Instructions for Large Scale Gold

Refining
By the Aqua Regia Acid Method
INTRODUCTION
It is known that the aqua-regia technique herein described is not the only method. It may
not be the best, and for the sake of experience electrolytic cells for refining precious
metals have been examined. These are probably very similar to the units used by many
large professional refiners. These electrolytic methods eliminate some of the problems of
the acid method described here, but they too have their own set of disadvantages and are
not as suitable for the small lots and needs of individual jewelry manufacturers.
The amounts which have been refined range from small lots of about 45 g (1 oz Troy)
of fine gold recovered from about l00 g (3 on Troy) of scrap, to more than 3 Kg (100 oz
Troy) from about 6 Kg (200 on Troy) of scrap. The latter is about the maximum that is
reasonably handled in the equipment herein described.
The usual refining lot in practice is 300-1000 g (10 to 30 oz Troy) of fine gold recovered
from 600-3000 g (20 to 100 oz Troy) of scrap.

OUTLINE OF THE PROCESS

The gold refining technique described here is the rather ancient wet chemical method
whereby the gold-bearing scrap is dissolved in aqua-regia. m is gold solution is then
filtered and the jewelers bench dirt, sandpaper grit, grinding wheel grains and similar
material remains on the filter as a solid sludge, together with any silver present which
will be in the form of silver chloride.
The filter and sludge are washed with repeated small amounts of water to wash all gold
chloride solution down through the filter. Other metals that were in the alloy or in the
scrap (nickel, zinc, copper, iron, etc.) are also in this solution which is usually green in
color. However, if no nickel or copper is present it will most likely be the characteristic
yellow of gold chloride.
The nitric acid from the excess of aqua-regia used in the digestion is removed either by
boiling or chemical reaction.
To recover the gold as metal a reducing chemical is added to selectively change the gold
chloride into solid gold particles and leave the other metal chlorides unchanged and in
solution. When tests show this to be complete the solution is filtered and the gold in the

filter thoroughly washed. The clean gold is then melted and poured into molds or made
into shot.
The acids used in the process are very corrosive and highly toxic fumes are produced.
Appendix No. 2 discusses safety precautions in the operation of the process and should
be carefully studied before any part of this work is undertaken.

REFINING PROCEDURE
Type of Scrap Considered and Preliminary Treatment
The gold scrap that is considered in this report is old jewelry and the material from
jewelry bench work, filings, clippings, scrap jewelry pieces, grinding wheel dust, casting
spills, sprues, strip pot sludge, etc. Such material has been found to contain from 20%
gold in fairly dirty bench scrap to more than 70% gold in pure strip pot sludge.
Experience indicates that most shops produce a bench scrap (lemel) that contains 30% to
40% gold.
The dust from floor sweepings or from polishing wheel vacuum collectors and similar
low-grade scrap requires extensive preliminary treatment which is not described here.
So-called "green gold" and some low carat white golds contain considerable silver and
are very difficult or impossible to dissolve in aqua-regia as an insoluble silver chloride
film is formed which prevents further action by the aqua-regia. Such golds or any high
silver alloy must be melted with several times their weight of copper or brass and shotted
to permit dissolution. (See later section on Gold Shot.)
If the scrap contains shellac, rubber wheel particles, rouge or similar material it is best
simmer it in lye (sodium hydroxide and water (a saturated solution) in a ratio of 10
volumes of lye/water to 1 volume of bench sweeps.
Massive pieces of metal take a very long time to digest in aqua-regia. Any such large
pieces should be shotted as described later. Strip pot sludge should be well washed with
water to remove cyanide residues before acid is added to it.
Mixing Aqua Regia
This and many of the operations described here should be carried out under an efficient
fume extraction hood. Details of the hood used in this work are given in Appendix No. 3.
Aqua regia is a combination of nitric acid and hydrochloric acid (the industrial grade of
hydrochloric is sometimes called muriatic acid), it is made by mixing 1 volume of
concentrated nitric acid with 4 volumes of concentrated hydrochloric acid.

If muriatic acid is used (it is usually less costly) the proportions are calculated to be: 1
nitric acid to 4.5 muriatic acid by volume. There are reasons to err on the side of using
more hydrochloric acid than theoretically necessary rather than too much nitric acid.
The precautions for mixing the acids are simple. Avoid splashes, protect eyes and work in
the open or under a fume hood. These acids mix quietly.
Both acids and especially hydrochloric emit acrid fumes. No heat is evolved when mixing
but the aqua-regia at once starts to emit chlorine gas, which evolves slowly for several
days. DO NOT STOPPER aqua-regia bottles. A closed aqua-regia vessel can develop
enough chlorine pressure to burst. Store in the open or in a fume hood.
The aqua-regia can be used immediately, days or weeks and probably months after
preparation.
Digesting the Scrap
The scrap gold is placed in the digesting vessel. Glass may be used for small batches.
Teflon plastic is also suitable for the strong oxidizing conditions of the aqua-regia and
Teflon will tolerate heating if done with care.
For many batches 6 liter Ehrlenmeyer flasks are used, or the glass jars illustrated are
suitable if heated with much care.
1 to 2 kilos of scrap material in a 6 liter Ehrlenmeyer flask is typical though up to as
much as 35 kilos have been
treated in these flasks. This quantity requires
considerable agitation and stirring to keep the
reaction going and is better done in a jar under a
fully ventilated fume hood.
The amount of aqua-regia required for a given batch
varies and depends on the proportion that is acid
soluble and the quality of the metal present. It has
been found that from 3.5 to 5 liters of aqua-regia are
required per kilo of scrap, most batches fall in the
range of 4 to 4.5 liters of aqua-regia per kilo of
scrap, in smaller amounts this is equivalent to 4 to
4.5 ml (cc) per-gram of scrap.
If the scrap is the usual filings and dust from
jewelers benches, the aqua-regia will react very
rapidly and may boil over, so the acid must be added
slowly and with care.

A slow flow or dropwise addition can be made from a bottle (with a bottom outlet) set on
a shelf above the flask. If the aqua-regia is several days old and is no longer producing
chlorine gas a siphon from a high container is also a convenient way of adding aqua-regia
slowly.
If the scrap is in the form of old jewelry or metal shot or other large pieces, the reaction
will be slower and a considerable amount of aqua-regia can just be poured onto the scrap.
Care is advised as the reaction is often quite slow to start and then after some minutes
becomes very, very active. The jar or flask may get quite hot which increases the reaction
speed.
When there is jewelry with diamonds, rubies, sapphires and similar acid-resistant gems
these can be left in place and recovered from the filter.
The reaction of the aqua-regia with the metals in the
scrap produces nitrogen oxides. Some of these are
red-brown in color, others are colorless but take up
oxygen as soon as they reach air and then turn redbrown .these fumes are acrid, choking and extremely
toxic; they dissolve quite easily in water and in
caustic solutions; they are heavier than air and the
aqua-regia digestion should be done under a good
fume hood.
The preferred practice is to add the aqua-regia to the
batch in two or three separate additions. Up to about
half or three-quarters of the expected amount of aquaregia is added and the mixture is allowed to stand for some time. Occasional agitation is
good, especially with finely divided material.
When brown fumes are no longer evolved and the bubbling of the solution is quiet a little
hydrochloric acid is usually added. Sometimes a further spurt of activity is seen, the
original hydrochloric acid having been depleted leaving some unused nitric acid
available, excess hydrochloric acid is not harmful. If no more brown fumes appear the
liquid is carefully poured off into a glass or plastic container, being careful not to disturb
the solid material in the reaction vessel.
More aqua-regia is then added as before and the cycle repeated until the addition of fresh
aqua-regia produces no reaction, i.e. brown fumes and bubbling.
Enough aqua-regia must be added to dissolve all of the gold, however the excess aquaregia that is required to accomplish this will later have to be removed so large excesses
should be avoided.
Toward the last the reaction is much slower and it is desirable to warm the solution and to
agitate it regularly, but the aqua-regia should not be heated to boiling. If heated too much

it will produce brown fumes merely because it is too hot, this wastes acid and obscures
the end of the solvent action.
The reaction also slows down near the end because of the amount of fine, sludge present
which tends to restrict the contact between aqua-regia and undissolved gold, so frequent
agitation is helpful.
When pieces of jewelry or larger pieces of metal are being dissolved it often seems that
the jewelry is not being attacked because it is still there in its original shape, however
such pieces usually crumble if crushed with a stirring rod. Most jewelry alloys contain
silver and the aqua-regia dissolves the gold and other alloying metals leaving insoluble
silver chloride as a residue in the original size and form. It is good to break these as there
may be a yet undissolved core that will dissolve more quickly if exposed.
When it is apparent that the reaction is complete the solution should be cooled to room
temperature.
Filtering
The aqua-regia now contains various metal chlorides (and perhaps nitrates) in solution
and insoluble silver chloride as well as a lot of unwanted material in the sludge, and this
mixture (when cooled) must be filtered. me reason for tooling is that silver chloride,
though quite insoluble in water, is slightly soluble in strong acids and this volubility is
lower in cold acids. Silver is probably the major non-gold constituent of gold refined by
this procedure; though few assays have been made, these have consistently shown
996/1000 gold.
The aqua-regia solutions are filtered with a Buchner filtering funnel and a 4 liter vacuum
filtering flask. Two sizes of funnel have been employed, a small one about 125 mm
diameter, and a larger one about 250 mm diameter.
Experience has shown that
the paper discs usually used
in these filters by chemists
tend to float away when the
filter is filled with liquid, but
coffee urn filters obtainable
from hotel and restaurant
supply shops have proved
very satisfactory. These
should be large enough to
line the bottom and sides of
the filter funnel, inserted dry,
wetted thoroughly with water
and firmly seated and
pressed into the corners to

avoid wrinkles and vacuum leaks. Two ply's of filter paper are used to help filtration and
avoid breakthrough.
Vacuum is produced by means of a water pump (aspirator) available from chemical
laboratory suppliers. Plastic, not metal pumps should be used, as the acid fumes from
aqua-regia filtering rapidly reduce the pumping ability of a metal pump. For the same
reason mechanical vacuum pumps are not recommended unless provided with efficient
acid vapor traps.
me filtrate is usually clean and clear. If, however, some solids come through at first the
filter should be stopped after a little while and the liquid poured back for refiltering.
Usually the liquid will then be clear and clean.
Filtering proceeds more rapidly if the clearest part of the aqua-regia is decanted onto the
filter first. When the sludge and solids get into the filter the process usually slows up. All
of the solids should be washed into the filter with a small stream of water, a wash bottle is
useful for this operation.
When the filtering is complete the paper and the sludge should be washer with repeated
small streams of water. This is to get all gold chloride solution out of the filter and sludge
into the liquid.
The filter paper now contains the unwanted material and also the silver chloride. It is
recommended that this should be dried and saved until at least a 30 gallon drumfull is
accumulated. The silver and any residual gold can be recovered separately or the material
sent to a refiner.
The filtered liquid is usually a rather handsome clear green color, due to nickel and
copper etc. If only gold chloride were present it would be yellowish.
The filtered solution is poured into a plastic container (plastic buckets or 5 gallon pickle
pails are suitable) for the next steps of eliminating excess nitric acid and precipitating the
gold.
Nitric Acid Elimination
The excess aqua-regia that was added to insure complete solution of all gold is, of course,
still in the solution at this stage and must be eliminated to allow the gold to be
precipitated.
The classic procedure for nitric acid elimination is repeated boiling to near dryness with
the addition of hydrochloric acid with some sulphuric acid near the end. This is a lengthy
and patience-trying process.
The best way is with urea. Add urea to the solution until there is no more fizzing.

Precipitating the Gold

The classic method of reducing gold chloride in solution to solid gold particles is to add
"copperas" to the solution. "Copperas" is an ancient name for ferrous sulphate, a rather
cheap chemical. A number of other chemicals will also 'reduce' the gold chloride but
Storm Precipitant (available from Shor) is better. In hot water, dissolve a weight of Storm
Precipitant equal to the weight of dissolved metal.
The precipitation of
gold can be seen as
a 'cloud' of particles
in the solution. The
end point of the
precipitation is
difficult to see,
some clues may be
noted in the density
of the 'cloud' of gold
particles. The
solution will be
clearer and
noticeably less
yellowish especially
if a drop is viewed
one white
chinaware surface.
m is is because the yellow gold chloride is gone and the green of the other chlorides
remains. Deliberate care during this gold precipitation work is advised. Observe the signs
and test the solution frequently to avoid large excesses of Storm Precipitant.
A rapid and satisfactory test for gold in solution is described in Appendix No. 4. sulphur.
Careful addition of Storm Precipitant and a slow approach to the end point can avoid this.
The sulphur dioxide odor, however, can be used as one of the signs that gold precipitation
is complete.
A problem that occurs when too much Storm Precipitant is added is that copper chloride,
which is very soluble in the cupric chloride (CuC12) form, is reduced by the excess
Storm Precipitant to the cuprous chloride (CuCl) state, which forms a white precipitate.
Limited experience with this contaminant has shown that it will reduce gold quality a
little and it will affect the color of the gold.
If cuprous chloride is present it will make the melting of the gold a memorable
experience. Dense clouds of choking white fumes will clear all persons out of the furnace
room quite quickly.

Some excess of Storm Precipitant is required to cause this undesirable side reaction, and
it is felt that the advantages of no boiling and little or no emission of brown fumes make
it worthwhile to use Storm Precipitant even though larger volumes of liquids are handled
and some care at the end point is needed.
If, through error, some cuprous chloride crystals are formed, they can be removed as
described later.
Filtration
Then the solution has been cleared of gold it should be allowed to stand for several hours.
Although gold is heavy and most settles quickly, some particles are very small and
require time to settle to the bottom.
Standing for a period, if possible overnight, facilitates the subsequent filtering operation.
In the interest of reducing the time and aggravation of filtering work the clear upper
portion of the barren solution can be decanted and only the bottom few inches near the
gold filtered. A simple siphon will remove the upper portion of the liquid quietly without
stirring up gold particles. The bottom few inches of liquid are then poured through the
filter keeping the gold largely in the pail.
The same Buchner filters and
the same kind of filter and
paper are used for gold as
previously used to filter
aqua-regia. me gold is
washed with repeated small
amounts of water until the
water coming through the
filter is quite clear and
colorless. The gold in the
pail is then just covered with
concentrated hydrochloric
acid and thoroughly stirred.
The acid is added to the filter
and the treatment repeated
several times, followed by repeated washing with water. This treatment will remove small
amounts of contaminants including cuprous chloride.
When the gold has been treated with hydrochloric acid and thoroughly washed it is then
ready to transfer into the filter.
A soft kitchen scraper helps move the gold into the filter and a small water jet is very
useful in clearing the last particles into the filter. Often there is gold adhering to the walls
of the pail, this can be scrubbed down with a stiff brush and washed into the filter with

the water jet. When the gold is all in the filter the vacuum should be run for a while to
get the gold as dry as possible.
Melting the Gold
Any gas or electric furnace normally used for gold melting or alloy production may be
used for melting the precipitated gold.
Crucible material is not critical and may be selected to suit the melting furnace.
Crucibles, however, should be clean and kept solely for melting pure gold.
An ordinary spoon is convenient to transfer the gold into the crucible and working over a
smooth clean surface so any spilled particles are not lost. When the filter is quite empty
of gold the paper and the last of the gold are transferred into the crucible. The paper
should be pressed down firmly.
When the gold is being melted the filter paper will burn and leave any adhering gold in
the crucible. It usually burns quite slowly because the furnace flame at gold melting
temperature does not have excess oxygen. With time it will burn away.
Fluxes are not usually needed, but if a surface film appears on the melt a very little borax
may be added. However, the filter paper can be burned out very quickly by adding a
small amount (say ~ teaspoonful or less) of sodium nitrate. When this is added to a hot
crucible the paper burns with an eye-dazzling flame.
Slag from such fluxes are often very liquid and cannot be easily skimmed off. If it is
necessary to skim, a graphite rod is used to dip out the slag as adhering lumps on the end
of the rod.
When the gold is well molten and 'quick' it can be poured into the mold. The mold should
be smoked with a sooty flame,` or sprayed with a silicone parting fluid. Mineral oil
coatings are sometimes recommence but they may discolor the gold.
The mold must be warm and dry. A cold would may have traces of moisture| and molten
metal poured onto traces of moisture can create the most amazing, costly and potentially
painful eruptions.

Making Gold Shot

If the gold is to be used for
alloying, making it is much
more convenient in the form of
shot rather than bars.
Various methods have been
employed for making shot or
grain. In this work the molten
gold is poured slowly into a
container of water which has a
fan-shaped jet of water passing
across the vessel about 1 inch
below the surface. Excess of
water flows out over the top of
the vessel.
To help slow the fall of the
metal and prevent
agglomeration of the shot a
sloping metal baffle is fitted.
The whole assembly is
illustrated to the right.