Magnesium oxide




World production of the metal is increasing by between 10 and 20 per cent annually to 420 000 tpa (primary form). About one-half is produced by the CIS and China.

Major uses are in aluminium alloying (47 per cent), die casting and steel desulfurisation (14%). 

Prices range around A$5 000 to A$6 000 per tonne. 

The common production of magnesium metal is by the electrolysis of magnesium chloride which co-produces chlorine. For example a typical 50 000 tonne per year plant, will co-produce about 150 000 tonnes of chlorine. This plant would cost around A$800 million and requires 60 Mwatt hours per tonne of metal produced of which 5 per cent is attributable to the preparation of lime (used to convert magnesium chloride to the hydroxide).


bulletSea water and lake brine MgCl2/MgSO4 (0.13 to 0.8%) bulletCarnalite MgCl2KCl.6H2O (8.6%) bulletDolomite MgCO3 CaCO3  (13 %) bulletSerpentine 3MgO SiO2 2H2O (24%) bulletMagnesite MgCO3 (28%)

There are several technologies which could impact on magnesium oxide production. The choice is also influenced by whether there is a market for chlorine and of course the raw material.


AMC aims to produce magnesium metal in Gladstone having established a pilot plant. It will recycle the chlorine required in the process using the magnesium carbonate (magnesite) deposits.

HCC in joint venture with with Multiplex is undertaking a feasibility study for carnalite from bitterns. Bitterns contain between 3 and 6 per cent magnesium (cf just 0.13% in sea water).

Leanora-Laverton district (with Anaconda) also reviewing prospects for a magnesite deposite.


The total energy required to produce one tonne of magnesium is around 35 to 40 MWhour for both the electrolytic and thermal routes. The electrolysis only step component ranges down to just 12 MWhour per tonne with the latest technology with the balance being thermal energy including that required for the production of lime, drying and billet production. The thermal reduction processes as described below requires a higher proportion of electrical energy with approximately the same amount of energy. The electrolytic route will benefit from integration with a petchem project that could consume the chlorine by-product. 

The requirements for electrolytic step in the production are proprietary and details of plant operation difficult to obtain. There is a theoretical minimum of 7.0 MWh per tonne of metal.

Note: 1 MWh is 3.6 gigajoules. At around one-third effiency of conversion efficiency, 1 MWh of power will require about 10 gigajoules of gas.

By implication of their proposed use, the other thermal (metalloreduction - currently only silicon though aluminium is proposed) processes are comparable in energy use. Where they differ is in the use of raw materials, and the energy consumed in the production of lime used to convert the chloride to the hydroxide intermediate (avoided eg by Qmag).

Producers of chlorine for sale (typically about two-thirds of the chlorine produced in electrolysis is used in preparation steps so that about one tonne of chlorine is produced available for sale per tonne of magnesium produced).

The first part of the report refers only to the reduction stage (production of the metal from the salt or oxide). The other, additional, energy consumers per tonne of magnesium depends on the process.

Electrolysis route

Additional thermal energy (per tonne metal produced):

Lime: if manufactured by metal producer at 10 gigajoules per tonne (3 MWh equivalent in power) to convert natural or magnesium chloride source


Production of hydrochloric acid: including syngas step to produce hydrogen (to react with liberated chlorine). Energy (gas) presumed about same as lime (5-10 gigajoules per tonne).


Magnesium chloride preparation: about 16 gigajoules (5 MWh equivalent in power).

Acid recovery: about 15 gigajoules per tonne (5 MWh per tonne).


If the chlorine is sold to a petchem project (eg. EDC/VCM/PVC) or for use in domestic water treatment, it provides a theoretical maximum credit of around 60 gigajoules per tonne (20MWh per tonne of magnesium). However this depends on the efficiency of chlorine use - with up to two-thirds consumed in the process and disposed as waste) so the credit reduces to as low as 6 MWh..

Reduction route

Process uses power. No breakdown available but total energy consumption approximately same as electrolysis route ie around 35 MWh per tonne of magnesium metal produced. There is no credit for chlorine as a user of magnesite or dolomite.

Includes thermal energy for preparation of dolime (or mixture of magnesia and lime) at 15 gigajoules (4 MW hours equivalent).



Dow Chemical had been operating a 100 000 tpa plant in Freeport Texas using magnesium chloride extracted from sea water. The process involved precipitation of the magnesium hydroxide with lime and the subsequent neutralisation with hydrochloric acid to produce magnesium chloride which is electrolysed to the metal. It is believed to have used about 18.5 MWhours per tonne of magnesium metal in the electrolysis stage. The project closed in 1998.

Norks Hydro

Norsk Hydro in Norway at Porsgrunn (55 000 tpa) and Becancour Quebec (50 000? tpa) use magnesium chloride brines containing 33 per cent magnesium chloride. Purification, concentration and drying at elevated temperatures with hydrogen chloride gas (to prevent decomposition - ie. little consumption) provides 100 per cent magnesium chloride with is then electrolysed.

It uses 28 MWh of electricity per tonne of magnesium.

Amax Brine

Amax Brine. Great Salt Lakes. Produces brine with 7.5 per cent magnesium which is concentrated to dry magnesium chloride which is purified including with chlorine (ie. little direct consumption).


Alcan Technology diaphragm cell by Osaka Titanium in Japan and Oremet Titanium in USA.

Highly efficient using magnesium chloride producing magnesium to 14MWh electricity per tonne and experimentally down to 11MWhours per tonne.

The AMC in Qld will use magnesite, dissolved in hydrochloric acid to be purified and dehydrated to be used in the electrolytic cell. It will use Alcan cells with technology developed by CSIRO. Energy consumption of 10-12 MWh per tonne is anticipated (plus thermal energy including syngas for hydrogen etc.) to a total of all energy of perhaps 30-35MWh per tonne.

MPLC Process (Magcan process)

Mineral Processing Licensing Corporation (MPLC) raw magnesite in reactor that produces anhydrous magnesium chloride in a single step. Currently not operational but production of 10/12 MWh per tonne anticipated.

Thermal reduction processes

Reduce magnesia with metals. Eg. metallothermic uses aluminium, calcium, carbon or silicon as reductant normally in DC transferred arc furnace. Presently only silicon is used (Pidgeon process). A carbon process is feasible at 12 MWh per tonne.


Operated by Pechiney in France and Northwest Alloys in the USA and by Magnachron in Serbia.

Similar to Pidgeon process with current through slag. Begins with dolomite (MgCa carbonate) though can use mixture of lime and magnesite. Reduction with ferrosilicon.

Operating down to below 10 MWh per tonne if operating continuously.


Heggie process proposed for Batchelor Northern Territory uses magnesia and aluminium reductant in DC transferred arc furnace.

Magnesium oxide production 

Also see 

bulletMagnesium metal technolgy and economics. bulletMagnesium in Australia

This is a brief review that should be integrated with an understanding of magnesium metal production given the production of the latter often involves the production of magnesium hydroxide and magnesium oxide.

The project should capture two sources of advantage, the raw materials (bitterns at Dampier Salt - containing 300 000 tonnes of magnesium) and natural gas helped by a regional market to provide transport cost savings. To the limited extent exposed in this report, the production of magnesium oxide captures some - the production of magnesium metal using alternative technologies should be evaluated in conjunction with magnesium oxide.

The W.A. market for magnesium oxide will be about 30 000 tonnes per year (though further work may establish other users such as for water treatment). However, there are different grades involved presenting some additional cost.

Magnesium oxide may be produced from salt field bitterns though in so doing involves the use of lime, which substantially increase the cost vis a vis its production from natural carbonate like Qmag. While there appears to be a prima-facie basis for furthering the evaluation of magnesium oxide production, its competitiveness would be enhanced by its production in conjunction with magnesium metal.

An interesting consideration for review is for the use of technology that produces magnesium oxide with hydrogen chloride, as a by-product, that could be used to liberate magnesium chloride from a known magnesium mineral deposit near Port Hedland. Another is for the chlorine liberated in the electrolytic magnesium production to be converted to hydrochloric acid (with hydrogen from syngas) and applied to that magnesium deposit.

What is magnesium oxide?

Magnesium oxide (magnesia, MgO) is a white powder broadly similar to calcium oxide (lime, CaO) and is rarely found in nature as such but more commonly as the carbonate form including the less common mineral complex with calcium carbonate (carnallite).

Magnesium oxide is traded commercially as light burned magnesia, dead-burned magnesia, calcined and electrofused forms (used as refractory lining). The amount of heating reduces the reactivity of magnesium oxide (including with carbon dioxide on storage) reflecting its surface area to mass (and density).


Magnesium oxide is the principal input for the magnesium compounds industry, refractories and for insulating elements in electric furnaces. In the USA, some 36 per cent is for animal feeds and fertilisers, 19 per cent for chemical processing, 18 per cent for metallurgical purposes (refractories, electrical, water treatment, gas scrubbing, 17 per cent of manufacturing aids. In Australia important uses are in the production of nickel and HBI DRI by BHP.

Process of manufacture

There are many processes for producing magnesium oxide all start with magnesium chloride from seawater, sometimes the mineral carnallite or a magnesium carbonate mineral. Most involve the thermal (not electric) decomposition of magnesium hydroxide in a rotary or shaft kiln.

One process reacts calcium oxide (lime) with magnesium chloride, such as from magnesium rich (bitterns) in water, to produce magnesium hydroxide and calcium chloride as waste. The magnesium hydroxide is readily converted to the oxide by heating.

MgCl2 + CaO > Mg(OH)2 and CaCl2

About 1.4 tonnes of lime is required for each tonne of MgO produced.

The magnesium hydroxide is insoluble and is separated from the calcium chloride and sodium chloride under controlled conditions. The magnesium hydroxide is then heated in a furnace (shaft or rotary kiln similar to that used for lime production) to produce the desired form of magnesium oxide.

Mg(OH)2 > MgO +H2O

This process is also commonly, but not exclusively, used when producing magnesium metal based on using magnesium chloride.

Another process begins with the burning (calcination) of magnesite (a magnesium carbonate) which is more direct but often produces a lower grade of the oxide (90 to 98 per cent) though Qmag, with an usually high quality magnesite can produce a high purity form (see later).

Dead Seas Periclase in Israel uses a concentrated magnesium rich brine from the Dead Sea with is sprayed into a reactor at about 1700șC to decompose the magnesium chloride to the oxide producing hydrochloric acid as a by-product. The oxide is converted to the hydroxide which is washed to purify and then calcined to yield the various forms of the oxide. A 99 per cent purity form is produced.

Worldwide about two-thirds of magnesium oxide (some 11 million tonnes per year) produced is from sea water (co-produced with salt production) and the rest from natural mineral deposits magnesite (magnesium carbonate) dolomite (calcium magnesium carbonate) and salt domes.

It is relevant to note that the production of magnesium metal commonly, but not exclusively, involves magnesium hydroxide as an intermediate (noting this is readily converted to the oxide). This could suggest the production of magnesium oxide from a chloride source is suited for integration with a magnesium metal producer as a side product.

The common process for production of magnesium reacts a solution of magnesium chloride (as found in sea water and bitterns) with lime (calcium oxide) to produce magnesium hydroxide and calcium chloride which is generally flushed to sea. Magnesium hydroxide is simply heated to drive off the water leaving magnesium oxide. The magnesium oxide is then chlorinated by reaction with hydrochloric acid (while the chlorine later recovered and converted to hydrochloric acid. (A Russian process for magnesium metal uses magnesium chloride in an impure form), while there are other technologies as well (including silicothermic and carbothermic processes). See end of report. These should be evaluated with the production of magnesium oxide.


There are two producers in Australia. Causmag at the inland town of Young in New South Wales and Queensland Metals Corporation, through QMAG at Gladstone, Queensland both use magnesium carbonate deposits. The QMAG operation has one of the world’s largest magnesite (magnesium carbonate) deposits of 1.2 billion tonnes containing 500 million tonnes.

It is notable for its high purity. The typical composition is

CaO 0.8 to 1.1%

Fe2O3 60-100 ppm

MnO 200-900 ppm

NiO <30 ppm

B absent

The company has the capacity to produce up to 120 000 tonnes of deadburned magnesia, 30 000 tonnes of electrofused magnesia and 30 000 tonnes of calcined product (ie. derived from 180 000 tonnes of light-burned magnesia). The calcined product is used in mineral processing (notably nickel), pulp and paper, waste and water treatment, agricultural (stockfeed and fertiliser) and the tanning industries.

The company indicates in its annual report that "has also pursued the development of other magnesite related projects and minerals exploration activities".

It has pursued a joint venture with a French company for a flame retardant (Flamemag) which is a high purity magnesium hydroxide used in polymers (to replace halogenated chemicals).

Some 30 000 tonnes of calcined magnesia will be sold by QMAG to New Zealand Steel, BHP DRI and Cawse Nickel by late 1999, early 2000.

It is relevant to note that QMAG is trialing the production of calcined magnesia on a rotary kiln operated by Pacific Lime at Rockhampton signalling the similarity of magnesia and lime chemically and in their process of manufacture.

MgO production economics

The production of lime involves the high temperature heating of calcium carbonate to calcium oxide in a shaft furnace or rotary kiln - basically identical to the production of magnesium oxide. Depending on scale, between 4.5 and 6.5 gigajoule of gas is required per tonne of lime. While accurate figures are presently not available to the consultant, it is estimated that magnesium oxide would require about 9 to 10 gigajoule per tonne of MgO produced depending on the degree of calcination.

It is relevant to note that lime is supplied at around $100 to $120 per tonne ex-plant in Kwinana (at import parity). The energy requirement in its production is with gas at $2 per gigajoule, would represent only between 9 and 13 per cent of the selling price (but perhaps 50 per cent of the marginal cost of production).

Inquiries in W.A. has shown that MgO contracts are being negotiated at $310 to $320 per tonne delivered to the nickel producers Cawse and Comet Resources. Based on this, energy (including for the production of lime see later) would represent about 10 per cent of the delivered value of the product. This introduces the relevance of transport costs to estimate the ex-factory price.

Transport of MgO and advantage of Dampier over Gladstone

The cost of transporting MgO by tanker from Dampier to Port Hedland is estimated at $18 per tonne (eg. to BHP); to Kalgoorlie region; (Cawse) at $112 per tonne; and Ravensthorpe (Comet) at $157 per tonne. If tippers can be used (plastic lined), costs will reduce by $45 per tonne - so for Kalgoorlie to $67 and Ravensthorpe to $112 per tonne. Transport from Gladstone (Qmag) or Newcastle (for Causmag) would be cheapest by ship ($120 line haul) to Fremantle and then road freighted (ca. $60 for all three destinations with some allowance for backloading). The transport cost from Gladstone is therefore at $180 per tonne, between $68 (Ravensthorpe) and $162 per tonne (to Kalgoorlie) more expensive to transport than if supplied from a Dampier-located MgO manufacturer.

The significance of transport is clearly important when the cost of magnesia is being quoted as being ca $310 per tonne delivered implying a ex-plant (Qmag) cost as low as $130 per tonne (ie. comparable to lime prices in the Perth region).

On the $130 tonne ex-factory price, energy would represent 23 per cent of the ex-plant cost (perhaps one-half of the marginal cost of production).

Magnesium supply from Dampier Salt

Dampier salt produces bitterns containing between 300 000 and 350 000 tonnes of magnesium as Mg (in other words containing up to 24 per cent magnesium chloride - a chemical which contains 25.5 per cent magnesium).

Magnesium oxide demand

Magnesium oxide is required by BHP to about 8 000 tonnes pa when in full production of 2.5million tonnes; by Cawse about 5 000 tonnes increasing to 15 000 tonnes and planned by Comet Resources at Ravensthorpe to 20 000 tonnes. Murrin Murrin and Bulong use different technology avoiding MgO.

The grade of MgO accepted varies from 95 per cent for use by Cawse (allowing 4 per cent calcium oxide) to Comet seeking 98.5 per cent minimum and less than 1 per cent calcium oxide (eg. Qmag’s EMD45). Comet Resources, subject to raising $700m in funding, will come into production from third quarter of 2001. BHP requires a soft-burned form of MgO different to that required by the other two companies. The differences will present a small (say 5 per cent) cost penalty in production.

Production cost profile

This section is only indicative and a 100 000 tpa plant is nominated representing a mid-scale project. It could supply the W.A. market leaving two-thirds for exports. The cost of a plant dedicated to produce 100 000 tonnes pa of magnesium oxide is estimated at $60 million. More research is required to quantify but this would provide for separation and kiln facilities.

Key cost inputs per 100 000 tonne of calcined MgO produced per year.


Per unit cost


Total cost pa.

/tonne MgO



1 000 000 tonnes





140 000 tonnes

$11 200 000




10 gj/tonne.

$2 000 000


Capital recovery

@7.5% pa

$60 m

$4 500 000


Total material costs (before eg. labour)




$177 + #

# Low value - bitterns are currently returned to the sea with some sales as a dust suppressant for roads implying a negative value.

A low value for lime has been assumed given the scale of demand, a production profile suggests it could be produced at this cost per tonne.


Clearly the lime is a key cost contributor (representing some two-thirds of raw material costs). Given same technology to MgO production from the carbonate, it could be manufactured on-site requiring some 5 gigajoules per tonne of lime produced or, at the consumption ratio for MgO production, some 7 gigajoules per tonne of magnesium oxide produced.

Qmag has substantial advantage avoiding the use of lime which has a significance about the value of its freight disadvantage for supply to W.A. vis a vis a supplier in Dampier.

Magnesia could be produced at $200 per tonne which would make for competitive supply to some users in WA. On an international market, Qmag would be a lower cost supplier.

It is useful to note that the rationalisation of cement production by the acquisition of Rugby of AdBri, it is quite probable some surplus lime production facilities will become available in W.A. It is not clear whether a lime plant could be relocated competitively.

An interesting consideration for review is for the use of technology that produces magnesium oxide with hydrogen chloride, as a by-product, that could be used to liberate magnesium chloride from a known magnesium mineral deposit near Port Hedland. Another is for the chlorine liberated in the electrolytic magnesium production to be converted to hydrochloric acid (with hydrogen from syngas) and applied to that magnesium deposit.

The concept is novel and could be usefully evaluated.


This report is NOT indicative of our private client reports.

It has been prepared simply to get you started. It is cheaper than you will find on the web for other detailed information and thus we rarely recover the cost of its preparation.

Given the value of your time, if only a small percentage of the information proves useful, then it should have paid for itself. Feel free to discuss the preparation of a confidential custom report.


Chemlink Pty Ltd ABN 71 007 034 022. Tel 61 8 9294 3254 Publications 1997. All contents Copyright © 1997. All rights reserved. Information in this document is subject to change without notice. Products and companies referred to are trademarks or registered trademarks of their respective companies or mark holders. URL: