GEOLOGICAL SURVEY CIRCULAR 544

Determination of Gold in

Geologic Materials by

Solvent Extraction and

Atomic-Absorption

Spectrometry

Determination of Gold in

Geologic Materials by.

Solvent Extraction and

Atomic-Absorption

Spectrometry

By Claude Huffman, Jr., J. D. Mensik, and

L. B. Riley

Ｍｾ＠

GEOLOGICAL SURVEY CIRCULAR 544

Washington 1967

United States Department of the Interior

STEWART L. UDALL, Secretary

Geological Survey

William T. Pecora, Director

First printing 1967

Second printing 1 968

Free on application to the U.S. Geological Survey, Washington, D.C. 20242

Abstract ——————————-Introduction

—————————-Reagents——————————-

Apparatus——————————

Cyanide

atomic-absorption procedure ——

Sample preparation——————–Standardization

———- ————–

CONTENTS

Page

1

1

1

2

3

3

4

Fire-assay atomic-absorption procedure —

Results ———————————

Sample dissolution, cyanide procedure—–Extraction

of gold with methyl isobutyl

ketone——————————–

Conclusions – – – – – – – – – – – – – – – – – – – – – – — – – – -References——————————

ILLUSTRATIONS

Page

4

5

5

5

6

6

Page

Figure 1. Water bath equipped for aeration and constant temperature—————————– 2

2. Constant-temperature water bath with magnetic stirrer——————————– 3

TABLES

Page

Table 1. Comparison of gold values obtained by atomic-absorption methods and by the

conventional gravimetric fire-assay method —————————————- 5

2. Comparison of gold values obtained by fire-assay atomic-absorption and

cyanide atomic-absorption methods———————————————— ·6

III

Determination of Gold in Geologic Materials by

Solvent Extracti n and Atomic-Absorption Spectrometry

Huffman, Jr., J. D. Mens’ik, and L. B. Riley

ABSTRACT

The two methods presented for the deter ination of traces of

gold in geologic materials are the cyani e atomic-absorption

method and the fire-assay atomic-absorption ethod. In the cyanide

method gold is 1 each e d with a sodium c y n ide solution. The

monovalent gold is then oxidized to the trival nt state and concentrated

by extracting into methyl isobutyl keto e prior to estimation

by atomic absorption. In the fire-assay atomi -absorption method,

the gold-silver bead obtained from fire assay· s dissolved in nitric

and hydrochloric acids. Gold is then conce trated by extracting

into methyl isobutyl ketone prior to determ ·nation by atomic absorption.

By either method concentrations a low as 50 parts per

billion of gold can be determined in a 15-gra sample.

INTRODUCTION

The U.S. Geological Survey heav metals program

has instituted reconnaissance studie requiring rapid,

precise, and sensitive methods for de ermining various

elements in geologic materials. The determination of

gold for this program has required ethods more sensitive

than the conventional fire-ass y method. Lakin

and Nakagawa (1965) described a sp ctrophotometric

method for determining traces o gold in geologic

materials; they found, however (oral commun., 1966),

that thallium and iron interfere. Wals ( 1955) has shown

that atomic-absorption spectrometry is extremely well

suited for direct determination of low concentrations of

certain elements in solution, but until recently no work

seems to have been undertaken forth determination of

gold. Olson (1965) and Simmons (19 5) used a cyanide

solution to leach the gold from silice us carbonate ores

prior to estimation by atomic absorpt on. Tindall ( 1965)

described an atomic-absorption met od for the determination

of gold in various types of amples dealt with

in the copper mining industry. He u ed an aqua regia

digestion followed by an organic ext action toconcentrate

the gold into a small volume Strelow, Feast,

Mathews, andVanZyl(1966)applied natomic-absorption

method to the determination o gold in cyanide

waste solutions. All these atomic-a sorption methods

have been used primarily for deter:nining gold in the

parts-per-million range.

This paper describes two rapid ｡ｮ､ｰｲ･｣ｩｳＬｾｴｯｭﾭ

absorption methods for gold determination currently

used in the U.S. Geological Survey’s Denver laboratory:

the cyanide atomic-absorption method and ｴｨｾ＠ fire-assay

atomic-absorption method. In developing these

methods, sensitivity was stressed to enable g?.ochemical

prospecting for gold at theparts-per-billionlevel.

Both methods can detect concentrations as low as 50

ppb (parts per billion) or 0.0015 ounce per ton of gold

in a 15-gram sample.

In the cyanide method the sample is roasted and then

treated with a sodium cyanide solution to dis·solve the

gold. The cyanide solution is acidified with tydrobromic

acid and bromine water to oxidize the m’)novalent

gold to the trivalent state. The gold is then con7.entrated

by extracting into a measured volume of methyl isobutyl

ketone (MIBK). The gold concentration is determined by

an atomic -absorption spectrophotometer.

In the classical fire-assay method, the sil·rer ofthe

gold-silver bead obtained by cupellation is dissolved

and the gold residue is determined gravim<?.trically.

The sensitivity of this method is approximately 0.02

troy ounce per ton (600 ppb) on a 15-g (half assay ton)

sample. In the proposed fire-assay atomic-absorption

method the gold-silver bead from fire assay is dissolved

in nitric and hydrochloric acids anc then extracted

into a small volume of organic solve·1t (MIBK)

prior to estimation by atomic absorption. This procedure

thus increases the sensitivity by about a factor

of 10 when a 15-g sample is used. The sensitivity can

be further increased by combining beads obtained from

two fire-assay fusions of 30 g each of the same sample.

With this technique about 15 ppb can be determined.

This is about the limit imposed by reagent blanks.

REAGENTS

Standard gold stock solution, 400 ppm (parts per million)

gold. Dissolve 0.200 g of gold metal in a covered

beaker with 30 milliliters of aqua regia on a steam

bath, remove cover, and evaporate to dryress. Heat

on a hotplate until fumes from HN03 are no longer

present. Add 250 ml of ( 1:1) HCl to the beaker and

digest on the steam bath to dissolve the gold salts.

Transfer the solution to a 500-ml volumetric flask

and dilute to volume with distilled water.

Sodium cyanide solution, 0.25 percent. Dissolve 5 g

NaCN in 500 ml distilled water, add 2 drops of a

10-percent solution of NaOH, and dilute to 2,000 ml.

The pH of this solution should be about 11.

Hydrobromic acid, 48-percent reagent grade.

Bromine water, saturated solution.

Calcium oxide powder, reagent grade.

Methyl isobutyl ketone (MIBK).

r· Excess air outlet

Santomerse S solution, wetting agent, (

volume;volume in vrater, obtained fr01

Chemical Co., St. Louis, Mo.

APPARATUS

Atomic-absorption ｳｰ･ｾｴｲｯｨｭＮ＠ A

Elmer model 303 ato:-nic-absorption spet

meter is used with the following instrum

meters:

Wavelength: 2428 Angstroms.

Slit: 5 (3 millimeters, 20 Angstroms).

Source: Gold hollow cathode. 14 milliampc

rent.

——

Water ｬ･ｶｾＭ r

Figure 1.-Water bath equipped for aeration and constant temperature.

2

Acetylene: 9 pounds per square inch.

Air: 25 pounds per square inch.

Flame adjustments: With the burner lit, adjust the

flow meters to approximately 2.8 on the acety-

Hotplate and magnetic stirrer

Figure 2.-Constant-temperature water bath with magnetic

stirrer.

3

lene and 3.0 on the air while aspirating water.

Minor adjustments may be necessary to obtain

a blue flame when the MIBK is aspirated.

Water bath equipped for aeration and constant temperature

(fig. 1 ). The aeration device constructed in

this laboratory consists of a compressor-vacuum

pump capable of delivering air at one-third cubic

foot per minute under a pressure of 15 po·.mds per

square inch. This amount of air is more than sufficient

for the 20 outlets shown in figure 1. The

compressor supplies a group of air regulator valves

(available from aquarium supply houses) which are

connected to 3/ 16-inch-diameter Pyrex tubes by

Tygon tubing. The valves are used to individually

adjust the air supply to -each sample. A steady air

flow to the tubes is maintained by means of a small

air tank and one or more bleed-off valves.

Constant-temperature water bath with magnetic

stirrer (fig. 2). Teflon-coated ｾＭｩｮ｣ｨ､｡ｭ･ｴｲ＠

magnetic stirring bar.

Pyrex culture tubes, 25 by 200 millimeters, with polyseal

closure screw cap.

CYANIDE ATOMIC-ABSORPTION PROCED’JRE

SAMPLE PREPARATION

Crush sample in a jaw crusher to less than ¥!-inch

size, then split and grind a portion on ceramic plates

set for less than 80 mesh. The sample is no'” sieved;

a large portion of the sample will be less than 200

mesh.

1. Weigh a 2- to 15-g sample and transfer to a porcelain

dish or crucible. Place the crucible in a

cold muffle furnace and heat for 1 hour after the

temperature reaches 600°C. Do not exceed 650°C

because samples of high clay content may tend

to sinter and thus trap the goldandprevent solution

during the cyanide step.

2. Cool. Add 0.25 g CaO powder and mix thcroughly;

transfer sample to a Pyrex culture tube. Add 30

ml of the 0.25-percent NaCN solution to the tube,

place cap on the tube, and shake by hand to wet

the entire sample. Place tube in an automatic

shaker and shake for 15 minutes; let stc.nd overnight.

3. Reshake to break up any cake formed in ｴｨｾ＠ bottom

of the tube. Wash down sides of the tube with about

10 ml of the NaCN reagent. Add 4 dro’Js of the

0.25-percent Santomerse S wetting agent to the

tube and heat for 30 minutes in the water bath at

80° -85°C while aerating the sample (fig. 1 ).

4. If a 2 -g sample is used, proceed to step 5. If a

sample of more than 2 g is used, centrifuge the

contents of the tube at 2, 000 revolutions per minute

for 5 minutes, then decant the liquid phase

into another culture tube containing a magnetic

stirring bar. Wash the residue left in the first

tube with 10 ml of the 0.25-percent NaCN solution

containing 4 drops of saturated bromine water.

Centrifuge again and combine wash solution with

the original solution.

5. Add 5 drops bromine water and mix. Then, working

under a well-ventilated hood, cautious 1 y add 5

ml of 48-percent HBr. (CAUTION: Hydrogen cyanide

gas is evolved.) Heat in a water bath equipped

for magnetic stirring (fig. 2) for 30 minutes at

90°C,

6. Cool. Remove the stir bar with another magnet held

outside the tube, add 10 ml MIBK, and shake for 1

minute. Centrifuge if step 4 has been omitted.

7. Measure the absorption of the MIBK phase along

with reference standards (see following section,

“Standardization”) on an atomic-absorption spectrophotometer

at a wavelength of2428 Angstroms.

Use a water -saturated so 1 uti on of MIBK as a

reference blank. To obtain a standard calibration

curve, plot the absorption readings against gold

concentration expressed as parts per million in

the organic phase. The gold content of the sample

in parts per million is equal to:

gold in ppm in organic phase X 10 ml

weight of sample in grams

If the sample has an absorbance greater than the

lO ppm standard, a dilution of the organic layer

can be made. The dilution method is illustrated

in the following example: Transfer a 2-ml aliquot

of the organic phase to a tube containing about

30 ml of water and 2 ml of 48-percent HBr. Add

8 ml of MIBK, shake, and measure the absorption.

This dilution factor of 5 will handle a 10-g

sample containing as much as 50 ppm Au.

STANDARDIZATION

Prepare reference standards containing 0.1, 0.2,

0.4, 0.8, 1.0, 2.0, 4.0, 8.0, and 10.0 ppm Au in the

organic phase for working curves by adding appropriate

aliquots of a gold standard solution to a series of

culture tubes. Add 2 ml of 48-percent HBr to each and

adjust the volume to about 40 ml with water. Add 10.0

ml of MIBK and shake for 1 minute to extract the gold

into the organic phase. Measure the absorption of these

standards on the 1X or lOX scales of the atomica

b s or p t i o n spectrophotometer. P e r i o d i c a 11 y,

standards can, and probably should, be checked by

being subjected to the procedure starting with step 2,

However, we have noted no difference in absorption

readings between the standards taken through the entire

procedure and those prepared as described.

FIRE-ASSAY ATOMIC-ABSORPTION PROCEDURE

The fire-assay method involves a fusion of ground

ore with fluxes to form a slag phase and a metallic

lead phase. The lead collects any gold, platinum metals.,

or silver and settles out at the bortmn when poured

into a mold. After cooling., tl’le slag is broken away

from the lead button which is then put into the furnace

in a cupel. The lead melts and is slowly oxidized by

the hot air. The lead oxide wets the semiporous cupel

material, generally bone ash, and is absorbed. The

cupel material is not wet by the molten lead nor by the

noble metals, which remain as a bead after the gradual

absorption of the lead oxide.

4

The procedure for the further treat11ent of the bead

is as follows:

1. Transfer the silver-gold bead obtained by fire assay

to a 50-ml beaker.

2. Add 5 ml of (1:1) HN03, cover, and digest on a

steam bath for 45 minutes. Uncover and evaporate

the solution to about 2.5 ml.

3. Add 6 ml of concentrated HCl, cover, and let stand

20 minutes. Place beaker on steam bath and digest

for 30 minutes, uncover, and evaporate to hard

dryness.

4. Add 7 ml of concentrated HCl, cover, and digest on

the steam bath until all salts are in solution. Uncover,

add about 15 ml of water, and digest an

additional 30 minutes.

5. Cool. Quantitatively transfer the solution to a Pyrex

culture tube by washing beaker with about 10 ml

of water. Add 2 ml of 48-percerr: HBr, mix, add

10 ml of MIBK, and shake tube for 1 minute.

6. Determine the absorbance of the organic phase along

with reference standards prepared in the same

manner as described in the cyc.nide procedure.

We routinely add 2 milligramsofsilverinthe fusion

step. Precut lengths of lead rod, eac’1 containing this

amount of silver, are obtainable commercially under

the trade name Inquarts. Silver is important in this

procedure for three reasons: (1) It has a protective

effect and thus reduces gold losses c’uring the cupellation

step (Bugbee, 1940, p. 107), (2) it allows trace

amounts of gold to be easily transferred from the

cupel to a suitable container for analysis, and (3) the

addition of silver provides a silver-gold bead that is

easily dissolved, The ratio of silver to gold should be

greater than 3:1 to allow dissolutioncfthebead; beads

having lower ratios of silver to gold may be difficult

to dissolve, even in aqua regia. For ｳｾｭｰｬ･＠ that contain

sufficient silver, such addition is unnecessary;

conversely, for samples than contain considerable

gold, the addition of more than 2 rrilligrams silver

may be necessary to attain this 3:1 ratio. The silvergold

bead is routinely weighed to allow an estimate

or approximate check of the combined gold and silver

content of the ore,

For a full discussion of the fire-assay method the

reader is referred to Bugbee ( 1940) or Beamish ( 1966 ).

Table 1.-Comparison of gold values obtained by ｡ｴｯｭｩ｣ｾ｢ｳｲｰｮ＠ methods and by the conventional gravimetric

ｦｩｲ･ｾ｡ｳｹ＠ method

[Gravimetric fire-assay by 0, M. Parker and L, B. Riley, Atomic-absorption determinations by

Claude Huffman, Jr., J.D. Mensik, and L. B. Riley]

Gold (ppm)

Sample Atomic absorption

Sample description

Gravimetric and locality

fire assay Fire assay Cyanide

96130 _______________

4,8 4.8

0223180 ———— 10,3 11

Dll9999 ———— 13,7 16

D100874 ———— 15.8 18

Dll9971 ———— 16.5 15

D119956 ———— 26.4 27

RESULTS

The accuracy of the two proposed atomic-absorption

methods was first tested by analyzing a suite of six

selected samples previously an a 1 y zed by the conventional

gravimetric fire-assay method. The results

(table l) show good agreement between the two atomic

absorption methods and the gravimetric fire-assay

method. These samples were selected because they

represent a wide variety of rock and ore types as

indicated by semiquantitative spectrographic analysis.

For example, sample Dll9956 contains 1.5 percent As,

7 percent Fe, and 0.07 percent Zn; sample Dl19971

contains 710 percent As, lOpercent Fe, and 0.03 percent

Sb; sample Dll9999 contains> lO percent Fe, 5

percent Cu, and 0.2 percent Mn. About 90 percent of

sample Dl00874 is estimated to be pyrite.

The cyanide atomic-absorption method has been

thoroughly tested by routinely analyzing several hundred

samples of varying composition. All samples that

contained at least 0. 5 ppm Au were reanalyzed by the

fire-assay atomic-absorption method; in addition,

many randomly selected samples containing less than

0.5 ppm Au were checked. Table 2 presents some of

the data on 20 samples from Alaska and Nevada areas.

The cyanide and fire-assay methods show concordant

results.

SAMPLE DISSOLUTION, CYANIDE PROCEDURE

Simmons (1965) observed that low results were obtained

on some samples if the ignition step was omitted

prior to the cyanide treatment. We have confirmed his

observation and routinely roast all samples.

The cyanide process has been used commercially for

many years. The commercial processes are based on

5

4.3 Sulfide ore, White Pine mine,

Jefferson County, Mont.

10 Channel sample, Gold Anchor

mine, Clear Creek County, Colo.

15 Pyrite, Billy Goat mine, Okanogan

County, Wash.

14 Quartz pyrite ore, Phoenix mine,

Clear Creek County, Colo.

16 Quartz vein, Okanogan County,

Wash.

25 Mineralized quartzite, Okanogan

County, Wash.

the following chemical reaction:

4Au+8NaCN+2Hz0+02 -‘)oo4AuCN·NaCN+4NaOH.

The amount of gold which can be dissolved in sodium

cyanide solution depends on the amount of oxyr-en present

(Jacobson, 1949, p. 554). The aeraticndevice

described in this paper supplies the oxygen for the

preceding reaction and furnishes agitation to a large

sample. It is desirable to keep the sample ir a single

tube during the entire analysis. In the ir.itial stage of

this investigation, 2 -g samples were aeratec, conveniently

by means of magnetic stirring. However, when

we attempted to increase our sample size to 15 g we

found that the magnetic stirrer would not sufficiently

agitate the larger sample. The aeration device provides

sufficient agitation of a 15-g sample.

EXTRACTION OF GOLD WITH METHYL ISOBUTYL

KETONE

Methyl isobutyl ketone (MIBK) has been used extensively

in atomic-absorption work to preconcentrate

certain elements that show insufficient ｳ･ｮｾｩｴｶｹ＠ in

aqueous solution. Many authors favor this reagent

because it has excellent flame characteristics and it

is only slightly soluble in water ( 1. 9 g per 100 ml).

We wish to emphasize that the gold is in the monovalent

state after the cyanide treatment and will not be

extracted quantitatively into the MIBK unlesE’ oxidized

to the trivalent state. In thi’s method the dicyanogold

(I) ion is oxidized to the bromaurate (III) ion by the

addition of bromine water and hydrobromic acid. The

reaction is as follows:

ｎ｡ａｵＨｃｎＩｺＫＲｈｂｲＫｂＲｾｂＳＫＲｈｃｎＫｎＮ＠

The bromaurate (III) ion is then extracted quan’:itatively

into MIBK prior to estimation by atomic ab”‘orption.

Table 2.-Comparison of gold values obtained by fireassay

atomic-absorption and cyanide atomic-absorption

methods

[Analyses by Claude Huffman, Jr., J. D.

Mensik, and L. B. Riley1

Gold (ppm)

Sample

Sample

Fire Cya- description

assay nide

D124654 __ 0,05 0.05 Altered rock.

ABH775 __ . 09 .07 Altered diorite .

D124913 __ .09 . 1 Altered zone .

ABH780 __ .4 . 3 Iron -stained quartz .

D124670 __ . 7 ,8 Altered zone .

D124667 __ . 9 ,8 Quartz vein .

D124665 __ 1 1 Altered zone.

Dl24943 __ 2 2 Massive sulfide

vein.

ABH923 __ 2 2 Stained silicified

quartzite.

D124916 __ 3 3 Sulfide replacement

lenses.

ABH926 __ 4 4 Drusy quartz and

iron oxide,

D12493l__ 7 5 Altered zone and

quartz vein,

ABH871 __ 9 8 Gossan zone in

quartzite,

D124676 __ 12 13 Quartz vein and

altered zone,

0124675 __ 17 16 Quartz vein.

D124920 __ 27 27 Sulfide -bearing

quartz -calcite

vein.

ABH854 __ 44 40 Quartz, calcite,

sulfides.

D124679 __ 52 58 Altered zone.

ABH832 __ 85 73 Iron-stained quartz.

D124681 __ 350 330 Altered zone.

Test solutions containing 1, 10, 50, and 100 micrograms

of monovalent gold in 40 ml of cyanide solution

were oxidized by adding 5 drops of saturated bromine

water and 5 ml of 48-percent HBr, heated to expel

hydrogen cyanide gas, cooled, and extracted with 10 ml

of MIBK. All these test solutions showed by atomicabsorption

spectrophotometry a recovery of greater

than 95 percent in the organic phase.

Recently. Strelow, Feast, Mathews, and Van Zyl

ｾ＠ 1966) ｣ｾｮｦＱｲｭ･､＠ our observation that monovalent gold

m cyamde solution is not directly extractable with

MIBK; they used KMn04 in HCl solution to oxidize the

gold to the trivalent state.

Atomic-absorption methods for the determination of

gold in solution are reasonably free from interferences.

No effect from diverse ions has been noted in

6

this work. In the fire-assay bead method, silver is

partially extracted into the MIBK, but haS” no effect on

the gold determination. Strelow, Feast, Mathews, and

Van Zyl (1966) also observed no interferences; they

stated that Tl(lll), Cr(Vl), Mo(Vl), As(V), ｾｉｊＨｖ＠ ), Sn(IV),

Se(l V ), Te(IV), and Fe(III) are partially extracted but

cause no interference with the determination of gold

by atomic absorption.

CONCLUSIONS

The fire-assay separation combined with atomic

absorption has two advantages over the classical fireassay

method: ( 1) It eliminates the tedious and critical

steps of transferring, parting, annealing, and then

weighing trace amounts of gold and (2) it provides

greater accuracy and precision in low and fractional

parts-per-million range.

We employ the more rapid cyanide method for scanning

large numbers of samples for their gold content.

Samples that show a gold content of greater than 1 ppm

are then reanalyzed by the fire-assay atcmic-absorption

method to confirm the results.

REFERENCES

Beamish, F. E., 1966, The analytical chemistry of the

noble metals in International series of monographs

in an a 1 y tic a 1 chemistry [1st ed. J: New York,

Pergamon Press, v. 24. 609 p.

Bugbee, E. E., 1940, A textbook of fire assaying [3d

ed.]: New York, John Wiley &Sonrc,Inc.,314p.

Jacobson, C. A., ed., 1949, Encyclopedia of chemical

reactions: New York, Reinhold Put. Corp., v. 3,

800 p.

Lakin, H. W., and Nakagawa, H. M., 1965, A spectrophotometric

method for the de t e r m i nat ion of

traces of gold in geologic materials, in Geological

Survey research 1965: U.S. Geol. Survey Prof.

Paper 525-C, p. Cl68-C171.

Olson, A. M., 196.5, Gold assay by atomic absorption

spectrophotometry: Perkin-Elmer Atomic Absorption

Newsletter, v. 4, no. 5, p. 278-280.

Simmons, E. C., 1965, Gold assay by atomic absorption

spectrophotometry: Perkin- Elmer Atomic

Absorption Newsletter, v. 4, no. 2, p. 281-287.

Strelow, E. C .. Feast, P. M., Mathews, C. J. C., and

Van Zyl, C. R., 1966, Determination of gold in

cyanide waste solutions by solvent extraction and

atomic a b so r p t ion spectrophotometry: Anal.

Chemistry, v. 38, p. 115-117.

Tindall, F. M., 1965, Silver and gold assay by atomic

absorption spectrophotometry: Perkin-Elmer

Atomic Absorption Newsletter, v. 4, no. 9, p. 339-

340.

Walsh, Ao, 1955, The application of atomic absorption

spectra to chemical analysis: Spectrochim. Acta,

v. 7, p. 108-117.

RU, S. GOVERNMENT PRINTING OFFICE : 1968 0 – 285-957