in flux 12. Fluo-

ride and chloride rich zones with different amounts of magnesium, sodium, potassium and calcium

are present as are areas with a enriched content ofbarium, sulphur and oxygen are.

Table 3: Chemical Analysis of different micrograph areas with EDX (Flux 12, refer to Fig. 7

4.2 Characterisation of molten commercial Salt (without additions)

Virgin flux l2 is molten at 660'C and samples are taken after each l0 minutes. The results of the

samples analysed by inductively coupled plasma spectrometry (lCP) are shown in Table 4.

Proceedings oJ EMC 200:

,lr'w'

402

(right))

Sample Mq t%l A t t % l Na I%l K lo/ol Ca t%l c t t % t F I%I BaI% S I % ) I %Point 1 1 q ? o . u 7.4 2 .7 öz.u 4.2 29.3Point 2 4.0 0.8 15 .3 1 6 . 6 44.7 18.2Point 3 1 6 . 8 1 . 4 2.3 0 .8 9.4 1 . 2 24.9 o.u J O . YPoint 4 7 . 1 0.7 25.2 4 .7 1 . 6 52.0 8.4

ra'''v)

Table 4: Mean analyses of mtirr:

a 20 0 8 :0 . 1 2 :0 0 8 i i

The results ofthe tests show thar tlu,.ing the first 40 minutes withour conr.i::

4.3 Characterisation of S

After charging 3 kg of magnesrum r:.samples of sludge were taken and anal,,

Table 5: sludge sample composnr,,-

The metal content is finely dispersc,c Rsalt content is nearly constant.

Figure 8 (left) shows a sludge probe :::ra magnesium droplet with a dtamercr .:inside and at the edge of the drop. sl:c,,.,:phases is shown in Figure g (right r

Proceedings ofEMC 2003

/7,'v'

- . .0p) left (enlarged l :50) and r ight

,: : . :n Figure 7 left .

a l%l S i I % I o l,k]2 7 . 8 1 . 6 38.7

- : ' : : i : : phases exist ing in f lux 12. F' luo-

j : - : ' ,::, sodium, potassium and calciunl

. - : - --: and oxygen are.

' - : . . i . i - t D X l F i u x 1 2 , r e t ' e r t o F i g . -

Salt (without additions)

-'-,;i'. l0 minutes. The results of tl'.,

,- P ' are shown in Table 4.

Proceedings oJ EMC 2t.)r i

,^','Y.)Recycling of Mg-sludge

Table 4: Mean analyses of molten salt based on l6 samples.

Flux 12 Mgnon o" Mgo, Alnon o, Alo, 7 n Na K L,A cl F Balancet%1 lokl Tol t%1 t%l t%] t%l t%l l"k] l"/"1 %1

0 min. 9.8 0 10.2710.444.3950.39J , / O 10.7510 m in . 8 . 1 2 2.37 0.08 0.04 0 .01 9 .21 9.44 4.7647.583.08 15.2620 min. 8 . 1 2 2.49 0 . 1 2 0.05 0 .01 9.50 9.79 F . 1 7 48.78 J . t o I z . ö J30 min . 8 .01 3.01 0.08 0.05 0.01 9.47 9.83 5.2948.753.79 1 1 7 140 min. 7.98 J . t z 0 . 1 0 0.05 0 .01 9 .31 9.65 5 .2148.403.58 13 .59

The results ofthe tests show that flux 12 basically does not change consistence during heating div-ing the first 40 minutes without contact with molten magnesium.

4.3 Characterisation of Sludge

After charging 3 kg of magnesium briquettes contaminated with oil into I kg molten salt severalsamples of sludge were taken and analysed (Table 5).

Table 5: sludge sample composition after charging magnesium (based on 7 samples)

The metal content is finely dispersed. Regarding the weight of salt components it is evident that thesalt content is nearly constant.

Figure 8 (left) shows a sludge probe after salt contact with magnesium plates (AZ9l). In the centrea magnesium droplet with a diameter of nearly 0.4 mm is shown. Remarkable are the bright areasinside and at the edge ofthe drop, showing different phases. A detailed picture ofthe inside brighterphases is shown in Figure 8 (right).

Proceedings of EMC 2003

Nierlerie, Frieclrich

Figure 8: SEMt metallographic investigation of sludge with magnesir.rn'r drop (charged materiaL

FIux 12 and magnesium platcs (4291); left : enlargement l : 100; r ight: drop centeenlargement I :100

The analysis of the marked points in Figure 8 (right) is shown in Table 6

Table 6: EDX-analysis of metallic phases in the Mg droplet rnarked rn Figure 8 (right)

Po in t Mqf%l At f%t znl%l Cul%l N i t% ot%t,] 63.0 25 .2 2 .3 7 . 4 0 . 7 ' 1 .5

2 64,9 29.2 2 .4 2 , 43 95 .5 1 . 8 '2 ,

The cietail picture of the edge area shown in Figure 8 and the chemical EDX-analysis of the phr..

intcrfacc betwcen magnesium metal and sludge is shown in Figure 9.

'.17

Thc magnes ium me t i r , , l

s is ts ofs i l icon, z i rconiunt : r ithree and four consists nrr , . .marker three and four are nt.r

A Ä+ . + Character isarr, , ; -

A large scale micrograph t r i . i Ishown ln Figure I0. Besic l , : i r . .seen. An bulk analysis of rh, , . ,

Fulther 2.6 % A1,2^4 '],;Na I 1

Figure 9: SE,M and E,DX analysis; enlargement 1: 500: s lLrdge ' r , i th

nrater ia l : f lux 12 and magnesium plates (AZ9l) ; inter fäce

and s ludge

magnesium drop; chr

between magnesium r

Figure 10: SEM; enlargemc.n,

Figure l l shows the phases t rminium, chlor ide, f luor ide ar .n i um, ex i s t a l so i n t h i s samp icFigure I l are magnesium. e t : i.rf the picture is pure iron thrr r.

oceedings ofEMC 2003.+0;+ Proceedings oJ EA4(

/ta,\v!

:::.icnesium droP (charged material:

: - j r ' f lent l : 100; r ight : droP center

- Tab le 6

':-:rkcd tn Figure 8 (right)

:mrcal EDX-analYsis of the Pha::

Point 2 Point 3 Point 438,82,2 0,4

1 . 50,2 a 1

1 . )

4 a a 10 ,51 2 R q

1 . 1 .1.4

1 3 , 3 0.346.4 1 . 640.2 36.8 36

- ; ;1 ' , r i th magnesium droP; charg ' -

: : : : i , lce between magnesium nle: '

Proceedings o/ EMC :

rIi\.vlRecycling o/ Mg-,sludge

The magnesium metal (marker 1) is surrounded by three phases. The white phase (marker 2) con-sists of silicon, zirconium and oxygen. This may be a drag in from the crucible smoothing. Markerthree and four consists mostly of magnesium, chloride, fluoride and oxygen. Differences befweenmarker three and four are mainly the content ofchloride and fluoride.

4.4 Characterisation of used Slag

A large scale micrograph of three times used salVslag after melting magnesium plates (A291) isshown in Figure 10. Beside the magnesium and the salt phase some dark and white areas can beseen. An bulk analysis of the surface shows that i t consists of 45 oÄMg20.6oÄO and 15.6% CLFurther 2.6 % AI,2.4 %Na, 3.4 %K,2.8 %oCa and 3.8 %F are detected.

Figure l0: SEM; enlargement l : 20; general micrograph of f lux 12 used three t imes.

Figure I I shows the phases in detail. The dark phase (point l) consists mostly of magnesium, alu-minium, chloride, fluoride and oxygen (see Table 7). Potassium and calcium, together with tita-nium, exist also in this sample, but only in smaller amounts. The main components of phase three inFigure I I are magnesium, chloride, iron and oxygen. The white phase near the centre (Fe/Point 2)ofthe picture is pure iron that possibly originates from the crucible.

Proceedings of EMC 2003

Nteclerie. Friedrich ^,

- . ' UCVC lOpl . ) t r , '

After f inishing the pre-tnr .processes (800 l) the amour:p loduc t i on . A b rgge r t u r r r . ,f t c t o r s . s t i n . r ng co r ro i t r cn .

Figure 13 shoivs a schelre 1, , , rstructed by Thernrostar. ..-llr ,

emeroe.hVdraulic tilting f l l n . L lsvstem T--l'

I computer aided

htocess "ontro,

Figure I l : SEM; cnlargement 1: 200; detai l picture of Figure 10

Table 7: EDX-analysis of phases shown in Figure 11

Sample Mq l% Al to/o' K f%t C a l % c | I % t t % l Ti t % o l %Poin t 1 33"8 7 . 6 2 A 1 , 6 o 2 36,8 0 , 6 7 , 4Po in t 3 3 1 , 4 1 , 4 0 ,4 1 7 . 9 6 ,0 4 1 . 6

A more detailed picture is shown in Figure l2 where the transitiorr betrveen salt (slucige'; on rhe

srcie and magnesium on the right side is mapped. E1,e-catching are fivc phases, Some phases cont.,

nearly 50 o/o manganese or iron. A summary of the h,DX analyses is also given.

Poini :^ a '

5,61 . 91 . aa t

9 ' )

15 :

0,9l 3 :

l i igure l2: SL,M;detai led picture of the bounciary between magnesiurr metal and sal t

.106 Proceetlings ttf Elul(

Figure i3: Project der. ,e lopcrt \ i

The 35 I prototype furnace. n, r;tures, e.g. a remote-contro l lca r . .end of thc cast ing channei r : r ,Fuf thermore, the saf-ety is cnJ.r , r . .top is an argon loop duct svstclde l i r r ed o f f gas duc t . Add r r i e rwi th cover gas. I t is eas\ . to .1, , . :mant l ing of the crucib le. I r

-Sency rut) out I \ tT la in lAi l te.

can be used. With a pou,er , , ; ,e lectucal power is a lso n.-e,r , ,ox idat ion. The cornputer_aic: ,

sary, e.g. , for dcoi l ing of \1r -

\ u x i l i a r v e q u i p m e r r r . r . ^ .' e rbu rne r .

can be u5ed , . r -

Proceeding.s of EI.IC :()(/.1

Ar loop wrrlno .

i.envcen salt (sludge) on the '

- phases . So t l t e Phases co ' '

: r Iso gtven.

rr rnetal and sal t

Proceedings ttf Etv-l(

!v,l Rccl'cling uf \ fg-:ltrdgc

5 Development of a new resistive heated Mg fumace at the IME

After finishing the pre-trials and comparing thc results of the gas fired 5 I furnace with industryprocesses (800 l) the amount of salt needed in a small flrnase rises above the amount used in realproduction. A bigger furnace with meeting the necessary process parameters, e.g. tempcrature, saltfactors, st in' ing eondit ions" u.as developed in order to sinrulate industry conditrons.

Figure 13 shows a scheme and a prcture of the new magnesiunr funrace dcveloped by IME and con-structed trv Therrnostar,,Aachen.

The 35 I prototype fi:rnace. hcated resistively with SiC bars, sewes a high number of special fea-

tures, e.g. a remote-controlled hydraulic tilting system. The furnace is constructed in a way, that the

end of the cast ing channel is p laced in the rotat ing axis. This a l lows an ezrsy and secure tapping.

I"urthemorc, the safety is enhanced through an also remote-contlolled lid. The characteristic of the

top is an argon loop duct system to protect the melt and two openings for a mechanical stilTer and a

defined offgas duct. Additional the atmosphere befween refiactory and iron crucible can be flooded

wtth cover gas. It is easy to control the sr,rrface of the crucible (safetyl) because of the simple dis-mantling of the crucible. If, contrary to expectations, the crucible cracks ciuring mcltrng a emer-gency run out is maintained at the bot tom using mcl t fuse. Instead of i ron. a lso a ceramic cmcibleoan be used. wi th a power of 45 kw, i t is possib le to heat the system up to 1500 "c. The excesselectr ical power is a lso needed to minimise the f i rs t mel t ing phase to minimize the possib i l i r_v ofoxidation. The computer-aided proccss control carries out special heating programs. This is neces-sary, e.g. , for deoi l ing of Mg tuming br iquet tes.

Auxi l iarv equipment, l ike a mechanical argon in ject ing st iner. a cont inuous mater ia l feecl or an af ' ,terburner, can be used on request.

Proceedings of EIVIC 2003 .1(i7

Figure l3: Project developed Argon-shie lded and resist ive heated Mg fumace (35 l )

Niederle. Friedrich

6 Perspectives

The results achieved in the gas fired fumace are promising. Instead of direct dumping sludge will bereused up to two times. Parallel to saving costs for dumping and new flux, the metal yield rises be-cause the magnesium content of sludge is recovered and the metal oxidation is minimised due to notexisting hydrate water.

lmportant to know is that the results of the gas fired fumace are different to industry practise. Mag-nesium recycler use 3 o% salt referring to charged clean scrap and only up to 20 %o salt refening tocharged scrap with a high level of impurities. Before transfening the parameters of the 5 I crucible

fumace to industry scale, tests in the new 35 I resistive heated fumace have to be conducted to

simulate real conditions where a gas fired fumace is not qualified for. Especially the optimiseö

cover gas equipment promises new effective results. Along with the powerful resistive heating sys-

tem and computer controlled temperature, new possibilities for a better process control and datir

logging can be reached. This way, methods of removing cooling lubricants fiom turnings or bn-quettes like water, emulsion or oil will be tested by using definite preheating parameters. By com-

bination of these methods the magnesium metal yield will be optimised. Because sludge of different

processes never has the same composition the produced sludge, subjected to each different chargeC

material, has to be characterised carefully. Furthermore the reuse and the recycling of industry pro-

duction sludge will be investigated. These test series are scheduled until end of2004.

I Acknowledgments:

We thank BMBF lbr funding this project. Special thanks to our partners:

- Norsk Hydro Bottrop, Magnesium Producer: providing material (Mg/sludge) and scale L.:

tests

- Rheinkalk HDW, Salt Producer: providing salt

- IFG Düsseldorf, Foundry lnstitute: tests with recycled magnesium

- WZL-RWTH Aachen, Machine Tools and Production Engrneering: safety metal cu:r ' -

of Mg, producer of dry tuming-briquettes or briquettes with oil, water or emulsion a.

bricants.

- GHI-RWTH Aachen for the EDX-analyses.

Proceedings ol EltlC

ilr

'lr''v,,

t l l

t2l

t3l

L4l

L " l

Literature

DrrzE A.; ScH,qnrbuch; A lumi r : , . . ^

DITZE A.. .

schaften. .\:. ... .Weinheim. l

N teopnr , .Contanrr.-,:

01"06.2002. \ rL

Jücsren . A . : D ip lom"r . .Magnesium; 2002

Hrsspr-N, G.; Diplomarbeit: Ennrl isierte und feintei l ige Mg_Rest.t

,,:,ceedings of EMC 2003408

l, lread of direct dumping sludge will be

h üd new flux, the metal yield rises be-

tclat oxidation is minimised due to not

E rrc different to industry practise. Mag-

q and only up to 20 o/o salt referring to

*rring the parameters of the 5 I crucible

' bd fumace have to be conducted to

p gnlified for. Especially the optimised

tl ri6 the powerful resistive heating sys-

}lr br a better process control and data

: @ling lubricants from turnings or bri-

;&finite preheating parameters. By com-

f tc optimised. Because sludge of different

Fdgp, subjected to each different charged

b rcuse and the recycling of industry pro-

*üled until end of 2004.

rDqrf parmers:

lriding material (Mg/sludge) and scale up

I

f1lcld magnesium

üion Engineering: safety metal cutting

liFfts with oil, water or emulsion as lu-

Proceedings of EMC 2003

t t l

I2l

t3l

I4l

t5l

Recycling of Mg-sludge

Literature

DtrzE A.; ScHenr C.; Recycling von Magnesiumlegierungen, Magnesium Taschen-

buch; Aluminiumzentrale Düsseldorf(Hrsg.) l. Auflage; Al-Verlag Düsseldorf 2000

DtwE A., ScHanr C.; "Recycling von Magnesiumlegierungen" in Magnesium - Eigen-

schaften, Anwendungen, Potenziale; DGM, K. U. Kainer (Hrsg.); Wiley- VCH Verlag,

Weinheim,2000

NIEDERLE, A.; FPJEDRIcH, B; Recycling Of Magnesium Production-Scrap With High

Contamination, Impurity And Oxidation Level; GDMB Congress M3;'29.05.2002 -

01.06.2002. Vienna

JücHrER, A.; Diplomarbeit: Entwicklung einer Versuchsanlage zum Recycling vonMagnesium;2002

HIBBELN, G.; Diplomarbeit: Entwicklung eines Recycling-Prozesses ftir niedrig metal-

lisierte und feinteilige Mg-Reststoffe; 2003

Proceedings of EMC 2003 409