THIOSULFATE DEGRADATION DURING GOLD LEACHING IN AMMONIACAL THIOSULFATE SOLUTIONS : A FOCUS ON TRITHIONATE By N O E L E N E A H E R N B . S c , University of Natal, 1994 B.Sc. (Hons), University of Natal, 1995 M.A .Sc , University of Cape Town, 1997 A thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Faculty of Graduate Studies (Materials Engineering) The University of British Columbia October 2005 © Noelene Ahern, 2005 ABSTRACT Thiosul fate has shown cons iderab le promise a s an alternative to cyan ide for gold leach ing. However , one of the main limitations of the thiosulfate sys tem is the high consumpt ion of thiosulfate. B e s i d e s increasing the cost of the p rocess , the degradat ion products of thiosulfate have been c la imed to pass iva te gold sur faces and the polythionates often produced are loaded onto resins p roposed for gold recovery. T h e thiosulfate degradat ion p rocess is not completely unders tood. O f the degradat ion products, trithionate is a concern in the resin recovery of gold and is persistent in gold leach solut ions. Ve ry little is known about the expected behaviour of trithionate, both with respect to its formation and its interaction with other solut ion spec ies . T h e focus of this work w a s therefore to further the understanding of the behaviour of trithionate in gold leach solut ions. Exper imenta l work w a s carr ied out to determine the kinet ics of trithionate degradat ion in sys tems resembl ing gold leaching solut ions, and a kinetic mode l w a s der ived for trithionate degradat ion. T h e rate of degradat ion of trithionate in aqueous ammon iaca l solut ions w a s exp ressed by Equat ion 1. -d [S 3 0 6 2 - ] /d t = (k 3 [NH 4 + ] + k 2 [NH 3 ] + k , [ O H l + k 0 ) [S 3 O 6 2 T [1] where k 0 = 0.012 h" \ = 0.74 IvV.h;1, k 2 = 0.0049 M" 1 .h ' 1 , k 3 = 0.01 M " 1 . ^ 1 . In s o m e c a s e s , the p resence of lower concentrat ions of thiosulfate cata lysed the react ion while e x c e s s thiosulfate inhibited it. However , under typical gold leach ing condi t ions, thiosulfate w a s not expec ted to have a signif icant effect s o w a s exc luded from Equat ion 1. Cupr i c copper w a s not found to have any signif icant effect on the rate of trithionate degradat ion under the condit ions tested. Th is observed trithionate degradat ion rate equat ion w a s integrated with known kinetic behaviour of thiosulfate and tetrathionate based on literature f indings to deve lop an overal l mode l for the thiosulfate degradat ion and the result ing solut ion speciat ion of the sulfur oxyan ions in the a b s e n c e of ores . T h e model w a s evaluated against exper imenta l data and its shor tcomings were identif ied. ii The model parameters were adjusted to obtain a best fit to the experimental data. It was found that the best-fit parameters varied with the experimental conditions, indicating inadequacies in the model. The main concern was that the understanding of the thiosulfate degradation reactions is limited, and the way in which thiosulfate degradation was described had a major impact on the model output. In particular, the effects of copper species and pH on thiosulfate degradation have not been adequately addressed in the literature. Even taking into consideration the limitations of the model, based on the model output, decreasing the cupric concentration and increasing the ammonia concentration should help to minimise thiosulfate degradation. Solution recycle can also be used to minimise thiosulfate degradation but can result in a build up of trithionate. Limiting the reaction time would also be useful. This investigation has led to an improved understanding of the behaviour of trithionate in gold leach solutions and the model of the thiosulfate degradation system is a first step in developing a useful assessment method for thiosulfate degradation and solution speciation under gold leaching conditions. Further research is required to refine the model, particularly with respect to thiosulfate degradation to trithionate and tetrathionate. iii TABLE OF CONTENTS Abstract " Tab le of Con ten ts . . . . iv List of Tab les viii List of F igures x List of S y m b o l s and Abbrev iat ions xviii Acknow ledgemen ts xix 1. Introduction 1 2. Literature Review 3 2.1 Introduction 3 2.2 T h e ammon iaca l thiosulfate gold leaching sys tem 3 2.3 Structure and thermodynamic propert ies of the sulfur oxyan ions 5 2.4 Thiosul fate degradat ion in gold leaching - exper imenta l observat ions 12 2.5 Thiosul fate chemistry and fundamental s tudies 18 2.5.1 Oxidat ive degradat ion in gold leaching sys tems 18 2.5.2 Disproport ionat ion and reductive degradat ion of thiosulfate 32 2.5.3 Thiosul fate degradat ion inhibitors 33 2.6 Trithionate degradat ion 34 2.6.1 Interaction with water 35 2.6.2 Interaction with hydroxide 36 2.6.3 Interaction with ammon ia 36 2.6.4 Interaction with thiosulfate 36 2.6.5 Interaction with copper 37 2.7 Tetrathionate degradat ion 41 2.7.1 Interaction with hydroxide 41 2.7.2 Interaction with ammon ia 4 3 2.7.3 Interaction with copper 44 2.7.4 Interaction with thiosulfate 44 2.8 R e m o v a l of polythionates from solut ion 4 7 2.9 S u m m a r y of literature f indings 4 8 2.10 S c o p e and object ives 4 9 iv 3. Analytical Methods and Synthesis 50 3.1 Introduction 50 3.2 Ana l ys i s of sulfur oxyan ions - ion chromatography 50 3.2.1 Descr ipt ion of method 50 3.2.2 Stabil i ty of s tandard solut ions 51 3.2.3 Effect of other solut ion components on ion chromatographic ana lys is . 53 3.3 Ana lys i s of su l famate 54 3.4 Ana l ys i s of total a m m o n i a 55 3.5 Syn thes is of sod ium trithionate 55 3.6 Character isat ion of sod ium trithionate 56 3.6.1 Tota l sulfur 56 3.6.2 Total sod ium content 57 3.6.3 Volat i les 57 3.6.4 Sul fate 5 8 3.6.5 Polyth ionates 59 3.6.6 Overa l l trithionate purity 61 3.7 T r a c e impurity ana lys is of chemica ls u s e d 6 2 4. Kinetics of Trithionate Degradation - Methodology 6 3 4.1 Introduction 6 3 4 .2 Reac t ion kinet ics theory 6 3 4 .3 Exper imenta l method 64 4 .4 Da ta ana lys is 66 5. Kinetics of Trithionate Degradation - Results 72 5.1 Introduction 7 2 5.2 Stoichiometry 72 5.3 Reproducibi l i ty 74 5.4 Wate r 7 5 5.5 Hydrox ide 77 5.6 Ionic strength 79 5.7 Carbona te and bicarbonate 82 5.8 A m m o n i u m and ammon ia 83 v 5.9 pH 88 5.10 Thiosul fate 93 5.11 O x y g e n exc lus ion 95 5.12 Cupr i c copper 97 5.13 Tetrathionate 99 5.14 E lementa l sulfur, sulfate and copper powder 100 5.15 Tempera ture 101 6. Kinetics of Trithionate Degradation - Discussion and Modelling 104 6.1 Qual i tat ive d iscuss ion 104 6.2 Model l ing of trithionate degradat ion 108 7. Modelling of Sulfur Oxyanion Speciation During Thiosulfate Degradation.... 119 7.1 Introduction 119 7.2 Mode l setup 119 7.2.1 R1 - Thiosul fate degradat ion to tetrathionate 120 7.2.2 R 2 - Tetrathionate degradat ion 121 7.2.3 R 3 - Trithionate degradat ion 122 7.2.4 R 4 - Thiosul fate degradat ion to trithionate 123 7.2.5 R 5 - Thiosul fate degradat ion directly to sulfate 125 7.2.6 R 6 - Thiosul fate degradat ion to sulf ide 126 7.2.7 Incorporation of the rate equat ions into a mode l - method and constra ints. . 126 7.3 M o d e l sensit ivity to model parameters 127 7.3.1 Proport ion of thiosulfate forming tetrathionate ve rsus trithionate 129 7.3.2 Ra te of react ion R1 - thiosulfate degradat ion to tetrathionate 133 7.3.3 R a t e of react ion R 4 - thiosulfate degradat ion to trithionate 134 7.3.4 Ra te of react ion R 2 - tetrathionate degradat ion 135 7.3.5 Ra te of react ion R 3 - trithionate degradat ion 137 7.3.6 S u m m a r y - effect of model parameters 138 7.4 C o m p a r i s o n of model output with exper imental results in the a b s e n c e of ore 139 7.4.1 Exper iments 139 7.4.2 Val idat ion method 140 vi 7.4.3 Tes t 1 - no copper 142 7.4.4 Tes ts 2 a and 2b - low copper , pH 10 143 7.4.5 Tes t 3 - low copper , pH 9, low a m m o n i a 148 7.4.6 Tes t 4 - low copper , pH 9, higher a m m o n i a 151 7.4.7 Tes t 5 - h i g h copper , pH 10 152 7.4.8 S u m m a r y 155 7.5 C o m p a r i s o n of exper imental results with and without ore 159 7.6 S c o p e of use of model 163 7.7 Impact of solut ion condit ions 163 7.7.1 C o p p e r concentrat ion 164 7.7.2 Total ammon ia concentrat ion 167 7.7.3 pH 170 7.7.4 D isso lved oxygen 172 7.7.5 Thiosul fate concentrat ion 173 8. Conclusions 180 9. Recommendations 187 10. References 189 Append ix 1: The rmodynamic va lues used to construct E h - p H d iagrams 199 Append ix 2: Thiosul fate degradat ion - literature data 200 Append ix 3: Determinat ion of sulfur oxyan ions by ion chromatography 205 Append ix 4 : C h e m i c a l impurity ana lys is 208 vii LIST O F T A B L E S 2.1 Se lec ted G i b b s free energ ies of formation (298 K) for spec ies of interest in the copper-ammoniacal - th iosul fa te sys tem 11 2.2 Se lec ted standard oxidation potentials (298 K) 11 2.3 S u m m a r y of exper imental observat ions for thiosulfate degradat ion 13 2.4 Literature data for trithionate and tetrathionate product ion 16 2.5 Format ion of trithionate from thiosulfate in the p resence of oxygen (oxygen pressure 725 mm Hg , 30 °C, [Cu(ll)] = 1 m M , [NH 3 ] = 0.2 M) 26 2.6 Trithionate degradat ion in water 38 2.7 Tri thionate degradat ion in ammon ia 39 2.8 Tri thionate degradat ion in the p resence of thiosulfate 4 0 2.9 Tetrathionate degradat ion in alkal ine solut ions 46 2.10 Potent ia l methods for polythionate removal 4 7 3.1 S u m m a r y of observat ions on stability of s tandard solut ions of thiosulfate, trithionate and/or tetrathionate 52 3.2 Effect of added spec ies on ana lys is of sulfur oxyan ions 54 3.3 S o d i u m trithionate character isat ion 61 5.1 Reproducibi l i ty in rate constant determinat ion for trithionate degradat ion for two buffer sys tems using the initial rate method 75 5.2 V a l u e s for the observed rate constant k 0 b S for trithionate degradat ion in water us ing the initial rate method (40 °C) 76 5.3 Effect of hydroxide concentrat ion on the observed rate constant k o b s for trithionate degradat ion using the initial rate method (40 °C) 77 5.4 V a l u e s for the observed rate constant k o b s for trithionate degradat ion in the p resence of var ious salts used to adjust the ionic strength using the initial rate method 80 5.5 Effect of ammon ium concentrat ion on observed rate constant k o b s for trithionate degradat ion 88 5.6 Effect of thiosulfate on the observed rate constant k 0 b S for trithionate degradat ion using the integrated rate method 94 viii 5.7 Effect of limiting d isso lved oxygen on the observed rate constant k o b s for trithionate degradat ion in 0.1 M hydroxide solut ion 95 5.8 Effect of limiting d isso lved oxygen on the observed rate constant k o b s for trithionate degradat ion ammon ia / ammon ium bicarbonate solut ion 96 5.9 Effect of limiting oxygen on the observed rate constant k o b s for trithionate degradat ion using the integrated rate method 97 5.10 Effect of cupr ic addit ion on the observed rate constant k 0 b S for trithionate degradat ion using the integrated rate method 97 5.11 Effect of cupr ic copper on the observed rate constant k 0 b S for trithionate degradat ion 98 5.12 Effect of e lementa l sulfur, sulfate and copper powder on the observed rate constant k o b s for trithionate degradat ion 100 5.13 Effect of temperature on the observed rate constant k 0 b S for trithionate degradat ion at pH 8.8 - 10.1 102 5.14 Effect of temperature on the observed rate constant k o b s for trithionate degradat ion using the integrated rate method 103 7.1 Mode l parameters used to test model sensit ivity 128 7.2 S tandard exper imental parameters used in model l ing 129 7.3 Exper imenta l condit ions used in tests for model val idat ion 140 7.4 Adjustment of model parameters found to give improved agreement between model output and exper imenta l data 156 7.5 A s s a y results for ore from P lace r D o m e used in leach tests 159 7.6 Effect of recycl ing on thiosulfate consumpt ion and solut ion speciat ion (Initial condit ions 0.2 M S 2 0 3 2 * , 30 mg/l C u , 0.4 M N H 3 , 25 °C, p H 10, with model parameters set as in Tab le 7.2 for Tes t 2) . . . .178 7.7 Effect of recycl ing on thiosulfate consumpt ion and solut ion spec ia t ion (Initial condit ions 0.2 M S 2 0 3 2 " , 100 mg/l C u , 0.3 M N H 3 , 25 °C, pH 10, with model parameters set a s in Tab le 7.2 for Tes t 5) . . . . 179 ix LIST OF FIGURES 2.1 E h - p H d iagram for the aqueous sulfur sys tem (1M sulfur, 25 °C) 6 2.2 B a s i c structures of se lec ted sulfur oxyan ions 7 2.3 E h - p H d iagram for the aqueous sulfur sys tem with the fol lowing spec ies omitted: S 0 4 2 " , H S 0 4 " , H 2 S 0 4 . x H 2 0 (1M sulfur, 25 °C) 9 2.4 E h - p H d iagram for the aqueous sulfur sys tem with the fol lowing spec ies omitted: S 0 4 2 " , H S 0 4 " , H 2 S 0 4 . x H 2 0 , S 0 3 , S 0 3 2 " , H S 0 3 " , S 2 0 6 2 " , HSOsf , S 2 0 8 2 - , S 2 0 5 2 ' (1M sulfur, 25 °C) 10 2.5 B imolecu lar nuc lear substitution at the sulfenyl sulfur of trithionate 35 3.1 T G A / D T A for sod ium trithionate batch 2 58 3.2 T G A / D T A for sod ium trithionate batch 3 58 3.3 C h a n g e of measu red trithionate (indicative va lues only) and tetrathionate concentrat ions with t ime with and without us ing deaera ted water in solution preparat ion 60 3.4 C h a n g e of measured trithionate (indicative va lues only) and tetrathionate concentrat ions with t ime in the p resence of 2 m M N H 4 O H 60 3.5 C h a n g e of measu red trithionate (indicative va lues only), tetrathionate and thiosulfate concentrat ions with time in the p resence of 10 mg/l S 2 0 3 2 " and 10 mg/l S 4 0 6 2 " 61 4.1 Concent ra t ions of trithionate and thiosulfate with time for an integrated rate method test 67 4.2 Determinat ion of react ion stoichiometry us ing the integrated rate method 68 4.3 Typ ica l concentrat ion profile for trithionate and thiosulfate us ing the initial rate method 69 4.4 Determinat ion of the reaction order for the rate of trithionate degradat ion with respect to trithionate 70 4 .5 Typ ica l plot of trithionate degradat ion rate ve rsus initial trithionate concentrat ion to determine the rate constant k 0 b S f rom the s lope 71 x 5.1 Concentrat ion profi les for trithionate and thiosulfate 74 5.2 Obse rved rate constant for trithionate degradat ion k 0 b S ve rsus [OHT for pH<11 79 5.3 Effect of ionic strength on observed rate constant k o b s for trithionate degradat ion using the initial rate method 82 5.4 Effect of ionic strength of N a H C 0 3 / N a 2 C 0 3 buffer on the observed rate constant k o b s for trithionate degradat ion us ing the initial rate method 83 5.5 Effect of ammon ium concentrat ion on the observed rate constant k 0 b S for trithionate degradat ion using the initial rate method 84 5.6 Effect of ammon ia concentrat ion on the observed rate constant k o b s for trithionate degradat ion using the initial rate method 85 5.7 Effect of ammon ium concentrat ion on the observed rate constant k 0 b S for trithionate degradat ion for ( N H 4 ) 2 S 0 4 / N H 3 and N H 4 H C 0 3 / N H 3 buffer sys tems using the initial rate method 87 5.8 Dependency of the observed rate constant k o b s for trithionate degradat ion on the ammon ium concentrat ion at var ious pH 89 5.9 Dependency of the observed rate constant k o b s for trithionate degradat ion on the ammon ia concentrat ion at var ious pH 89 5.10 O b s e r v e d rate constant k o b s for trithionate degradat ion ve rsus p H for a range of ammon ia / ammon ium concentrat ions 91 5.11 O b s e r v e d rate constant k o b s for trithionate degradat ion ve rsus pH a t 2 5 ° C 91 5.12 Effect of p H on observed rate constant k o b s for trithionate degradat ion in sod ium carbonate / b icarbonate med ium 92 5.13 Effect of thiosulfate on the observed rate constant k o b s for trithionate degradat ion using the initial rate method 93 5.14 Effect of sulfur, sulfate or copper powder on trithionate concentrat ion profile 101 6.1 Concent ra t ion profi les for ammon ia , ammon ium ions and hydroxide ions at 40 °C with varying p H , using arbitrary concentrat ion units 106 xi 6.2 Plot of observed rate constant k o b s for trithionate degradat ion at 40 °C for an ammon ia + ammon ium concentrat ion of 0 M against p H , with model led trend super imposed (using k2= 0.0081 M" 1.h" 1) 111 6.3 Plot of observed rate constant k o b s for trithionate degradat ion at 4 0 °C for an ammon ia + ammon ium concentrat ion of 0.5 M against p H , with model led trend super imposed (using k2= 0.0081 M" 1.h" 1) 112 6.4 Plot of observed rate constant k o b s for trithionate degradat ion at 40 °C for an ammon ia + ammon ium concentrat ion of 0.9 - 1.2 M against p H , with model led trend super imposed (using k2 = 0.0081 M- 1.h" 1) 112 6.5 Plot of ca lcu la ted rate constant kcaic ve rsus observed rate constant k o b s for trithionate degradat ion at 4 0 °C for all da ta , us ing k2= 0.0081 M- 1 .h- 1 113 6.6 Plot of observed rate constant k o b s for trithionate degradat ion at 40 °C for an ammon ia + ammon ium concentrat ion of 0 M against p H , with model led trend super imposed (using k2= 0 .0049 M" 1.h" 1) 114 6.7 Plot of observed rate constant k o b s for trithionate degradat ion at 4 0 °C for an ammon ia + ammon ium concentrat ion of 0.1 - 0.2 M against p H , with model led trend super imposed (using k2 = 0.0049 M' 1 .h ' 1 ) 114 6.8 Plot of observed rate constant k o b s for trithionate degradat ion at 40 °C for an ammon ia + ammon ium concentrat ion of 0.3 - 0.4 M against p H , with model led trend super imposed (using k2 = 0.0049 IvrVh-1) 115 6.9 Plot of observed rate constant k o b s for trithionate degradat ion at 4 0 °C for an ammon ia + ammon ium concentrat ion of 0.5 M against p H , with model led trend super imposed (using k2 = 0.0049 IVr1.ir1) 115 6.10 Plot of observed rate constant k o b s for trithionate degradat ion at 4 0 °C for an ammon ia + ammon ium concentrat ion of 0.7 - 0.8 M against p H , with model led trend super imposed (using k2 = 0.0049 M- 1 . lr 1) 116 xii 6.11 Plot of observed rate constant kobs for trithionate degradation at 40 °C for an ammonia + ammonium concentration of 0.9 - 1.2 M against pH, with modelled trend superimposed (using k2 = 0.0049 M-1.h-1) 116 6.12 Plot of observed rate constant k0bS for trithionate degradation at 40 °C for an ammonia + ammonium concentration of 1.8 - 2.6 M against pH, with modelled trend superimposed (using k2 = 0.0049 lvr1.h-1) 117 6.13 Plot of calculated rate constant k c ai c versus observed rate constant k 0bS for trithionate degradation at 40 °C for all data, using k2= 0.0049 M'Vh"1 117 7.1 Schematic showing the basis for modelling 120 7.2 Logarithm of the oxygen consumption rate versus the logarithm of ammonia concentration for thiosulfate degradation to trithionate in the presence of oxygen 124 7.3 Logarithm of the oxygen consumption rate versus the logarithm of thiosulfate concentration for thiosulfate degradation to trithionate in the presence of oxygen 124 7.4 Modelled output for sulfur oxyanion speciation during thiosulfate degradation. Assumed all thiosulfate degradation is via Reaction R1 130 7.5 Modelled output for sulfur oxyanion speciation during thiosulfate degradation. Assumed all thiosulfate degradation is via Reaction R4 131 7.6 Modelled output for sulfur oxyanion speciation during thiosulfate degradation. Assumed thiosulfate degradation is via Reaction R1 and Reaction R4 (a = 50 %, b = 50 %) 132 7.7 Modelled output for sulfur oxyanion speciation during thiosulfate degradation. Assumed thiosulfate degradation is via Reaction R1 and Reaction R4 (a = 80 %, b = 20 %) 133 7.8 Modelled output for sulfur oxyanion speciation during thiosulfate degradation - kR1 increased by a factor 10 134 xiii 7.9 Mode l led output for sulfur oxyanion spec ia t ion during thiosulfate degradat ion - k R 4 i nc reased by a factor 10 135 7.10 Mode l led output for sulfur oxyanion spec ia t ion during 5 thiosulfate degradat ion - k R 2 -o inc reased by a factor 10 136 7.11 Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion - k R 2 - i i nc reased by a factor 10 137 7.12 Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion - k 0, ku k 2 and k 3 i nc reased by a factor 10 138 7.13 Mode l output ve rsus exper imental data for sulfur oxyan ion speciat ion during thiosulfate degradat ion 143 7.14 Mode l output ve rsus exper imenta l data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k R 2 - 1 x, k R 3 - 1 x, b k R 4 - 0 x 144 7.15 Mode l output versus exper imental data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 10 x, k R 2 - 1 x, k R 3 - 1 x, b k R 4 - 0 x 145 7.16 Mode l output ve rsus exper imenta l data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 14 x, k R 2 - 8 x, k R 3 - 1 x, b k R 4 - 0 x 146 7.17 Mode l output ve rsus exper imenta l data for sulfur oxyan ion spec ia t ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 15 x, k ^ - 8 x, k R 3 - 5 x, b k R 4 - 0 . 0 5 x 147 7.18 Mode l output versus exper imental data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k R 2 - 1 x, k R 3 - 1 x, b k R 4 - 0 x 149 7.19 Mode l output ve rsus exper imenta l data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k ^ - 10 x, k R 3 - 1 x, b k R 4 - 0 x 150 xiv 7.20 Mode l output ve rsus exper imenta l data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, - 10 x, k R 3 - 5 x, b k R 4 - 0 . 1 x 151 7.21 Mode l output ve rsus exper imenta l data for sulfur oxyan ion spec ia t ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1.5 x, k R 2 - 10 x, k R 3 - 5 x, b k R 4 - 0 x 152 7.22 Mode l output ve rsus exper imental data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 16 x, k ^ - 8 x, k R 3 - 1 x, b k R 4 - 0 x 153 7.23 Mode l output ve rsus exper imenta l data for sulfur oxyan ion ^ spec ia t ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 12 x, k R 2 - 8 x, k R 3 - 1 x, b k R 4 - 0 . 1 5 x 154 7.24 Mode l output versus exper imental data for sulfur oxyan ion spec ia t ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 12 x, kp^ - 8 x, k R 3 - 5 x, b k R 4 - 0 . 1 7 x 155 7.25 Mode l output ve rsus exper imental data for sulfur oxyan ion speciat ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 15 x, k R 2 - 8 x, k R 3 - 5 x, b k R 4 - 0 . 0 5 x 161 7.26 Mode l output ve rsus exper imenta l data for sulfur oxyan ion spec ia t ion during thiosulfate degradat ion. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 12 x, k ^ - 8 x, k R 3 - 5 x, b k R 4 - 0 . 1 7 x 162 7.27 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at vary ing copper concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 15 x, k R 2 - 8 x, k R 3 - 5 x, b k R 4 - 0 . 0 5 x 165 xv 7.28 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at varying copper concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k R 2 - 10 x, k R 3 - 5 x, b k R 4 - 0.1 X . . . . 1 6 6 7.29 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at varying copper concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 12 x, k R 2 - 8 x, k R 3 - 5 x, b k R 4 - 0 . 1 7 x 167 7.30 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at vary ing ammon ia concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 15 x, k R 2 - 8 x, k R 3 - 5 x, b k R 4 - 0 . 0 5 x 168 7.31 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at varying ammon ia concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k R 2 - 10 x, k R 3 - 5 x, b k R 4 - 0.1 x . . . 169 7.32 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at varying ammon ia concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 12 x, k R 2 - 8 x, k R 3 - 5 x, b k R 4 - 0 . 1 7 x 170 7.33 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at vary ing p H . Mode l parameters adjusted by the fol lowing factors: a k R 1 - 15 x, k R 2 - 8 x, k R 3 - 5 x, b k R 4 - 0.05 x 171 7.34 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at varying p H . Mode l parameters adjusted by the fol lowing factors: a k R i - 1 x, kpu - 10 x, k R 3 - 5 x, b k R 4 - 0.1 x 172 7.35 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at varying initial thiosulfate concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 15 x, - 8 x, k R 3 - 5 x, b k R 4 - 0 . 0 5 x 173 7.36 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at vary ing initial thiosulfate concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k R 2 - 10 x, k R 3 - 5 x, b k R 4 - 0 . 1 x 174 xvi 7.37 Mode l led thiosulfate and trithionate concentrat ion after 24 hours at varying initial thiosulfate concentrat ions. Mode l parameters adjusted by the fol lowing factors: a k R i - 12 x, k R 2 - 8 x, k R 3 - 5 x, b k R 4 - 0 . 1 7 x 175 7.38 Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion over 7 days 176 7.39 Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion with thiosulfate rep len ished every 24 hours 177 8.1 Typ ica l mode l output against exper imenta l data 183 xvii LIST OF SYMBOLS AND ABBREVIATIONS D0 2 Dissolved oxygen E° Standard potential, volts E h Redox potential, volts E a Activation energy, kJ/mol F Faraday's constant, 96500 C/mol A G 0 Standard free energy of formation, kJ/mol HPLC High performance liquid chromatography ICP Inductively coupled plasma spectrophotometry I Ionic strength, mol/l k Rate constant M Molarity, mol/l n Number of electrons RT Room temperature a Standard deviation SHE Standard hydrogen electrode t Time X V l l l ACKNOWLEDGEMENTS I w ish to acknowledge support of this project by P l a c e r D o m e , Ang logo ld - Ashan t i , T e c k C o m i n c o and Barr ick G o l d , a s wel l a s the Nat ional S c i e n c e and Eng ineer ing R e s e a r c h Counc i l ( N S E R C ) . I would like to thank my superv isors David Dre is inger and G u s V a n Weer t for their insightful adv ice . T h a n k s to the Hydro group, espec ia l ly Ani ta L a m and Berend W a s s i n k , and all the staff in the Mater ia ls Engineer ing Department who have made life eas ie r for me . Thank you to Dr M ichae l Wol f for the loan of the cool ing reactor. I thank my parents for their encouragement , in this and all the other endeavours I have under taken. I a m espec ia l ly grateful to my husband Chr is who knew that I wanted to go back to university s o m e day and actual ly got me to do someth ing about it! H e is my inspirat ion. xix 1. INTRODUCTION Cyanidat ion is a wel l es tab l ished and effective p rocess for gold recovery from ores and concent ra tes. However , there are increasing environmental conce rns pertaining to the use of cyan ide , and it is not sui table for certain ore types, particularly copper containing ores and ca rbonaceous refractory ores where cyan ide consumpt ion can b e c o m e uneconomica l and/or gold recovery low. Thiosul fate has shown cons iderab le promise a s an alternative to cyan ide for gold leach ing. However , one of the main l imitations of the thiosulfate sys tem is the high consumpt ion of thiosulfate and the lack of robustness in its applicabil i ty to a large variety of ores. B e s i d e s the fact that high reagent consumpt ion inc reases the cost of the p rocess , the degradat ion products of thiosulfate have been c la imed to pass iva te gold sur faces and the polythionates often produced are loaded onto resins p roposed for gold recovery (Muir and Ay lmore , 2002). Degradat ion products may a lso facil itate the precipitation of go ld , copper and si lver sul f ides, and insufficient thiosulfate lixiviant c a n c a u s e the precipitation of metal l ic gold and silver. Whi le thiosulfate consumpt ion is usual ly monitored and reported, the degradat ion p rocess is not complete ly unders tood. It is known that trithionate, tetrathionate and sulfate are common ly present in thiosulfate leach solut ions a s degradat ion products, but the factors inf luencing the formation and interactions of these spec ies are not c lear. Sul fate is a thermodynamical ly stable degradat ion product, but trithionate and tetrathionate are meta-stable. Tetrathionate is less stable at the alkal ine pH va lues used in gold leaching but trithionate is persistent in gold leach solut ions. A l s o , of the spec ies of interest, very little is known about the expec ted behaviour of trithionate, both with respect to its formation and its interaction with other solut ion spec ies . T h e focus of this work w a s therefore to further the understanding of the behaviour of trithionate in gold leach solut ions. Exper imenta l work w a s carr ied out to determine the kinet ics of trithionate degradat ion in sys tems resembl ing gold leaching solut ions, and a kinetic model w a s der ived for trithionate degradat ion. Th is model w a s integrated with known kinetic behaviour of thiosulfate and tetrathionate based on literature f indings to deve lop an overal l mode l for the thiosulfate degradat ion and the result ing solut ion 1 speciat ion of the sulfur oxyanions . The model was evaluated against experimental data and its shortcomings were identified. In this thesis , a literature review of thiosulfate degradation is g iven in Chapter 2. Analyt ica l and synthes is methods used are d i scussed in Chapter 3. The methodology, results and d iscuss ion and modell ing of trithionate degradation kinetics are d i scussed in Chapters 4, 5 and 6 respectively . In Chapter 7, a model of thiosulfate degradation and the resulting solution speciat ion is set up, based on literature data and on the findings in this work. Conc lus ions and recommendat ions resulting from this work are given in Chapters 8 and 9 respectively . 2 2. LITERATURE REVIEW 2.1 INTRODUCTION In this review, thiosulfate degradat ion in gold leaching sys tems and the stability of tetrathionate and particularly trithionate in aqueous sys tems are d i s c u s s e d to consol idate the current understanding of sulfur oxyan ion behaviour under condi t ions relevant to gold leaching. After a genera l introduction to the prob lem of thiosulfate degradat ion in gold leach ing, s o m e genera l structural and thermodynamic propert ies of the sulfur oxyan ions are summar i zed . Next, the thiosulfate degradat ion and formation of other sulfur oxyan ions observed exper imental ly in gold leaching sys tems are d i s c u s s e d . Th is is fol lowed by more detai led d iscuss ions on the chemistry of thiosulfate, trithionate and tetrathionate in sys tems relevant to gold leaching. Final ly, the overal l re levance of the avai lable literature to the understanding of the sulfur chemistry in gold leaching and the a reas of shor tcoming are d i scussed to justify the work carr ied out in this thesis. 2.2 THE AMMONIACAL THIOSULFATE GOLD LEACHING SYSTEM Cyan ide has b e c o m e the industry s tandard lixiviant for go ld . Cyan idat ion is a robust and wel l understood p rocess . However , the ineffect iveness of cyan ide for gold recovery from preg-robbing ores and the high cyan ide consumpt ions exper ienced with copper contain ing ores has lead to a search for an alternative lixiviant. C y a n i d e a lso has a negat ive publ ic image with respect to envi ronmental and safety concerns . However , there is an industry commitment of both gold producers and cyan ide manufacturers to the International C y a n i d e Managemen t C o d e ( E M J , 2004 , Mining M a g a z i n e , 2004) that promotes best pract ice in the use and managemen t of cyan ide , exceed ing the requirements of most governments and regulatory agenc ies . E v e n s o , there remain envi ronmental conce rns with the use of cyan ide, though these may be founded more on publ ic opinion than scienti f ic fact. Thiosul fate has been shown to be a promising alternative lixiviant for gold and si lver recovery and much research has been devoted to the ammon iaca l thiosulfate sys tem in recent years . Thiosul fate w a s first used in prec ious metals recovery in the 1 9 t h century in the Pa te ra P r o c e s s for si lver recovery (Mo l leman, 1998). A n a tmospher ic ammon iaca l thiosulfate 3 l each to recover gold and si lver f rom ammon iaca l oxidat ive p ressure leach res idues of copper sulf ide concentra tes w a s deve loped in 1978 (Berezowsky and Gorme ly , 1978), renewing interest in the thiosulfate lixiviant. S i n c e then, industrial and a c a d e m i c interest in the thiosulfate leaching sys tem has grown t remendously , with the main focus on f inding ways to e n h a n c e gold leaching and min imise degradat ion of the thiosulfate reagent. T o leach go ld , a sui table oxidant and a sui table complex ing lixiviant is required. Us ing thiosulfate a s a lixiviant in ammon iaca l solut ions, it has been found that the p resence of copper as an oxidant greatly enhanced the rate of react ion (Aylmore and Muir, 2001a) . T h e chemistry of the sys tem is complex and not yet fully unders tood. A simpl i f ied representat ion of the gold leaching sys tem is g iven in Equat ions 2.1 and 2.2, with the overal l react ion g iven in Equat ion 2.3 (Abruzzese et a l . , 1995, W a n 1997). Cupr i c copper or more speci f ical ly cupr ic tet raammine behaves a s the oxidant for go ld . Thiosul fate comp lexes the gold ion and stabi l ises the gold ion in solut ion. In the p rocess of gold leach ing, the cupr ic is reduced to cuprous, s o oxygen is required for its regenerat ion. A u + 5 S 2 0 3 2 " + C u ( N H 3 ) 4 2 + -> A u ( S 2 0 3 ) 2 3 " + 4 N H 3 + C u ( S 2 0 3 ) 3 5 " [2.1] 2 C u ( S 2 0 3 ) 3 5 ' + 8 N H 3 + 1 / 2 0 2 + H 2 0 -> 2 C u ( N H 3 ) 4 2 + + 2 O H " + 6 S 2 0 3 2 - [2.2] Overa l l : 2 A u + 4 S 2 0 3 2 " + 1 / 2 0 2 + H 2 0 ^ 2 A u ( S 2 0 3 ) 2 3 - + 2 O H " [2.3] O n e of the major obs tac les to thiosulfate leaching becoming a commerc ia l ly v iable p rocess is the rapid degradat ion of the thiosulfate reagent, wh ich is a concern economica l ly , technical ly and environmental ly. Thiosul fate is a meta-stable spec ies and can be ox id ized ultimately to sulfate through a number of react ion paths. T h e p resence of cupr ic copper , cons idered necessa ry by many to facil itate gold leach ing, enhances thiosulfate oxidat ion. T h e decompos i t ion products formed during gold leaching general ly include trithionate ( S 3 0 6 2 " ) . tetrathionate ( S 4 0 6 2 ) and sulfate. 4 E v e n though tetrathionate, trithionate and sulfate have very little effect on the gold oxidat ion react ion (Chu et a l . , 2003) , the formation of trithionate and sulfate in part icular represent an irreversible loss of thiosulfate under typical gold leach ing condi t ions. S i n c e act ivated carbon is not suitable for recover ing gold f rom thiosulfate solut ions, res ins have been proposed as a favourable alternative (F leming et a l . , 2000 , W e s t - S e l l s et a l , 2003 , J i et a l , 2003) . However , even low concentrat ions of trithionate and tetrathionate (0.01 M) load strongly onto anion exchange res ins, reducing gold loading (F leming et a l . , 2003 , Muir and Ay lmore , 2002). It has been found that concentrat ions of 0.05 M can effectively reduce gold loading to zero (Nicol and O 'Ma l ley , 2002) . T h e s e s p e c i e s were found not to affect the initial rate of gold loading but b e c a m e important over longer t imes when loaded gold could be d isp laced (Nicol and O 'Ma l ley , 2002) . Th is adsorpt ion is a ser ious issue for more aggress ive leaching condi t ions where a cons iderab le amount of thiosulfate is c o n s u m e d during leaching and a cons iderab le quantity of polythionates is p roduced. In one c a s e it w a s c la imed that mild leaching condi t ions (low thiosulfate, copper and ammon ia concentrat ions, neutral p H , short t ime) gave relatively low concentrat ions of tetrathionate and trithionate (F leming et a l . , 2003) . However , the authors a lso acknowledged that in genera l , l each l iquors still conta ined tetrathionate and trithionate after a few hours. T h e p resence of trithionate and tetrathionate is a lso of concern in effluent d i sposa l , a s these spec ies have the potential for ac id generat ion on comple te oxidat ion to sulfate, generat ing 1.3 mo les of ac id per mole of sulfur in trithionate and 1.5 moles of ac id per mole of sulfur in tetrathionate compared with the 1 mole produced per mole of sulfur in thiosulfate (Smith and Hi tchen, 1976). 2.3 STRUCTURE AND THERMODYNAMIC PROPERTIES OF THE SULFUR OXYANIONS Thermodynamica l ly , thiosulfate is a metastable ion. L ike the other metastable sulfur oxyan ions , thiosulfate needs to lose or ga in e lectrons to reach sulfate or sulf ide wh ich are stable (Wi l l iamson and Rimstidt, 1992). A typical E h - p H d iag ram for a q u e o u s sulfur 5 sys tems , shown in F igure 2 .1 , will therefore not include thiosulfate. A l l the E h - p H d iagrams shown in this thesis were generated using H S C Chemis t ry for W i n d o w s software (Roine, 1994, vers ion 5.0), but spec ies with thermodynamic data predicted by calculat ion in work by Wi l l iamson and Rimstidt (1992) were not inc luded a s they were acknow ledged by the authors of H S C Chemis t ry for W i n d o w s to be unrel iable and gave large stability reg imes for lesser known ions such a s H S 7 0 3 " ) . T h e data used to generate these d iagrams is shown in Append ix 1. E h (Volts) 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 H S 0 4 ---S 0 4 2 --H 2 S H S - s2-8 10 12 14 P H Figure 2.1 : E h - p H d iagram for the a q u e o u s sulfur sys tem (1 M sulfur, 25 °C) Thiosul fate c a n be ox id ized through tetrathionate ( S 4 0 6 2 " ) . trithionate ( S 3 0 6 2 ) , sulfite and other sulfur oxyan ions to sulfate a s the potential i nc reases , and is reduced to sulfur in ac id or bisulf ide in neutral or alkal ine solut ions (Aylmore and Muir, 2001b) . Under a lkal ine condi t ions, a number of metastable sulfur spec ies occur , for examp le sulfur oxyan ions and polysul f ides (S n 2 ")- E x a m p l e s of the sulfur oxyan ions are shown d iagrammat ica l ly be low in Figure 2.2. Sulfur is multivalent and therefore eas i ly forms 6 o o - s - s o Thiosulfate S 2 0 3 2 _ o o O-S-S-S-S-0 o o Tetrathionate S 4 0 6 2 - * o - s = o o Sulfate S 0 4 2 - * o o = s - s = o o Dithionite S 2 0 4 2 -O O " I I I I O-S-O-O-S-0 o o Persulfate S 2 0 8 2 -o o I I I I O-S-S-S-0 o o 2-Trithionate S 3 0 6 2 _ o o O-S-S-S-S-S -0 2-o o Pentathionate S 5 0 6 2 -o - s o 2-Sulfite S 0 3 2 " " 0 0 " O - S - S - 0 o o 2-Dithionate S 2 0 6 2 " F iqure 2.2 : B a s i c structures of se lec ted sulfur oxyan ions * indicates spec ies easi ly quantif iable by current methods at U B C 7 this large range of ions, a s wel l a s polysul f ides and col loidal precipitates (Aylmore and Muir, 2001b) . It is wel l known that the propert ies of the sulfur a toms in many of the spec ies shown in F igure 2.2 are not equivalent (Ames and Wi l lard, 1951). T h e propert ies of sulfur in thiosulfate sugges t a poss ib le combinat ion of sulfur with sulfite or sulf ide with sulfur d iox ide, with the effective oxidat ion state of +2 per sulfur (Suzuk i , 1999). T h e polythionates have two types of sulfur atom - sulfenyl sulfur and sul fonate sulfur. T h e rate of the metastable sulfur oxyan ions decompos i t ion to thermodynamical ly stable sulfur ions is often less than predicted. Thiosul fate has been found to be produced under alkal ine condi t ions, during pyrite leaching. N o sulfite w a s detected, implying a kinetic h indrance of further thiosulfate oxidation (Webster , 1984). T h e fact that thiosulfate has a signif icant stability under certain condi t ions m a k e s it useful a s a lixiviant for go ld . Tri thionate is often found to persist in gold leach l iquors a s a degradat ion product, even though it is thermodynamical ly unstable (Lam, 2002 , Nico l and O 'Ma l ley , 2002 , J i et a l . , 2001) . A n examp le of an E h - p H d iagram where the stable sulfate ion has been omitted to show the metastable doma in of thiosulfate and other spec ies is shown in F igure 2.3. T h e d iagram shows that thiosulfate is 'stable ' in a p H range of greater than 6 and redox potential - 0 . 2 V to 0.2 V ( S H E ) , in the a b s e n c e of other s p e c i e s s u c h a s complex ing agents . Tetrathionate exhibits metastabil i ty at lower p H . 8 It has been found that sulfite and dithionate (S 20 6 2") did not appear in certain gold leach liquors (Aylmore and Muir, 2001b), so these species and other related species were omitted from the E h-pH diagram, giving the diagram shown in Figure 2.4. In this diagram, trithionate is the most stable polythionate at high E h and high pH. Based on this diagram, direct formation of trithionate from thiosulfate should be possible, and this has indeed been suggested (see Section 2.5). However, tetrathionate is often reported in gold leach liquors implying that the kinetics of tetrathionate formation are favourable. The presence of oxygen (hence an increase in potential) should favour the formation of trithionate. 9 E h (Volte) 1.5 1.0 0.5 -i 1 1 1 r 1 1 1 1 1 1 r -1.0 -1.5 -2.0 s3o62-10 12 1 - ^ V ^ --=====—-^ JW" -H2S HS- S 2 ; 14 P H Figure 2.4 : E h - p H d iagram for the metastable aqueous sulfur sys tem with the fol lowing spec ies omitted - SO/'. H S O V . H ? S 0 4 . x H ? 0 . S C X S O / , H S O V . S ? Q« 2 ' . H S O V . S?0« 2 ' . s 2g£ (1 M sulfur, 25 °C) Fo r reference, a select ion of G i b b s free energ ies of formation for spec ies of interest in the gold leaching sys tem is shown in Tab le 2 .1 . F rom these , the standard potentials for a few pertinent thiosulfate oxidation react ions could be calculated (from the relation A G 0 = -nFE°) and are shown in Tab le 2.2. 10 Tab le 2.1 : Se lec ted G i b b s free energ ies of formation (298 K) for spec ies of interest in the copper - ammon iaca l - thiosulfate sys tem (Aylmore and Muir, 2001b) S p e c i e s AG° (kJ/mol) S p e c i e s AG 0 (kJ/mol) S p e c i e s AG 0 (kJ/mol) S 0 C u 0 A u 0 SG-32- -486.5 C u + 50.2 A u + 163.2 S 0 4 2 - * -742.0 C u2 + 65.0 A u 3 + 433 .5 S 2 0 3 2 ' -532.2 C u ( S 2 0 3 ) 35 ' -1624.7 A u ( S 2 0 3 ) 2 3 - -1050.2 s2o62- -966.0 C u ( S 2 0 3 ) 2 3 _ -1084.1 A u ( N H 3 ) 4 3 + 64.4 s3o62- -958.0 C u ( S 2 0 3 ) - -541.0 A u ( N H 3 ) 2 + -41.4 s4o62- -1040.4 C u ( N H 3 ) + -10.3 s5o62- -956.0 C u ( N H 3 ) 2 + 14.5 O H * -157.3 s2o42- -600.6 C u ( N H 3 ) 2 2 + -32 .3 H 2 0 -237.2 S 2 0 8 2 " -1115.0 C u ( N H 3 ) 32 + -73.2 N H 3 -26.7 C u ( N H 3 ) 4 2 + -113.0 * F rom W e a s t , 1975. Tab le 2.2 : Se lec ted standard oxidation potentials (298 K) Half Reac t ion E 0 o x i d (V) 2 S 2 0 3 2 - -> S 4 0 6 2 - + 2 e" 0.12 3 S 2 0 3 2 - + 3 H 2 0 -» 2 S 3 0 6 2 - + 6 H + +8 e" 0.51 S 3 0 6 2 " + 6 H 2 0 -> 3 S 0 4 2 ' + 12 H + +8 e 0.20 S 2 0 3 2 - + 5 O H " ^ 2 S 0 4 2 " + 5 H + + 2 e" -0.86 T h e s tandard potentials imply that degradat ion of thiosulfate directly to sulfate should occur readily under the condit ions required for gold leach ing, and a lso expla in why tetrathionate is likely to be formed, a s the standard potentials for these oxidat ion half react ions are general ly lower than those required to leach go ld . T o form trithionate f rom thiosulfate requires a higher potential (0.51 V ) . The data in Tab le 2.2 highlights the fact that the thermodynamics of the sys tem a lone cannot be used a s an indicative tool, a s kinetically we s e e the formation and pers is tence of trithionate, and sulfate is only formed in any signif icant quantity after long t imes or a s a product of polythionate degradat ion, 11 contrary to the predictions of thermodynamics. There appear to be significant kinetic hindrances of trithionate degradation to sulfate in the gold leaching system. 2.4 TH IOSULFATE DEGRADATION IN G O L D LEACHING - EXPERIMENTAL OBSERVATIONS It has been found that thiosulfate degradation in the gold leaching system is often not reported in detail in the literature. Reasons for this could be that much of the reported literature has focussed on gold recovery, or that suitable analytical methods were not available to easily measure thiosulfate consumptions. However, from the literature data available, it has been attempted to compile thiosulfate degradation data in this section to identify the main factors influencing the degradation. The table in Appendix 2 shows a compilation of thiosulfate consumption data. In many cases, the available data was not completely quantitative or was given as a range. Qualitative information is not included in the appendix but is given in the summary in Table 2.3. The observations in Appendix 2 and in Table 2.3 were for a large range of conditions and in many instances it was not clear why particular conditions were selected. Due to the interdependency of the various parameters and the difficulty in comparing data from different sources with incomplete data sets and different experimental regimes, only general trends are presented. There was some contradiction in the assessment of how different parameters affect thiosulfate degradation. While it is acknowledged that the concentrations of thiosulfate, ammonia and copper have a significant effect, there is not complete agreement in how these parameters affect the degradation and not much discussion as to how the combination of these species affects degradation. It is generally agreed that an increase in temperature and reaction time, and the presence of ore increases the thiosulfate degradation. The effect of sulfite and sulfate addition has not been clearly established. A more detailed discussion of the chemistry of the thiosulfate degradation inferred in Table 2.3 is given in Section 2.5. 12 Table 2.3 : Summary of experimental observations for thiosulfate degradation Parameter Effect References General Comment Temperature Inc => increase in degradation A, B, C, D Often implied by decrease in gold leaching, assumed to be due to loss of cupric and thiosulfate reagents. Ammonia concentration Inc => increase in degradation c Ammonia acts as stabiliser for cupric copper, hence it has often been postulated that ammonia inhibits the thiosulfate - copper reaction. Inc => decrease in degradation D, E F G H Optimum required A Copper concentration Inc => decrease in degradation G Postulated that once a certain level of cupric tetraammine is reached, excess cupric oxidizes the leach reagent rather than catalyzing the leach reaction. Inc => increase in degradation C, H, 1 Thiosulfate concentration Inc => increase in degradation B, C, H, J The dependency has been found to vary with other solution parameters e.g. copper concentration. Inc => decrease in degradation G, 1 Ore mineralogy Presence of sulfides => increase in degradation K, L, M Important aspect, but often difficult to quantify and difficult to compare data. Ultrafine milling of pyrite => no significant effect G Time Inc => increase in degradation G, K Expected trend as thiosulfate is not a thermodynamically stable species. Sulfite addition If present => decrease in degradation C, N,0 Likely to be highly dependent on ore type, and also known to decrease extent of leaching. If present => inconclusive P 13 Tab le 2.3 (cont inued) : S u m m a r y of exper imental observat ions for thiosulfate degradat ion Paramete r Effect Re fe rences G e n e r a l C o m m e n t P H pH > 11.4 required for min imum degradat ion B Inconclusive. Inc pH 8 . 5 - 1 0 ^ increase in degradat ion i Inc pH 9 . 8 - 1 1 . 4 ^ d e c r e a s e in degradat ion Q, R O x y g e n concentrat ion Inc => inc rease in degradat ion Q An ions (phosphate) Inc => dec rease in degradat ion Q, R Certa in an ions sa id to inhibit thiosulfate coordinat ion with copper (II) ammine comp lexes Re fe rences to Tab le 2.3 A A b b r u z z e s e et a l . , 1995 B Breuer and Jeffrey, 2000 C Re t t e ta l . , 1983 D Jeffrey, 2001 E Ay lmore and Muir, 2001a F Byer ley e t a l . , 1973a G Ay lmore , 2001 H Y e n e t a l . , 1999 I Langhans e t a l . , 1992 J Z ipper ian and R a g h a v a n , 1988 K W a n , 1997 L X u and S c h o o n e n , 1995 M F e n g and V a n Deventer , 2002b N K e r l e y a n d Bernard , 1981 O J i et a l . , 2001 P Li e t a l . , 1996 Q Breuer and Jeffrey, 2003a R Breuer and Jeffrey, 2003b A signif icant amount of thiosulfate degradat ion data w a s g iven by Ay lmore and Muir (2001a). A l s o reported by them w a s a compar ison of the consumpt ion of copper with t ime (copper measu red by ICP) and thiosulfate (measured by H P L C ) . (Cupr ic copper is known to ox id ize thiosulfate - s e e Sect ion 2.5.) Th is is not reproduced in this summary . However , it is interesting that al though copper consumpt ion general ly inc reased as thiosulfate consumpt ion inc reased , the two reagents were not c o n s u m e d at a constant ratio. Th is could imply precipitation of copper or oxidat ion of thiosulfate by another oxid iz ing agent such a s oxygen , a s wel l as by copper . In other recent work (Breuer and 14 Jeffrey, 2 0 0 3 a , b) it w a s a lso real ised that thiosulfate consumpt ion and cupr ic copper concentrat ion shou ld be cons idered separate ly . However , in o lder work, the thiosulfate degradat ion w a s implied by the measu red copper concentrat ion (Byer ley et a l . , 1973a) and even though under certain condit ions this can give a good approximat ion, it may not a lways be entirely appropr iate. A number of effects on thiosulfate degradat ion were d i s c u s s e d in a si lver leaching sys tem (Flett et a l . , 1983). In leach tests on synthet ic argenti te, both the thiosulfate and tetrathionate in solut ion were measu red . A l though it is common ly accep ted or a s s u m e d that tetrathionate is the major degradat ion product of thiosulfate under typical gold and si lver leach condi t ions, it w a s found that the loss of thiosulfate w a s slightly h igher than the ga in in tetrathionate, implying the formation of other sulfur oxyan ion spec ies . In many c a s e s only tetrathionate w a s measu red , which could give mis lead ing conc lus ions . T h e s e observat ions highlight the importance of appropriate ana lys is and that s o m e reported results and conc lus ions are not based on direct ana lys is . Ve ry little exper imenta l data on trithionate and tetrathionate product ion is reported pertaining to gold extract ion. It may be that in most reported c a s e s , only the change in thiosulfate concentrat ion w a s measu red , without determining the speciat ion of sulfur oxyan ion degradat ion products. A few va lues for trithionate and tetrathionate concentrat ions observed in gold leach l iquors are shown in Tab le 2.4. Wh i l e pentathionate has occas iona l ly been noted (qualitatively) in gold leach solut ions (De J o n g , 2004) , it is not a persistent spec ies under these condi t ions. 15 Tab le 2.4 - Literature data for trithionate and tetrathionate product ion L e a c h condi t ions S 2 0 3 2 - ( m M ) S 4 0 6 2 - (mM) S 3 0 62 " (mM) 0.05 M S 2 0 3 2 " 0.2 M N H 3 20 mg/l e a c h of P b , Z n , C u , A g , A u 48 hrs (Nicol and O 'Ma l ley , 2002) 10 Cont inuous R I P initial 0.05 M S 2 0 3 2 " 0.8 M N H 3 1 m M S 0 3 2 " pH 9.5 (Nicol and O 'Ma l ley , 2002) 0.089 0.1 L e a c h d ischarge slurry p H 6.9 (Ji et a l . , 2001) 75 2.3 3 Pregnant leach solut ion pH 10 (Ji e t a l . , 2001) 70 3.8 7.7 In leach tests at the Universi ty of Brit ish Co lumb ia (Lam, 2002) at room temperature, us ing initial solut ion concentrat ions of 0.1 - 0.6 M S 2 0 3 2 " , 0.16 - 4.7 m M C u (II) and 0.3 -0.5 M N H 3 , typical concentrat ions of the sulfur oxyan ions in the leach solut ions after 24 to 48 hours were: S 2 0 3 2 " 1 4 - 5 6 7 m M S 3 0 6 2 - 2 9 - 1 4 0 m M S 4 0 6 2 " 0 - 1 7 m M S 0 4 2 - 6 - 1 3 9 m M Trithionate in the f inal leach l iquor accounted for 7 - 8 3 % of the initial thiosulfate added . T h e formation of trithionate and tetrathionate in sys tems resembl ing gold leaching sys tems to examine the effect of speci f ic minerals on the thiosulfate degradat ion kinetics has recently been reported (De J o n g , 2004). Var ious minerals were a l lowed to react with an ammon iaca l thiosulfate solut ion (0.2 M ( N H 4 ) 2 S 2 0 3 at pH 10) with and without added copper . T h e author p roposed that trithionate w a s formed by hydrolysis of the tetrathionate produced during thiosulfate degradat ion rather than directly f rom thiosulfate 16 (see Section 2.5). This was not proven as the thiosulfate degradation kinetics were not measured directly in this work due to an unreliable analysis method. Without added copper, the formation of trithionate increased in the order no mineral ~ hematite ~ galena ~ arsenopyrite < chalcopyrite < pyrrhotite < pyrite < chalcocite. It is not clear why trithionate formation would be enhanced in the presence of minerals in the absence of copper, but is likely to be a complex effect of surface catalysis, pH, redox potential and possibly the presence of soluble trace elements. The formation of tetrathionate increased in the order no mineral ~ hematite ~ galena ~ arsenopyrite ~ chalcocite < pyrite < pyrrhotite < chalcopyrite. The authors proposed that certain minerals enhanced the degradation of tetrathionate to trithionate, for example chalcocite. Where copper was initially present, all the minerals except for chalcocite showed less trithionate formation than the no-mineral condition. The formation of trithionate decreased in the order chalcocite > no mineral ~ hematite > pyrite > arsenopyrite > galena > pyrrhotite ~ chalcopyrite. The formation of tetrathionate increased in the order chalcocite < no mineral ~ hematite < arsenopyrite < pyrite ~ galena ~ chalcopyrite < pyrrhotite. Both with and without initial copper, the presence of chalcocite caused the formation of a significant amount of trithionate. The initial rapid formation of trithionate corresponded to rapid initial copper extraction from the chalcocite. The tetrathionate concentration in this case was very low. Since the author had postulated that trithionate formed from tetrathionate, this observation lead to the suggestion that copper accelerates the decomposition of tetrathionate to trithionate. To better understand the effects of minerals a more comprehensive sulfur species analysis to include thiosulfate and sulfate would be required. More fundamental studies investigating the degradation of thiosulfate in the absence of ores are discussed in Section 2.5. 17 2.5 THIOSULFATE CHEMISTRY AND FUNDAMENTAL STUDIES E v e n though the stability of thiosulfate in aqueous solut ions is affected by many factors (Dhawale , 1993), thiosulfate in solut ion, prepared in freshly boi led doub le distil led water, is very stable stored in air tight bottles (Aylmore and Muir, 2001a) . A i r oxidat ion of thiosulfate at normal temperature and pressure is very s low. A t p H 7 solut ions aerated for 4 months under sterile condit ions had less than 10 % change in the thiosalt concentrat ion. There is signif icant oxidation at higher temperature and air or oxygen pressure , indicating kinetic control. (Rol ia and Chakrabar t i , 1982). T h e degradat ion of thiosulfate that can be expected in gold leach sys tems is d i s cussed below. 2.5.1 Oxidative Degradation in Gold Leaching Systems Acco rd ing to thermodynamics , one would expect that under oxidiz ing condi t ions, thiosulfate would eventual ly be ox id ized to sulfate (see F igure 2.1). Th is is demonst ra ted in Equat ion 2.4. S 2 0 3 2 " + 2 0 2 + H 2 0 -» 2 S 0 4 2 " + 2 H + [2.4] However , a number of metastable oxidation products can be expec ted . In the gold leaching sys tem, where thiosulfate, copper and a m m o n i a are present, tetrathionate has often been quoted a s a pr imary oxidation product. Thiosul fate oxidat ion to tetrathionate by oxygen can be demonst ra ted by Equat ion 2.5 (Wan , 1997, Li et a l . , 1996, Ay lmore and Muir, 2001a) . 2 S 2 0 3 2 " + H 2 0 + 1 / 2 0 2 -» S 4 0 6 2 - + 2 O H " [2.5] T h e thiosulfate leach sys tem is compl icated by the reduct ion of copper by thiosulfate (Byer ley et a l . , 1973a, Breuer and Jeffrey, 2000). T h e react ion of the cupr ic tet raammine 18 comp lex with thiosulfate to form tetrathionate is shown in Equat ion 2.6 (Aylmore and Muir , 2 0 0 1 a , Breuer and Jeffrey, 2000). 2 C u ( N H 3 ) 4 2 + + 8 S 2 0 3 2 - -» 2 C u ( S 2 0 3 ) 3 5 - + S 4 0 6 2 - + 8 N H 3 [2.6] T h e cupr ic te t raammine comp lex is used in the react ion equat ion rather than s imply the cupr ic ion a s it has been often reported that it is the comp lex that is respons ib le for react ing with thiosulfate (Lam, 2001 , Breuer and Jeffrey, 2000) . However , it has a lso been sugges ted that C u ( N H 3 ) 3 2 + rather than C u ( N H 3 ) 4 2 + ox id izes S 2 0 3 2 _ to S 4 0 6 2 _ (Byer ley et a l . , 1973a). Th is is d i s cussed later in this sect ion. T h e react ion of thiosulfate with cupr ic copper is rapid in aqueous solution but s lower with a m m o n i a present (Aylmore and Muir, 2001a) . In alkal ine solut ion, the react ion between cupr ic copper and thiosulfate d o e s not require oxygen (Aylmore and Muir, 2001a) . However , oxygen still p lays an important role in the overal l extent and rate of thiosulfate degradat ion. In the gold leaching sys tem, oxygen is used to convert cuprous copper to cupr ic, but depend ing on the amount of oxygen , s o m e direct oxidation of thiosulfate to tetrathionate and trithionate can a l so occur . In the p resence of oxygen , the redox potential r ises and the oxidat ion of cuprous to cupr ic is more rapid (as is the oxidat ion of thiosulfate, where the react ion rate has been quoted to be at least forty t imes higher in the p resence of oxygen (Byer ley et a l . , 1973b)). E v e n though in many cited examp les , tetrathionate is the main oxidat ion product referred to (Wan 1997, Muir and Ay lmore , 2002) , it must be noted that somet imes it has only been a s s u m e d that a loss in thiosulfate impl ies the product ion of tetrathionate, or somet imes tetrathionate w a s the only degradat ion product measu red , leading to poss ib le misinterpretat ion. Trithionate and sulfate have a lso been reported (Byer ley et a l . , 1973b, Muir and Ay lmore , 2002). B e s i d e s copper and oxygen , other oxidants, if present, can a lso degrade thiosulfate. If iron is present, it can d isso lve at pH less than 9.5 and degrade thiosulfate, producing tetrathionate which has no lixiviating act ion for gold or si lver (Aylmore and Muir, 2 0 0 1 a , P e r e z and Ga lav i z , 1987). H e n c e a high pH stabi l izes thiosulfate by minimizing iron dissolut ion (Pe rez and Ga lav i z , 1987). A l s o A g + , H g 2 + and cyan ide can degrade 19 thiosulfate (Kelly and W o o d , 1994). T h e addit ion of the metal ions of N i , Z n , P b , C d , C o and C r to the gold leaching sys tem all had the effect of dec reas ing the free thiosulfate concentrat ion, with the largest effect by C r and least by Ni (Feng and Deventer , 2002a) . It w a s d e d u c e d that oxidat ion (especial ly for C r 6 * and C o 3 + ) a s wel l a s complexat ion (especia l ly for C d 2 + ) p layed a role. F e w studies have been done to determine the mechan i sm or kinet ics of thiosulfate degradat ion, particularly in the gold leaching sys tem. A signif icant amount of fundamenta l work w a s done in the 1970s by Byer ley et a l . (1973a, 1973b, 1975) and a lso very recently by Breuer and Jeffrey (2003a, 2003b) . Copper (II) Oxidation of Thiosulfate in Absence of Oxygen Byer ley et a l . (1973a) studied the mechan i sm of thiosulfate degradat ion by cupr ic copper in ammon ium hydroxide in the a b s e n c e of oxygen . T h e m e c h a n i s m proposed w a s for a sys tem slightly different to that expected under gold leaching condit ions and w a s primarily related to sulf ide leaching in ammon ia solut ions, where thiosulfate w a s p roduced . W h e n thiosulfate w a s added to a copper -ammon ia sys tem, there w a s an immediate inc rease in the abso rbance of the solut ion prior to the c o m m e n c e m e n t of the react ion. But the posit ion of the absorpt ion max imum did not change at this initial t ime, nor during the react ion. T h e increase in abso rbance w a s interpreted a s an assoc ia t ion between the thiosulfate and the copper a m m o n i a comp lexes . T h e assoc ia t ion w a s sugges ted a s being at one of the axial posi t ions, therefore giving a strong polar iz ing effect of cupr ic copper on thiosulfate. T h e equatorial square planar structure of the copper complex remained essent ia l ly unchanged . T h e progress of the react ion w a s measu red by measur ing the cupr ic concentrat ion. Solu t ions ana lysed for thiosulfate by iodometr ic titration and qualitatively for tetrathionate directly after the react ion (when it w a s observed that all cupr ic had been consumed) showed that overal l one mole of copper (II) w a s c o n s u m e d for e a c h mole of thiosulfate, matching the stoichiometry of Equat ion 2.7. 20 2 C u 2 + + 2 S 2 0 3 2 - S 4 0 6 2 - + 2 Cu + [2.7] If higher copper (-0.03 M) or thiosulfate (-0.15 M) concentrations were used, or if the solutions were stored, trithionate and regenerated thiosulfate were produced as products of tetrathionate degradation (Byerley et al., 1973a). Tetrathionate chemistry is discussed later. At pH > 10, the cupric-ammine complex exists in equilibrium with hydroxo species, so some tri-ammine cupric species co-exist with the tetraammine species. Since the rate of the copper thiosulfate reaction was faster in aqueous solution, it was assumed that the tri-ammine cupric complex was more reactive than the tetraammine complex (Byerley et al., 1973a). Hence, a mechanism suggested from kinetic results implied substitution of thiosulfate into the co-ordination sphere of a copper (II) tri-ammine complex prior to the electron transfer step. An electron transfer from thiosulfate to cupric copper in an intermediate tri-ammine cupric-thiosulfate complex gives cuprous copper and thiosulfate, which dimerises to give tetrathionate. The mechanism outline is shown in Equations 2.8 to 2.11. It was determined that the involvement of free radicals in the rate determining step was unlikely as addition of a free radical inhibitor, mannitol, had no effect on the reaction rate (Byerley et al., 1973a). In Byerley's work plots of log [Cu2+] vs time were linear with the same slope, indicating first order dependency, assuming ammonia and thiosulfate concentrations were essentially constant during the reaction. Thiosulfate concentrations up to 0.10 M gave a rate of first order dependence. At higher thiosulfate concentrations (above 0.15 M), the reaction order was higher. There was a linear dependence of the rate on 1/[NH3] for ammonia concentrations of 0.1 - 1 M. Cu(NH 3) 4 2 + + H 2 0 - Cu(NH 3) 3(H 20) + NH 3 Cu(NH 3) 3(H 20) 2 ++ S 2 0 3 2 " -> Cu(NH 3) 3(S 20 3)+ H 2 0 [2.9] [2.10] [2.11] [2.8] Cu(NH 3) 3(S 20 3) + e Cu(NH 3) 3 + + S 2 0 3 : 2 S 2 0 3 2 - S 4 0 6 2 " + 2e 2-21 Under these condi t ions, the rate of dec rease of cupr ic copper concentrat ion by react ion with thiosulfate in the a b s e n c e of oxygen w a s g iven by Equat ion 2.12. -d[Cu 2 + ] /dt = k [Cu 2 + ] [S 2 0 3 2 1 / [NH 3 ] [2.12] where k = 8.5 x 10" 4 s ' 1 at 30 °C. T h e activation energy w a s 102.5 kJ /mo l (based on thermodynamic parameters at 30 °C). More recent work by Breuer and Jeffrey has deve loped a further understanding of the sys tem (Breuer and Jeffrey, 2000 , 2003b) . T h e main a im for this study of the factors affecting the gold leaching kinetics under anaerob ic condi t ions w a s to al low for the effect of dec reas ing copper (II) (and hence cupr ic to cuprous ratio) to be invest igated with a minimal dec rease in thiosulfate concentrat ion. Al l the componen ts except for copper (II) and a m m o n i a were purged with argon, then a concentrated copper (II) - ammon ia solut ion w a s injected into the v e s s e l . After mix ing, a samp le w a s transferred to a sea led U V cel l , with no air contact at any t ime. T h e genera l condit ions used were 0.1 M sod ium thiosulfate, 0.4 M a m m o n i a and 10 m M copper sulfate at 30 °C, p H 11.4. Comp le te exc lus ion of air w a s found to be essent ia l as even trace quanti t ies were shown to inc rease the copper (II) reduct ion rate. T h e cupr ic ammine comp lexes absorb ing at 605 nm were measu red using UV-v is ib le spectrophotometry. P lo ts of the cupr ic ammine comp lex concentrat ion ve rsus t ime showed that the relation of 1/[Cu(ll)] ve rsus t ime w a s fairly l inear, whi le that of log[Cu(ll)] ve rsus t ime w a s not. Th is would imply that the rate determining step w a s s e c o n d order, not first order a s sugges ted by Byer ley et a l . (1973a). A deviat ion f rom linearity at longer t imes w a s attributed to further copper (II) reduction by tetrathionate. T h e addit ion of up to 0.8 M sulfate w a s found to d e c r e a s e the rate of copper (II) reduct ion considerably . B y chang ing the exper imenta l procedure to inject a thiosulfate solut ion into a solut ion of copper (II), ammon ia and sulfate, further information on the m e c h a n i s m of the act ion of sulfate ions could be obta ined. In this instance, a not iceable induction per iod occur red , after which the react ion rate inc reased to reach a simi lar va lue to that obtained w h e n a copper (II) - ammon ia solut ion w a s injected into a thiosulfate -sulfate solut ion. 22 It w a s noted by Byer ley et a l . (1973a) and Breuer and Jeffrey (2003b) that in the react ion of thiosulfate with copper (II), it is likely that thiosulfate comp lexes with the cupr ic ammine comp lex at an axial si te, based on abso rbance measurements . T h e react ion thus occurs v ia an inner sphere mechan i sm. It w a s postulated that sulfate competes with thiosulfate for co-ordinat ion, hence s lowing the copper (II) reduct ion rate by thiosulfate. W h e r e thiosulfate w a s added to a copper (II) a m m o n i a sulfate solut ion, the sulfate w a s ab le to co-ordinate before the addit ion of the thiosulfate, giving an induction period for thiosulfate to replace sulfate before the react ion could occur . T h e effect of other an ions (chloride and nitrate) and espec ia l ly phosphate w a s a lso to reduce the copper (II) reduct ion rate, which is consistent with an inner sphere reduct ion react ion. Phospha te in part icular is thought to readily comp lex with copper (II) at the axial posi t ion, inhibiting the substitution of thiosulfate into the inner sphere . T h e effect of p H on the react ion mechan i sm w a s invest igated by us ing ammon ium thiosulfate instead of sod ium thiosulfate, with the s a m e concentrat ion of ammon ia and s a m e ionic strength. Th is had the effect of reducing the pH f rom 11.4 to 9.8. T h e react ion order of copper (II) reduction at pH 9.8 showed a first order rate limiting step, not s e c o n d order a s found at pH 11.4, which correlates with the observat ion by Byer ley et a l . (1973a). T h e dec rease in react ion order f rom two to one at lower pH could be due to the lack of an ions (hydroxide) to compete with thiosulfate for axial coordinat ion. Th is v iew w a s suppor ted exper imental ly by an observed return to s e c o n d order kinet ics when sulfate w a s added at p H 9.8. A rate law w a s not g iven in this work. It w a s sugges ted that the mechan i sm w a s different and more compl icated than that sugges ted by Byer ley et a l (1973a). Ano ther m e c h a n i s m related to the leaching of si lver sulf ide in the a b s e n c e of air has been postulated, but with little detail g iven (Kel ly and W o o d , 1994). First the reduct ion of copper by thiosulfate in the bulk solut ion (Equat ion 2.13) w a s postu lated, fol lowed by the substitut ion of copper for si lver in the sulf ide molecu le , within the mineral particle (Equat ion 2.14). 23 5 ( N H 4 ) 2 S 2 0 3 + 2 CuSG -4 -» C u 2 S 2 0 3 . 2 ( N H 4 ) 2 S 2 0 3 + 2 ( N H 4 ) 2 S 0 4 + ( N H 4 ) 2 S 4 0 6 [2.13] C u 2 S 2 0 3 . 2 ( N H 4 ) 2 S 2 0 3 + A g 2 S -» C u 2 S + A g 2 S 2 0 3 . 2 ( N H 4 ) 2 S 2 0 3 [2.14] Copper (II) Oxidation of Thiosulfate in Presence of Oxygen T h e react ion between copper ammine solut ions and thiosulfate in the p resence of oxygen but the a b s e n c e of o res w a s studies by Byer ley et al (1975). They found that rather than tetrathionate being formed (as sugges ted above (De J o n g , 2004) and a s noted w h e n no oxygen w a s present (Byer ley et al . ,1973a)), trithionate and sulfate were fo rmed, in ratios varying with the initial thiosulfate concentrat ion and p H . S i n c e both the trithionate and sulfate formed were stable with respect to further oxidat ion under the condit ions tested, it w a s proposed that the trithionate w a s not an intermediate to sulfate formation. It w a s p roposed that two react ions were occurr ing, a s in Equat ions 2.15 and 2.16. C o p p e r d o e s not appea r in these equat ions as it w a s proposed that copper ac ted a s a catalyst, a s d i s cussed later. 3 S 2 0 3 2 " + 2 0 2 + H 2 0 -> 2 S 3 0 6 2 - + 2 O H " [2.15] S 2 0 3 2 - + 2 0 2 + 2 O H " -> 2 S 0 4 2 " + H 2 0 [2.16] T h e react ion rate w a s monitored by monitoring the oxygen consumpt ion . T h e rate of oxygen consumpt ion showed an initial induction per iod, increas ing react ion rate then a l inear max imum rate. T h e initial thiosulfate concentrat ion had a signif icant inf luence on the attainment of the max imum rate. A l s o , the rate of oxygen consumpt ion w a s virtually identical for a copper (I) or a copper (II) a m m o n i a thiosulfate sys tem, s o it w a s a s s u m e d that in the former, copper (I) w a s rapidly ox id ized to copper (II) and then the sys tems were essent ia l ly equivalent. T h e initial oxygen consumpt ion region w a s postulated to involve a build up of a catalyt ical ly act ive copper - oxygen complex , which eventual ly reached a s teady concentrat ion at the max imum react ion rate. It w a s found that at different parts of the oxygen consumpt ion curve, different y ie lds of trithionate and sulfate were obta ined. T h e 24 highest yields of trithionate were observed in the initial part of the oxygen consumption curve, and for the highest initial thiosulfate concentrations (> 0.02 M) and lowest pH values (<10). Under gold leaching conditions, one would expect a predominantly trithionate product. At the maximum oxygen consumption rate both trithionate and sulfate were formed concurrently. The rate of sulfate formation was calculated using the total oxygen consumption minus the rate of oxygen consumption for trithionate formation. The percentage conversion of thiosulfate to trithionate at the end of oxidation is shown in Table 2.5. As the initial thiosulfate concentration increased, the relative yield of trithionate versus sulfate increased. Also as the pH decreased, the yield of trithionate increased. It was also shown that the initial thiosulfate concentration had an effect on the oxygen consumption rate (Byerley et al., 1975). However, in more recent work where the thiosulfate oxidation rate was measured directly, the rate was found to be independent of the thiosulfate concentration (Breuer and Jeffrey, 2003a). It is likely that the oxygen consumption rate is not directly related to thiosulfate oxidation (Breuer and Jeffrey, 2003a). While Byerley et al. discussed possible mechanisms for the formation of trithionate from thiosulfate, these were postulated to involve Cu(ll) - oxygen complexes as intermediates. This is not discussed here as it has since been shown that the existence of such complexes is not likely (Breuer and Jeffrey, 2003a). While no rate law was given, it was noted by Byerley et al., that the rate of oxygen consumption was proportional to the copper and oxygen concentrations. 25 Tab le 2.5 - Format ion of trithionate from thiosulfate in the p resence of oxygen (oxygen pressure 725 m m H g . 30 °C. fCu(ll)1 = 1 m M . TNH?1 = 0.2 M) (Bver lev et a l . . 1975) Initial S 2 0 3 2 _ (mM) P H % S conver ted to s3o62-5 11.2 30.6 15 11.2 42 .2 25 11.2 50.4 50 11.2 64.1 75 11.2 66.6 100 11.2 75.6 25 11.2 54.0 25 10.0 64.6 25 9.6 81.0 25 9.3 100.0 Breuer and Jeffrey (2003a) invest igated the sys tem in a different way. Tes ts were done in a f low through sys tem, monitoring the copper (II) ammine comp lexes using UV-v is ib le spectrophotometry. A concentrated copper ( l l ) -ammonia solut ion w a s injected into the v e s s e l and after a mixing time the solut ion w a s continual ly pumped through a U V cel l . T h e cel l w a s temperature control led, and abso rbance w a s measu red at 6 0 5 nm. T h e thiosulfate concentrat ion w a s determined by removing a samp le and ana lys ing it us ing a rotating e lect rochemica l quartz crystal microba lance, by measur ing the m a s s changes to a si lver e lectrode (Breuer et a l . , 2002) . T h e standard initial reagent concentrat ions were 0.4 M ammon ia , 0.1 M sod ium thiosulfate and 10 m M copper sulfate. W h e r e copper (I) w a s invest igated, an ammon ia solution w a s injected into a solut ion of thiosulfate and copper (I). Mixtures of copper (II) and thiosulfate, with and without air sparg ing , showed an initial d e c r e a s e in copper (II) concentrat ion with t ime, more s o in the p resence of air, whi le the thiosulfate concentrat ion d e c r e a s e d . Th is result w a s cons idered surpr is ing a s oxygen readi ly ox id izes copper (I) to copper (II). H e n c e it w a s shown that copper exists a s both 26 copper (I) and copper (II) during the oxidation of thiosulfate. Th is contrasts with the m e c h a n i s m by Byer ley et a l . (1973a) where it w a s c la imed that there w a s no dec rease in copper (II) concentrat ion. In Breuer and Jeffrey 's work, s ince both the rate of thiosulfate oxidat ion and the rate of copper( l l ) reduction were inc reased in the p resence of oxygen , an alternative mechan i sm to Byer ley et a l . 's w a s cons idered necessary . T h e oxidat ion of thiosulfate w a s slightly faster when the initial copper w a s present a s copper (I), and the initial increase of copper (II) w a s rapid. Th is sugges ted that the react ion products of copper (I) oxidation may be involved in the oxidat ion of thiosulfate. T h u s the m e c h a n i s m s of the copper (I) oxidation by oxygen were cons ide red . It is sa id to be wel l es tab l ished that the first s teps in copper (I) oxidat ion by oxygen involve the formation of copper (II) and peroxide, a s demonst ra ted in Equat ion 2.17. 2 C u + + 0 2 ^ 2 C u 2 + + H 0 2 " + O H - [2.17] However , the m e c h a n i s m s for this p rocess are not c lear. T h e kinet ics s h o w that the autooxidat ion of Cu( l ) comp lexes occurs v ia inner sphere mechan i sms involving d i -oxygen adducts formed accord ing to Equat ion 2.18. C u + + 0 2 ^ C u + 0 2 [2.18] Alternat ively this may be v iewed a s a copper (II) superox ide comp lex C u 2 + 0 2 " . T h e fact that no superox ide has been identified during copper (I) autooxidat ion is not surpr is ing g iven that its reactivity is c lose to diffusion rates Tes ts to invest igate the effect of peroxide on thiosulfate oxidat ion were therefore conduc ted . Addi t ion of 50 m M hydrogen peroxide to 0.1 M thiosulfate and 0.4 M a m m o n i a (without copper) gave less than 30 % thiosulfate oxidat ion after 1 hour. However , in the p resence of copper (II) a much faster react ion occur red. A rapid react ion a lso occur red when peroxide w a s added to a thiosulfate solut ion containing copper (I) and to a thiosulfate, copper (II) and a m m o n i a leach solut ion. H e n c e it w a s implied that the intermediates of peroxide decompos i t ion by copper ox id ize thiosulfate, and not peroxide directly. Perox ide has been cons idered previously to be involved in the 27 formation of tetrathionate, for examp le in the electrolytic oxidat ion of thiosulfate where peroxide w a s a s s u m e d to be act ive at the anode (G lass tone and Hick l ing, 1954). A react ion s c h e m e w a s sugges ted to descr ibe the copper ca ta lysed decompos i t ion of peroxide taking part in thiosulfate oxidat ion. Th is is descr ibed in Equat ions 2.19 to 2.26. T w o thiosulfate oxidat ion paths were predicted - the copper( l l ) thiosulfate react ion ca ta lysed by oxygen , evident by the higher initial rate of copper (II) reduct ion in the p resence of oxygen , and the oxidation of thiosulfate by superox ide or hydroxide radicals a s a result of copper (I) oxidat ion. C u 2 + + S 2 0 3 2 " (+ 0 2 ) -> C u + + X (X = unspeci f ied sulfur containing spec ies ) [2.19] C u + + 0 2 C u 2 + + G Y [2.20] 2 0 2 " + H 2 0 -> 0 2 + H 0 2 " + O H " [2.21] C u 2 + + H 0 2 " + O H ' ~ C u + + 0 2 " + H 2 0 [2.22] C u + + H 0 2 - + H 2 0 ^ C u 2 + + O H + 2 O H " [2.23] C u + + O H -» C u 2 + + O H " [2.24] O H + H 0 2 " -> H 2 0 + 0 2 " [2.25] R + S 2 0 3 2 - -» P (R = 0 2 " or O H , P unspeci f ied) [2.26] T h e effect of oxygen concentrat ion w a s invest igated by sparg ing the solut ion with oxygen , air or a g a s mixture containing 1.9 % oxygen in nitrogen. Increased oxygen concentrat ion inc reased the rate of thiosulfate oxidation and copper (I) oxidat ion, and assoc ia ted react ions accord ing to the peroxide react ion s c h e m e . However , the copper (II) to copper (I) ratio w a s hardly af fected. A l s o the rate of thiosulfate oxidat ion inc reased with inc reased air sparge rate, and it w a s sugges ted that the react ion w a s likely to be limited by the rate of m a s s transfer of oxygen into solut ion. A n inc rease in thiosulfate concentrat ion from 0.05 M to 0.15 M showed simi lar max imum thiosulfate oxidation rates. However , the cor responding copper (II) concentrat ion profi les s h o w e d that increas ing the thiosulfate concentrat ion inc reased the rate of copper (II) reduct ion and reduced the copper (II) to copper (I) ratio in solut ion. 28 A d e c r e a s e in copper concentrat ion by a factor five d e c r e a s e d the thiosulfate oxidation rate by a factor two. However it a lso had a signif icant effect on the solut ion potential, wh ich d e c r e a s e d as the copper concentrat ion d e c r e a s e d . T h e thiosulfate oxidat ion rate w a s simi lar for 0.2 M or 0.4 M a m m o n i a but the initial rate of copper (II) reduct ion increased with decreas ing a m m o n i a . T h e effect of p H (at constant ammon ia concentrat ion) w a s shown by compar ing react ion rates using ammon ium thiosulfate and sod ium thiosulfate, with the s a m e a m m o n i a concentrat ion. It w a s shown (Breuer and Jeffrey, 2003d) that at low pH va lues , even though the a m m o n i a concentrat ion w a s constant , copper (II) w a s much more reactive towards thiosulfate. S o it w a s not surpr is ing that the copper (II) concentrat ion dropped to a lower min imum level at lower p H . Thiosul fate oxidation w a s a lso faster at pH 9.8 than 11.4. T h e p resence of an ions w a s shown to reduce the rate of copper (II) reduct ion by thiosulfate with phosphate being the most effective. Th is w a s shown for the sys tem with or without oxygen present (Breuer and Jeffrey, 2 0 0 3 a , b). Th is is consistent with the theory of an ions compet ing for coordinat ion at the copper (II) axia l s i tes and thus reducing the rate of copper (II) reduction by thiosulfate involving oxygen . N o kinetic law w a s g iven in this c a s e . Copper (II) Oxidation of Thiosulfate - General Considerations It w a s p roposed by Byer ley et a l . (1973a, b, 1975) and Breuer and Jeffrey (2003b) that in the oxidat ion of thiosulfate by cupr ic copper , thiosulfate b e c a m e co-ordinated to the cupr ic -ammine complex . S e n a n a y a k e (2004) recently revisited this react ion, us ing literature data and thermodynamic and kinetic ana lys is . B a s e d on publ ished stability constants , he cons idered the spec ies distribution of the cupr ic - a m m o n i a - thiosulfate -hydroxide sys tem over a range of pH and ammon ia and thiosulfate concentrat ions. Overa l l , f rom p H 7 to 12, the spec ies C u ( S 2 0 3 ) 2 2 - , C u ( N H 3 ) 3 2 + , C u ( N H 3 ) 3 O H + and C u ( N H 3 ) 4 2 + were predominant. Th is supports Byer ley et a l 's v iew that a mixed comp lex of the type C u ( N H 3 ) 3 ( S 2 0 3 ) ° can be involved in thiosulfate oxidat ion. 29 A s s u m i n g the genera l rate law in Equat ion 2.27, the author determined the react ion orders q , r and I to be 1, -1.1 and 1.2 respect ively, based on an ana lys is of literature data . -d[Cu 2 + ] /dt = k C u [Cu 2 + ] q [NH 3 ] r [S 2 03 2 - ] ' [2.27] However , Breuer and Jeffrey (2003b) p roposed that the cupr ic - thiosulfate react ion could be s e c o n d order with respect to cupr ic concentrat ion. It w a s sugges ted by S e n a n a y a k e (2004) that the react ion could be either first or s e c o n d order with one dominat ing under certain condi t ions. B a s e d on work done in the a b s e n c e of ammon ia , S e n a n a y a k e found that the oxidat ion of thiosulfate by cupr ic copper took p lace via the mixed comp lex C u ( S 2 0 3 ) n ( H 2 0 ) p - 2 ( n - 1 ) and the rate determining step w a s the decompos i t ion of this comp lex to products. H e n c e two scenar ios were cons idered in the p resence of a m m o n i a - a first order decompos i t ion and a s e c o n d order decompos i t ion a s the rate determining s tep. Fo r the first order scenar io , the react ion in Equat ion 2.28 w a s cons ide red . Vary ing the va lues of p (2-3) and n (1-2) and compar ing the p roposed kinetic law with literature data , k C u w a s found to be 4 x 10" 4 s" 1 for C u ( N H 3 ) 3 ( S 2 0 3 ) ° and C u ( N H 3 ) 2 ( S 2 0 3 ) 2 2~, and k C u = 17 x 10" 4 s" 1 for C u ( N H 3 ) 3 ( S 2 0 3 ) 2 2'. T h e s e va lues are s imi lar to that obtained by Byer ley et al (1973a). Fo r the s e c o n d order scenar io , the reaction in Equat ion 2.29 w a s cons idered a s the rate determining step. 2 Cu(NH 3 ) p (S 2 0 3 ) n - 2 ( r v 1 ) -» products [2.29] Fo r this react ion the rate constant k C u w a s 0.1 - 0.2 M" 1 .s" 1 for the decompos i t ion of C u ( N H 3 ) 3 ( S 2 0 3 ) ° , C u ( N H 3 ) 2 ( S 2 0 3 ) 2 2 - a n d C u ( N H 3 ) 2 ( S 2 0 3 ) 0 . Cu(NH 3 )p(S 2 0 3 )n- 2 ( n - 1 ) products [2.28] 30 Pyrite Catalysed Thiosulfate Oxidation in Presence of Oxygen Cata lys is by pyrite has a l so been used to explain thiosulfate oxidat ion. In the a b s e n c e of copper , the rate of tetrathionate formation in a thiosulfate leach sys tem w a s directly proport ional to the pyrite sur face concentrat ion (Xu and S c h o o n e n , 1995). T h e dependency on pyrite sur face concentrat ion impl ies a sur face mediated react ion m e c h a n i s m . For pyrite to act as a catalyst, there had to be interaction of thiosulfate, pyrite and oxygen . Th is w a s d i scussed in terms of molecu lar orbital theory. S ince negl igible tetrathionate w a s formed in a sys tem of thiosulfate held under nitrogen, it w a s a s s u m e d that oxygen w a s the terminal e lectron acceptor in the oxidat ion react ion. A l s o , negl igible thiosulfate oxidat ion to tetrathionate occur red without pyrite present. S i n c e both pyrite and thiosulfate have been found to react s lowly with molecu lar oxygen , it w a s postulated that a sur face complex with interaction between the molecu lar orbitals of thiosulfate, pyrite and oxygen w a s required for the react ion to p roceed . A m e c h a n i s m w a s d e d u c e d based on the fol lowing three condit ions used to determine whether a redox react ion is likely to occu r accord ing to molecu lar orbital theory. 1. T h e energy of the lowest unoccup ied molecu lar orbital ( L U M O ) must be less than that of the highest occup ied molecu lar orbital ( H O M O ) or within 6 e V above that of the H O M O . 2. T h e symmet r ies of the H O M O and L U M O must be the s a m e to ensure proper over lap of the orbitals. 3. T h e electron transfer must yield a stable end product. T h e react ion mechan i sm postulated involved three electron transfer s teps . T h e first transfer w a s f rom thiosulfate to an anod ic site on pyrite, the s e c o n d w a s from an anod ic site to a cathodic site on the pyrite sur face via its conduct ion band , and the third w a s f rom the cathodic site to the terminal e lectron acceptor , oxygen (Xu and S c h o o n e n , 1995). 31 2.5.2 Disproportionation and Reductive Degradation of Thiosulfate Disproport ionat ion of thiosulfate to sulfur and sulfate, or sulf ide and sulfite, is represented in Equat ions 2.30 and 2 .31. Disproport ionat ion is expec ted to occur in oxygen deficient or low potential solut ions, or where there is a high copper concentrat ion (Aylmore and Muir, 2001a) . Th is type of decompos i t ion of thiosulfate leads to precipitation of e lementa l sulfur, go ld , or copper , gold or si lver sul f ides (Li et a l , 1996). E lementa l sulfur and cupr ic sulf ide have been obse rved exper imental ly in the gold thiosulfate leach sys tem (Wan , 1997, Ay lmore and Muir , 2001a) . Reduct ive decompos i t ion c a n be represented by Equat ions 2.32, or 2.33 and 2.34, in the p resence of copper (Li et a l , 1996). S 2 0 3 2 - + 8 H + + 8 e 2 H S " + 3 H 2 0 [2.32] C u 2 + + S 2 0 3 2 " + 6 H + + 6 e ^ S + C u S + 3 H 2 0 [2.33] 2 C u + + S 2 0 3 2 - + 6 H + + 6 e ^ S + C u 2 S + 3 H 2 0 [2.34] B e s i d e s the loss of the thiosulfate lixiviant and copper , increasing the operat ing cost , this react ion may lead to precipitation of si lver or block the sur face for further leaching (Li et a l , 1996). T h e pathway for the disproport ionat ion react ion of thiosulfate in the p resence of copper has been sugges ted (Muir and Ay lmore , 2002). It w a s sugges ted that a fast redox react ion between cupr ic copper and thiosulfate w a s fol lowed by s lower s ide react ions of the cuprous p roduced , forming C u 2 S , a s in Equat ion 2.35. 2 C u + + S 2 0 3 2 - + H 2 0 C u 2 S + S 0 4 2 " + 2 H + [2.35] 3 S 2 0 3 2 - + H 2 0 ^ 2 SO42- + 4 S + 2 O H " or 3 S 2 0 3 2 - + 6 O H " -» 4 S 0 3 2 " + 2 S 2 ' + 3 H 2 0 [2.30] [2.31] 32 2.5.3 Thiosulfate Degradation Inhibitors B a s e d on Equat ion 2 .31, it has been sugges ted that sulfite inhibits thiosulfate decompos i t ion (Ker ley and Bernard , 1981, Hemmat i et a l , 1989). Sulf i te is c la imed to prevent the formation of free sulf ide (S 2") and hence precipitation of gold or si lver (Ker ley and Bernard , 1981, Z ipper ian and R a g h a v a n , 1988). Ve ry low concentrat ions ( -0 .05 %) of sulfite have been used to stabi l ise thiosulfate but sulfite addit ion a lso lowers the potential, wh ich reduces cupr ic in solut ion (Wan , 1997, Ay lmore and Muir, 2 0 0 1 a , J i et a l , 2001) . Sulf i te c a n a lso be ox id ized by cupr ic copper or oxygen to sulfate and dithionate, depend ing on the condi t ions. Without sulfite, the fol lowing can occur in the p resence of the ox ides of ca lc ium, a s shown in Equat ion 2.36, and of iron, a lumin ium, m a n g a n e s e and copper (Ker ley and Bernard , 1981). C a O + A g 2 S 2 0 3 -» A g 2 S + C a S 0 4 [2.36] Sulf i te is a lso known to react with polythionates v ia the genera l react ion in Equat ion 2.37. Genera l l y this react ion equi l ibr ium lies to the right (Foss and Kr inglebotn, 1961). Equat ion 2.38 shows how ammon ium sulfite reacts with tetrathionate to produce thiosulfate and sulfate (Flett et a l , 1983) thus regenerat ing thiosulfate. Tri thionate, however , is stable in the p resence of sulfite (F leming et a l , 2003) . Sul fate has been sugges ted a s alternative to sulfite accord ing to Equat ion 2.39 (Hu and G o n g , 1991). S 0 3 2 _ + S x 0 6 2 - S^Oe 2 - + S 2 0 3 : 2- [2.37] ( N H 4 ) 2 S 0 3 + 2 N H 4 O H + ( N H 4 ) 2 S 4 0 6 -» 2 ( N H 4 ) 2 S 2 0 3 + ( N H 4 ) 2 S 0 4 + H 2 0 [2.38] S 0 4 2 - + S 2 - + H 2 0 S 2 0 3 2 - + 2 0 H " [2.39] Th is is unlikely a s sulfate is very stable (Aymore and Muir, 2001a) . 33 Conve rs ion of trithionate and tetrathionate to thiosulfate us ing sulf ide (e.g. N a H S ) or polysul f ides after removal o f ' go ld from the solut ion has a l so been sugges ted (Ji et a l , 2 0 0 1 , F leming et a l , 2003). Th is is demonst ra ted in Equat ions 2.40 and 2.41 (F leming et a l , 2003) . Conve rs ion rates of up to 99 % of both trithionate and tetrathionate have been ach ieved (Ji et a l , 2001). Sul f ide addit ion w a s advoca ted above sulfite addit ion a s sulfite cannot reduce trithionate at c o m m o n operat ing temperatures and pressures and sulf ide p roduces no sulfur containing by-products. A l s o sparg ing of a non-oxidis ing g a s s u c h a s H 2 S or S 0 2 has a lso been used to control the formation of polythionates. Alternat ive reductants s u c h a s hydrogen, fine reactive e lementa l sulfur and carbon monox ide have a l so been sugges ted (Ji et a l , 2001). Al l of these required prior gold removal to ensure no gold l osses by precipitat ion. Other degradat ion inhibitors such as phosphate and ethy lenediamine tetraacetic ac id ( E D T A ) have a lso been ment ioned briefly (Breuer and Jeffrey, 2003a) . T h e use of s u c h addit ives is expec ted to be limited to tank leaching appl icat ions and not to be of part icular use in heap leaching (Wan , 1997). 2.6 TRITHIONATE DEGRADATION T h e polythionates are wel l known for the e a s e with wh ich they undergo heterolytic c leavage in react ions with nucleophi l ic reagents, usual ly at a sul fenyl sulfur a tom due to the electrophi l ic character of the divalent sulfur a toms of the cha ins (Ritter and Krueger 1970). Depend ing on the propert ies of the nucleophi le , trithionate c a n undergo nucleophi l ic attack at either the sulfenyl sulfur a tom or the sul fonate sulfur a tom. Direct b imolecular nucleophi l ic substitut ion ( S N 2 ) at the sulfenyl sulfur a tom is shown in F igure 2.5 (Ritter and Krueger , 1970). It has been c la imed that of the two potential s i tes for nucleophi l ic attack, the sul fenyl sulfur a tom is much more react ive than the sul fonate S 3 0 6 2 - + S 2 - » 2 S 2 0 3 2 -4 S 4 0 6 2 - + 2 S 2 - + 6 O H " -» 9 S 2 0 3 2 - + 3 H z O [2.40] [2.41] 34 sulfur a tom. T h e sul fonate sulfur atom resists losing its sul fur-oxygen bonds (Sseka lo and B a m u w a m y e , 1993) and is general ly only subject to attack by hydroxide or other hard b a s e s (Ritter and Krueger , 1970). Soft polar izable b a s e s are expec ted to attack trithionate at the sulfenyl sulfur a tom. Hard b a s e s of low polarizabil i ty and high electronegativi ty (e.g. O H ' , ( C 2 H 5 ) 3 N ) are much less react ive towards trithionate (Ritter and Krueger , 1970). ~Nu O V 5+ o o = s - s - s = o I I _ O O J o II 0 = S - S - N u + ll o ' o II o - s I o Figure 2.5 - B imolecu lar nucleophi l ic substitution at sul fenyl sulfur of trithionate Spec i f i c examp les of nucleophi l ic substitution react ions relevant to gold leaching are d i s c u s s e d in fol lowing sect ions. 2.6.1 Interaction with Water Kur tenacker et al (1935) showed that the fol lowing two equat ions held for the decompos i t ion of trithionate in aqueous solut ions. S 3 0 6 2 - + H 2 0 -» S 2 0 3 2 _ + S 0 4 2 _ + 2 H + [2.42] 2 S 3 0 6 2 - + 3 H 2 0 -> S 2 0 3 2 - + 4 S 0 3 2 - + 6 H + [2.43] Equat ion 2.42 w a s found to occur f rom pH 5.6 to p H 12 at 50 °C whi le Equat ion 2.43 w a s found to occur at pH 13.4 at 50 °C. Whi le these equat ions have been wel l accep ted in the literature, the effects, if any, of the spec ies used for p H adjustments (e.g. hydroxide or carbonate ions) were not g iven any considerat ion. Kinet ic testwork on trithionate degradat ion in water is summar i zed in Tab le 2.6. 35 2.6.2 Interaction with Hydroxide B a s e d on Kur tenacker et al 's f indings (1935) the fact that trithionate degradat ion in water fol lows a different react ion at high pH (see Equat ions 2.42 and 2.43) impl ies that the hydroxide concentrat ion has an effect on the degradat ion path. In his work, the degradat ion rate w a s higher for pH va lues where Equat ion 2.43 w a s val id , than for Equat ion 2.42. Ro l ia and Chakrabar t i (1982) did not notice any effect of hydroxide concentrat ion, but the highest pH that they used w a s pH 11 (at 70 °C), lower than that where Equat ion 2.43 is expec ted to predominate. It has been sugges ted that hydroxide would attack trithionate directly at the sulfonate sulfur a tom (Ritter and Krueger , 1970). 2.6.3 Interaction with Ammonia It has been sugges ted that Equat ion 2.44 holds for the interaction between trithionate and a m m o n i a (Naito et a l , 1975). T h e p resence of sul famate w a s not conf i rmed in these tests. T h e stoichiometry of the ammono lys is react ion w a s verif ied by the authors in anhydrous liquid a m m o n i a , but not publ ished. Prev ious ly su l famate had been measu red during the degradat ion of trithionate at 110 °C at ammon ia concentrat ions of 4 to 8 N (Sh ieh et a l , 1965). B a s e d on this, it w a s a s s u m e d that su l famate would be a likely react ion product at 4 0 °C to 80 °C and an ammon ia concentrat ion between about 0.5 and 2.75 M. It is not known how likely it is for ammono lys is to occur in a q u e o u s a m m o n i a solut ions of low concentrat ion. Kinet ic testwork on trithionate degradat ion in a m m o n i a solut ions is summar i zed in Tab le 2.7. S 3 0 6 2 - + 2 N H 3 -> S 2 0 3 2 ' + N H 2 S 0 3 " + N H 4 + [2.44] 2.6.4 Interaction with Thiosulfate E x c h a n g e react ions are c o m m o n amongst the sulfur oxyan ions . T h e kinet ics of the exchange between trithionate and thiosulfate is summar i sed in Tab le 2.8, a long with the kinet ics of the effect of thiosulfate on trithionate degradat ion in water and aqueous a m m o n i a solut ions. T h e rate of the exchange react ion between trithionate and thiosulfate w a s found to inc rease with the concentrat ion and charge of the posit ive ions, hence it w a s p roposed that the interaction w a s between an ionic comp lex of thiosulfate, 36 trithionate or both, rather than between free ions (Fava and Pa jaro , 1954). A s a nucleophi le , thiosulfate is expec ted to interact with trithionate through the sul fenyl sulfur of trithionate. 2.6.5 Interaction with Copper Cupr i c copper has been reported to degrade trithionate (Kel ly and W o o d , 1994). S * s h o w s a radioact ively marked isotope of sulfur to give an indication of the react ion pathway, in Equat ion 2.45. T h e condit ions for this react ion were not g iven. [ 0 3 S - S * - S 0 3 ] 2 - + C u 2 + + 2 H 2 0 - C u S * + 2 S 0 4 2 " + 4 H + [2.45] In a study of the react ion between copper (II) and thiosulfate in ammon iaca l solut ions in the a b s e n c e of oxygen , it w a s proposed that the thiosulfate ion b e c a m e co-ordinated to the copper - a m m o n i a comp lex before copper reduct ion (Breuer and Jeffrey, 2003b) . Addi t ion of trithionate to this sys tem inc reased the copper reduct ion rate, s o it w a s a s s u m e d that trithionate can reduce copper but the poss ib le react ion occurr ing under these condi t ions w a s not d i s c u s s e d . Addi t ion of sulfate at the start of the test greatly reduced the react ion rate. It w a s deduced that the copper - trithionate react ion p roceeded v ia an inner sphere mechan i sm , where trithionate must first d isp lace sulfate f rom the copper co-ordinat ion sphere in order to react. It w a s c la imed that the react ions between cupr ic copper and trithionate or tetrathionate are faster than thiosulfate. However , we still s e e signif icant amounts of trithionate in gold leach solut ions. Th is may be an indication that there is insufficient copper (II) avai lab le or that the react ion between copper and trithionate is not as rapid a s sugges ted , or has a more comp lex dependency on other condi t ions (e.g. thiosulfate concentrat ion). A l s o , in recent work investigating the effects of minerals on thiosulfate degradat ion (De J o n g , 2004) , the formation rate of trithionate general ly inc reased in the p resence of copper , wh ich is contrary to what one would expect if copper reacted faster with trithionate than thiosulfate. 37 Tab le 2.6 - Trithionate degradat ion in water Ref. Exper imenta l Kinet ics Act ivat ion energy Mechan i sm D iscuss ion 1975 Naito U s e d ethanol / ?i?h water mixtures to vary [H 2 0 ] . s3o62-degradat ion determined by measur ing [S 2 0 3 2 1 and assuming the react ion in Equat ion 2.42. - d [ S 3 C V - ]/dt = k w [ H 2 0 ] [ S 3 C V T k w = 3.56 x 1CV6 M - 1 . m i n 1 at 40 °C k w = 9.80 x 10" 6 M - 1 .mirV 1 at 50 °C k w = 7.04 x 10" 5 M - 1 . m i n 1 at 70 °C Addi t ion of sal ts gave an average va lue of k w =1.08 x 10" 6 M - 1 . m i n 1 at 50 °C. Th is effect w a s not d i scussed by the authors. ?7n 3 f ln^ S 3 0 6 2 - + H 2 0 ^ S 3 O e . H 2 0 2 - P H not g iven - the - ou o ) react ion a s s u m e d to be S 3 0 6 . H 2 0 2 " v a , i d m a v not have -» S 2 0 3 2 - + S 0 4 2 - + 2 H + b e e n appropr iate. Indirect determinat ion of 2 n d s tep rate determining, kinet ics. Wate r w a s expec ted to attack the sul fonate sulfur a tom directly. Rolia Constant p H fjji tests using i a B ^ N a O H addit ion. - d [ S 3 0 6 2 - ]/dt = k w [H 2 0 ] [S 3 0 6 2 - ] (assumed) k w = 6 . 7 8 x 1 0 - 5 M - 1 . m i n 1 91.7 kJ/mol A s s u m e d Naito 's (70 - 85 °C) mechan ism. (70 °C, p H 11) MC r\ 2- 1U« — Temperatures higher than gold leach ing. Hof-mann-Bang, 1950 Decompos i t ion in sod ium acetate, p H 5 -9. S 2 0 3 2 " measured to infer S 3 0 6 2 _ degradat ion from Equat ion 2.42. - d [ S 3 0 6 " - ]/dt = kw[H20][S30 S 2 0 3 2 - + N H 2 S 0 3 + H + 2 step rate determining. Ammono lys i s reaction proceeds in competit ion with hydrolysis react ion. T h e stoichiometry w a s not veri f ied under the condit ions u s e d . T h e p H o f t h e ammono lys i s tests w a s not indicated. Indirect determinat ion of kinet ics. Tab le 2.8 - Trithionate degradat ion in the p resence of thiosulfate Ref. Exper imenta l K inet ics Act ivat ion energy Mechan i sm D iscuss ion - O a S - S - S O y + S ^ S O a ' -~o 3s-s-s*-so 3+so 3 2-^ 0 3 - S - S * - S 0 3 + S - S O Fava and Pajaro, 1954 E x c h a n g e react ion between S 2 0 3 2 _ and S 3 0 6 2 ' measu red using radioact ively marked S * o 3s-s-so 3 2+s*-so 3 2-O3S -S ' - S 0 3 2 + S - S 0 3 2 " Rate = k [ S 2 0 3 ] [ S 3 0 6 ] Ra te not apprec iab ly affected by pH (7 to 10) or by a 10-fold increase in the sur face to vo lume ratio. k = var ied with concentrat ion of inert sa l ts , espec ia l ly concentrat ion and charge of posit ive ions. 56 kJ /mol (25 - 51 °C) Independent of the type or concentrat ion of inert salt. 2-Proposed that reaction w a s between trithionate and ionic complex of thiosulfate (due to effect of positive ions). -1^ o Naito et al, 1975 S 3 0 6 degradat ion determined by measur ing [ S 2 0 3 2 T and assuming the react ion in Equat ion 2.42. - d l S a O ^ / d t = ( k w [ H 2 0 ] + k a [NH 3 ] + k t [ S 2 0 3 2 l ) [ S 3 0 6 2 - ] k, = 4 . 1 5 x 1 0 - * M- 1 .min- 1 at 40 °C k, = 2 . 0 5 x 10" 2 M " 1 . m i n 1 at 6 0 ° C 53 kJ /mol (40 - 80 °C) Thiosulfate essent ial ly behaves as a catalyst. s3o6 2"+s2o3 2-^ S3O6 . S 2 0 3 S 3 0 6 2 " . S 2 0 3 2 " + H 2 0 - * 2 S 2 0 3 2 - + S 0 4 2 " + 2 H + 1 s t step rate determining. T h e mechan i sm proposed would involve a large transit ion state - it w a s not expla ined why this would be feas ib le or why the react ion rate would inc rease. Rolia et al, 1982 Constant p H tests, pH 5.5 - 8, 85 °C S 2 0 3 2 " (up to 10 mM) increased S 3 0 6 2 ' degradat ion rate. Naito et a l 's mechan ism a s s u m e d (1975) 2.7 TETRATHIONATE DEGRADATION A s d i scussed in the sect ion on trithionate (Sect ion 2.6) the polythionates are wel l known for their react ions with nucleophi l ic reagents. Th is sect ion examines s o m e of the react ions between tetrathionate and other solution components . 2.7.1 Interaction with Hydroxide Tetrathionate is known to be unstable in alkal ine solut ions. Tetrathionate can be d e c o m p o s e d to var ious products by increasing the pH (Muir and Ay lmore , 2002 , Ro l ia and Chakrabar t i , 1982, Nai to et a l , 1970b, Lyons and N ick less , 1968). Degradat ion of tetrathionate in aqueous ammon ia showed the fol lowing (Naito et a l , 1970b, Smi th and Hi tchen, 1976). pH 8.9 2 S 4 0 6 2 - -> S 3 0 6 2 - + S 5 0 6 2 ' [2.46] pH 11.5 4 S 4 0 6 2 - + 6 O H " ^ 2 S 3 0 6 2 " + 5 S 2 0 3 2 " + 3 H 2 0 [2.47] Further decompos i t ion of the trithionate and pentathionate formed is shown in Equat ions 2.48 to 2.50 (Rol ia and Chakrabar t i , 1982, Naito et a l , 1970b). pH 12 S 3 0 6 2 - + 2 0 H " -> S 2 0 3 2 " + S 0 4 2 " + H 2 0 [2.48] pH 13 2 S 3 0 6 2 " + 6 O H " -» S 2 0 3 2 " + 4 S 0 3 2 " + 3 H 2 0 [2.49] H e n c e , if tetrathionate is aged in strongly alkal ine solut ions, it will f inally form thiosulfate and sulfite (Naito et a l , 1970b) a s shown in Equat ion 2 .51. Other react ion s c h e m e s for the alkal ine decompos i t ion of tetrathionate have a lso been sugges ted . Tetrathionate has been c la imed to degrade to thiosulfate, sulfite and sulfoxyl ic ac id (S (OH) 2 ) , with further degradat ion of the sulfoxyl ic ac id to sulf ide (Muira and K o h , 1983). T h e react ion equat ions are shown below. It would be expec ted that the sulf ide and sulfite in Equat ion 2.54 would combine to produce thiosulfate. 2 S 5 0 6 2 " + 6 O H " ^ 5 S 2 0 3 2 " + 3 H 2 0 [2.50] 2 S 4 0 6 2 " + 6 0 H " 3 S 2 0 3 2 " + 2 S 0 3 2 " + 3 H 2 0 [2.51] 41 10 S 4 0 6 2 - + 20 O H " -» 10 S 2 0 3 2 " + 10 S ( O H ) 2 + 10 S 0 3 ; 6 S ( O H ) 2 + 6 O H - > 3 S 2 0 3 2 - + 9 H 2 0 4 S ( O H ) 2 + 8 O H " -» % S 2 - + 8 / 3 S 0 3 2 " + 8 H 2 0 2- [2.52] [2.53] [2.54] There have been s o m e studies on the kinetics of tetrathionate decompos i t ion in alkal ine sys tems , with detai ls g iven in Tab le 2.9. In the a b s e n c e of ammon ia , copper and oxygen , the rate of the alkal ine decompos i t ion of tetrathionate w a s descr ibed by Z h a n g and Dre is inger (2002). No build-up of trithionate w a s found in this testwork, implying that either trithionate w a s not a part of the main react ion m e c h a n i s m or that it d e c o m p o s e d at a simi lar rate to tetrathionate. T h e poss ib le catalyt ic effect of the thiosulfate product on further tetrathionate degradat ion w a s not cons idered . Ro l ia and Chakrabar t i (1982) did similar work but in the p resence of oxygen , and found that whi le the rate law could be exp ressed in the s a m e way a s that of Zhang and Dre is inger (see Tab le 2.9), the rate constant w a s more than 10 t imes smal ler . It w a s p roposed by Z h a n g and Dreis inger that the p resence of oxygen in the work of Ro l ia et a l . and the a b s e n c e of oxygen in their work w a s probably the reason for this d isc repancy , and that oxidants could poss ib ly retard the rate of tetrathionate degradat ion. However , s imi lar work by Breuer and Jeffrey (2004) showed that tetrathionate degradat ion under air-saturated condi t ions gave no measurab le di f ference in kinet ics to tests under ni trogen. The i r results were simi lar to those of Z h a n g and Dreis inger. It w a s sugges ted that rather than the p resence of oxidants being respons ib le for the di f ference in kinet ics measured by the two previous groups , variat ions in the ionic strength of the test solut ions w a s responsib le (Breuer and Jeffrey, 2004) . Increasing the ionic strength w a s found to have a signif icant effect on the initial rate of tetrathionate decompos i t ion . 42 2.7.2 Interaction with Ammonia T h e effect of ammon ia on the rate of tetrathionate degradat ion w a s measu red at 50 to 80 °C by Naito et al (1970b). T h e ammon ia concentrat ion w a s found to have a signif icant effect on the rate, with tetrathionate decompos ing much faster at higher a m m o n i a concentrat ions. In addit ion to the trithionate, thiosulfate and pentathionate formed during the alkal ine decomposi t ion of tetrathionate, the formation of su l famate ( S 0 3 N H 2 " ) has a lso been proposed in the p resence of a m m o n i a (without copper) . After trithionate w a s produced by the alkal ine degradat ion of tetrathionate (Equat ion 2.47), su l famate w a s reportedly produced (Equat ion 2.55) (Naito et a l , 1970b, 1975). S 3 0 6 2 ' + N H 3 + O H " S 0 3 N H 2 - + S 2 0 3 2 " + H z O [2.55] T h e sul famate yield depended on the ammon ia concentrat ion. If the a m m o n i a concentrat ion w a s less than 2 N, hardly any su l famate formed. If there w a s not enough a m m o n i a , the solut ion b e c a m e neutral or weak ly ac id ic with the precipitation of sulfur and the formation of higher polythionates. Wi th e x c e s s free ammon ia , the overal l react ion gave sul famate instead of sulfate. (Equat ions 2.56 and 2.57) 4 S 4 0 6 2 ' + 10 O H " 7 S 2 0 3 2 " + 2 S 0 4 2 " + 5 H 2 0 [2.56] 4 S 4 0 6 2 _ + 2 N H 3 + 8 O H " -» 7 S 2 0 3 2 - + 2 S 0 3 N H 2 - + 5 H 2 0 [2.57] In the p resence of air or oxygen a simi lar set of equat ions could be written (Equat ions 2.58 and 2.59). 3 S 4 0 6 2 - + 10 O H " + % 0 2 -> 4 S 2 0 3 2 " + 4 S 0 4 2 " + 5 H 2 0 [2.58] 3 S 4 0 6 2 " + 4 N H 3 + 6 O H " + % 0 2 ^ 4 S 2 0 3 2 " + 4 S 0 3 N H 2 + 5 H 2 0 [2.59] Breuer and Jeffrey (2004) found that tetrathionate decompos i t ion in the p resence of an a m m o n i u m sulfate / ammon ia buffer w a s faster than for a carbonate buffer at the s a m e p H , even though the ionic strength w a s lower. A simi lar observat ion w a s made for phosphate buffers in compar ison with an ammon iaca l sys tem. T h e s e authors had p roposed that an increase in ionic strength would increase the rate of tetrathionate degradat ion. Th is observat ion w a s not exp la ined. 43 2.7.3 Interaction with Copper In an investigation of the react ion between thiosulfate and cupr ic copper in the p resence of oxygen (Breuer and Jeffrey, 2003b) , the effect of tetrathionate and trithionate on the react ion were a l so invest igated briefly. Addi t ion of tetrathionate inc reased the copper (II) reduct ion rate, implying that tetrathionate w a s a lso ox id ized by copper (II). However , depend ing on the initial condi t ions, the tetrathionate concentrat ion w a s expec ted to reach a s teady state, when its product ion (by thiosulfate oxidation) and consumpt ion (by copper (II) reduction) were equa l . If sulfate w a s added with tetrathionate to the thiosulfate sys tem, the initial rate of copper (II) reduct ion w a s not af fected, hence it w a s conc luded that tetrathionate did not need to complex with copper (II) for reduct ion to occur . S i n c e the addit ion of sulfate had been found to s low the react ion between copper (II) and thiosulfate (and hence the product ion of tetrathionate), then the amount of tetrathionate at s teady state would be lower in this c a s e . Simi lar ly, if addit ion of ammon ium ions (reduction of pH) increased the copper (II) reduct ion react ion, the s teady state concentrat ion of tetrathionate would be expec ted to be higher. S o m e sol id copper compounds have been sa id to inc rease the rate of tetrathionate oxidat ion. ( C h a n d a and R e m p e l , 1986) Chalcopyr i te , covell i te and chalcoc i te have all been sugges ted to cata lyse oxidat ion of tetrathionate by air. In the c a s e of cuprous ox ide cata lys is , only sulfate w a s formed a s a product. N o react ion equat ion w a s g iven. 2.7.4 Interaction with Thiosulfate T h e d isp lacement react ion between thiosulfate and the polythionates is wel l known (Foss , 1961, F o s s and Kr ingelbotn, 1961, F a v a and B resado la , 1955). Tetrathionate can undergo nucleophi l ic attack at the sulfenyl sulfur, with either sulfite or thiosulfate d i sp laced , a s in Equat ions 2.60 and 2.61 (where S * denotes a marked sulfur atom). Simi lar ly sulfite can react with tetrathionate, d isp lac ing thiosulfate, a s in Equat ion 2.62. [ 0 3 S S S 2 0 3 ] 2 " + [ S * S 0 3 ] 2 " -» [ 0 3 S S * S S 2 0 3 ] 2 " + S 0 3 : [ 0 3 S S S 2 0 3 ] 2 - + [ S * S 0 3 ] 2 - [ 0 3 S S * S 0 3 ] 2 - + [s2o3r |2-2- [2.60] [2.61] 44 [o3sss2o3]2- + so 3 2- -> [ 0 3 S S S 0 3 ] 2 - +[ s2o3]2- [2.62] T h e kinet ics of Equat ion 2.60 were invest igated in the p resence of formaldehyde as a sulfite acceptor (Foss and Kr ingelbotn, 1961), at neutral p H . A summary of the kinetics f indings is shown in Tab le 2.9. T h e posit ive inf luence on the rate by posit ive ions implied that the intermediate complex for the react ion w a s likely formed not between free ions but ionic comp lexes of tetrathionate, thiosulfate or both. T h e equat ions above (2.60 and 2.62) can be used to expla in the catalyt ic effect of thiosulfate on tetrathionate degradat ion. Under non-oxidis ing condi t ions, s o m e thiosulfate (about 62 .5 %) can be regenerated from the decompos i t ion of tetrathionate to higher or lower polythionates through the formation of trithionate (Aylmore and Muir, 2 0 0 1 a , M a r s d e n and H o u s e , 1992). Th is is highly ca ta lysed by thiosulfate, a s represented in Equat ions 2.63 to 2.66 (Byer ley et a l , 1973a , G e l v e s et a l , 1996). S 4 0 6 2 - + S 2 0 3 2 - S 5 0 6 2 - + S 0 3 2 - [2.63] S 0 3 2 - + S 4 0 6 2 - -» S 3 0 6 2 ' + S 2 0 3 2 - [2.64] T h e effect of thiosulfate w a s add ressed in a study by Ro l ia and Chakrabar t i (1982) who der ived a rate equat ion for tetrathionate degradat ion in the a b s e n c e of a m m o n i a at pH 11 . Thiosul fate w a s found to increase the rate of tetrathionate degradat ion. T h e kinetics results are shown in Tab le 2.9. However , Breuer and Jeffrey (2004) p roposed that the apparent inc rease in rate when thiosulfate w a s present w a s due to the increase in ionic strength that thiosulfate addit ion wou ld represent rather than the thiosulfate itself. S 5 0 6 2 - + 3 O H " -» 5 / 2 S 2 0 3 2 - + % H 2 0 Overa l l : 2 S 4 0 6 2 - + 3 O H " % S 2 0 3 2 " + S 3 0 6 2 " + % H 2 0 [2.65] [2.66] 45 Tab le 2.9 - Tetrathionate degradat ion in neutral to alkal ine solut ions Act ivat ion energy Mechan ism Ref. Exper imenta l K inet ics D iscuss ion 98.5 kJ/mol Sugges ted presence of (22-40 °C) d issolved oxygen and/or copper had a role in mechan ism. Z h a n g Alka l ine et a l , degradat ion. 2002 A b s e n c e of oxygen . In b icarbonate/ hydroxide or H P C V h y d r o x i d e buffer, deaerated with N 2 , under N 2 -d[S4cV- ]/dt = k [ O H l [ S 4 0 6 2 " ] k = 0 . 3 8 x 1 0 3 M" 1 .s" 1 (22 °C, pH 10 -11 .5 ) pH and temperature dramatical ly affect degradat ion rate in alkal ine solut ion. Ro l ia Alka l ine - d [ S 4 0 6 2 " ]/dt 115.5 kJ /mol None given. Rate constant 10 et a l , degradat ion in = (k 1 + k 2 [S 2 0 3 2 1) [OH1 [S 4 0 6 2 l (15-45 °C, t imes smal ler than 1982 p resence of pH 11) Zhang . thiosulfate. = 0.022 M \ s " 1 Constan t pH k 2 = 2.77 M" 2 . s 1 tests us ing (25 °C, pH 11) N a O H addit ion. O x y g e n present. 0\ Naito et a l , 1970b A m m o n i a c a l degradat ion at 50 - 8 0 °C None given In p resence of >2M ammon ia , sul famate formed f rom further degradat ion of trithionate product. F o s s et a l , 1961 Thiosul fate exchange react ions - d [ S 4 0 6 2 - ]/dt = k [S 2 0 3 2 1 [S 4 0 6 2 T k = 1.3 x IO" 3 M" 1 . s ' 1 (25 °C, l = 1.15 M) 50.2 kJ /mol ( 2 0 - 4 0 °C, I = 1.15 M) S 4 0 6 2 " + s2o32- s5o62-+ S 0 3 2 ' v ia nucleophi l ic d isp lacement 2.8 REMOVAL OF POLYTHIONATES FROM SOLUTION Polythionates are detrimental in gold recovery from thiosulfate leach solutions using adsorbants such as resins. Besides the reactions already discussed to remove polythionates from solution, by hydrolysis, ammonolysis or interchange between various sulfur species, it is possible to remove polythionates in other ways. While many of these methods do not allow for recovery of thiosulfate, this may not be of importance in cases where reduction of thiosulfate consumption is not critical. There is also an environmental concern of discharge of polythionates, as these can produce acid on further oxidation to sulfate under suitable conditions. A few potential methods are given in Table 2.10. This summary simply gives an indication of the types of processes available and has not been studied in any detail for this review. Table 2.10 : Potential Methods for Polythionate Removal Method Reference Applicability Bacterial oxidation Hansford and Vargas, 2001 De Jong et al, 1997 Sand etal, 1995 A number of bacteria, generally requiring acidic conditions, can oxidize various polythionates to sulfate. Mn0 2 oxidation Schlppers and Jorgenson, 2001 Oxidation to sulfate. HOCI oxidation Horvarth and Nagypal, 2000 Oxidation to sulfate. Sulfite addition Kerley and Bernard, 1981 Ji etal, 2001 Reversal to thiosulfate and other lower polythionate. Sulfide addition Ji etal, 2001 Reversal to thiosulfate. Alkaline treatment Zhang and Dreisinger, 2002 Reversal to thiosulfate and other polythionates. Cyanide Nor and Tabatabai, 1975 Koh,1990 Mizoguchi and Okabe, 1975 Thiosulfate and sulfate produced, and other unwanted products. Selective resin adsorption Wassink, 2002 Anion exchange resins can be used to adsorb thiosulfate and polythionates, with selective elution of thiosulfate. 47 2.9 SUMMARY OF LITERATURE FINDINGS There is a limited understanding of the degradat ion of thiosulfate in gold leaching sys tems . Wh i le thiosulfate consumpt ion is usual ly documented in leach tests, the degradat ion products are not often identified or quanti f ied. There have been a number of fundamenta l s tudies on the degradat ion of thiosulfate to predominant ly trithionate and/or tetrathionate. It s e e m s that one can expect trithionate to form directly when oxygen is present, rather than as a decomposi t ion product of tetrathionate, whi le tetrathionate is expec ted in the a b s e n c e of oxygen . T h e m e c h a n i s m s by which these s p e c i e s form and the w a y s in which the solut ion condi t ions affect their rates of formation and their proport ions formed relative to other sulfur oxyan ions is not understood. However , there is currently a fair amount of research being devoted to the study of thiosulfate degradat ion under gold leaching condit ions. T h e subsequen t behaviour of the thiosulfate degradat ion products is not well understood. Tetrathionate has been identified in gold leach solut ions, but it is usual ly quick to degrade under the alkal ine condit ions used in gold leaching. Thiosul fate ca ta lyses this degradat ion, and it is postulated that copper can a lso inc rease the degradat ion rate. Tetrathionate is not a s persistent as trithionate under many gold leaching condi t ions. Tri thionate is more persistent in gold leaching solut ions but very little is known about this spec ies in the context of gold leaching. M u c h of the avai lable literature concern ing trithionate has not been directly related to gold leach ing: the older literature is fundamenta l in nature and other studies have been publ ished in the context of sulfur oxidat ion pathways during the ammon iaca l treatment of sul f ides. T o be able to a s s e s s thiosulfate degradat ion in a larger context, it is essent ia l to deve lop a better understanding of the behaviour of trithionate. v Al leviat ing the problem of thiosulfate degradat ion is critical to the s u c c e s s of the thiosulfate leaching p rocess for gold. Th is review has shown that whi le the degradat ion of thiosulfate is not fully unders tood, there is a fair amount of research currently underway to invest igate fundamenta ls of this degradat ion p rocess . However , to be ab le to understand the sys tem in its entirety, it is necessa ry to understand the behaviour of the other sulfur oxyan ions expec ted to be present, mainly trithionate and tetrathionate. 48 Of these two spec ies , trithionate often appears to be more persistent in the alkal ine solut ions u s e d , but very little is understood about this spec ies . T h e fol lowing chapters thus d i scuss exper imenta l work carr ied out to better understand the behaviour of trithionate, fol lowed by the integration of the f indings with the expec ted behaviour of thiosulfate and tetrathionate based on literature observat ions. 2.10 SCOPE AND OBJECTIVES T h e speci f ic s c o p e and object ives of this study are outl ined below: • T o further the understanding of trithionate solut ion chemistry under condit ions relevant to gold leaching by thiosulfate by exper imental ly determining the kinetics of trithionate degradat ion and identifying the effects of var ious solut ion condit ions on the kinet ics. • T o incorporate the f indings into a mode l to predict trithionate degradat ion kinet ics. • T o use literature data and rate equat ions descr ib ing thiosulfate and tetrathionate react ions, a s wel l a s the exper imental ly der ived trithionate degradat ion rate equat ion to deve lop a s imple kinetic model to determine the expec ted sulphur oxyan ion solut ion speciat ion during gold leaching us ing thiosulfate. • T o evaluate the sensit ivity of the model to parameters in the rate equat ions. • T o evaluate the ability of the model to adequate ly descr ibe exper imenta l data . • T o identify l imitations of the model to determine where further research is required to improve the mode l . • T o sugges t w a y s in which thiosulfate degradat ion or the formation of polythionates can be min imised during gold leach ing. 49 3 ANALYTICAL METHODS AND SYNTHESIS 3.1 INTRODUCTION In this chapter, the analyt ical methods used in this work are descr ibed . A l s o , the synthes is and character isat ion of the sod ium trithionate used in the kinetic work and for analyt ical cal ibrat ion purposes is descr ibed . T h e exper imenta l procedure for the kinetic testwork is descr ibed in Chap te r 4. 3.2 ANALYSIS OF SULFUR OXYANIONS - ION CHROMATOGRAPHY 3.2.1 Description of Method Solut ion samp les were ana lysed for thiosulfate, trithionate, tetrathionate and sulfate us ing high per formance liquid chromatography ( H P L C ) using a D ionex Se r i es 600 sys tem. Ana lys i s of thiosulfate, trithionate and tetrathionate involved separat ion of the spec ies on an O m n i P a c P A X - 1 0 0 co lumn, and ana lys is of the separa ted spec ies us ing UV-v is ib le absorpt ion spectrometry at 205 nm. Sul fate w a s separa ted from the other sulfur oxyan ions us ing an l o n P a c A S 4 A - S C co lumn and measu red by conductivity detect ion. B e c a u s e the determinat ion of sulfate w a s by a different method, it w a s not poss ib le to ana lyse a s ingle samp le for all the spec ies of interest s imul taneously. More detai ls of the chromatography methods are given in Append ix 3. Chromatography requires the use of s tandard solut ions to cal ibrate the instrument. T h e preparat ion and s torage of these solut ions is d i s cussed in Sect ion 3.2.2. A l l s a m p l e s and s tandards were diluted to the required concentrat ion range (general ly less than 20 mg/l) using ultra-pure de ion ised water and ana lysed immediately. lodometr ic titration is often used to determine thiosulfate, but it is difficult to determine thiosulfate and the polythionates individually in mixtures. C o p p e r a lso interferes with this method (Wass ink , 2002). 50 3.2.2 Stability of Standard Solutions Experimental Work Cal ibrat ion s tandard solut ions of sod ium thiosulfate ( S I G M A , anhydrous , >99% pure), sod ium trithionate (synthes ised, s e e Sect ion 3.5) and sod ium tetrathionate ( S I G M A , dihydrate) were prepared in de- ion ised water. T h e salts were made up to 1000 mg/l solut ions of e a c h an ion, then diluted to 100 mg/l solut ions, and then to 2, 10, 15 and 20 mg/l solut ions to be used in cal ibrat ion. T h e cal ibrat ion s tandards were ana lysed within two hours of preparat ion. T o test the stability of s tandard solut ions over longer t imes and under different s torage condi t ions, 100 mg/l and 10 mg/l solut ions of thiosulfate, trithionate and tetrathionate (individually and in combinat ions) were made up in the s a m e way a s the cal ibrat ion s tandards . T h e 100 mg/l solut ions were diluted to 10 mg/l immediately before ana lys is . S a m p l e s of the var ious 100 mg/l and 10 mg/l solut ions were stored in g lass bottles at room temperature on the laboratory bench , in the dark, with the head s p a c e purged with ni trogen, in the fr idge and in the f reezer (defrosted at room temperature immediately prior to analys is) . S o m e solut ions were a l so stored in plast ic bott les at room temperature on the laboratory bench . After set t imes up to about three w e e k s , the solut ions were ana l yzed for thiosulfate, trithionate and tetrathionate. Results Resu l ts are genera l and quantitative. S o m e instrumental p rob lems were being exper ienced at the t ime of this work giving spur ious results in c a s e s . T h e overal l t rends are summar i sed in Tab le 3.1. 51 Tab le 3.1 : S u m m a r y of observat ions on stability of s tandard solut ions of thiosulfate, trithionate and/or tetrathionate Initial solut ion concentrat ion C h a n g e s to thiosulfate C h a n g e s to tetrathionate C h a n g e s to trithionate 10 mg/l S 2 0 3 2 " < 10 % degradat ion over 22 days . Slightly more stable in f reezer. Fo rmed a s S 2 0 3 2 " degradat ion product. None . 100 mg/l S 2 0 3 2 " < 10 % degradat ion over 22 days . Simi lar to 10 mg/l solut ion. S e e n a s S 2 0 3 2 " degradat ion product None . 1 0 m g / I S 4 O 6 2 - Negl ig ib le. < 10 % degradat ion over 18 days . F reez ing enhanced degradat ion to - 2 0 % in 18 days . Negl ig ib le except for that formed f rom S 4 0 6 2 " degradat ion for f rozen samp le in 1:1 ratio. 100 mg/l S 4 0 6 z - Negl ig ib le. < 10 % degradat ion over 18 days . F reez ing more stable than for 10 mg/l solut ion - - 1 3 % degradat ion in 18 days . Negl ig ib le except for that formed f rom S 4 0 6 2 " degradat ion for f rozen samp le in 1:1 ratio. 10 mg/l S 3 0 6 2 " Fo rmed as degradat ion product cor responding to 1:1 ratio of S 3 0 6 2 " deg raded . Sma l l amount present. A n exchange between S 3 0 6 2 " and S 4 0 6 2 " w a s noted for solut ions of S 3 0 6 2 " in water (see Sec t ion 3.6). - 4 0 % degradat ion in 14 days . Stabi l ised slightly in fridge to g ive - 2 0 % degradat ion in 14 days . 10 mg/l S 2 0 3 2 " 10 mg/l S 4 0 6 2 " Destabi l ised by p resence of S 4 0 6 2 \ Typical ly 30 - 50 % degradat ion in 22 days . F reez ing stabi l ised. Increased accord ing to S 2 0 3 2 " degradat ion. F reez ing c a u s e d rapid degradat ion of S 4 0 6 2 " by - 6 0 % in 8 days . Fo rmed f rom S 4 0 6 2 " degradat ion in 1:1 ratio for f rozen samp le . 100 mg/l S 2 0 3 2 " 100 mg/l S 4 0 6 2 -Negl igible degradat ion. Negl ig ib le change except for f reezing where - 5 0 % degraded in 8 days . Fo rmed f rom S 4 0 6 2 " degradat ion in 1:1 ratio for f rozen samp le . 10 mg/l S 2 0 3 2 " 10 mg/l S 4 0 6 2 -10 mg/l S 3 0 6 2 " Degradat ion in simi lar range as other solut ions. Stab le on f reez ing. Increases except for f reezing where - 20 % degraded in 1 day. D e g r a d e s at s imi lar rate to S 3 0 6 2 " a lone. Increases on f reezing f rom S 4 0 6 2 " degradat ion in 1:1 ratio. 100 mg/l S 2 0 3 2 " 100 mg/l S 4 0 6 2 -100 mg/l S 3 0 6 2 " S tab le . Simi lar to 10 mg/l solut ion. Sl ightly more stable than 10 mg/l solut ion. Increases on f reezing from S 4 0 6 2 " degradat ion in 1:1 ratio. 52 In genera l , the condit ions under which the sulfur oxyan ion solut ions were stored did not signif icantly affect their stability, with the except ion of refrigeration which stabi l ised trithionate, and f reezing which signif icantly enhanced tetrathionate degradat ion. In genera l , the more concentrated solut ions were more stable. It w a s conc luded that cal ibration standard solut ions should be used immediately after preparat ion for the best accuracy . 3.2.3 Effect of Other Solution Components on Ion Chromatographic Analysis A number of non-sulfur spec ies expec ted to be present in the solut ions in the kinetic tests were added to solut ions of thiosulfate, trithionate and tetrathionate to determine the effect of these components on the ion chromatographic method. Usua l ly it is r ecommended to use 'matrix matching ' of s tandard solut ions used for ion chromatography cal ibrat ion, however , in this c a s e many of the componen ts tested were known to or suspec ted to react with the analyte spec ies , s o matrix matching could not u s e d . Typ ica l test solut ions in the kinetic testwork (Chapter 4) were expec ted to contain 2-7 g/l trithionate, 0.1-0.4 g/l thiosulfate and negl igible tetrathionate. Other solut ion componen ts expec ted to be present (individually or in combinat ion) were ammon ia , a m m o n i u m , b icarbonate, carbonate, chlor ide and potass ium (from potass ium chlor ide added) . Fo r the purposes of testing the effects of these componen ts on the ana lys is of the sulfur oxyan ions , a solut ion of 100 mg/l e a c h of thiosulfate, trithionate and tetrathionate w a s prepared. F rom this solut ion a set of solut ions of 10 mg/l of e a c h sulfur oxyanion w a s prepared, containing the componen ts of interest at the concentrat ions shown in Tab le 3.2. T h e concentrat ions c h o s e n represent typical levels of e a c h component relative to levels expec ted in the test solut ions (taking into account the dilution required for ion chromatography). E a c h samp le w a s ana lysed at least three t imes s ince it is recogn ised that there is s o m e variabil ity in the ion chromatographic method. T h e m e a n and s tandard deviat ion for e a c h analyte is shown in Tab le 3.2. T h e largest di f ference in measu red concentrat ion compared with a s tandard solut ion with no extra componen ts added w a s 2.7 % for 53 thiosulfate, 1.7 % for trithionate and 2.4 % for tetrathionate. The error expected in preparing the diluted samples was expected to be larger than this, hence it was concluded that the addition of ammonia, ammonium, bicarbonate, carbonate or potassium chloride to the levels indicated in Table 3.2 did not affect the analysis of thiosulfate, trithionate and tetrathionate. The effect of these species on the analysis of sulfate was not tested as sulfate was only measured occasionally and was not used directly in deriving the kinetic results shown in Chapter 5. The effect of copper on analysis was not tested as it is known that cupric copper reacts readily with thiosulfate and that cuprous copper is readily oxidized to cupric in the presence of oxygen. Table 3.2 : Effect of added species on analysis of sulfur oxyanions Species added Concentration of added species (mM) Mean measured value and standard deviation for nominal 10 mg/l of each of S 2 0 3 2 ' , S 3 0 6 2 ' and S 4 0 6 2 ' s2o3*- S30^ S 4 0 6 2 " mean 0 mean a mean 0 None - 10.0581 0.0169 10.2042 0.0111 10.1115 0.0151 NH 3 1 10.2037 0.0292 10.3763 0.0670 10.4557 0.1214 NH 4 + as (NH 4 ) 2 S0 4 1 10.0196 0.0469 10.2162 0.0448 9.8734 0.0846 HCCV as NaHC0 3 1.3 10.0502 0.0067 10.1683 0.0287 10.1256 0.0151 C 0 3 2 ' as N a 2 C 0 3 0.9 10.3319 0.0432 10.3315 0.0193 10.2516 0.0929 KCI 1.9 10.0702 0.0224 10.1519 0.0169 10.0504 0.0154 3.3 ANALYSIS OF SULFAMATE Sulfamate (NH 2S0 3 ') is a potential species of interest in ammoniacal thiosulfate systems. Two methods were considered to analyse for sulfamate. A titration method used by Sherritt (Liebovitch, 2005) was found to be unsuitable as unreliable results were expected for sulfamate levels below 1 g/l (which is much higher than anticipated levels in the test solutions) and thiosulfate is known to interfere with the determination. It is 54 suspec ted that trithionate would a lso interfere, espec ia l l y in the solut ions of interest where it is expec ted to be present in e x c e s s compared with sul famate. A chromatograph ic method using conductivity detect ion (under the s a m e condit ions as that for sulfate determination) w a s found to be suitable to determine sul famate and sulfate s imul taneously , but in the p resence of ammon ia the sul famate peak w a s obscu red . In addit ion, s ince the sul famate concentrat ion in the test solut ions w a s expec ted to be very low (or even non-existent) the l ikel ihood of being ab le to opt imise the ion chromatograph ic method to determine sul famate w a s cons idered low. 3.4 A N A L Y S I S O F T O T A L A M M O N I A A s tandard distil lation method w a s used to ana lyse for total ammon ia . S o d i u m hydroxide (1M) w a s transferred to a round-bottomed f lask connec ted to a condenso r with a spray tube. A suitable al iquot of ammonia-conta in ing solution and ant i -bumping granu les were added to the f lask. T h e f lask w a s heated and the ammon ia g a s produced w a s col lected in a hydrochlor ic ac id solut ion (1M HCI). Disti l lation w a s cont inued until the vo lume in the round-bot tomed f lask w a s reduced by about half and no more g a s bubbles were observed to be enter ing the ac id solut ion. T h e e x c e s s ac id w a s titrated against a s tandard sod ium hydroxide solut ion to determine the total a m m o n i a concentrat ion. T h e a m m o n i a concentrat ion w a s determined by subtract ing the known ammon ium concentrat ion of the original solut ion from the total a m m o n i a content determined by disti l lation. Rep l ica te tests involving dilution of a stock a m m o n i a solut ion fol lowed by ana lys is of the diluted solut ion for total ammon ia concentrat ion gave up to a 10 % range in results. 3.5 S Y N T H E S I S O F S O D I U M T R I T H I O N A T E S o d i u m trithionate is not commerc ia l ly ava i lab le , s o w a s syn thes ised accord ing to a method descr ibed by Kel ly and W o o d (1994). T h e method involved oxidat ion of sod ium thiosulfate us ing hydrogen peroxide at low temperature (near 0 °C). A number of ba tches were synthes ized accord ing to the fol lowing method, based on Equat ion 3.1. 2 S 2 0 3 2 ' + 4 H 2 0 2 S 3 0 6 2 - + S 0 4 2 " + 4 H 2 0 [3.1] 55 S o d i u m thiosulfate pentahydrate (150 g) w a s d isso lved in 90 ml de- ion ised water in a g lass beaker. T h e beaker w a s p laced in a cool ing reactor to reduce the temperature to about 1 °C. Wi th cont inuous stirring, 140 ml 30 % (w/v) hydrogen perox ide w a s added dropwise (over a few hours), taking care that the temperature remained below 20 °C. Stirring w a s c e a s e d and the beaker w a s left at about 0 °C for 1 to 2 hours, al lowing for crystal l izat ion of sod ium sulfate. T h e sod ium sulfate w a s removed by filtration through a W h a t m a n No. 1 filter paper. T h e sulfate w a s w a s h e d on the filter with 100 ml ethanol wh ich w a s a l lowed to mix with the filtrate. T h e filtrate w a s transferred to a beaker at about 3 °C, 250 ml ice-cold ethanol w a s added and the solut ion left at 0 - 3 °C for one hour. T h e result ing precipitate (mostly sulfate) w a s aga in removed by filtration, and w a s h e d on the filter with 200 ml ice-cold ethanol , which w a s a l lowed to mix with the filtrate. T h e filtrate w a s transferred to a beaker containing 1 I ice-cold e thanol , and 100 ml ethanol w a s used to r inse the filtrate from the f lask into the beaker . T h e mixture w a s stirred thoroughly and left at 0 - 3 °C for 1 - 2 hours. S o d i u m trithionate formed and w a s separa ted by filtration and w a s h e d with 50 ml ethanol , 50 ml ace tone and dried in a des iccator . A white crystal l ine material w a s produced and stored in a g lass bottle in a refrigerator. Stor ing of s u c h sol ids at low temperature has been recommended (Miura, 2003) . 3.6 CHARACTERISATION OF SODIUM TRITHIONATE 3.6.1 Total Sulfur T h e total sulfur content w a s determined by complete ly oxid is ing all the sulfur spec ies to sulfate and analys ing sulfate. Comp le te oxidation of trithionate to sulfate by boil ing in peroxide w a s cons idered quest ionable (Tan and Ro l ia , 1985) and instead Fenton 's reagent w a s used to facilitate oxidat ion (Druschel , 2003). Fenton 's reagent a l lows for hydroxy radicals to be produced in situ from the react ion of hydrogen peroxide on ferrous ions. Fer rous chlor ide (1.35 g F e C I 2 . x H 2 0 ) w a s d isso lved in hydrochlor ic ac id (50 ml 0.1 M HCI). A n y und isso lved iron w a s removed using a syr inge filter. A known m a s s of the sod ium trithionate (about 0.2 g) w a s added to the ferrous solut ion and hydrogen perox ide (10 ml 30 % (w/v)) w a s added . T h e react ion w a s left to go to complet ion and 56 the solut ion w a s diluted to 100ml. T h e diluted solut ion w a s then diluted further a s appropr iate and ana lysed for sulfate by ion chromatography. 3.6.2 Total Sodium Content A known m a s s of sod ium trithionate w a s heated to about 790 °C in a furnace until no more fumes were emit ted, and then left at that temperature for another hour. T h e trithionate w a s conver ted to sod ium sulfate (Rol ia and Chakrabar t i , 1982) accord ing to Equat ion 3.2, and the sod ium content determined by the m a s s change . A correct ion w a s made for the quantity of any volat i les and any sod ium sulfate present a s an impurity in the trithionate. N a 2 S 3 0 6 + 0 2 -> N a 2 S 0 4 + 2 S 0 2 [3.2] 3.6.3 Volatiles Thermograv imetr ic ana lys is ( T G A ) w a s done on two trithionate ba tches . T h e samp le w a s heated at 10 °C per minute in an a lumina crucible in hel ium a tmosphere to 800 °C. T h e results are shown graphical ly for two batches of trithionate in F igures 3.1 and 3.2. T h e m a s s drop below 100 °C w a s likely due to a loss of ethanol adsorbed to the sol id . (Ethanol w a s used extensively in the synthesis. ) T h e m a s s change at around 250 °C could be due to decompos i t ion of the trithionate. 57 0 200 400 600 800 Temperature (°C) Figure 3.1 : TGA/DTA for sodium trithionate batch 2 0 200 400 600 800 Temperature (°C) Figure 3.2 : TGA/DTA for sodium trithionate batch 3 3.6.4 Sulfate A solution of the trithionate was analysed for sulfate by ion chromatography. 3.6.5 Polythionates A solut ion of the trithionate w a s ana lysed for polythionates by ion chromatography. T h e solut ions were ana lysed immediately after dilution and were ana lysed a number of t imes in s u c c e s s i o n . T h e concentrat ion profile with time is shown in F igure 3.3 for solut ions prepared using deaera ted water and in water without deaerat ion. A l though the m e a s u r e d trithionate concentrat ion w a s in quest ion here, the va lues on the graph are still indicated a s mg/l trithionate, based on ana lys is against a mixed trithionate, tetrathionate and thiosulfate standard made using a different batch of trithionate. T h e actual trithionate concentrat ions are indicative only. Wh i l e no thiosulfate w a s found to be present, the initial ana lys is at t ime zero showed a signif icant concentrat ion of tetrathionate (up to 3 mg/l for an est imated 10 mg/l trithionate concentrat ion). Repea t ana l yses of the s a m e solut ion showed a continual ly dropping tetrathionate concentrat ion and a cor responding inc rease in the trithionate concentrat ion. T h e c a u s e of this phenomenon is not c lear. However , it w a s found that ana lys ing trithionate in alkal ine solut ion (containing 2 m M a m m o n i a for 10 mg/l S 3 0 6 2 _ , s e e F igure 3.4) or where thiosulfate and tetrathionate were present (Figure 3.5), this phenomenon did not occur . It is poss ib le that dilute solut ions of trithionate in such an unbuffered sys tem (water) were more subject to variat ions in spec ia t ion. 59 10 e 9 E CN CD o CO CO II A * " A A 0 H A • i • CO o I 3 CQ 0 100 200 Time (min) 300 A Trithionate (with nitrogen) A Trithionate (no nitrogen) • Tetrathionate (with nitrogen) • Tetrathionate (no nitrogen) Figure 3.3 : Change of measured trithionate (indicative values onlv) and tetrathionate concentrations with time with and without using deaerated water in solution preparation • CM CD o CO CO 11 10 9 8 7 6 CO O o to 3 CQ 20 40 60 80 Time (min) 100 • Trithionate • Tetrathionate Figure 3.4 : Change of measured trithionate (indicative values onlv) and tetrathionate concentrations with time in the presence of 2mM N h U O H 60 CM CO o CN CO CN co o •* CO CD o CO CO 11 10 9 8 7 6 20 A _8_ • A • A 40 60 Time (min) 80 A Trithionate • Tetrathionate x Thiosulfate 100 Figure 3.5 : Change of measured trithionate (indicative values onlv), tetrathionate and thiosulfate concentrations with time in the presence of 10 mg/l SgOg2' and 10 mg/l S 40g-3.6.6 Overall Trithionate Purity Table 3.3 gives comparative data for all the trithionate batches. The purity was calculated as follows: % purity = (total S (%) - sulfate S (%)) x 100 / 40.3 Table 3.3 : Sodium trithionate characterisation Batch S0 4 2 - Na STOT Volatiles Purity (based (%) (%) (%) (%) on total S and S0 4 2 ) (%) Theoretical 0 19.3 40.3 100 Batch 0 1.34 40.1 98.4 Batch 1 1.69 16.8 36.1 87.2 Batch 2 0.67 36.0 10 88.8 Batch 2a 4.27 39.1 93.4 Batch 3 1.13 40.0 3 98.3 Batch 4 96.0* * Standardised against Batch 2 61 3.7 TRACE IMPURITY ANALYSIS OF CHEMICALS USED Solut ions of the main chemica ls used in this testwork, namely sod ium trithionate (Batch 4), sod ium thiosulfate, a m m o n i a and ammon ium bicarbonate, were ana lysed by inductively coup led p l asma spectrophotometry to s c a n for the p resence of t race e lement impurit ies. T h e data is shown in Append ix 4. 62 4 KINETICS OF TRITHIONATE DEGRADATION - METHODOLOGY 4.1 INTRODUCTION T w o standard methods are used to determine react ion kinetics - the integrated rate method and the initial rate method (Brezonik, 1994). Both methods were used in this work to determine the kinetics of trithionate degradat ion. In this chapter , the bas ic theory behind e a c h method is summar i zed , and the exper imenta l methods used to col lect the necessa ry data are descr ibed . Examp le data and the interpretation thereof a re g iven for e a c h method. 4.2 REACTION KINETICS THEORY For a gener ic irreversible react ion (Equat ion 4.1), the rate equat ion is typical ly g iven by Equat ion 4.2. A + B y p r o d u c t s [4.1] E lementary react ion kinetics a l lows for the determinat ion of the react ion orders with respect to the var ious reactants (a and b in Equat ion 4.2) and the rate constant (k). In the c a s e of trithionate degradat ion, Equat ion 4.2 can be exp ressed a s Equat ion 4 .3 . where k 0 b S is the observed rate constant and is expec ted to be dependent on var ious react ion condit ions and other reactant concentrat ions. T h e initial rate method involves measur ing the rate of a react ion over short t imes, before any signif icant changes in concentrat ion of the reactants occur . It is a s s u m e d that the react ion rate is constant over the initial t ime per iod, i.e. the concentrat ion ve rsus t ime profile for the reactant (trithionate) is l inear. T h e react ion order is determined by compar ing the initial rates measu red at different initial trithionate concentrat ions. F rom Equat ion 4 .3 , if only the concentrat ion of Ra te = k[A] a [B] b [4.2] -d [S 3 0 6 2 - ] /d t = k o b s [ S 3 0 6 2 T [4.3] 63 trithionate is chang ing , the react ion order with respect to trithionate can be exp ressed by Equat ion 4.4. Reac t ion order = In ( r 1 / r 2 ) / ln ( [S 3 0 6 2 - M S 3 0 6 2 - ]2) [4.4] where ^ is the initial rate at initial concentrat ion [ S 3 0 6 2 ' ]i and r 2 is the initial rate at initial concentrat ion [ S 3 0 6 2 ' ] 2. A plot of ln(r) ve rsus l n [ S 3 0 6 2 ' ] g ives the react ion order a s the s lope . T h e rate constant can be determined by plotting the measu red initial rate against the initial concentrat ion of trithionate (raised to the power of the react ion order). T h e gradient of such a plot g ives the observed rate constant, accord ing to Equat ion 4 .3 . T h e integrated rate method involves data col lect ion over longer t imes. A s s u c h , the concentrat ion of reactants and products changes signif icantly over the durat ion of the measurements . T o account for the continual ly chang ing concentrat ions, an integrated form of Equat ion 4 .3 is used . For a react ion rate first order with respect to the trithionate concentrat ion, the equivalent integrated rate equat ion is shown in Equat ion 4 .5 . In [ S 3 0 6 2 1 , = -k o b s t + In [ S 3 O 6 2 - ] 0 [4.5] T h u s for a first order dependency , a plot of In [ S ^ 2 ^ ve rsus t ime g ives a straight line with the observed rate constant a s the s lope and In [S3062-]o (the initial trithionate concentrat ion) a s the intercept. S u c h a plot can be used a s a d iagnost ic tool in determining the react ion order. T h e theory for s e c o n d and higher order react ions is not d i s c u s s e d here, but is readily avai lable in the literature (Brezonik, 1994). 4.3 EXPERIMENTAL METHOD S o d i u m trithionate w a s synthes ized accord ing to the method in Chap te r 3 for use in these tests. T h e a im of the exper imental p rogramme w a s to establ ish the factors inf luencing trithionate degradat ion kinetics. The kinet ics were examined in s imple a q u e o u s solut ions with var ious components present. Most tests were done in a buffer of 64 ammonium bicarbonate / ammonia, since the buffer pH of this system was maintained between pH 9.3 and 10.3, which is the pH range of interest. An ammonium sulfate / ammonia buffer was also considered, but in that system small concentrations of sulfate produced during the degradation of trithionate could not be accurately measured. The test solutions were made up to the required concentrations using deionised water. integrated Rate Method The ammonium bicarbonate or ammonium sulfate solution was made up (as above) and aqueous ammonia was added to adjust the pH. The amount of ammonia added was not measured. Most tests were done at room temperature unless otherwise stated. In tests where the temperature was controlled, the solutions were brought to the required temperature in a water bath. For each test, the required mass of sodium trithionate was diluted using the buffer to give a nominal concentration of 8.5 g/l or 4.5 g/l trithionate, and in some cases sodium thiosulfate and/or sodium tetrathionate were added to give 10 g/l thiosulfate and 2 g/l or 10 g/l tetrathionate. These concentrations were selected based on concentrations typically found in gold leach liquors (see Chapter 2). The solutions were transferred to glass vessels and stoppered with rubber stoppers. The ionic strength was not adjusted. At set times, samples were withdrawn by pipette and diluted for analysis for thiosulfate, trithionate and tetrathionate concentration by ion chromatography. The solutions were also analysed for sulfate level at the termination of the test. The tests were typically run for 36 to 54 days. Initial Rate Method Three or four aliquots (50 ml) of each test solution containing all solution components except sodium trithionate were placed in sealed flasks in a water bath for at least 45 minutes to reach the required temperature. Tests were done at 40 °C unless otherwise stated. Solid sodium trithionate was added to each of three conical flasks and at time zero, one aliquot of test solution was added to each flask and mixed. The amount of 65 sod ium trithionate added w a s to give nominal concentrat ions of 2 g/l, 5 g/l and 7 g/l trithionate, based on expec ted va lues in gold leach solut ions (see Chap te r 2). In s o m e c a s e s a 'blank' test w a s done where trithionate w a s exc luded to determine the pH c h a n g e in the a b s e n c e of trithionate. T h e f lasks were sea led with rubber sep tums and kept at constant temperature in a water bath. S a m p l e s were removed at set t imes either by removing the rubber sep tum or by using a syr inge to samp le through the sep tum if a signif icant amount of ammon ia w a s present or if the sys tem w a s under nitrogen a tmosphere . S a m p l e s were immediately diluted a s appropr iate and ana lysed for trithionate and thiosulfate (and occas iona l ly sulfate) concentrat ion using the ion chromatography methods d i scussed earl ier. Total a m m o n i a w a s ana lysed accord ing to the method in Chapte r 3. T h e tests were typically run for 3 to 4 hours. 4.4 DATA ANALYSIS Integrated Rate Method T h e data evaluat ion for a typical data set obta ined using the integrated rate method is exp la ined in this sect ion by example , for a test where 9 g/l S 3 0 6 2 ' and 10 g/l S 2 0 3 2 ' were initially present. T h e data f rom all tests were treated in a s imi lar w a y and the f indings are summar i zed in Chap te r 5. T h e trithionate and thiosulfate concentrat ions as wel l a s the logarithm of the trithionate concentrat ion are plotted in F igure 4 .1 . There w a s a good straight line fit to the l n [S 3 0 6 2 1 versus t ime data points. E lementary reaction kinetics shows that a relation of this type cor responds to pseudo first order kinet ics, shown in Equat ion 4 .5 , where k o b s is the observed rate constant (Brezonik, 1994). T h e expec ted relat ionships for other react ion orders did not fit the exper imental data, s o it w a s deduced that the degradat ion of trithionate fol lows first order kinetics. T h e rate constant w a s determined by us ing the best fit curve of this type. 66 20 eg 6 1 2 CM CO ° 8 • CM o CO 4 f D •— 4 6 ° o o U O f 0 A m I X 3 2 1 ™ CD o co CO 500 1000 T i m e (hr) -2 1500 A Tri thionate 0 Thiosul fate * In (trithionate) F igure 4.1 : Concent ra t ions of trithionate and thiosulfate with t ime for an integrated rate method test ( R o o m temperature. N H 4 H C Q 3 = 0.03 M. NIK, added to adjust initial DH to - 9 . 7 ) Plott ing [S203 2 -] or [S0 4 2 ' ] against the dec rease in trithionate concentrat ion gave the stoichiometr ic co-eff icients of thiosulfate and sulfate in the react ion equat ion a s the gradients. T h e data w a s plotted in this manner in Figure 4 .2 , using exper imental ly observed concentrat ions for thiosulfate and sulfate, and using the best-fit exponent ia l equat ion (from Figure 4.1) to determine the trithionate concentrat ions. 67 0.15 o E, • CM o CO CM o CM CO 0.10 h o . 0.05 0.00 0.00 o - OO - io" o 0.01 0.02 0.03 0.04 o Thiosulfate! • Sulfate 0.05 Decrease in S 3 0 6 " (mol/l) Figure 4 .2 : Determinat ion of reaction stoichiometry using the integrated rate method ( R o o m temperature. N h L H C O g = 0.03 M. N H 3 added to adjust initial pH to - 9 . 7 ) Da ta from the other tests were treated in a similar way. W h e r e tetrathionate w a s initially added (to s imulate a poss ib le recycle solution), the tetrathionate w a s found to degrade very quickly under the alkal ine condi t ions, caus ing an initial inc rease in trithionate and thiosulfate concentrat ions. In these c a s e s , the best fit of the exponent ia l equat ion for trithionate degradat ion exc luded the initial data where tetrathionate w a s still present, and the extrapolated trithionate concentrat ion at t ime ze ro w a s der ived from the trithionate concentrat ion at the t ime when the initial tetrathionate had d isappeared . Initial Rate Method Typ ica l profi les of trithionate concentrat ion and thiosulfate concentrat ion with t ime are shown in Figure 4 .3 for the initial rate method. A l inear d e c r e a s e in trithionate concentrat ion and a corresponding l inear increase in thiosulfate concentrat ion were found, a s expec ted for this method. Ove r the duration of the test, typically less than 10 % of the trithionate had reacted. 68 Q 1 2 3 I a Thiosul fate T i m e (h) Figure 4 .3 : Typ ica l concentrat ion profile for trithionate and thiosulfate us ing the initial rate method (40 °C. pH 9.9. 0.05 M N h U H C O a . 0.687 M N h k total ionic strength 1 M) T h e react ion order w a s determined graphical ly using the relation in Equat ion 4.4. A n examp le showing this dependency is g iven in Figure 4 .4 , where the average s lope w a s 1. T h e rate equat ion for the degradat ion of trithionate c a n thus be exp ressed a s in Equat ion 4.6, where the observed rate constant, k o b s , is expec ted to be a function of the concentrat ions of other solut ion components affecting the trithionate degradat ion rate. - d [ S 3 0 6 2 - ]/dt = k o b s [S 3 0 6 2 1 [4.6] 69 0.5 1.5 i 2 -2 y = 1 . 0 x - 4 . 2 in s 3 cy-Figure 4.4 : Determinat ion of the reaction order for the rate of trithionate degradat ion with respect to trithionate (40 °C. NHJHCOg = 0.2 M. initial p H 9.2) T h e concentrat ion of trithionate versus t ime w a s plotted and using a trendline, the best line fit through the data w a s determined, giving the initial react ion rate as the s lope and the initial trithionate concentrat ion a s the intercept. A s tandard statistical method (Miller and Mil ler, 1988) to determine the standard deviat ion in the s lope and intercept of a regress ion line w a s used to determine the standard deviat ions of the initial rate and initial trithionate concentrat ion. T h e initial rate w a s plotted against the initial trithionate concentrat ion for e a c h test ser ies , usual ly compr ised of three data sets , to determine the ave rage observed rate constant, accord ing to Equat ion 4 .6 . T o account for exper imenta l variabil ity, the est imated max imum rate (initial rate plus the s tandard deviat ion) w a s plotted against the min imum initial concentrat ion (initial concentrat ion minus the s tandard deviat ion), and the min imum rate against the max imum concentrat ion. Reg ress i on l ines were generated for all three plots, and forced through the origin. T h e s lope of the regress ion l ines represented the average, max imum and min imum observed rate constants . A typical plot to determine the rate constant k is shown in Figure 4 .5 . In all subsequen t f igures showing the observed rate constant , the error bars indicate the est imated min imum and max imum rate constants determined in this way. 70 0.16 • Average rate constant • Min imum rate! constant A Max imum rate constant Initial S3O6 " Concentration (g/l) Figure 4 .5 : Typ ica l plot of trithionate degradat ion rate ve rsus initial trithionate concentrat ion to determine the rate constant knhs f rom the s lope (40 °C. PH 9.9. 0.05 M N H 4 H C O 3 , 0.687 M NH A . total ionic strength 1 M) T h e react ion stoichiometry w a s determined in the s a m e w a y as for the integrated rate method. 71 5. K I N E T I C S O F T R I T H I O N A T E D E G R A D A T I O N - R E S U L T S 5.1 I N T R O D U C T I O N Both the integrated rate method and the initial rate method were used to investigate the kinetics of trithionate degradation . The experimental methods and the ways in which the data were interpreted using these methods were d i scussed in Chapter 4. The initial rate method is expected to give more reliable data than the integrated rate method, as by the nature of this method, the reaction conditions are held more constant and the reactions of products are expected to be minimized. However, the integrated rate method is useful to give an indication of the factors expected to be of signif icance in the trithionate degradation reactions, and also to serve as confirmation of effects noted using the initial rate method. In this chapter, the trithionate degradation kinetics using both methods are d i s cussed together. Most of the results were obtained using the initial rate method and the quantitative conc lus ions were based primarily on this data . Resu lts from the integrated rate method are indicative and confirmatory. T h e a im of this work was to establ ish the factors influencing trithionate degradation, starting with s imply trithionate in aqueous solution and examining solution components systematical ly to identify individual effects . The effects were manifested in changes to the observed rate constant, k0bS. in Equation 5.1. -d[S3Oe2- ]/dt = k o b s [S30621 [5.1] 5.2 S T O I C H I O M E T R Y In much of the research reported in the literature, the rate of reaction of various sulfur oxyan ions has been deduced based on analysis of thiosulfate and an a s sumed reaction stoichiometry, rather than direct measurement of the a s s u m e d reaction products. In this work, both thiosulfate and trithionate were measured at the s a m e time. O n e would expect, based on Equation 5.2 (based on prior work, s ee Chapter 2), that the rate of trithionate degradation would be equal to the rate of thiosulfate production. S 3 0 6 2 - + 2 0 H " S 2 0 3 2 " + S 0 4 2 " + H 2 0 [5.2] 72 For both the initial rate method and the integrated rate method, the ratio of thiosulfate formed to trithionate degraded was measured at 40 °C and at room temperature. In genera l , a ratio of 1:1, corresponding to Equation 5.2 was found . The limited data for sulfate a lso gave c lose to a 1:1 stoichiometric relationship with trithionate, which is a lso consistent with the reaction in Equat ion 5.2. Spec if i c c a s e s where a different stoichiometry from Equation 5.2 was observed are d i s cussed separately . There were anomal ies noted in s ome initial rate tests . In these tests, the trithionate and thiosulfate concentrations were measured , but in most cases , not the sulfate concentration, a s a different ion chromatographic method was needed , s ince sulfate could not be determined simultaneously . In genera l it was found that the thiosulfate formation rate was l inear except for the initial sampl ing point. It is poss ib le that slight impurities in the trithionate solution or the necessary re-equilibration of the ion chromatography co lumn when the first samp le of the ser ies was injected may have had some influence here. It is relevant to note that especia l ly at the start of the reaction, the concentration of thiosulfate was negligible compared with trithionate, but it was necessary to ana lyse for both spec ies at the s ame time using the s ame dilution factor for analysis . Th is is not ideal for accurate analys is of the low concentration of thiosulfate. Excluding this initial point, the thiosulfate formation stoichiometry was as expected . T h e sulfate formation stoichiometry was a lso as expected for the limited data avai lable . In addition, the sulfur ba lances for the measured spec ies of trithionate, thiosulfate and sulfate were typically between 95% and 105%, implying that these spec ies were the only sulfur spec ies present in significant quantities and no other products (e.g. pentathionate) needed considerat ion . It has been suggested (Naito et al . , 1975) but not proven that su lfamate could form in the presence of ammon ia under similar conditions to those tested in this work. If su lfamate was present as a reaction product from trithionate degradation, its concentration would be expected to be trace under the conditions tested here, and it could not be determined in the samp le matrix. 73 5.3 REPRODUCIBILITY Integrated Rate Method T h e concentrat ion profi les for two very similar tests are shown in F igure 5.1. Wh i l e the test condi t ions were nominal ly the s a m e , there w a s a smal l di f ference in the initial trithionate concentrat ion and the tests were done at room temperature at different t imes, s o there were likely a l so temperature var iat ions. Within these constraints, the results were reproducible - the observed rate constants for the tests were 0.0016 h"1 and 0.0017 h" 1. • CM CO o CM CO CM CD o CO CO 12 10 8 6 4 2 i i '.A A A A A A A A A A * A A A o A A A O ^ " A A A A A 4M-o8»°° i • l I I i A Trithionate • Thiosul fate 0 200 400 600 800 1000 1200 T i m e (hr) F igure 5.1 : Concentrat ion profi les for trithionate and thiosulfate (room temperature. NH4HCO3 /NH3 buffer. f N H / 1 = 0.03 M. initial PH 9.7. open symbo ls represent one test, c losed symbo ls represent a repeated test) Initial Rate Method S i n c e two different buffer sys tems were used predominant ly in this work (the a m m o n i a -a m m o n i u m bicarbonate sys tem, and the sod ium bicarbonate - sod ium carbonate sys tem) the reproducibil i ty of the initial rate method w a s tested for e a c h buffer sys tem at 74 4 0 °C . T h e observed rate constants (and their potential error ranges) are shown in Tab le 5.1. For both buffer systems , the potential range in the rate constant measured for each test w a s significant. This is to be expected s ince determination of the rate involved measur ing smal l changes of concentration over time, and e a c h rate constant was determined from only four independent data points, as d i scussed in the methodology sect ion . For this reason , the data is a lways shown including the potential min imum and max imum values, shown as horizontal bars in all graphs . However , the agreement between tests for each buffer system was well within the error range for observed for e a c h rate constant. Tak ing into account the appropriate ranges, the initial rate method provided reproducible results for the observed rate constants . Tab le 5.1 : Reproducibi l ity in rate constant determination for trithionate degradation for two buffer systems using the initial rate method Buffer system Total ionic strength (M) Buffer contribution to ionic strength (M) PH Repl icate no. [NH 3 ] (M) k0bs Oh"1) kobs max (h- 1) kobs min (h- 1) N a 2 C 0 3 / 1 0 0 .0103 0 .0125 0 .0084 0 .5 0.15 8.5 N a H C 0 3 2 0 0 .0110 0 .0129 0 .0093 NH4HCO3 / N H 3 1 0 .032 10.2 1 2 0 .84 0 .89 0 .0168 0 .0159 0.0181 0 .0174 0 .0155 0 .0144 5.4 WATER T h e degradation of trithionate in water was investigated at 4 0 °C using the initial rate method only. E ither sod ium perchlorate or potass ium chlor ide was u sed to adjust the ionic strength (I). Resu lts are shown in Tab le 5 .2 . The max imum and minimum rate constants are not shown in this c a s e . B e c a u s e the reaction rate was very s l ow and few data points were taken, the calculated deviation of the initial trithionate concentration 75 w a s often unrealist ical ly large. However , the observed rate constants were consistent for the four tests carr ied out. Tab le 5.2 : V a l u e s for the observed rate constant l w for trithionate degradat ion in water us ing the initial rate method (40 °C) 1 adjusted with: l ( M ) kobs (h"1) KCI 0.1 0.010 1.0 0.012 N a C I 0 4 0.1 0.012 1.0 0.013 T h e use of KCI or N a C I 0 4 f rom 0.1 M to 1.0 M did not have a signif icant effect. T h e observed rate constants in water were very simi lar to those extrapolated f rom work by Nai to et al (1975). Us ing their rate constant at 40 °C and assum ing a water concentrat ion of 55.5 M, a va lue for k o b s of 0.012 h' 1 w a s obta ined, very c lose to the va lues measu red here. Whi le it has been proposed that the concentrat ion of water has an inf luence on the rate (Naito et a l . , 1975), this w a s not tested in this work. It w a s proposed by Naito et a l . (1975) that water w a s expec ted to attack the sul fonate sulfur a tom of trithionate directly. Th is w a s d i scussed in the literature review in Chap te r 2. It is important to note that the natural pH of trithionate in water w a s around 4.5 (at 4 0 °C), and d e c r e a s e d with t ime to around pH 2 after 24 hours at 4 0 °C. T h e react ions occurr ing at these low pH va lues are not necessar i ly the s a m e a s those at the higher pH va lues relevant to gold leaching. T h e overal l stoichiometry of thiosulfate formed to trithionate reacted w a s typically 0.5-0.6 : 1. At neutral and higher pH the stoichiometry is expec ted to be 1:1 accord ing to Equat ion 5.2 (Kurtenacker, 1935) but at the low p H va lues observed here, it is likely that s o m e of the thiosulfate formed w a s further degraded to e lementa l sulfur, which w a s observed in the solut ions. 76 Th is data shows that trithionate will undergo a s low degradat ion in a q u e o u s solut ions. Th is inherent instability is c o m m o n to the polythionates. In summary , the trithionate degradat ion rate in water at 40 °C can be exp ressed by Equat ion 5.3. -d [S 3 0 6 2 - ] /d t = 0.012h" 1 [S 3 0 6 2 - ] = k 0 [ S 3 0 6 2 ' ] [5.3] 5.5 HYDROXIDE Trithionate degradat ion in sod ium hydroxide solut ion w a s invest igated using the initial rate method. Al l tests were done at 40 °C and a total ionic strength of 1 M, adjusted us ing sod ium perchlorate. T h e results are shown in Tab le 5.3. Tab le 5.3 : Effect of hydroxide concentrat ion on the observed rate constant k„hg for trithionate degradat ion using the initial rate method (40 °C) [NaOH] (M) A v e r a g e initial pH Rat io S 2 0 3 2 " formed : S 3 0 6 2 " degraded (M:M) kobs Ch"1) kobs min (h-1) k0bs m a x (h-1) 0.5 12.62 0.2-0.4 0.729 * * 0.1 11.96 0.3-0.4 0.143 0.132 0.156 0.01 11.06 0.7-0.8 0.0205 0.0184 0.0224 0.006 10.88 0.8-0.9 0.0146 0 .0133 0.0159 0.001 9.76 0.8-0.9 0.0123 0.0098 0.0149 *Fo r the test at 0.5 M N a O H , the react ion rate w a s very high, s o the initial rate method could not be used to determine the observed rate constant. Instead, an exponent ia l fit consistent with first order kinetics w a s used to der ive the rate constant (i.e. the integrated rate method). The error margin w a s not determined, hence no min imum or max imum value for the rate constant is reported. A t h igher hydroxide concentrat ions (> 0.1 M) and cor responding higher p H s ( -12 and higher), the degradat ion of trithionate was very fast. A l s o the ratio of thiosulfate formed to trithionate reacted cor responds (within error) to literature observat ions of the react ion of trithionate at high p H accord ing to Equat ion 5.4 (Kur tenacker , 1935) for react ion at 50 °C and pH 1 3 . 4 - 1 3 . 7 . 77 2 S 3 0 6 2 - + 3 H 2 0 ^ S 2 0 3 2 ' + 4 S O , 2 " + 6 H + [5.4] In the context of gold leaching, slightly lower p H s are of more interest. T h e expected trithionate degradat ion for pH ~6 to -11 can be exp ressed by Equat ion 5.2 and the measu red react ion stoichiometry under these condit ions supports this. T h e sulfate concentrat ion w a s not measu red . It should be noted that the tests using 0.001 M N a O H had a signif icant pH drop of 3 to 6 pH units over the duration of the test. However this did not appear to affect the react ion stoichiometry. T h e data- for pH va lues of around 11 and less cor responding to Equat ion 5.2 were interpreted in the fol lowing way. It w a s shown previously in Sec t ion 5.4 that trithionate hydro lyses in water at a rate corresponding to Equat ion 5.5 at 40 °C. It is p roposed that the overal l observed rate constant in the p resence of hydroxide cons is ts of a component for the degradat ion in water plus a component for degradat ion inf luenced by the p resence of e x c e s s hydroxide. T h e graph in Figure 5.2 shows a plot of the observed rate constant versus the hydroxide concentrat ion for tests with a hydroxide concentrat ion of < 0.01 M, including tests in water only. B a s e d on this g raph, the observed rate constant can be exp ressed as Equat ion 5.6 at 40 °C. T h e intercept is in agreement with the expec ted rate constant in water. It is noted f rom Tab le 5.3 that the observed stoichiometry of thiosulfate formed to trithionate reacted w a s slightly lower than one would expect , based on either Equat ion 5.4 or Equat ion 5.2. It is poss ib le that part of the thiosulfate formed g o e s on to react itself. A l s o it shou ld be noted that the concentrat ion of the thiosulfate in the solut ions is very much lower than that of the trithionate, and it is poss ib le for signif icant analyt ical errors to occur in this range. -d [S 3 0 6 2 - ] /d t = 0.012 h" 1 [S 3 0 6 2 - ] = k 0 [S 3 0 6 2 - ] [5.5] -d [S 3 0 6 2 - ] /d t = (0.0121V 1 + 0 .74M" 1 l r 1 [OH- ] ) [S 3 0 6 2 - ] = (k 0 + k 1 LOH"]) [ S 3 0 6 2 ^ [5.6] 78 0.04 0.03 0.01 0 H — i , : 1 0 0.005 0.01 0.015 [OH"] M Figure 5.2 : O b s e r v e d rate constant for trithionate degradat ion I w ve rsus TOHI for pH<11 In the literature, Ro l ia and Chakrabar t i (1982) c la imed that for trithionate degradat ion between p H 10 and 11 at 70 to 85 °C, the hydroxide concentrat ion did not affect the trithionate degradat ion rate. However , no results were g iven to support this asser t ion , and the condit ions used were quite different f rom those used in this work and of re levance to gold leaching. It has been sugges ted (Ritter and Krueger , 1970) that hydroxide would attack trithionate directly at the sul fonate sulfur a tom as it is a hard base . Th is w a s d i scussed in the literature review in Chap te r 2. 5.6 IONIC S T R E N G T H It is wel l known that in aqueous react ions, the ionic strength (I) of the solut ion can affect the react ion rate (Brezonik, 1994), so it w a s cons idered important to invest igate the ionic strength in this work. T h e cho ice of salt to be used in sett ing the ionic strength w a s invest igated using the initial rate method. It w a s shown (Sect ion 5.4) that for a solut ion of trithionate in water, the p resence of var iable amounts of KCI or N a C I 0 4 did not affect the observed trithionate degradat ion rate constant appreciably . However , in the 79 p resence of ammon ia , the cho ice and concentrat ion of salt had an effect on the rate of degradat ion. A ser ies of tests at 40 °C w a s done in the p resence of ammon ia , us ing KCI , NaCI or N a C I 0 4 to adjust the ionic strength. No ammon ium w a s added and the pH (>10.6) w a s not adjusted. Resu l t s are shown in Tab le 5.4. Tab le 5.4 : V a l u e s for the observed rate constant I w for trithionate degradat ion in the p resence of var ious sal ts used to adjust the ionic strength using the initial rate method I adjusted with: [NH 3 ] (M) (meas) l ( M ) kobs (h-1) kobs min (h-1) kobs max Ch"1) Naito et a l . est kobs (h - i r KCI 0.15 0.1 1.0 0.0166 0.0232 0.0141 0.0205 0.0192 0.0260 0.013 0.013 NaCIC-4 0.13 0.1 1.0 0.0175 0.0160 0.0158 0.0137 0.0192 0.0185 0.013 0.013 0.39 0.1 0.0116 0.0098 0.0134 0.014 KCI 0.37 0.5 0.0173 0.0161 0.0186 0.014 0.39 1.0 0.0236 0.0208 0.0264 0.014 NaCI 0.41 1.0 0.0169 0.0142 0.0197 0.014 NaCICU 0.49 1.0 0.0171 0.0154 0.0188 0.014 * T h e s e observed rate constants were est imated only, a s this data w a s not suppl ied by Naito et al (1975). Instead they used an est imation of the water concentrat ion (not given) to exp ress k o b s in terms of k w , k a and k t for hydrolysis, ammono lys i s and the react ion with thiosulfate respect ively. Genera l l y the results obtained gave higher rate constants than those extrapolated from Nai to et al 's rate equat ion, introduced in Chapte r 2. However , there were a number of di f ferences between the method used by Naito et a l . and in this work. A t a low ionic strength (0.1 M) , the use of KCI gave rate constants simi lar to solut ions containing N a C I 0 4 , but when higher levels of KCI were u s e d , the trithionate degradat ion rate 80 i nc reased signif icantly. NaCI did not have any effect on the rate at an ionic strength of 1 M , implying that rather than the chloride an ion being responsib le for the increased rate w h e n KCI w a s u s e d , the po tass ium ion w a s respons ib le . Po tass i um affected the trithionate degradat ion rate only in the p resence of ammon ia . However , in the a b s e n c e of ammon ia (see Sect ion 5.4), the pH w a s much lower and the react ion stoichiometry w a s different, so this situation cannot be directly compared with the sys tem of interest, in the p resence of ammon ia , except to note that po tass ium d o e s not affect the base l ine degradat ion in water. T h e inf luence of posit ive ions on the react ion rate impl ies the formation of an act ivated complex between the posit ive ion and one of the reactants. In this set of tests, the only spec ies present were sod ium trithionate and ammon ia , and the spec ies used to set the ionic strength. S i n c e the concentrat ion of sod ium did not have any signif icant effect it would s e e m that an e x c e s s of sod ium where the trithionate w a s a l ready assoc ia ted with sod ium did not provide any compet ing interaction. However , the addit ion of po tass ium, wh ich has a larger ionic radius than sod ium, introduced a compet ing cation for complexat ion with trithionate. T h e p resence of po tass ium made the trithionate more amenab le to degradat ion in the p resence of ammon ia . T h e exchange react ion between trithionate and thiosulfate has been found to be inf luenced by the concentrat ion and charge of cat ions present (Fava and Pajaro , 1954) but it w a s not evident whether trithionate or thiosulfate w a s the most likely to form a comp lex with the posit ive ion. S i n c e ionic comp lexes of thiosulfate had been reported previously, it w a s a s s u m e d by F a v a and Pa jaro (1954) that trithionate reacted with an ionic comp lex of thiosulfate in this c a s e . T h e results reported in this thesis imply that trithionate too can form ionic comp lexes with cat ions (e.g. M S 3 0 6 ' ) . Th is is d i s cussed further in Chap te r 6. B a s e d on these results, N a C I 0 4 w a s used in all further tests to adjust the ionic strength. It is important to real ise that in Naito et al . 's work, the use of KCI for ionic strength adjustments may have had a signif icant inf luence on the results and w a s not taken into cons iderat ion. T h e effect of ionic strength on trithionate degradat ion w a s invest igated further us ing sod ium perchlorate at slightly lower pH where both ammon ia and ammon ium were 81 present. Trithionate solut ions in N H 4 H C 0 3 / N H 3 at pH 9 - 9.2 at 40 °C were not affected by ionic strength adjustments up to about 1.8 M, as s e e n in Figure 5.3. F igure 5.3 : Effect of ionic strength on observed rate constant I w for trithionate degradat ion using the initial rate method (0.15 M NHdHCOa at pH 9-9.2 bv N H a addit ion. 40 °C) 5.7 C A R B O N A T E A N D B I C A R B O N A T E S i n c e most tests were carr ied out in an ammon ium bicarbonate - ammon ia buffer sys tem, it w a s n e c e s s a r y to ensure that the b icarbonate and carbonate ions did not apprec iab ly affect the rate. It had a l ready been estab l ished that the concentrat ion of sod ium in the form of sod ium perchlorate did not affect the trithionate degradat ion rate (Figure 5.3). T h e effect of chang ing the concentrat ion of a sod ium bicarbonate / sod ium carbonate buffer w a s invest igated using the initial rate method. T h e results are shown in F igure 5.4 for pH 9.0-9.2 at 40 °C. There appeared to be no inf luence of the concentrat ion of carbonate / b icarbonate under these condi t ions. S imi lar t rends have been noted for sod ium bicarbonate solut ions at pH ~8 (not shown) where increasing the concentrat ion f rom 0.5 M to 0.7 M bicarbonate had no inf luence on the rate of trithionate degradat ion. T h e use of ammon ium bicarbonate w a s thus d e e m e d acceptab le in 82 investigating the sys tem, particularly to invest igate the effect of ammon ium concentrat ion. F igure 5.4 : Effect of ionic strength of NaHCOg / Na?COg buffer on the observed rate constant I w for trithionate degradat ion using the initial rate method (40 °C. pH 9.03 - 9.20. total ionic strength 0.51 M c losed symbo ls . 1 M open symbol ) 5.8 AMMONIUM AND AMMONIA Initial Rate Method T h e trithionate degradat ion rate in solut ions of ammon ium bicarbonate and ammon ium sulfate of var ious concentrat ions w a s determined to find the effect of ammon ium concentrat ion on the rate. A plot of the observed rate constant ve rsus the a m m o n i u m concentrat ion for the two sys tems is shown in Figure 5.5. T h e observed rate constant inc reased linearly with a m m o n i u m concentrat ion for both sys tems , but w a s greater at higher ammon ium concentrat ions in the sulfate med ium than the b icarbonate med ium. It shou ld be pointed out that in the ammon ium bicarbonate sys tem, the pH w a s fairly stable (around neutral) whi le in the ammon ium sulfate sys tem, the initial pH w a s 4 to 5, dropping very rapidly throughout the test to as low a s pH 1.1. 83 At pH values as low as this, other factors could come into play and the observed dependency on ammonium concentration was probably not an isolated effect. The data measured in the ammonium bicarbonate system is more suitable to judge the effect of ammonium concentration. S o 0.04 0.03 0.02 0.01 0.00 — o o u o • a J i 0.5 1 1 1 1.5 [NH4 +] (M) i 2 o pH 7.4 ammonium bicarbonate! • pH4 ammonium sulfate Figure 5.5 : Effect of ammonium concentration on the observed rate constant l w for trithionate degradation using the initial rate method (40 °C. total ionic strength 1 to 2 M. natural pH) Taking into consideration the degradation of trithionate in water and the gradient of the ammonium bicarbonate data in Figure 5.5, the extent of the dependency on the ammonium concentration in the bicarbonate system could be expressed as Equation 5.7 at 40 °C. This expression of the rate equation is similar to that for hydroxide solutions, discussed in Section 5.5. -d[S3062-]/dt = (0.012h1 + 0.01M-1h-1[NH4+])[S3O62T [5.7] The ammonium ion had a very similar effect to the potassium ion, discussed in Section 5.6. These ions have similar radii, with ammonium having a radius of 0.143 nm and potassium having a radius of 0.138 nm (Brodbelt and Liou, 1993). This is discussed further in Chapter 6. Since gold leaching by thiosulfate generally takes place in ammoniacal medium, the effect of ammonium was quantified as in Equation 5.7. 84 T h e effect of ammon ia on the trithionate degradation rate was measured , in the absence of ammon ium ions at high pH (>10.5), using the initial rate method. The dependency of the observed rate constant on the ammon ia concentration is shown in F igure 5.6. T h e data in Figure 5.6 shows the observed rate constant adjusted for the effect of hydroxide concentration, a s determined in Sect ion 5.5 . 0.04 Q 0 .03 i X 0.02 N-O i S 0.01 J 0.00 0 0 .5 [ N H 3 ] ( M ) 1 1-5 Figure 5.6 : Effect of ammon ia concentration on the observed rate constant I w for trithionate degradation using the initial rate method (40 °C . total ionic strength 1M using N a C l O j ) T h e data shows a slight increase in the observed rate constant with ammon ia concentration . Two trend lines are indicated on the graph. The best-fit trend line implies that the observed rate constant can be expressed as Equation 5.8 (at 4 0 °C). kobs = 0.0156h" 1 + 0 . 0 0 4 9 M - V [ N H 3 ] + 0 .74 i V T V [OH] [5.8] However , the basel ine degradation rate of trithionate in water (Equation 5.3) would imply an intercept of 0 .012 h" 1. By fixing the intercept of the graph in Figure 5.6 to 0 .012 h"\ the observed rate constant can be expressed as Equat ion 5.9 (at 40 °C) . y = 0 . 0 0 8 1 x + 0 . 0 1 2 85 k o b s = 0.012h"1 + 0.0081 M-1h"1[NH3] + 0.74 i V T V [OH"] [5.9] A relatively small change in the intercept resulted in a larger variation to the gradient. It should be noted that the results indicating a dependency on ammonia concentration are strongly dependent on the accuracy of the ammonia analysis method. Since the ammonia concentration was found by difference between a total ammonia plus ammonium concentration, and an initial ammonium value, this method is prone to some error. Naito et al. (1975) assumed that ammonia was involved in a nucleophilic substitution reaction at the sulfonate sulfur atom of trithionate, producing thiosulfate and sulfamate as products. As discussed in the literature review (Chapter 2) the presence of sulfamate was not confirmed by these authors and this deduction was based on reactions known to occur under very different reaction conditions. A suitable method to analyse sulfamate in the presence of trithionate and thiosulfate could not be found (see Chapter 3 on analytical methods) so this assumption could not be confirmed in this work. Figure 5.7 shows the observed rate constant plotted against the ammonium concentration for both the ammonium bicarbonate / ammonia and the ammonium sulfate / ammonia systems at a constant pH of 9.1 - 9.2 (at 40 °C). Since the pH was constant, an increase in ammonium concentration also implied an increase in ammonia concentration. The data show that the two systems behave fairly similarly, within the error margins for each data point. It is assumed that the observed rate constant can be interpreted as being representative of the sum of various individual components, including the baseline degradation rate of trithionate in water and an enhanced degradation seen either in the presence of ammonia or ammonium (as noted above). The effect of hydroxide concentration was not included in this case as under these conditions it was expected to contribute to less than 0.1 % ofthe observed rate constant. The observed rate constant was thus expressed as a sum of its components, as in Equation 5.10. 86 0.04 0.03 0.02 0.01 y = 0.022x +0.013 0.2 NH4+(M) 4 -y = 0.014x + 0.012 0.4 0.6 0.8 • Ammonium sulfate / ammonia • Ammonium bicarbonate/ammonia Figure 5.7 : Effect of ammonium concentration on the observed rate constant l w for trithionate degradation for (NhUgSCu / NH 3 and NH 4 HCOa / NH 3 buffer systems using the initial rate method (40 °C, pH 9.13 - 9.19, ionic strength 0.51 M closed symbols. 1 M open symbols) k o b s = 0.0121V1 + k2[NH3] + k3[NH4 +] [5.10] For the data in Figure 5.7, at pH 9.1 to 9.2, the concentration of ammonia was typically 2.1 times greater than that of ammonium, given in Equation 5.11. k o b s = 0.012h"1 +2.1 k2[NH4 +] + k3[NH4+] [5.11] The slope in Figure 5.7 was thus 2.1 k2 + k3. Taking k3 as 0.01M"1.h"1 (see earlier) and k2 as either 0.0049M"V 1 (Equation 5.8) or 0.0081 M " V (Equation 5.9), one would expect the slope in Figure 5.7 to be between 0.020 and 0.027 M"1h"1. This is in agreement with the slope obtained for the data obtained in bicarbonate medium (0.022 M"1h"1). 87 Integrated Rate Method Chang ing the ammon ium and ammon ia concentrat ion for the standard ammon ium bicarbonate / ammon ia buffer sys tem a s well a s for an ammon ium sulfate / ammon ia buffer sys tem w a s a lso invest igated at 25 °C using the integrated rate method. Resu l ts a re shown in Tab le 5.5. Tab le 5.5 : Effect of ammon ium concentrat ion on observed rate constant l w for trithionate degradat ion (25 °C. nominal ly 0.023 M S a Q f i 2 ' . ammon ia added to adjust pH to - 1 0 . 3 ) Buffer sys tem [NH 4 + ] (M) kobs (h"1) ( N H 4 ) 2 S 0 4 / N H 3 0.02 0.0022 0.09 0.0027 N H 4 H C 0 3 / N H 3 0.03 0.0026 0.15 0.0034 For either buffer sys tem, increasing the concentrat ion of ammon ium (and hence ammon ia ) inc reased the observed rate constant. Wh i le the effects of a m m o n i a and a m m o n i u m were not examined individually us ing this method, the trends match those found at 40 °C using the initial rate method. 5.9 pH A number of comparat ive tests were done at constant pH using the initial rate method. T h e total a m m o n i a and ammon ium concentrat ions were var ied but kept at a constant ratio to maintain the des i red p H . T h e s a m e data is shown in two different w a y s - a s the observed rate constant plotted against the ammon ium concentrat ion and against the a m m o n i a concentrat ion. S i n c e it has been shown that the hydroxide concentrat ion in f luences the react ion rate, the effect of hydroxide ion concentrat ion should be taken into account as the p H var ies. However , the va lue of this contribution w a s general ly negl igible, and even at pH > 10.5 w a s less than 4 % of the observed rate constant , wh ich is less than the level of uncertainty in the measurements . T h e adjustment w a s thus not made in the f igures. 88 JD O 0.04 0.03 0.02 0.01 0.00 - I ~ 0.5 1 1.5 [ N H 4 + ] (M) o pH 7.4 • pH 8.8 • pH 9.0 A p H 9.2 o p H 9.9 x p H 1 0 . 2 F igure 5.8 : Dependency of the observed rate constant Iw for trithionate degradat ion on ammon ium concentrat ion at var ious pH (40 °C) 0.04 0.03 o 0.02 0.01 0.00 • • • _ A I • — • J A y x * o " - _ " -o s i - - X £ 0 0.5 1 [NH 3 ] (M) i 1.5 SI pH 8.8 • pH 9.0 A p H 9 . 2 O p H 9 . 9 X p H 1 0 . 2 • p H > 1 0 . 5 F igure 5.9 : Dependency of the observed rate constant Iw for trithionate degradat ion on a m m o n i a concentrat ion at var ious pH (40 °C) 89 It is difficult to s e e any variat ions in the s lope of the observed rate constant ve rsus a m m o n i a or ammon ium concentrat ions with pH f rom Figure 5.8 or 5.9. T h e s a m e data w a s therefore plotted in a different way, a s sets of data at constant a m m o n i a plus a m m o n i u m concentrat ion against the p H , a s in Figure 5.10. S i n c e the effect of changing the ionic strength had not been found to be signif icant when N a C I 0 4 w a s used to adjust it, the data in Figure 5.10 represents a range in ionic strengths. E a c h data ser ies represents a smal l range of total ammon ia concentrat ion as indicated on the graph. Error bars were exc luded for clarity. There appears to be a min imum degradat ion rate around pH 10. T h e fol lowing genera l observat ions were made , taking into considerat ion that the uncertainty levels for each data point were signif icant (see Sect ion 5.3): • For any p H , a s the total ammon ia plus ammon ium inc reased , the observed rate constant inc reased. • There w a s a min imum in the trithionate degradat ion rate at around pH 10. A t pH va lues higher than around 10 (at 40 °C), the observed rate constant increased a s the p H inc reased . At pH va lues lower than around 10 (at 4 0 °C), the observed rate constant inc reased as the pH dec reased . Th is effect is attributed to the relative amounts of ammon ia and ammon ium ions, rather than the pH (hydroxide concentrat ion) directly in this pH range, a s d i scussed in Chap te r 6. • There w a s a min imum trithionate degradat ion cor responding to the degradat ion in water. 90 CO n o 0.04 0.03 0.02 0.01 0.00 A o X A - X — x m 11 13 N H 3 + NH4+ concentrat ion x O M • 0 .1-0 .2M A 0 .3-0 .4M • 0 .5M x 0 .7-0 .8M A 0 .9-1 .2M o 1.5-1.7M o 1.8-2.6M pH Figure 5.10 : O b s e r v e d rate constant l w for trithionate degradat ion ve rsus pH for a range of a m m o n i a / ammon ium concentrat ions (40 °C, var ious ionic strength) T h e observed rate constant at 25 °C showed very little change with an inc rease in p H . However , only three data points were avai lable and the magni tude of the observed rate constants were much less than at 40 °C, s o it would be much more difficult to notice t rends. F igure 5.11 : O b s e r v e d rate constant l w for trithionate degradat ion ve rsus pH at 25 °C (Total ionic strength = 1.1 M, 1 M N H / . ammon ium bicarbonate / ammon ia sys tem) 91 In order to invest igate pH effects in the a b s e n c e of compl icat ing effects from a m m o n i a and a m m o n i u m , a sod ium bicarbonate / sod ium carbonate sys tem w a s used using the initial rate method at 40 °C. It w a s shown (Sect ion 5.7) that the b icarbonate and carbonate spec ies did not inf luence the trithionate react ion rate at pH 9-9.2. T h e b icarbonate to carbonate ratio w a s var ied to invest igate the effect of p H without having to introduce ammon ia spec ies to the sys tem. T h e observed rate constant ve rsus pH is shown in F igure 5.12. B e c a u s e the concentrat ions of b icarbonate and carbonate did not have an effect, the data shown in Figure 5.12 include data where varying concentrat ions of b icarbonate/carbonate were used (with the ionic strength from the b icarbonate / carbonate buffer ranging between 0.05 and 0.9 M). N o obv ious trend in react ion rate with pH w a s observed in this p H range. Th is impl ies that rather than pH it is the relative concentrat ions of the ammon ium and a m m o n i a which inf luence the trithionate degradat ion. F igure 5.12 : Effect of pH on observed rate constant I w for trithionate degradat ion in sod ium carbonate/b icarbonate med ium (40 °C, Ionic strength 0.5 M for c losed symbo ls . 1 M for open symbols ) 92 5.10 THIOSULFATE Initial Rate Method T h e effect of thiosulfate on the trithionate degradat ion rate w a s invest igated in ammon iaca l b icarbonate buffer using a range of ammon ium concentrat ions and p H . T h e tests were done at 4 0 °C with aqueous ammon ia added to adjust the pH to between 9 and 10. Resu l ts are shown in Figure 5.13. S o 0.04 0.03 0.02 0.01 0.00 A t A 0.1 0.2 0.3 0.4 0.5 [S 2 0 3 2 -] (M) • 1 M ammonium pH 9.4 • 0.5M ammonium pH 9.4 o 0.15 M ammonium pH9 A 0.024 M ammonium pH 10 Figure 5.13 : Effect of thiosulfate on the observed rate constant I w for trithionate degradat ion using the initial rate method (40 °C) (Total ionic strength for 1M ammon ium tests w a s 1M (open symbols ) or 2 M (c losed symbols ) . Total ionic strength for 0.5 M ammon ium tests w a s 0.6 M (open symbol ) or 1.8 M (c losed symbols) . Total ionic strength for all other data points w a s 1 M.) A t low ammon ium concentrat ion (0.024 M) at pH 10, the addit ion of thiosulfate had a negl igible effect on the trithionate degradat ion, perhaps caus ing a slight inhibition of the degradat ion. However , f rom the observat ions of trithionate degradat ion at different p H , it w a s found that near pH 10, the trithionate degradat ion w a s at a min imum (see Sec t ion 5.9). A n y effects of thiosulfate may have been min imised at this p H . 93 At 0.15 M a m m o n i u m , pH 9, addit ion of thiosulfate inc reased the trithionate degradat ion rate but c a u s e d inhibition at higher levels. Increasing the ammon ium concentrat ion to 0.5 M (and the pH to 9.4) showed a simi lar trend where thiosulfate enhanced the rate at lower levels but inhibited it at higher levels. Similar ly, for 1 M a m m o n i u m (pH 9.4) addit ion of thiosulfate inc reased the trithionate degradat ion rate. A t the levels of thiosulfate tested, no inhibition w a s noted for this ser ies . Poss ib l e reasons for these observat ions are d i scussed in Chapte r 6. Integrated Rate Method T h e effect of thiosulfate addit ion at room temperature is shown in Tab le 5.6 for two sets of tests, us ing the integrated rate method. It is c lear that within the variabil ity of the exper imenta l condi t ions, thiosulfate at 0.09 M (almost double the initial amount of trithionate and three t imes the initial ammon ium concentrat ion) did not have any signif icant effect. T h e react ion stoichiometry w a s not affected by the p resence of initial thiosulfate. However at room temperature, any di f ferences in observed rate constant are expec ted to be much less than at 40 °C, s o it is poss ib le that the effects were s imply not measurab le at this temperature. Tab le 5.6 : Effect of thiosulfate on the observed rate constant l w for trithionate degradat ion using the integrated rate method (room temperature. rNH 4HCO ?l = 0 .03M. NHs added to pH - 9 . 7 ) Initial S 3 0 6 2 - (M) Initial S 2 0 3 2 _ (M) kobs (h"1) 0.043 0 0.0017 0.041 0.083 0.0021 0.054 0 0.0016 0.053 0.091 0.0017 94 5.11 OXYGEN EXCLUSION Initial Rate Method T h e effect of limiting the amount of oxygen in solut ion on the trithionate degradat ion rate w a s invest igated. Solut ions were sparged with nitrogen and a nitrogen-fi l led g love bag w a s used during preparat ion of the test. S a m p l e s were removed by syr inge through a sep tum. Wh i le this method may not have complete ly removed all oxygen from the sys tem, the amount of oxygen present would have been signif icantly l imited. T h e eff ic iency of oxygen removal w a s not determined. T h e effect of limiting oxygen in solution w a s invest igated in the a b s e n c e of a m m o n i a for both the low alkal ine pH react ion (Equat ion 5.2) and the high alkal ine react ion (Equat ion 5.4). At high p H (pH ~ 12 at 40 °C) 0.1 M N a O H w a s used . T h e results are shown in Tab le 5.7 where it can be s e e n that limiting the oxygen in solut ion gave no signif icant change in the rate constant or the react ion stoichiometry with respect to thiosulfate under these condi t ions, implying that d isso lved oxygen has no role in this degradat ion. Tab le 5.7 : Effect of limiting d isso lved oxygen on the observed rate constant knhg for trithionate degradat ion in 0.1 M hydroxide solut ion A tmosphe re A v e initial pH Rat io S 2 0 3 2 " formed : s3o62-decayed kobs (h-1) k o b s min (h-1) k 0 b s max Ch"1) air 11.96 0.3-0.4 0.143 0.132 0.156 N 2 ( 0 2 l imited) 12.09 0.2-0.4 0.133 0.124 0.141 T h e effect of nitrogen sparg ing w a s a lso tested at a two different p H s in the a m m o n i a / ammon ium bicarbonate med ium. Resu l ts are shown in Tab le 5.8, where aga in there w a s no signif icant effect. Th is is not surpr is ing a s the ant ic ipated react ion, shown in Equat ion 5.2, d o e s not involve oxygen . There is likely to be an effect noted after longer 95 t imes, as the thiosulfate produced from trithionate degradat ion can oxid ise s lowly in the p resence of oxygen . Tab le 5.8 : Effect of limiting oxygen on the observed rate constant l w for trithionate degradat ion in ammon ia / ammon ium bicarbonate solut ion A tmosphere [ N H 4 H C O 3 ] (M) [NH 3 ] (M) A v e initial P H k o b s ( I T 1 ) k o b s rnin Ch"1) k o b s max (h-1) air 0.5 0 7.4 0 .0182 0.0163 0.0203 N 2 ( 0 2 l imited) 0.5 0 7.9 0 .0185 0.0178 0.0192 air 0 0.492 10.9 0.0171 0.0154 0.0188 N 2 ( 0 2 l imited) 0 0.625 11.0 0.0190 0.0180 0.0200 Integrated Rate Method In s imi lar tests us ing the integrated rate method, the oxygen concentrat ion w a s limited by us ing water that had been sparged with nitrogen and filling the test f lask head s p a c e with nitrogen at the start of the test and at each sampl ing point. Resu l ts for trithionate degradat ion both in the p resence and a b s e n c e of initial thiosulfate are shown in Tab le 5.9. Reduc ing the level of oxygen had no effect on the trithionate degradat ion, conf irming the f indings using the initial rate method. 96 Tab le 5.9 : Effect of limiting oxygen on the observed rate constant I w for trithionate degradat ion using the integrated rate method (room temperature. rNhUHCOJ = 0 .03M. N H s added to pH - 9 . 7 ) Test A tmosphere Initial S 3 0 6 2 " (M) Initial S 2 0 32 " (M) kobs (h"1) 14 air 0.054 0 0.0016 16 nitrogen 0.048 0 0.0016 15 air 0 .053 0.091 0.0017 17 nitrogen 0.048 0.089 0.0016 5.12 CUPRIC COPPER Cupr i c copper is cons idered necessa ry by many to cata lyse the thiosulfate leaching of go ld , but it is known to enhance the degradat ion of thiosulfate. It w a s therefore cons idered important to establ ish the effect of copper on trithionate degradat ion. Integrated Rate Method Cupr i c sulfate (0.0045 M) w a s added to a trithionate solution (0.049 M) to determine the effect on the trithionate degradat ion rate. Tab le 5.10 shows the effect on the observed rate constant and on the react ion stoichiometry. Tab le 5.10 : Effect of cupr ic addit ion on the observed rate constant k„hg for trithionate degradat ion using the integrated rate method (room temperature. N H j H C O g / N H g buffer. TNH/1 = 0.03 M. initial pH -9 .7 ) Tes t C u 2 + present (M) pH after 432 hrs pH after 600 hrs kobs (h"1) Stoichiometry S 2 0 3 2 " : S 3 0 6 2 " S 0 4 2 " : S 3 0 6 2 " 14 0 - 9.13 0.0017 0.9 0.9 20 0.0045 8.21 - 0.0018 0 3.3 97 E v e n with copper present, the trithionate degradat ion react ion still fol lowed first order react ion kinet ics. A l though the rate constant w a s simi lar either with or without copper present, the pH drop w a s much more signif icant with copper present (based on the limited data avai lable) , and no thiosulfate formed. E v e n though the rate constants were similar, they cannot be directly compared b e c a u s e over these long t imes different react ions occurred with and without copper present. Initial Rate Method T h e effect of cupr ic copper (in the form of cupric sulfate) on the trithionate degradat ion rate in water and in ammon ia / ammon ium bicarbonate buffer w a s invest igated. A i r w a s not exc luded from these tests, as for all other tests except for those reported in Sect ion 5 .11 , and the temperature w a s maintained at 40 °C. T h e results are shown in Tab le 5 .11. It should be noted that in s o m e tests 0.01 M copper w a s a d d e d . T h e trithionate concentrat ion used in these tests w a s 0.01 M to 0.04 M, of which only about 10 % is expec ted to degrade over the duration of the test. H e n c e the ratio of copper to trithionate w a s high compared to typical gold leaching condi t ions. However , the a im w a s to identify an effect, if any, even in exaggera ted form. Tab le 5.11 : Effect of cupr ic copper on the observed rate constant k n h g for trithionate degradat ion [Cu 2 + ] (M) [NH4HCO3] (M) A v e initial pH kobs (h"1) k min (hf1) k max (IV1) 0 0 4 0.012 * 0.01 0 4 0.027 0.021 0.029 0 0.54 9.4 0.022 0.021 0.023 0.001 0.5 9.7 0.025 0.023 0.026 0.01 0.5 9.6 0.026 0.025 0.027 0 0.02 9.6 0.013 0.011 0.015 0.01 0.02 10.2 0.013 0.012 0.014 * Not determined - s e e Sect ion 5.4 98 In water, there w a s a signif icant effect of copper on the trithionate degradat ion rate. Precipi tat ion of copper sul f ides w a s a lso evident, so this result cannot be directly compared to that in the a b s e n c e of copper or to higher pH tests where no precipitation w a s noted. In the p resence of an a m m o n i a / ammon ium buffer, copper did not have any signif icant effect on the observed rate constant. T h e stoichiometry w a s the s a m e a s in the a b s e n c e of copper and no precipitation w a s evident. Whi le the observed rate constant w a s not affected for the integrated rate method either, a different react ion stoichiometry and precipitation w a s observed . It may be that in the integrated rate method tests, the much longer test durat ions were responsib le for this. It is poss ib le that over these t imes, the signif icant pH drop in combinat ion with the p resence of react ion products in the solut ion could have affected the overal l react ion stoichiometry. T h e initial rate method results are cons idered to be more reliable. It has been proposed by Jeffrey and his coworkers that trithionate reacts with cupr ic copper (Breuer and Jeffrey, 2003b) . In work focusing on the react ion between cupr ic copper and thiosulfate, they measured the rate of reduction of cupr ic copper , inferring f rom this the thiosulfate degradat ion rate. O n addit ion of trithionate to the sys tem, the cupr ic copper w a s reduced faster, s o it w a s deduced that cupr ic copper reacts with trithionate in a simi lar way to with thiosulfate. That observat ion d o e s not agree with that found in these results. It w a s found that for the thiosul fate-copper react ion, the amount of oxygen present is very important (Breuer and Jeffrey, 2003a) . In the current tests on trithionate, oxygen w a s not control led. It may be necessa ry to careful ly control the levels of d isso lved oxygen to notice any effect such as that p roposed by Jeffrey et a l . 5.13 T E T R A T H I O N A T E T h e effect of tetrathionate w a s only examined using the integrated rate method. W h e n tetrathionate w a s present at the start of a test, it w a s found to very rapidly degrade to lower than the ana lys is detect ion limit, caus ing a cor responding smal l increase in the 99 trithionate and thiosulfate levels. Th is w a s expec ted a s tetrathionate is known to degrade rapidly under alkal ine condit ions, as shown in Equat ion 5.12. S 4 0 6 2 ' + % O H ' -> 5 / 4 S 2 0 3 2 ' + 1 / 2 S 3 0 6 2 " + % H 2 0 [5.12] B e s i d e s this smal l initial effect, the kinetics of trithionate degradat ion were not affected in any other way by the p resence of tetrathionate. 5.14 E L E M E N T A L S U L F U R , S U L F A T E A N D C O P P E R P O W D E R A few exploratory tests were carr ied out to investigate the effect of the p resence of e lementa l sulfur (hydrophobic), sulfate and e lementa l copper on the trithionate degradat ion rate, using the integrated rate method. Initial condit ions of 0.05 M S 3 0 6 2 , 0.08 M S 2 0 3 2 ' and 0.009 M S 4 0 6 2 ' were used . T h e change of trithionate concentrat ion with time for e a c h test is shown in Figure 5.14. T h e observed rate constants cor responding to Figure 5.14 determined from the point when all the tetrathionate had degraded are shown in Tab le 5.12. Tab le 5.12 : Effect of e lementa l sulfur, sulfate and copper powder on the observed rate constant knhs for trithionate degradat ion (room temperature, fNH 4 HCQ 3 l = 0 .03M, NHs added to pH - 9 . 7 , nominal ly 0.05 M S?OR2'. 0.08 M S?Oa 2 ' . 0.009 M S 4 Q F I 2 ) S p e c i e s added kobs (h"1) none 0.0021 S (4.95 g/l) 0 .0015 S 0 4 2 " ( 1 5 g / I ) 0.0021 C u (8 g/l) 0 .0010 Wh i le the addit ion of sulfate did not affect the observed rate, e lementa l sulfur may have slightly retarded the react ion, whi le metal l ic copper powder dec reased the rate signif icantly. W h e r e copper powder w a s used , the copper powder b e c a m e black, probably due to the formation of sul f ides. Insufficient sulfate ana l yses were avai lable to draw a definitive conc lus ion regarding the degradat ion stoichiometry in the p resence of 100 copper powder, but indicat ions are that higher levels of sulfate than expec ted accord ing to Equat ion 5.2 were obta ined. T h e s e effects were not pursued further in this work, but future work may be warranted. Ui 10 9 8 7 6 5 5 £ 4 3 2 1 0 I T T i X A A No added s p e c i e s A Sulfur A Sulfate x C o p p e r powder 100 200 300 400 500 T i m e (hr) 600 Figure 5.14 : Effect of sulfur, sulfate or copper powder on trithionate concentrat ion profile (room temperature, pH ~ 9.7) 5.15 T E M P E R A T U R E Init ial R a t e M e t h o d Most of the kinetic tests were done at 40 °C, to ensure that the react ion rates were high enough for the initial rate method to be used with a reasonab le degree of accuracy and to limit the scatter in the data. A few comparat ive tests were done at 25 °C to quantify the effect of temperature on the degradat ion rate. T h e observed rate constants for tests done in an ammon ium bicarbonate / ammon ia buffer of ionic strength 1.1 M for three initial pH va lues at 25 °C and 40 °C are shown in Tab le 5.13. B a s e d on this data, the activation energ ies were calcu lated using the Ar rhen ius equat ion. The apparent activation energ ies ranged between 72.9 and 78.9 v kJ/mol between pH 8.8 and 10.1. T h e perceived increase in activation energy with p H 101 w a s within the uncertainty of the measurements of the rate constants. Act ivat ion energ ies calcu lated under similar condit ions in the literature ranged from 81 - 87 kJ /mol for a temperature range of 40 - 80 °C (Naito et a l . , 1975), to 91.7 kJ /mol measu red at 70 - 85 °C (Rol ia and Chakrabar t i , 1982). Th is implies the react ion is under chemica l react ion control. Tab le 5.13 : Effect of temperature on the observed rate constant k ^ for trithionate degradat ion at pH 8.8 - 10.1 (NhUHCOq added to give I = 1.1 M. var iable rNH3l) P H Temperature kobs (h"1) Activat ion (°C) energy (kJ/mol) 25 0.0054 8.8 72.9 40 0.0221 25 0.0070 9.4 74.2 40 0.0294 25 0.0067 10.1 78.9 40 0.0308 Integrated Rate Method T h e s a m e temperature effect was found in the p resence of thiosulfate, using the integrated rate method. T h e observed rate constants and calcu lated activation energy are shown in Tab le 5.14. It is important to note that b e c a u s e of the speed of the react ion, very few data points were measured at 40 °C, s o the derivation of the rate constant w a s not likely to be as accurate as using the initial rate method. A l s o , for the purposes of the Ar rhen ius calculat ion, it w a s a s s u m e d that the room temperature tests were at 22 °C. T h e est imated activation energy w a s about 71 kJ /mo l . 102 Tab le 5.14 : Effect of temperature on the observed rate constant I w for trithionate degradat ion using the integrated rate method d N H ^ H C O g l = 0 .03M. N H a added to pH - 9 . 7 ) Tempera ture Initial Initial kobs (h"1) Est imated E a S 3 0 6 2 " s 2o 3 2- (kJ /mol) (M) (M) R T 0.041 0.083 0.0021 70.6 40 °C 0.047 0.090 0.0110 103 6. KINETICS OF THE DEGRADATION OF TRITHIONATE - DISCUSSION AND MODELLING 6.1 QUALITATIVE DISCUSSION T h e observed degradat ion of trithionate in the p resence of water or hydroxide ions conf i rmed the literature. It is likely to p roceed v ia nucleophi l ic attack at the sul fonate sulfur a tom (Naito et a l . , 1975). A n increase in the concentrat ion of hydroxide ions (or pH) inc reased the trithionate degradat ion rate, as expec ted , s ince hydroxide is known to react directly with trithionate (Equat ion 6.1). S 3 0 6 2 - + 2 0 H - S 2 0 3 2 - + S 0 4 2 " + H 2 0 [6.1] A sod ium bicarbonate/carbonate buffer had no effect on the rate, at constant pH (varying concentrat ion) or with varying pH (varying carbonate/b icarbonate ratio) f rom around pH 7.8 to 11 (at 40 °C). A l though the hydroxide concentrat ion inc reased the rate, this effect w a s very smal l at pH va lues less than 11, s o varying the pH in this range did not have much effect. T h e bicarbonate and carbonate an ions were inert in this sys tem. Th is is not surpr is ing a s posit ive ions, not negat ive ions, are likely to have an effect on react ions between an ions, i.e. trithionate and hydroxide in Equat ion 6 .1 . Increasing the ionic strength using N a C I 0 4 did not have any effect on the rate of trithionate degradat ion in the p resence of ammon ia . However , addit ion of potass ium or ammon ium ions general ly increased the rate. Whi le no data is avai lable for trithionate, ion pair ing of thiosulfate with cat ions is known. In fact, it has been shown that in the p resence of sod ium, po tass ium or ammon ium ions, the concentrat ion of the complexed thiosulfate ion ( M S 2 0 3 _ ) is the s a m e as or higher than that of the free thiosulfate ion (Senanayake , 2005). Th is researcher found that there w a s a l inear relat ionship between the logari thms of the assoc ia t ion constants for the spec ies M S 2 0 3 _ and M S 0 4 _ for the cat ions (M + ) N a + , K + and H + . B a s e d on this relat ionship and the assoc ia t ion constant of N H 4 S 0 4 _ , the assoc ia t ion constant of N H 4 S 2 0 3 _ w a s calcu lated and w a s found to be s imi lar to that for K S 2 0 3 \ T h e similarity between the behaviour of po tass ium ions and ammon ium ions is likely due to their simi lar ionic radii. 104 It has been shown that in the sulfur exchange reaction between thiosulfate and trithionate, an increase in charge and concentration of positive ions increases the reaction rate (Fava and Pajaro, 1954). The authors therefore postulated this reaction to be between ionic complexes of thiosulfate, trithionate or both, rather than between free ions. Since the trithionate degradation rate varied with the concentration of positive ions, it is reasonable to believe that cations influence the reaction, possibly by forming ionic complexes with trithionate. An increase in sodium ion concentration did not give an increase in the trithionate degradation rate. However, since the trithionate salt was in the sodium form, sufficient sodium was always present to ensure complex ions with sodium, so one would only expect to notice an effect when a competing ion (potassium or ammonium) was introduced. It was interesting that the trithionate degradation rate continued to increase with the ammonium or potassium ion concentrations even when the cation concentration was much higher than the trithionate concentration, and did not reach a maximum over the range of cation concentrations tested. The presence of potassium ions in acidic solutions (the natural pH obtained for a solution of sodium trithionate in water) did not increase the trithionate degradation rate, but the reaction stoichiometry was not necessarily the same as in alkaline systems. With no ammonium present, an increase in ammonia concentration increased the rate of trithionate degradation. While it could not be proven that the reaction in Equation 6.2 forming sulfamate proceeded, it is possible that this reaction occurred to a small extent, or that the interaction between ammonia and trithionate facilitated the reaction between hydroxide ions and trithionate by altering the electronic properties of the trithionate. S 3 0 6 2 " + 2 NH 3 S 2 0 3 2 - + NH 2 S0 3 " + NH 4 + [6.2] Based on the above discussion, the overall rate of degradation of trithionate is influenced by water, hydroxide and ammonia, and the reaction is enhanced by changing the electronic properties of a complex ion of trithionate by changing the cation type or concentration. It was assumed that the water concentration was constant in this case and that no thiosulfate was originally present. In an ammoniacal medium, which is of most interest to this study, the relative concentrations of ammonia, hydroxide ions and 105 ammon ium ions are shown diagrammat ical ly in Figure 6.1 with varying p H . S i n c e the degradat ion rate inc reases with the concentrat ion of e a c h of these spec ies (but to different extents), it is feasib le based on the d iagram in Figure 6.1 that the overal l rate would vary with pH in an ammon iaca l sys tem, and there could be a pH of min imum degradat ion rate. T h e way in which these qualitative effects can be combined quantitatively is d i s cussed in Sect ion 6.2. 4 7 10 13 P H Figure 6.1 : Concentrat ion profi les for ammon ia , ammon ium ions and hydroxide ions at 40 °C with varying p H , using arbitrary concentrat ion units It w a s shown that the effect of thiosulfate on the trithionate degradat ion rate w a s dependent on the ammon ium concentrat ion. W h e r e the ammon ium concentrat ion w a s higher than the thiosulfate concentrat ion, the trithionate degradat ion rate general ly inc reased with an increase in thiosulfate concentrat ion. However , as the thiosulfate concentrat ion e x c e e d e d the ammon ium concentrat ion, the trithionate degradat ion rate w a s unaffected or even dec reased with a further increase in thiosulfate. Thiosul fate is known to form comp lex ions, and it has been postulated that the ion N H 4 S 2 0 3 " ex ists (Senananyake , 2005). Other react ions have been proposed to occur via thiosulfate comp lex ions (Fava and Pajaro, 1954, C h a n d r a and Jeffrey, 2004). At high ammon ium concentrat ions, the concentrat ion of the ammon ium comp lex ion is expec ted to be higher, facilitating the interaction between thiosulfate and trithionate. However , a s the thiosulfate concentrat ion b e c o m e s comparab le with the ammon ium ion concentrat ion, 106 the possibil i ty of free thiosulfate ions (or the sod ium complex , s ince thiosulfate w a s added as sod ium thiosulfate) exist ing inc reases . If free or sod ium thiosulfate ions do not readily interact with trithionate, one would expect the trithionate degradat ion rate to plateau with any further increase in thiosulfate concentrat ion. T h e fact that a further inc rease in thiosulfate concentrat ion actually d e c r e a s e s the trithionate degradat ion rate impl ies a h indrance of the react ion. If thiosulfate interacts v ia the sulfenyl sulfur a tom of trithionate, where it acts to alter the electronic propert ies of trithionate to facilitate its react ion with hydroxide and ammon ia , then it is poss ib le that too much thiosulfate could sterical ly hinder these react ions. T h e effect of thiosulfate on the trithionate degradat ion rate w a s not completely in agreement with that found by Naito et al (1975). T h e range of thiosulfate concentrat ions tested w a s based on those to be expected during gold leach ing, and were much higher than those used by Naito et a l . (typically 0.02 M). T h e trithionate degradat ion rate dec reased at higher levels of thiosulfate, rather than increasing a s expec ted from Naito et al . 's rate equat ion. Naito et a l . (1975) expla ined the catalytic effect they observed when thiosulfate w a s present by proposing that thiosulfate formed a comp lex with trithionate and the complex reacted with water more rapidly than trithionate a lone. In the p resence of ammon ia , cupr ic copper did not affect the react ion rate. In the c a s e s tested, the ammon ia w a s present in a molar ratio of more than 4:1 to cupr ic copper , so it is expec ted that a signif icant amount of the cupr ic copper w a s present a s the tet raammine complex . T h e p resence of ammon ia is known to stabi l ise cupr ic copper and hence inhibit the react ion between cupr ic and thiosulfate. It has been proposed that thiosulfate jo ins the inner co-ordinat ion sphere of cupr ic tet raammine for the react ion to p roceed (Byer ley et a l , 1973a , Breuer and Jeffrey, 2000 , 2003b) . A similar mechan i sm may be relevant for the interaction between trithionate and copper , but due to steric h indrances, co-ordinat ion of trithionate to cupric tet raammine may not be as favourable (trithionate is larger than thiosulfate). W h e r e copper is not primarily present as the cupr ic tet raammine complex, there may be an inf luence on the trithionate degradat ion rate. 107 6.2 MODELLING OF TRITHIONATE DEGRADATION T w o types of exper imental method were used to invest igate the react ion kinetics of the degradat ion of trithionate: the integrated rate method and the initial rate method. T h e integrated rate method gave useful information about the sys tem and showed trends consis tent with those found using the initial rate method. The ana lys is which fol lows is based primarily on results der ived from the initial rate method, as in these tests the exper imenta l var iables could be better control led over the short duration of e a c h test. T h e effects noted in the testwork on react ion kinetics were combined to derive a mathemat ica l model to express the trithionate degradat ion rate as a function of the solut ion condi t ions. It w a s found that the rate of trithionate degradat ion w a s first order with respect to the trithionate concentrat ion. T h e observed rate could thus be exp ressed by Equat ion 6.3. T h e observed rate constant , k0bS> w a s found to depend on other solut ion condit ions and the structure of this dependency w a s determined by testing one solut ion var iable at a t ime. Fo r a solut ion of sod ium trithionate in water at 40 °C, the observed rate constant w a s found to have the magni tude 0.012 h _ 1 (see Sect ion 5.4). T h e observed rate constant kobs in Equat ion 6.3 can thus be exp ressed a s Equat ion 6.4 under these condit ions. T h e p resence of hydroxide at pH -11 and less (at 4 0 °C) increased the trithionate degradat ion rate. In this c a s e , the dependency of the observed rate constant on the hydroxide concentrat ion could be exp ressed by Equat ion 6.5 (see Sect ion 5.5). -d [S 3 0 6 2 - ] /d t = k o b s [S 3 0 6 2 - ] [6.3] kobs = k 0 = 0 .012 h -1 [6.4] k o b s = M O K ] + k 0 = 0.74 M" 1.h- 1[OH-] + 0.012 h -1 [6.5] 108 T h e p resence of ammon ia and ammon ium, wh ich are of importance in the ammon iaca l gold leaching sys tem, were found to inf luence the trithionate degradat ion rate. Addi t ion of a m m o n i a to the sys tem a lso increased the p H . In deriving the dependency of the observed rate constant on the ammon ia concentrat ion, it w a s thus necessa ry to correct for the dependency on the hydroxide concentrat ion determined earl ier. In a plot of the corrected rate constant against the ammon ia concentrat ion, fixing the intercept as the base l ine degradat ion rate in water, the observed rate constant could be exp ressed by Equat ion 6.6. (See Sect ion 5.8) kobs = k 2 [NH 3 ] + M O H ] + k 0 = 0.0081 M- 1 .h- 1 [NH 3 ] + 0.74 M' 1 .h - 1 [OH] + 0.012 rf 1 [6.6] T h e equivalent dependency on ammon ia concentrat ion der ived by Naito et a l . exp ressed in a simi lar way as Equat ion 6.6 showed a dependency a s in Equat ion 6.7. T h e s e authors did not state the pH of their work nor did they invest igate any effects of hydroxide concentrat ion or pH on the degradat ion rate. Thei r ammon ia dependency w a s much less than that obtained in this work. However , the value of the gradient in the plot of the adjusted observed rate constant versus the a m m o n i a concentrat ion w a s very sensi t ive to the intercept va lue se lec ted , and within the error range of each data point, a number of trend l ines could be fit to the data. The ammon ia dependency w a s of the s a m e order of magni tude. T h e inf luence of the ammon ium ion w a s to increase the rate of trithionate degradat ion. T h e potass ium ion gave a very similar effect and the two ions are of a simi lar s i ze , s o it w a s proposed that the ammon ium ion's interaction w a s based on its assoc ia t ion with the trithionate ion (see Sect ion 6.1). At neutral pH in the a b s e n c e of ammon ia , the observed rate constant could be exp ressed by Equat ion 6.8. In this c a s e , no correct ion for hydroxide concentrat ion w a s necessa ry a s the pH w a s neutral. kobs = 0.0031 M- 1 .h- 1 [NH 3 ] + 0.012 h ,-1 [6.7] kobs = k 3 [ N H 4 + ] + k 0 = 0 . 0 1 M - 1 . h - 1 [ N H 4 + ] + 0.012 h ,-1 [6.8] 109 Other solut ion components like b icarbonate, carbonate, sulfate and oxygen were shown to have a negligible effect on the trithionate degradat ion rate. T h e interaction between trithionate and thiosulfate w a s more complex , depend ing on both the thiosulfate concentrat ion and the ammon ium concentrat ion. For the purposes of model l ing the trithionate kinet ics, the condit ions expected to most c lose ly match typical gold leaching condit ions were cons idered . Typ ica l thiosulfate concentrat ions used in gold leaching are around 0.2 M, and this concentrat ion d e c r e a s e s with t ime. At the high p H s used in gold leaching, most of the total a m m o n i a present (typically 0.2 -0.4 M) is expec ted to be present a s ammon ia , not a m m o n i u m ions, s o typically the ammon ium ion concentrat ion is expected to be less than 0.2 M. Under these condi t ions, the effect of thiosulfate on the observed rate constant for trithionate degradat ion w a s very smal l , and for the purposes of model l ing trithionate kinetics, the effect of thiosulfate w a s cons idered negl igible. N o interaction of trithionate with copper w a s accounted for, a s in the p resence of sufficient ammon ia , cupr ic copper w a s found to have no signif icant effect on the trithionate degradat ion rate. T h e degradat ion of trithionate was exp ressed as the s u m of var ious interact ions, occurr ing in paral lel . Us ing the data at 40 °C the var ious interactions were combined a s in Equat ion 6.9, using Equat ions 6.5, 6.6 and 6.8. kobs = k 3 [NH 4 + ] + k 2 [NH 3 ] + k,[OH] + k 0 [6.9] = 0.01M" 1 .h" 1 [NH 4 + ] + 0.0081 M ' 1 . h 1 [ N H 3 ] + 0.74 M- 1 .h- 1 [OrT] + 0.012 h"1 T h e s a m e approach of parallel interactions was used by Naito et a l . (1975) but they only examined the effects of water, ammon ia and thiosulfate, and at concentrat ions that were not speci f ical ly relevant to gold leaching. T h e data for the measured observed rate constants for the initial rate tests are shown plotted against pH in F igures 6.2 to 6.4 for three representat ive ranges in total ammon ia plus ammon ium concentrat ions. T h e observed rate constant a s calcu lated by Equat ion 110 6.9 is shown super imposed for e a c h . T o calculate k o b s f rom Equat ion 6.9, the pH w a s conver ted to [ O H ] , and a lso used to calculate the ratio of a m m o n i a to ammon ium, g iven the total ammon ia plus ammon ium and using the p K a va lue for ammon ia of 8.8 at 40 °C (Dean , 1992). Quali tat ively, the model led va lues match the exper imental t rends, showing a local min imum in degradat ion rate at pH around 10, but the exper imenta l data shows a lower min imum degradat ion rate at high ammon ia / ammon ium concentrat ions than the model predicts (i.e. the min imum rate is lower). F igure 6.5 shows the calcu lated rate constant plotted against the observed rate constant for all the data on a logarithmic sca le . In s o m e c a s e s the ammon ia concentrat ion w a s not measured but est imated based on the ammon ium concentrat ion and p H . Th is is indicated in F igure 6.5 where the correlat ion coeff icient, R 2 , is equa l to 0.6. Whi le the scatter is signif icant, the gradient is one, showing an overal l qualitative match. T h e typical 10 % error margins in the exper imental data are thought to contribute to this scatter significantly. 0.04 Nhb + NH 4 concentration x 0 M ko =0.012 k i = 0.74 k 2 = 0.0081 k a = 0.01 Figure 6.2 : Plot of observed rate constant knhg for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0 M against p H , with model led trend super imposed (using k? = 0.0081 M 1 . h 1 ) 111 0.04 0.03 V 0.02 0.01 0.00 5 - 1 ~ 7 PH 11 N H + NH 4 concentration 13 x 0.5 M Ko = 0.012 ki = 0.74 k2 = 0.0081 ka = 0.01 F igure 6.3 : Plot of observed rate constant k^. for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0.5 M against p H , with model led trend super imposed (using k ? = 0.0081 M' 1 .h ' 1 ) 0.04 0.03 0.02 0.01 0.00 i 5 i 7 PH 11 Ni-fe + 1NH4 concentration 13 x 1 M ko = 0.012 ki = 0.74 k2 = 0.0081 ka = 0.01 F igure 6.4 : Plot of observed rate constant knhs for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0.9 - 1.2 M against p H , with model led trend super imposed (using k ? = 0.0081 M' 1 .h ' 1 ) 112 Figure 6.5 : Plot of ca lcu lated rate constant k R a i r ve rsus observed rate constant I w for trithionate degradat ion at 40 °C for all da ta , using k? = 0.0081 M ' 1 . h ' 1 If the dependency of the observed rate constant on a m m o n i a (k 2) is changed to 0.0049 M" 1 .h" 1 (as indicated by the best fit line through the ammon ia dependency data points, F igure 5.6, Sec t ion 5.8), the fit of the model appears better, as is s e e n for different total a m m o n i a concentrat ions in F igures 6.6 to 6.12. However , in the plot of the calcu lated ve rsus the observed rate constant in Figure 6.13, the correlat ion coeff icient R 2 is aga in 0.6. A s d i scussed earl ier, the s lope of the line in F igure 5.6 for the a m m o n i a dependency w a s very sensi t ive to the c h o s e n intercept and g iven the error margins for e a c h data point, there were l ines of many s lopes that were consistent with the data. If we refer back to Equat ion 5.11 where the dependency of the rate constant on both the a m m o n i a and ammon ium concentrat ions w a s measu red , a value for k 2 of 0 .0049 M" 1 .h" 1 w a s more consistent with the exper imental data. 113 0.04 NFfe + NH.4 concentration x O M Ko = 0.012 ki = 0.74 k 2 = 0.0049 ks = 0.01 Figure 6.6 : Plot of observed rate constant k ^ for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0 M against p H , with model led trend super imposed (using k? = 0.0049 M' 1 .h ' 1 ) 0.04 0.03 0.02 0.01 0.00 Nhb + NH 4 concentration x 0.15 M ko = 0.012 ki = 0.74 k 2 = 0.0049 ka = 0.01 Figure 6.7 : Plot of observed rate constant k ^ for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0.1 - 0.2 M against p H , with model led trend super imposed (using k? = 0.0049 M' 1 .h ' 1 ) 114 0.04 0.03 V 0.02 0.01 0.00 7 9 PH 11 Nhfe + NH 4 concentratbn 13 x 0.35 M ko =0.012 ki = 0.74 k 2 = 0.0049 ks = 0.01 F igure 6.8 : Plot of observed rate constant k^ for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0.3 - 0.4 M against p H , with model led trend super imposed (using k ? = 0.0049 M" 1.h" 1) 0.04 0.03 V 0.02 .a 0.01 0.00 7 9 PH X j 11 Nhb + NH 4 concentratbn x 0.5 M 13 ko = 0.012 ki = 0.74 k 2 = 0.0049 ka = 0.01 F igure 6.9 : Plot of observed rate constant k n h g for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0.5 M against p H , with model led trend super imposed (using k ? = 0.0049 M" 1.h" 1) 115 0.04 0.00 PH 11 NH3 + NH 4 concentration 13 x 0.75 M Ko = 0.012 ki = 0.74 k 2 = 0.0049 kg = 0.01 Figure 6.10 : Plot of observed rate constant knhg for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 0.7 - 0.8 M against p H , with model led trend super imposed (using k 2 = 0.0049 M 1 . h 1 ) 0.04 0.03 T 0.02 0.01 0.00 N H 3 + N H 4 concentration x 1 M ko = 0.012 ki = 0.74 k 2 = 0.0049 KB = 0.01 Figure 6.11 : Plot of observed rate constant k ^ for trithionate degradat ion at 40 °C for an ammon ia + ammon ium concentrat ion of 0.9 - 1.2 M against p H , with model led trend super imposed (using k ? = 0.0049 M' 1 .h ' 1 ) 116 0.04 0.03 V 0.02 0.01 0.00 NHs + NH4 concentration x2.1 M ko = 0.012 k i = 0.74 k 2 = 0.0049 kg = 0.01 Figure 6.12 : Plot of observed rate constant k n h g for trithionate degradat ion at 40 °C for an a m m o n i a + ammon ium concentrat ion of 1.8 - 2.6 M against p H , with model led trend super imposed (using k ? = 0.0049 M" 1.h" 1) 117 T o summar ize , based on the exper imental work, the degradat ion of trithionate to thiosulfate and sulfate at 40 °C could be adequate ly model led using the rate equat ion and parameters exp ressed in Equat ion 6.10. T h e s low react ion of trithionate with water, represented by k 0, will predominate in typical gold leaching solut ions. Th is rate equat ion w a s used in combinat ion with rate equat ions for other sulfur spec ies to model the overal l sulfur speciat ion expec ted during thiosulfate degradat ion during gold leaching. Th is is d i s cussed in Chapte r 7, where it will be shown that al though the model for trithionate degradat ion is prone to a fair amount of variability, this does not affect the overal l sys tem speciat ion to any signif icant degree. -d[S 3 0 6 2 - ] /d t = (k 3 [NH 4 + ] + k 2 [NH 3 ] + k,[OH] + k 0 ) [ S 3 O 6 2 ] [6.10] where at 40 °C k 0 = 0.012 h' 1 ^ = 0.74 iVr 1.h' 1 k 2 = 0.0049 IVT.IY1 k 3 = 0.01 ivr1.h-1 118 7 MODELLING OF SULFUR OXYANION SPECIATION DURING THIOSULFATE DEGRADATION 7.1 INTRODUCTION T h e degradat ion of thiosulfate in gold leaching sys tems , which a lso involve ammon iaca l and copper components , is complex and not fully understood. It has been proposed that thiosulfate can degrade directly to tetrathionate, trithionate or sulfate, but tetrathionate and trithionate themse lves undergo further degradat ion. No publ ished model for this sys tem has been found in the public domain . Th is chapter shows how a s imple model w a s set up for thiosulfate degradat ion in the a b s e n c e of ores and shows the model sensit ivity to both the model parameters and exper imenta l condi t ions. T h e model w a s tested against a set of exper imental data in the a b s e n c e of ore and compared with the exper imenta l behaviour in the p resence of an ore. B a s e d on this evaluat ion, the shor tcomings and s c o p e of use of the model were identif ied. T h e model w a s used to show s o m e expected effects of changing the solution condit ions on thiosulfate degradat ion. 7.2 MODEL SETUP Figure 7.1 shows a simpli f ied schemat ic of s o m e of the poss ib le react ion pathways for thiosulfate in an ammon iaca l , copper-contain ing solut ion. Th is schemat i c w a s used to set up a bas ic kinetic model of the sys tem, using rate equat ions der ived from the literature or f rom exper imental work in this thesis (for trithionate). It should be noted that this schemat i c is not exhaust ive but the react ions represented are expected to character ise the sys tem sufficiently wel l to be able to use the model to identify the factors playing the most signif icant role in thiosulfate degradat ion and sulfur oxyanion spec ia t ion. In the schemat ic the react ions are label led R1 to R 6 , with fract ions a , b, c and d represent ing the fraction of thiosulfate at any time reacting through e a c h of the pathways shown . It is recogn ised that the va lues for a , b, c and d could change with t ime, but this w a s not incorporated into the mode l . 119 R1 S 2 0 3 2 - + C u 2 + -> 1 / 2 S 4 0 6 2 - + C u + •a s 2 o 3 2 -t R2 S 4 0 6 2 - + 3 / 2 OH- 5 / 4 S 2 0 3 2 - + 1 / 2 S 3 0 6 2 " + 3 / 4 H 2 0 R3 S 3 O e 2 - + 2 OH" -> S 2 0 3 2 " + S 0 4 2 " + H 2 0 R4 \ \ S 2 0 3 2 " + V 3 H 2 0 + 2 / 3 0 2 -» 2 / 3 S 3 0 6 2 " + 2 / 3 OH" d \ S 2 0 3 2 " + 2 0 2 + 2 OH" -> 2 S 0 4 2 " + H 2 0 V R6 S 2 0 3 2 " + 2 OH- + 2 / 3 C u 2 + -> 4 / 3 S 0 3 2 " + H 2 0 + 2 / 3 C u S Figure 7.1 : Schemat i c showing the bas is for model l ing T h e way in which e a c h of react ions R1 to R 6 were handled in setting up the model is d i s cussed in turn below. 7.2.1 R1 - Thiosulfate Degradation to Tetrathionate s2o32- + C u 2 + -» y2 s4o6 2' + c u + [7.1] T h e rate equat ion for the react ion between thiosulfate and copper w a s taken from work by Byer ley et al (1973a). T h e rate of dec rease of cupr ic copper concentrat ion by react ion with thiosulfate in the a b s e n c e of oxygen w a s g iven by Equat ion 7.2. -d[Cu 2 + ] /dt = k R 1 [Cu 2 + ] [S 2 0 3 2 - ] / [NH 3 ] [7.2] where k R 1 = 8.5 x 10" 4 s" 1 at 30 °C. Us ing the Arrhenius equat ion and the activation energy of 102.5 kJ /mo l , the rate constant at 25 °C w a s extrapolated to be 4.2 x 10" 4 s ' 1 . B a s e d on the react ion stoichiometry for Reac t ion R 1 , the rate of thiosulfate degradat ion at 25 °C could be exp ressed by Equat ion 7.3. 120 - d [ S 2 0 3 2 l / d t = 4.2 x 10" 4 [Cu 2 + ] [S 2 0 3 2 " ] / [NH 3 ] M.s ' 1 [7.3] T h e rate constant w a s found to increase up to 40 t imes in the p resence of oxygen , but it is a lso known that other react ions can occur in the p resence of oxygen (e.g. R 4 and R5) . 7.2.2 R2 - Te t ra th ionate D e g r a d a t i o n S 4 0 6 2 - + % O H " 5 / 4 S 2 0 3 2 - + V2 S 3 0 6 2 " + 3 / 4 H 2 0 [7.4] In the a b s e n c e of ammon ia , copper and oxygen , the rate of the alkal ine decompos i t ion of tetrathionate w a s descr ibed by Zhang and Dreis inger (2002) in Equat ion 7.5. -d[S 40 6 2-]/dt = k [ S 4 0 6 2 ] [ O H ] [7.5] where k = 1.38 x 1 0 3 M" 1 .h" 1 at 22 ° C and the activation energy w a s 98.5 kJ /mo l . H e n c e at 25 ° C , k = 5 . 7 4 x 1 0 - 1 IvrVs - 1 . T h e effect of copper w a s not cons idered by Z h a n g and Dreis inger in the derivation of this rate equat ion and copper has been sugges ted to react with tetrathionate at a signif icant rate (Breuer and Jeffrey, 2003b) . A l s o , the poss ib le catalytic effect of thiosulfate on the degradat ion w a s not cons idered. Whi le Naito et a l . (1970b) examined the react ions of tetrathionate in ammon iaca l sys tems , no rate equat ion w a s der ived. The effect of thiosulfate w a s add ressed in a study by Ro l ia and Chakrabar t i (1982) who der ived a rate equat ion for tetrathionate degradat ion in the a b s e n c e of ammon ia and copper at pH 11. T h e rate equat ion in the p resence of thiosulfate at 25 °C w a s exp ressed as Equat ion 7.6. -d [S 4 0 6 2 " ] /d t = (k R 2 . 0 + k R 2 . 1 [ S 2 0 3 2 l ) [ S 4 0 6 2 l [ O H ] [7.6] where k R 2 . 0 = 0.022 M" 1 .s" 1 and k R 2 . i = 2.77 M" 2 .s" 1 . T h e rate constants were determined using the publ ished rate constants at 35 °C and the activation energy in the a b s e n c e of thiosulfate (115.5 kJ/mol) . 121 In the a b s e n c e of thiosulfate, the rate equat ion in Equat ion 7.6 is reduced to Equat ion 7.5, but it should be noted that the rate constant kR 2-o is more than 10 t imes smal ler than that found by Zhang and Dreis inger. It w a s proposed by Z h a n g and Dreis inger that the p resence of oxygen in the work of Rol ia and Chakrabar t i and the a b s e n c e of oxygen in their own work w a s probably the reason for this d isc repancy , and that oxidants could possib ly retard the rate of tetrathionate degradat ion. Breuer and Jeffrey (2004) found that the ionic strength had an effect on tetrathionate degradat ion and proposed that the di f ference between the rate constants of Zhang and Dreis inger and Ro l ia and Chakrabar t i w a s b e c a u s e the solution ionic strengths were different. S ince the model led sys tem w a s for thiosulfate degradat ion in the p resence of oxygen and thiosulfate, the rate equat ion p roposed by Ro l ia and Chakrabar t i (Equat ion 7.6) w a s used in the mode l . 7.2.3 R3 - Trithionate Degradation Whi le limited literature data is avai lable for the degradat ion of trithionate in ammon iaca l solut ions, the methodology and condit ions used to derive the kinet ics were not entirely sui table for gold leaching condit ions. T h e rate equat ion der ived in this thesis (see Chap te r 6) w a s used in model l ing the sys tem, as shown in Equat ion 7.8. T h e limitations of this rate equat ion were d i scussed in Chapte r 6. Th is rate equat ion is most sui table for pH va lues where the ammon ium to thiosulfate ratio is smal l , a s d i s cussed in Chap te rs 5 and 6. It shou ld be noted that the rate equat ion in Equat ion 7.8 inc ludes the hydrolysis of trithionate, represented by k 0. S 3 0 6 2 " + 2 O H " -> S 2 0 3 2 " + S 0 4 2 " + H 2 0 [7.7] - d [ S 3 0 6 2 " ]/dt = (k 0 + HOH] + k 2 [NH 3 ] + k 3 [ N H 4 + ] ) [ S 3 0 6 2 - ] [7.8] where at 4 0 °C k 0 = 3 . 3 x 1 0 - 6 s 1 k, = 2.1 x l O " 4 M ' 1 . s k 2 = 1.3 x 10" 6 IvT.s k 3 = 2 . 7 x 1 0 " 6 IVTVS .-1 .-1 ,-1 122 Us ing the activation energy of 75.3 kJ /mo l , the parameters at 25 °C were calcu lated to be k 0 = 1.4 x 1 0 " 8 s " 1 ki = 5 . 8 x 1 0 - 5 M 1 . s " 1 k 2 = 3.1 x 1 0 " 7 M - 1 . s 1 k 3 = 6 . 4 x 1 0 - 7 M - 1 . s 1 7.2.4 R4 - Thiosulfate Degradation to Trithionate S 2 0 3 2 " + 1 / 3 H 2 0 + 2 / 3 0 2 -» 2 / 3 S 3 0 6 2 " + 2 / 3 O H " [7.9] T h e degradat ion of thiosulfate directly to trithionate in the p resence of copper ammines and oxygen a s p roposed by Byer ley et a l . (1975) w a s d i scussed in the literature review in Chap te r 2. T h e react ion kinetics of thiosulfate degradat ion were measured by measur ing the oxygen consumpt ion. It w a s found that under certain condi t ions, the initial rate of oxygen consumpt ion general ly cor responded to the formation of trithionate only, whi le after longer t imes when the max imum oxygen consumpt ion w a s reached , both trithionate and sulfate were formed. Byer ley et a l . did\not deduce a comprehens ive rate equat ion using their data, as the sys tem appeared quite complex. However , for the pu rposes of this mode l , the kinetic data for the initial rate of oxygen consumpt ion measu red by Byer ley et a l . w a s used to derive an approx imate rate equat ion, as fol lows. A s noted by Byer ley et a l . , the rate of oxygen consumpt ion w a s found to be proportional to the copper and oxygen concentrat ions. T h e dependency on thiosulfate and ammon ia concentrat ions w a s not straightforward. A s s u m i n g that the rate of thiosulfate degradat ion w a s equal to the rate of oxygen consumpt ion (as a s s u m e d by Byer ley et al.) the logarithm of this rate w a s plotted against the logarithm of ammon ia concentrat ion in F igure 7.2 and against the logarithm of the thiosulfate concentrat ion in Figure 7.3. 123 -11.0 CD -t—» ro S -11.5 Q. E CO c o o CM o -12.0 -12.5 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 In NH, Figure 7.2 : Logar i thm of the oxygen consumpt ion rate versus the logarithm of ammon ia concentrat ion for thiosulfate degradat ion to trithionate in the p resence of oxygen - Data from Byer ley et al (1975) -10.0 c -12.0 H 1 . 1 1 -4.0 -3.5 -3.0 -2.5 -2.0 In S 2 0 3 2 -Figure 7.3 : Logar i thm of the oxygen consumpt ion rate versus the logarithm of thiosulfate concentrat ion for thiosulfate degradat ion to trithionate in the p resence of oxygen - Data from Bver lev et al (1975) 124 It w a s found that the rate of oxygen consumpt ion had a react ion order of approximately -1 with respect to thiosulfate (for concentrat ions greater than 0.025 M) and approximately zero with respect to ammon ia (in the range of about 0.2 M to 0.6 M ammonia) . T h e rate w a s independent of the a m m o n i a concentrat ion for ammon ia concentrat ions where the cupr ic tetraammine spec ies w a s expec ted to be at a max imum concentrat ion relative to the cupr ic concentrat ions used in this work, and dec reased at higher or lower a m m o n i a concentrat ions. Therefore Byer ley et a l . p roposed that the cupr ic tet raammine spec ies is the act ive cupr ic spec ies in the formation of trithionate from thiosulfate. B a s e d on this prel iminary evaluat ion of Byer ley et al 's data , the overal l rate equat ion was exp ressed as Equat ion 7.10. - d [S 2 0 3 2 " ] / d t = k R 4 [Cu 2 + ] [0 2 ] / [S 2 0 3 2 - ] [7.10] T h e rate constant k R 4 w a s est imated for each set of data, based on the measured rate and the test condi t ions, and the average rate constant w a s found to be 1.03 + 0.17 s" 1 (at pH 11.2, 30 °C at wh ich most of the data w a s avai lable) . No activation energy w a s measu red but g iven the limited accuracy of determining the rate us ing oxygen consumpt ion and of deriving the rate equat ion shown in Equat ion 7.10, the rate constant at 25 °C is expec ted to be of the s a m e order of magni tude as that at 30 °C. H e n c e using the rate constant est imated at 30 °C w a s cons idered sufficiently accura te within these limitations. A concern in using this rate equat ion is the fact that the oxygen consumpt ion rate is not likely to be a direct indicator of the thiosulfate degradat ion rate. A l s o , the rate equat ion and rate constant were der ived under very speci f ic condit ions which may not a lways be appl icable to thiosulfate leaching condit ions. 7.2.5 R5 - Thiosulfate Degradation Directly to Sulfate S 2 0 3 2 " + 2 0 2 + 2 O H 2 S 0 4 2 " + H 2 0 [7.11] 125 It has been found that sulfate can form directly f rom thiosulfate in the p resence of oxygen (Byerley et a l , 1975). However , under condit ions typical in gold leach ing, the proportion of sulfate formed (via React ion R5) to trithionate formed is expec ted to be very low, as documented in the literature review in Chap te r 2. S ince the extent of this react ion is expec ted to be negligible compared with compet ing react ions and no rate equat ion w a s avai lab le, Reac t ion R 5 w a s exc luded from the mode l . 7.2.6 R6 - Thiosulfate Degradation to Sulfide S 2 0 3 2 " + 2 O H " + 2 / 3 C u 2 + % S 0 3 2 " + H 2 0 + 2 / 3 C u S [7.12] Wh i le react ion R 6 showing the formation of copper sulf ide from the react ion between copper and thiosulfate w a s included in Figure 7.1 for comp le teness , kinetic data on this react ion w a s not avai lable. A l s o during gold leaching, based on m a s s ba lances of the sulfur spec ies , sulf ide does not appear to be formed in any signif icant quanti t ies, if at all (Lam, 2002). H e n c e this reaction w a s not included in the mode l . 7.2.7 Incorporation of the Rate Equations into a Model - Method and Constraints T h e model w a s set up in a s imple sp readshee t format. T h e concentrat ion profi les of the var ious sulfur oxyanion spec ies with time were determined iteratively us ing time intervals of 0.05 hours over 24 hours. T h e change in concentrat ion of e a c h of the spec ies thiosulfate, trithionate, tetrathionate and sulfate w a s determined using the rate equat ions in Equat ions 7.3, 7.6, 7.8 and 7.10 and the react ion stoichiometr ies of Equat ions 7.1, 7.4, 7.7 and 7.9. T h e s e equat ions accounted for all the sulfur spec ies in the model s o the total sulfur present remained constant. T h e concentrat ions of the sulfur oxyan ions at t ime t as a function of the solut ion condi t ions, react ion rates and react ion stoichiometry are shown in Equat ions 7.13 to 7.16. 126 [ S 2 0 3 2 - ] t = [S 2 0 3 2 - ] t - i + (% r2 t + r3t - M t - r4 t)At [ S 4 0 6 2 - ] , = [ S 4 0 6 2 ^ - i + ( V l , - r2,)At [ S 3 0 6 2 " ] t = [S 3 0e 2 l t . i + ( 1 / 2 r2 t + 2 / 3 r 4 t - r3 t)At [S0 4 2 " ] t = [ S 0 4 2 - ] M + (r3t)At [7.13] [7.14] [7.15] [7.16] where M , = a k R 1 [SzOg^t-iICult-i/ENHalM r2 t = (k R 2^, + k ^ I S z O a ^ t - i ) [OHIt-i [S 4 0 6 2 - ] t - i r3t = (k 0 + k ^ O H j , , + k 2 [ N H 3 ] „ + k 3 [ N H 4 ] M ) [ S 3 0 6 2 - ] „ r4, = b k R 4 [CulnIOzlt- i / ISaOa 2 ! , . , [7.17] [7.18] [7.19] [7.20] Standard va lues for the rate constants in Equat ions 7.17 to 7.20 and initial condit ions are d i s c u s s e d in Sect ion 7.3. T h e model w a s subject to the fol lowing constraints: T h e only chemica l react ions a s s u m e d to be occurr ing were those shown in Figure 7.1 a s d i scussed above . It w a s a s s u m e d that any cuprous copper formed w a s immediately re-oxid ised to cupric, s o that the cupric concentrat ion remained constant. It w a s a s s u m e d that the solut ions conta ined 10 mg/l d isso lved oxygen . T h e a m m o n i a concentrat ion, pH and d isso lved oxygen concentrat ion were a s s u m e d to remain constant. T h e pH w a s used to calculate the hydroxide ion concentrat ion and the ratio of ammon ia to ammon ium given the total ammon ia concentrat ion. T h e effects of ore were ignored. T h e effects of other an ions (e.g. sulfate) were ignored. 7.3 MODEL SENSITIVITY TO MODEL PARAMETERS T h e model output is dependent on the form of the rate equat ions, the va lues of the rate constants and the solut ion condit ions. Before testing the validity of the model against 127 exper imenta l data , the sensit ivity of the model to changing the model parameters w a s invest igated. It w a s ant ic ipated that the model as set up in Sect ion 7.2 may not a lways be suitable to descr ibe exper imental observat ions during gold leaching as the condit ions under which the var ious model rate equat ions were der ived were not a lways related to gold leaching. By identifying the effects of the var ious parameters , a better understanding of poss ib le shor tcomings of the model in descr ib ing exper imenta l si tuat ions could be ga ined. In this sect ion, the forms of the rate equat ions were not changed , but the qualitative effects of chang ing the rate constants for e a c h rate equat ion (see Tab le 7.1) and a lso of vary ing the proportion of thiosulfate reacting to form tetrathionate ve rsus trithionate (React ion R1 versus Reac t ion R4) were invest igated. A s a s tandard condit ion, the rate constants d i s cussed in Sect ion 7.2 were used , and it w a s a s s u m e d that 80 % of the thiosulfate reacted to form tetrathionate v ia Reac t ion R1 and 20 % reacted to form trithionate via Reac t ion R 4 . T h e s e percentages were se lec ted s imply a s examp les , and are consistent with the observat ion (see later in Sect ion 7.4) that a greater percentage of thiosulfate appea rs to react v ia Reac t ion R1 to form tetrathionate. Tab le 7.1 : Mode l parameters used to test model sensit ivity Paramete r S tandard value R a n g e tested a (proportion of thiosulfate react ing via R1 to form tetrathionate) 8 0 % 0 - 1 0 0 % b (proportion of thiosulfate reacting via R 4 to form trithionate) 2 0 % 0 - 1 0 0 % km 4 . 2 x 1 0 ^ s" 1 4.2 X10" 4 s 1 x 1 0 kp.2-0 0.022 M ^ . s1 0.022 M " 1 . s - 1 x 1 0 kp.2-1 2.77 M" 2 .s" 1 2.77 M " 2 . s 1 x 10 k 0 1.4 x 10"8 s 1 1.4 x 10" 8 s" 1 x 10 k i : 5 . 8 x 1 0 - 5 M - 1 . s - 1 5 . 8 x 1 0 - 5 M - 1 . s ' 1 x 1 0 k 2 3.1 x 1 0 "7 M 1 . s 1 3.1 x 1 0 7 M^ .s " 1 x 10 k 3 6 . 4 x l 6 '7 M - 1 . s 1 6 . 4 x IO" 7 M" 1 .s" 1 x 10 kp.4 1.03 s" 1 1.03 s 1 x 10 128 T h e solut ion condit ions in Tab le 7.2 were se lec ted as standard condit ions for the pu rposes of this evaluat ion. Tab le 7.2 : S tandard exper imental parameters used in model l ing L e a c h condit ions Standard va lue Temperature 25 °C [ S 2 0 3 2 ] 0 0.2 M [NH 3 ] t 0 . 0.4 M [Cu 2 + ] 30 mg/l P H 10 D 0 2 10 mg/l 7.3.1 Proportion of Thiosulfate Forming Tetrathionate versus Trithionate T h e proportion of thiosulfate reacting directly to form tetrathionate (React ion R1) versus that reacting directly to form trithionate (React ion R4) (i.e. a and b in Tab le 7.2) w a s cons ide red . In Figure 7.4 it is a s s u m e d that thiosulfate does not form trithionate directly, only tetrathionate (i.e. a = 100%) while in F igure 7.5, only trithionate is formed (i.e. b = 100%). T w o intermediate combinat ions of these two react ion paths are shown in F igures 7.6 and 7.7. 129 12 Time (hrs) 18 0.10 0.08 0 CD —\ GO 0.06 c 0.04 0.02 0.00 24 co "O CD o CD' CO •Thiosulfate - Tetrathionate | -Trithionate Sulfate Figure 7.4 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion • A s s u m e d all thiosulfate degradat ion is via Reac t ion R1 (a = 100 %, b = 0%). • Initial condit ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. 130 12 Time (hrs) 18 0.10 0.08 4 - 0.04 0.02 0.00 24 CD —i co 0.06 g. co T3 CD o CD" co -Thiosulfate Tetrathionate Trithionate - Sulfate Figure 7.5 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion • A s s u m e d all thiosulfate degradat ion is via Reac t ion R 4 (a = 0 %, b = 100 %). • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. F igures 7.4 and 7.5 show the ex t remes of the model with respect to the thiosulfate degradat ion pathway. F rom this it is obv ious that the model output is very sensi t ive to the ratio of thiosulfate a s s u m e d to react to form tetrathionate ve rsus the thiosulfate that forms trithionate directly. Quali tat ively, where Reac t ion R1 to tetrathionate is favoured, the thiosulfate degradat ion curve is concave . Whi le this is not immediately c lear f rom Figure 7.4, it is apparent if the thiosulfate concentrat ion axis is changed . W h e r e Reac t ion R 4 to trithionate is favoured, the thiosulfate degradat ion curve is convex and the trithionate formation curve is much more concave . T h e s e observat ions were used to a s s e s s the model against exper imental data in Sect ion 7.4. 131 Figure 7.6 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion • A s s u m e d thiosulfate degradat ion is v ia both Reac t ion R1 and React ion R 4 (a = 50 %, b = 50 %). • Initial condit ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. 132 0.10 0.08 0.06 4- 0.04 0.02 12 T ime (hrs) 1 0.00 24 CD co c co TD CD O CD' CO -Thiosulfate - Tetrathionate I - Trithionate Sulfate Figure 7.7 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion • A s s u m e d thiosulfate degradat ion is via both React ion R1 and Reac t ion R 4 (a = 80 %, b = 20 %). • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. For the purposes of demonstrat ing the sensit ivity of the model to other parameters, it w a s a s s u m e d that around 80 % of the thiosulfate reacts via Reac t ion R1 and 20 % reacts via Reac t ion R 2 , s imply a s an example . Al l other f igures showing the sensit ivity of the model to the model parameters can be compared to Figure 7.7. 7.3.2 Rate of Reaction R1 - Thiosulfate Degradation to Tetrathionate T h e rate constant for React ion R 1 , k R 1 , w a s increased by a factor of 10 and the model output is shown in Figure 7.8. W h e n compared with Figure 7.7, it can be s e e n that increas ing the va lue of k R 1 has a signif icant effect on the model output, increasing the rate of thiosulfate degradat ion, the rate of trithionate formation, and the tetrathionate concentrat ion. T h e standard value for k R 1 shown in Tab le 7.1 w a s der ived in the 133 a b s e n c e of oxygen and it is known that the rate of this react ion can increase up to 40 t imes in the p resence of oxygen (Byerley et a l , 1973b). 12 Time (hrs) 18 0.05 24 •Thiosulfate Tetrathionate I Trithionate - Sulfate Figure 7.8 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion - kgi increased by a factor 10 • A s s u m e d thiosulfate degradat ion is via both Reac t ion R1 and React ion R 4 (a = 80 %, b = 20 %). • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. 7.3.3 Rate of Reaction R 4 - Thiosulfate Degradation to Trithionate T h e rate constant for Reac t ion R 4 , k R 4 , w a s increased by a factor of 10 and the model output is shown in Figure 7.9. Th is has a t remendous effect on the output, giving very rapid thiosulfate degradat ion and trithionate formation, accentuat ing the convex nature of the thiosulfate degradat ion curve and concave nature of the trithionate formation curve. Ve ry little tetrathionate is formed. Note that the d a s h e d line for the thiosulfate concentrat ion profile after about 16 hours indicates that the rapid decl ine of thiosulfate to zero concentrat ion may not be reliable and is a manifestat ion of the inverse dependency of the thiosulfate degradat ion rate on the thiosulfate concentrat ion (Equat ion 7.10). T h e 134 rate constant descr ib ing this reaction w a s increased by a factor of 10 in Figure 7.9, which is much higher than would be expec ted based on compar ison with exper imenta l data (see Sect ion 7.4). F igure 7.9 : Model led output for sulfur oxyanion speciat ion during thiosulfate degradat ion - kp4 increased by a factor 10 • A s s u m e d thiosulfate degradat ion is via both React ion R1 and React ion R 4 (a = 80 %, b = 20 %). • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. 7.3.4 Rate of Reaction R 2 - Tetrathionate Degradation T h e rate constants for Reac t ion R 2 , kR2-o and k R 2 . i , were increased by a factor of 10 (separately) and the model output for each situation is shown in F igures 7.10 and 7.11. T h e sca le has been changed to show the dif ferences in tetrathionate concentrat ion more clearly. Wh i le adjust ing the value of k R 2 . 0 did not have much effect, the tetrathionate concentrat ion dec reased signif icantly by increasing k R 2 . i . 135 Figure 7.10 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion - k R ? ^ increased bv a factor 1 0 • A s s u m e d thiosulfate degradat ion is via both Reac t ion R1 and Reac t ion R 4 (a = 80 %, b = 20 %). • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. 136 Figure 7.11 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion - kR ?.i increased by a factor 10 • A s s u m e d thiosulfate degradat ion is via both Reac t ion R1 and Reac t ion R 4 (a = 80 %, b = 20 %). • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. 7.3.5 Rate of Reaction R3 - Trithionate Degradation T h e rate constants for React ion R 3 , k 0 , k i , k 2 and k 3 , were all i nc reased by a factor of 10, and the model output is shown in Figure 7.12. Increasing the rate of trithionate degradat ion had the overal l effect of lowering the rate of trithionate formation, and the formation of sulfate w a s enhanced . T h e trithionate concentrat ion curve b e c a m e more convex. 137 0.05 0.04 4- 0.03 0.02 4- 0.01 0.00 CD cn tz cn TD CD o CD' cn -Thiosulfate - Tetrathionate I - Trithionate Sulfate Figure 7.12 : Mode l led output for sulfur oxyanion speciat ion during thiosulfate degradat ion - kn. ki. k? a n d k 3 i n c r e a s e d b v a f a c t o r 1 0 A s s u m e d thiosulfate degradat ion is v ia both Reac t ion R1 and React ion R 4 (a = 8 0 % , b = 2 0 % ) . Initial condit ions: 0.2 M S 2 0 3 2 " , 0.4 M N H 3 , 30 mg/l C u 2 + , pH 10, 25 °C. 7.3.6 Summary - Effect of Model Parameters T h e percentage of thiosulfate react ing to form tetrathionate ve rsus that react ing to form trithionate had a signif icant impact on the model output, as did the magni tude of the rate constants for these react ions. W h e r e tetrathionate formation w a s favoured, the thiosulfate concentrat ion profile was more concave , whi le where trithionate formation w a s favoured it w a s convex. Chang ing the rate constants for tetrathionate degradat ion had a signif icant effect on the tetrathionate concentrat ion, a lso affecting the thiosulfate and trithionate concentrat ions slightly, a s these spec ies are products of that degradat ion. T h e rate of degradat ion of trithionate affected the trithionate concentrat ion, as well as having an impact on the thiosulfate and sulfate concentrat ions. Increasing the rate of trithionate degradat ion c a u s e d the trithionate concentrat ion profile to b e c o m e more convex. Overa l l , the model w a s most sensi t ive to the rate of degradat ion of thiosulfate to 138 tetrathionate and/or trithionate. A compar ison of the model output with exper imenta l data (Sect ion 7.4) w a s used to better est imate these model parameters . 7.4 COMPARISON OF MODEL OUTPUT WITH EXPERIMENTAL RESULTS IN THE ABSENCE OF ORE T h e a im of this sect ion w a s to determine whether the model qualitatively and quantitatively w a s able to descr ibe exper imental data in the a b s e n c e of ore, and to identify shor tcomings in the model . 7.4.1 Experiments A few exper iments were carr ied out to provide a data set against which to test the mode l . A solut ion containing ammon ia and copper (added as cupr ic sulfate) w a s made up and the pH adjusted using sulphuric ac id (to pH 9 or 10). T h e solut ion w a s transferred to a beaker in a water jacket and brought to 25 °C. A n aliquot of sod ium thiosulfate solut ion w a s added s o that the final solut ion would have the required concentrat ion of thiosulfate (0.2 M), copper (30 or 100 mg/l) and ammon ia (0.3 - 0.4 M). T h e h e a d s p a c e above the solut ion w a s purged with oxygen and the beaker sea led with plast ic fi lm. T h e solut ion w a s stirred using a magnet ic stirrer and kept at 25 °C. T h e test condit ions are shown in Tab le 7.3. S a m p l e s were removed periodical ly by syr inge connec ted to a tube extending into the solut ion and diluted as required for ana lys is for thiosulfate, trithionate and tetrathionate concentrat ions by ion chromatography, as d i scussed in Chap te r 3. Sul fate w a s not measu red a s it could not be determined at the s a m e t ime as the other sulfur oxyan ions and storing of the samp le for later ana lys is of sulfate w a s not cons idered due to the somet imes rapid change in solut ion concentrat ions. Therefore no sulfur ba lance could be carr ied out. T h e total ammon ia concentrat ion w a s ana lysed by titration, as d i scussed in Chapte r 3. 139 Tab le 7.3 : Exper imenta l condit ions used in tests for model val idation Test Initial thiosulfate C o p p e r Total a m m o n i a pH (M) (mg/l) (M) 1 0.2 0 0.42 10 2 a 0.2 32 0.42 10 2b 0.2 32 0.38 10 3 0.2 30 0.33 9 4 0.2 30 0.40 9 5 0.2 101 0.30 10 T h e solut ion w a s kept under oxygen s ince the model a s s u m e d that there w a s sufficient oxygen present s o that all cuprous copper w a s re-oxidised to cupr ic copper . In a more e laborate mode l , it would be necessa ry to take copper speciat ion into considerat ion. A s s u m i n g that only thiosulfate w a s likely to react with cupr ic copper (and not any trithionate or tetrathionate that formed), the max imum oxygen requirement to maintain all the copper in the cupric form could be determined based on the total thiosulfate concentrat ion. Accord ing to Equat ion 7.1, each mole of thiosulfate ox id ized requires one mole of cupr ic copper . H e n c e for 100 ml 0.2 M thiosulfate solut ions, 20 mmol cupr ic copper would be required for complete oxidation of thiosulfate to tetrathionate. E v e n with low levels of copper present (30 - 100 mg/l) cont inual replenishment of the cupric could facil itate this. T h e amount of oxygen required to oxid ise 20 mmol of cuprous copper is 5 mmol . B a s e d on the standard g a s vo lumes at a tmospher ic pressure, this is equivalent to about 112 ml oxygen . T h e h e a d s p a c e in the vesse l used w a s approximately 160 ml, al lowing for sufficient oxygen to be present s o as not be limiting. 7.4.2 V a l i d a t i o n M e t h o d Wh i le the forms of the rate equat ions were not changed , the magni tude of the rate constants and the proportion of thiosulfate react ing to form tetrathionate or trithionate (i.e. a or b in F igure 7.1) were adjusted to fit exper imenta l data to identify shor tcomings of the mode l . B e c a u s e the change of thiosulfate concentrat ion w a s g iven by the appropr iate rate equat ion multiplied by the percentage of thiosulfate react ing either v ia Reac t ion R1 or R 4 , the product of this percentage and the appropr iate rate constant for 140 e a c h of these react ions w a s cons idered as a s ingle parameter in this evaluat ion. H e n c e the fol lowing four parameters were adjusted a s necessa ry to attempt to find a good fit of the model to the exper imental data: a x k R 1 for Reac t ion R 1 , where a = 1 for the standard c a s e k R 2- 0 and k R 2 . i (adjusted together by the s a m e factor a s k R 2 ) for Reac t ion R 2 k 0, k 1 t k 2 and k 3 (adjusted together by the s a m e factor a s k R 3 ) for Reac t ion R 3 b x k R 4 for Reac t ion R 4 , where b = 1 for the standard c a s e T h e s e parameters were multiplied by a factor as necessa ry to give the best fit. H e n c e a factor for a k R 1 of 10 would imply that the percentage of thiosulfate reacting via Reac t ion R1 multiplied by the rate constant would be 10 t imes higher than if all the thiosulfate reacted via Reac t ion R1 (a = 1) and the rate constant k R 1 w a s its s tandard va lue as in Tab le 7.1. Th is method cannot dist inguish whether an overal l inc rease of, 10 t imes (for a k R 1 ) is due to a 10 fold increase in the rate constant only, for examp le , or a 20 fold inc rease in the rate constant but with only 50 % of the thiosulfate react ing via this react ion. T w o app roaches were cons idered in compar ing the exper imental data to the model output. In the first approach , it w a s attempted to use a mathemat ica l method to adjust the model parameters to obtain the best fit. T h e s u m of the squares of the errors between the exper imenta l data points and the model led va lues w a s min imised by adjust ing the model parameters. However , this approach gave a large variation in the best fit mode l parameters , giving what appeared to be mean ing less scenar ios at t imes. A l s o the final ou tcome of the adjusted parameters w a s strongly dependent on the starting va lues se lec ted . T h e concentrat ions of thiosulfate and trithionate found exper imental ly were general ly higher than those of tetrathionate and sulfate. H e n c e minimising the s u m of the squares of errors for all spec ies could be easi ly b iased depend ing on which spec ies were present in higher concentrat ions and on whether the absolute or relative errors were used . T h e s e c o n d approach used v isual inspect ion and adjustment of parameters and gave more realist ic results. Th is method used what w a s learnt in examin ing the sensit ivity of the model to the model parameters in Sect ion 7.3 and the parameters were adjusted based on the expected shor tcomings of parts of the model and a v isual best fit. T h e 141 concave and convex natures of the concentrat ion profi les were used to est imate whether the sys tem w a s b iased towards thiosulfate degradat ion to tetrathionate or trithionate, and the magni tudes of the rate constants were adjusted to al low for an improved model fit to match both the measured concentrat ions and the qualitative s h a p e of the concentrat ion profi les. T h e results obtained were more realist ic and consistent using this app roach . T h e five sets of test results are first cons idered separate ly and the overal l f indings are summar i sed thereafter. In all f igures, the model output is shown by cont inuous l ines whi le exper imenta l data is shown as individual data points. S i n c e no exper imental data for sulfate concentrat ions were avai lable, the model led sulfate concentrat ions are not shown . 7.4.3 Test 1 - No Copper T h e graph in Figure 7.13 shows the model output and exper imenta l data points for a test with no copper present. Ove r the duration of the test there w a s no signif icant thiosulfate degradat ion. T h e model a s s u m e s that copper is present. T h e s e results show that it is appropr iate to require copper in the model to have any signif icant thiosulfate degradat ion, for these condit ions and test duration. 142 0 0.2 0.04 o ~ 0.1 CD cz 2 O ' o O CN CO 0.0 » T • 1 1 1 1 1 1 1 r 0 2 4 6 8 10 12 14 16 18 20 22 24 T ime (hrs) 0.00 -Thiosulfate Trithionate Tetrathionate Figure 7.13 : Mode l output versus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.42 M N H 3 , 0 mg/l C u 2 + , pH 10, 25 °C, sea led vesse l with 0 2 in head s p a c e 7.4.4 T e s t s 2a a n d 2b - L o w C o p p e r , p H 10 T h e fol lowing condit ions were used in Tes t 2 (a and b): C u 2 + (added a s C u S 0 4 ) - 32 mg/ l , N H 3 (total) - 0.38 - 0.42 M, pH 10. T h e graph in Figure 7.14 shows the test data. T h e thiosulfate concentrat ion curve is concave whi le the trithionate concentrat ion curve is convex . T h e initial data point for the thiosulfate concentrat ion is lower than expec ted , probably due to poor mixing. T h e tetrathionate concentrat ion reaches a max imum within the first 2 hours of reaction then slowly degrades . B a s e d on these qualitative trends and the qualitative behaviour of the mode l , it is reasonab le to a s s u m e that most of the thiosulfate reacts via react ion R1 to form tetrathionate. T h e model output assuming that all thiosulfate reacts via this pathway and none reacts to form trithionate directly is a lso shown in F igure 7.14. In Figure 7.14, the model predicts much s lower thiosulfate 143 degradat ion than observed exper imental ly and correspondingly much s lower trithionate and tetrathionate formation. 0.2 4 > c o CD V 0.1 C O 0.0 • •<> o 0.06 0.05 0.04 CD CO rz 0.03 % CD O CD' CO 4- 0.02 <=; 0.01 0.00 0 2 4 6 8 10 12 14 16 18 20 22 24 26 T ime (hrs) •Thiosulfate o • -Trithionate A • Tetrathionate Figure 7.14 : Mode l output versus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Mode l parameters adjusted by the following factors: a k R 1 - 1 x, k R 2 - 1 x, k R 3 - 1 x, b k R 4 - O x . • Initial condi t ions: 0.2 M S 2 0 3 2 \ 0.42 M N H 3 (c losed symbols ) or 0.38 M N H 3 (open symbols) , 32 mg/l C u 2 + , pH 10, 25 °C, sea led vesse l with 0 2 in head space . • A s s u m e d that all thiosulfate reacts v ia react ion R 1 . Mult iplying the rate of thiosulfate degradat ion v ia Reac t ion R1 (ak R 1 ) by a factor of 10 al lowed a much better fit to the thiosulfate degradat ion data. It is known that this react ion p roceeds m u c h faster in the p resence of oxygen than in its a b s e n c e (for wh ich the rate equat ion w a s derived), so increasing the rate is justif iable. Th is effect is shown in F igure 7.15. However , al though the thiosulfate data w a s adequate ly descr ibed by the 144 mode l , the tetrathionate concentrat ion predicted w a s too high, and the trithionate concentrat ion too low. 0.06 -Thiosulfate -Trithionate - Tetrathionate Figure 7.15 : Mode l output versus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Mode l parameters adjusted by the fol lowing factors: akR1 -10 x, k R 2 - 1 x, k R 3 - 1 x, b k R 4 - 0 x. • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.42 M N H 3 (c losed symbols ) or 0.38 M N H 3 (open symbols) , 32 mg/l C u 2 + , pH 10, 25 °C, sea led vesse l with 0 2 in head s p a c e • A s s u m e d that all thiosulfate reacts via react ion R 1 . T h e rate of tetrathionate degradat ion predicted by the model w a s too low. However , increas ing the rate of tetrathionate degradat ion would affect the net thiosulfate concentrat ion, a s thiosulfate is a product of tetrathionate degradat ion. Hence , increasing the rate of tetrathionate degradat ion would a lso require a further increase of the thiosulfate degradat ion rate to maintain a good fit for the thiosulfate data. F igure 7.16 shows the model output where the tetrathionate degradat ion rate (k R 2 ) is increased by a factor of 8 and the thiosulfate degradat ion rate to tetrathionate (ak R 1 ) is increased by a 145 factor of 14. This produces a better fit in general, however, the trithionate concentration curve is not quite convex enough to fit the experimental data. 0.06 —Thiosulfate — Trithionate - - Tetrathionate Figure 7.16 : Model output versus experimental data for sulfur oxyanion speciation during thiosulfate degradation • Model parameters adjusted by the following factors: akR1 - 1 4 x, kR2 - 8 x, kR 3 -1 x, bkR 4 - 0 x. • Initial conditions: 0.2 M S 2 0 3 2 \ 0.42 M NH 3 (closed symbols) or 0.38 M NH 3 (open symbols), 32 mg/l Cu 2 + , pH 10, 25 °C, sealed vessel with 0 2 in head space • Assumed that all thiosulfate reacts via reaction R1. It was shown in determining the model sensitivity to the model parameters that increasing the rate of trithionate degradation would give a more convex trithionate concentration profile. Thiosulfate is produced during the degradation of trithionate, so increasing the trithionate degradation rate would required an increase in the rate of thiosulfate degradation to ensure a satisfactory fit remains for the thiosulfate data. Also, since it has been shown that trithionate forms from thiosulfate in the presence of copper 146 and oxygen , it is not realist ic to a s s u m e that no thiosulfate reacts v ia Reac t ion R 4 . A l lowing for a smal l fraction of thiosulfate to react v ia this route ( b k R 4 = 0.05) e n h a n c e s the model fit slightly. In Figure 7.17, the model output is shown where the rate of thiosulfate degradat ion is increased by a factor of 15, the rate of tetrathionate degradat ion is inc reased by a factor of 8, the rate of trithionate degradat ion is increased by a factor of 5, and s o m e react ion of thiosulfate to trithionate is a l lowed. Th is p roduces a slightly more sat isfactory fit to the exper imenta l data than in Figure 7.16. 0.06 o | c CD O rz o o CO o CM CO "l 1 1 1 1 i 1 1 1 r 0 2 4 6 8 10 12 14 16 18 20 2 2 24 26 T ime (hrs) -Thiosulfate o • -Trithionate A A Tetrathionate | o Figure 7.17 : Mode l led ve rsus exper imental data for sulfur oxyanion spec ia t ion during thiosulfate degradat ion • Mode l parameters adjusted by the following factors: akR1 -15 x, kR2 - 8 x, kR3 -5 x, bkR4 - 0.05 x. • Initial condit ions: 0.2 M S 2 0 3 2 ~ , 0.42 M N H 3 (c losed symbols) or 0.38 M N H 3 (open symbols) , 32 mg/l C u 2 + , pH 10, 25 °C, sea led vesse l with 0 2 in head s p a c e • A s s u m e d that thiosulfate reacts v ia reaction R1 and R 4 . 147 7.4.5 Test 3 - Low Copper, pH 9, Low Ammonia The following conditions were used in Test 3: C u (added as C u S 0 4 ) - 30 mg/l, N H 3 (total) - 0 .33 M, pH 9. The graph in Figure 7.18 shows that the thiosulfate concentration curve is concave while the trithionate concentration curve is convex, but less so than for Test 2 at higher p H . The overall thiosulfate degradation was less and less trithionate was formed . The max imum tetrathionate concentration reached was higher than that at higher p H . The model output shown in Figure 7.18 a s s u m e s that the thiosulfate reacts only via React ion R1 to form tetrathionate, based on the qualitative trends of the data . Whi le the thiosulfate degradation predicted adequately fits the data, especia l ly at shorter times, the tetrathionate concentrations predicted were too high while the trithionate concentrations were too low, as was found in Test 2. Increasing the tetrathionate degradation rate provided a better fit, as shown in Figure 7.19. However, the trithionate concentration curve in Figure 7.19 was concave instead of convex. T o alter the shape of the trithionate concentration curve, the rate of trithionate degradation could be increased . Increasing the rate of trithionate degradation in this c a s e would improve the shape of the model led trithionate profile, but the va lues would be too low compared with experimental data . T o increase the trithionate concentration, the reaction of thiosulfate degradation to trithionate was ass igned a non-zero value. F igure 7 .20 shows the model output where it is a s sumed that part of the thiosulfate reacts v ia React ion R 4 to form trithionate, and a lso that the trithionate degradation rate is 5 t imes higher than predicted by the original model . The fit to the experimental data is much improved. 148 0.02 CZ o 2 -#—* rz CD O c o o CN CO o CM CO 0.1 H 0.0 0.01 o zr CD —I co rz CO "a CD o CD" CO — i 1 i i 1 1 i i 1 1 1 1— 0 2 4 6 8 10 12 14 16 18 20 22 24 26 0.00 T ime (hrs) -Thiosulfate - Trithionate Tetrathionate Figure 7.18 : Mode l output versus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k R 2 - 1 x, k R 3 - 1 x, b k R 4 - 0 x. • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.33 M N H 3 , 30 mg/l C u 2 + , pH 9, 25 °C, sea led vesse l with 0 2 in head s p a c e • A s s u m e d that all thiosulfate reacts via react ion R 1 . 149 c o 2 -»—' fZ cu o cz o o C O 0.02 0.01 CD Cfl rz tn TD CD O CD' 0.00 Time (hrs) -Thiosulfate - Trithionate Tetrathionate Figure 7.19 : Model output versus experimental data for sulfur oxyanion speciation during thiosulfate degradation • Model parameters adjusted by the following factors: ak R 1 - 1 x, kR2 - 10 x, kR 3 - 1 x, bkR 4 - 0 x. • Initial conditions: 0.2 M S 2 0 3 2 " , 0.33 M NH 3 , 30 mg/l Cu 2 + , pH 9, 25 °C, sealed vessel with 0 2 in head space. • Assumed that all thiosulfate reacts via reaction R1. 150 0.02 c o « CO c 0 o c o o C O CD CO c CO • a CD o CD CO 0 2 4 6 8 10 12 14 16 18 20 22 24 26 T i m e (hrs) •Thiosulfate Trithionate — - — - Tetrathionate F igure 7.20 : Mode l output versus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 x, k R 2 - 10 x, k R 3 - 5 x, b k R 4 - 0.1 x. • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.33 M N H 3 , 30 mg/l C u 2 + , pH 9, 25 °C, sea led vesse l with 0 2 in head space . • A s s u m e d that all thiosulfate reacts via reaction R 1 . 7.4.6 Test 4 - Low Copper, pH 9, Higher Ammonia Tes t 4 w a s very simi lar to Test 3, except that the a m m o n i a concentrat ion w a s higher at 0.40 M. T h e fol lowing condit ions were used in Test 4: C u 2 + (added a s C u S 0 4 ) - 30 mg/ l , N H 3 (total) - 0.40 M, pH 9. Fitting of the model required simi lar adjustments as for Tes t 3, and a satisfactory fit is shown in Figure 7.21. 151 0.02 —Thiosulfate — Trithionate - - Tetrathionate Figure 7.21 : Mode l output ve rsus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Mode l parameters adjusted by the following factors: a k R 1 - 1.5 x, k R 2 - 10 x, k R 3 -5 x, b k R 4 - 0.1 x. • Initial condi t ions: 0.2 M S 2 0 3 2 \ 0.40 M N H 3 , 30 mg/l C u 2 + , pH 9, 25 °C, sea led vesse l with 0 2 in head s p a c e • A s s u m e d that all thiosulfate reacts v ia react ion R 1 . 7 . 4 . 7 Test 5 - H i g h Copper, pH 1 0 T h e fol lowing condit ions were used in Tes t 5: C u 2 + (added a s C u S 0 4 ) - 101 mg/ l , N H 3 (total) - 0.30 M, pH 10. A s shown in Figure 7.22, the thiosulfate concentrat ion curve is initially concave while the trithionate concentrat ion curve is initially convex . T h e higher copper concentrat ion c a u s e d a much more rapid thiosulfate degradat ion and trithionate product ion than in Test 2. T h e tetrathionate reached a max imum concentrat ion then degraded much more rapidly than for Test 2 at lower copper concentrat ion. B a s e d on 152 the bas ic s h a p e of the curves, it w a s first a s s u m e d that no thiosulfate reacted via Reac t ion R 4 to form trithionate directly. A s in the previous examp les , it w a s necessa ry to inc rease the rate of thiosulfate degradat ion to tetrathionate and of tetrathionate degradat ion to improve the model fit. Th is model output is shown in Figure 7.22. Whi le the model fit w a s adequate at shorter t imes, it did not descr ibe the data at longer t imes. Thiosulfate Trithionate Tetrathionate F igure 7.22 : Mode l output ve rsus exper imental data for sulfur oxyanion speciat ion dur ing thiosulfate degradat ion • Mode l parameters adjusted by the fol lowing factors: akR1 -16 x, kR2 - 8 x, k R 3 -1 x, b k R 4 - 0 x. • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.30 M N H 3 , 101 mg/l C u 2 + , pH 10, 25 °C, sea led vesse l with 0 2 in head s p a c e • A s s u m e d that all thiosulfate reacts via reaction R 1 . A more l inear trend for both thiosulfate concentrat ion and trithionate concentrat ion w a s required. A l lowing for thiosulfate degradat ion via both Reac t ion R1 and React ion R 4 w a s found to give this type of trend. F igure 7.23 shows the model output where this is taken into account . 153 Thiosulfate Trithionate Tetrathionate F igure 7.23 : Mode l output versus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Mode l parameters adjusted by the fol lowing factors: a k R 1 - 12x, kR2 - 8 x, k R 3 - 1 x, b k R 4 - 0.15 x. • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.30 M N H 3 , 101 mg/l C u 2 + , pH 10, 25 °C, sea led vesse l with 0 2 in head space . • A s s u m e d that all thiosulfate reacts via react ion R 1 . Simi lar output is obtainable where the rate of trithionate degradat ion is inc reased 5 fold to be consistent with that required for Tes ts 2, 3 and 4. Th is is shown in Figure 7.24. 154 Thiosulfate Trithionate Tetrathionate F igure 7.24 : Mode l output versus exper imental data for sulfur oxyanion speciat ion during thiosulfate degradat ion • Mode l parameters adjusted by the fol lowing factors: a k R 1 - 1 2 x , k R 2 - 8 x , k R 3 -5 x , b k R 4 - 0 . 1 7 x . • Initial condi t ions: 0.2 M S 2 0 3 2 " , 0.30 M N H 3 , 101 mg/l C u 2 + , pH 10, 25 °C, sea led vesse l with 0 2 in head s p a c e • A s s u m e d that thiosulfate reacts via reaction R1 and R 4 . 7.4.8 S u m m a r y Wh i le adjustments to the model were based only on v isual inspect ion, the extent of such adjustments required to give a better fit to the exper imental data give useful insight into the shor tcomings of the mode l . E x a m p l e s of adjustments found to improve the agreement between the model and exper imental data are shown in Tab le 7.4. It shou ld be noted that these model parameter adjustments are not unique. 155 Tab le 7.4 : Adjustment of model parameters found to give improved agreement between model output and exper imenta l data Tes t [Cu] (mg/l) [NH 3 ] (M) P H Multipl ication factor for a k R 1 Multipl ication factor for k R 2 ( k R 2 . i and l D O CD 3 a x o v> c_ ST r* CD o CD a CD a. I r~ CO 0) ro o o • 1 M Au ore, quartz, muscovite, 90 % <74 Stirred tank 25,40, 60 3 0.125-2 Na 0 1 - 8 30-60 8 . 5 -10.5 40 % solids -80 40-50% lodometric titration B Quartz Stirred tank 25 24 0.25 NH 4 0.25 M S 0 4 2 ' 1 6 With NH4OH 20 42 3% lodometric titration with Pyrite Stirred tank 25 24 0.25 NH 4 0.25 M so 4 2 " 1 6 With NH4OH 20 392 28% acetic acid added to Arsenopyrite Stirred tank 25 24 0.25 NH 4 0.25 M S O 4 2 " 1 6 20 252 18% eliminate Cu -NH 3 complex Chalcopyrite Stirred tank 25 24 0.25 NH* 0.25 M so 4 2 -1 6 With NH4OH 20 168 12% ettect. pyrrhotite Stirred tank 25 24 0.25 NH 4 0.25 M S O 4 2 " 1 6 With NH4OH 20 140 10% C Untreated pyrite cone Stirred tank RT? 96 0.4 Na 0 4 50 10.2 80-180 417 61 57% Immediate HPLC 96 0.5 4 50 52 39% 24 96 0.8 0.8 4 50 30 56 14% 26% 24 96 0.8 1.5 50 39 128 18% 59% 24 96 0.8 3 50 99 115 46% 54% 24 96 0.8 4 12.5 90 166 42% 77% 24 96 0.8 4 50 22 99 10% 46% 24 96 0.8 4 37.5 41 137 19% 64% 24 96 0.8 4 62.5 40 80 19% 37% to leached D Synthetic Ag 2S in air -20 24 0.072 NH 4 0 16 9.92 Ag86.7 192 43% IX, also for s4o62-in air -20 24 0.073 NH 4 0 15 9.92 Ag81.7 164 36% in N 2 -20 24 0.073 NH« 0 15 9.92 Ag 85.3 114 25% in air -20 24 0.073 NH 4 3.5 mM so3 15 9.92 Ag66.4 114 25% in N 2 -20 24 0.073 NH 4 3.5 mM S 0 3 15 9.92 Ag 70.3 68 15% in air -20 24 0.073 NH 4 14 mM S 0 3 15 9.92 Ag50.7 68 15% in N 2 -20 24 0.073 NH 4 14 mM S 0 3 15 9.92 Ag44.7 25 5% E Low grade oxidized Au ore Bottle roll ambient 48 0.2 Na 0-0.0125 M not specified 1 83 0.40 IC for SjOa2", S0 3 2 " . S 0 4 2 ' Bottle roll ambient 94 0.2 Na 0-0.0125 M not specified 1 90 1.71 Bottle roll ambient 94 0.125 Na 0-0.0125 M not specified 1 85.8 3.60 Bottle roll ambient 94 0.05 Na 0-0.0125 Mnot specified 1 <11 4.61 to E Rhyolite and andesite breccia matrix, 80 % < 200 mesh, Mn Stirred tank 25 28 0.15 NH 4 -31 8.5 40 % solids 8 32% lodometric titration Stirred tank 25 28 0.23 NH 4 -31 8.5 40 % solids 12 32% Stirred tank 25 28 0.38 1NH4 -31 8.5 40 % solids 15 23% Stirred tank 25 28 0.15 NH 4 -31 9.5 40 % solids 13 52% Stirred tank 25 28 0.23 NH 4 -31 9.5 40 % solids 21 55% Stirred tank 25 28 0.38 NH 4 -31 9.5 40 % solids 33 52% Stirred tank 25 28 0.12 N1H4 -31 10 40 % solids 9 36% Stirred tank 25 28 0.23 NH 4 -31 10 40 % solids 19 50% Stirred tank 25 28 0.38 NH 4 -31 10 40 % solids 35 55% F Mn ore, 2.1 % Mn, difficult to teach See text 18%? NH 4 3% (NH4)2 SO3 2 % NH4OH - 4 g/l 86.7 3.6 G Sulfide gold flotation cone with Cu, 90% -200 mesh Stirred tank 60 0.5 0.14 1NH4 0.8 M (NH4fc S 0 4 4 47 10.2 500 2.2 7% unspecified Stirred tank 60 1 0.14 NH 4 0.8 M (NH4)2 SO4 4 47 10.2 500 2.2 7% Stirred tank 60 1.5 0.14 NH 4 0.8 M (NH4)2 SO« 4 47 10.2 500 0.0 0% Stirred tank 60 2 0.14 NH 4 0.8 M (NH4)2 SO4 4 47 10.2 500 6.6 21% to o H Refractory gold ore Bottle roll RT 24 0.5 NH 4 - 6 50 10.2 45 % solids 72 52.0 H Refractory gold ore Column RT 50 d 0.1 NH 4 - 3 30 10.2 3000 45 15.5 Column RT 50 d 0.3 NH 4 - 3 30 10.2 3000 64.2 27.5 Column RT 50 d 0.5 NH 4 - 3 30 10.2 3000 64.9 30.3 Column RT 50 d 1 NH 4 - 3 30 10.2 3000 71.9 36.8 Column RT 50 d 0.3 maintained NH 4 - 1 50 10.2 3000 57 35.8 Column RT 50 d 0.3 maintained NH 4 - 3 50 10.2 3000 60.7 30.4 Column RT 50 d 0.3 maintained NH 4 - 6 50 10.2 3000 71.7 28.1 Column RT 50 d 0.3 maintained NH 4 - 3 50 10.2 3000 67.7 27.5 Column RT 50 d 0.3 maintained NH 4 - 3 50 10.2 3000 60.7 30.5 Column RT 50 d 0.3 maintained NH 4 - 3 50 10.2 3000 56.7 35.8 Column RT 50 d 0.5 NH 4 - 3 30 10.2 830 38.6 Column RT 50 d 0.5 NH 4 - 3 30 10.2 5000 15.9 lodometric titration References for Appendix 2 A Abbruzzese et al, 1995 B Fend and Van Deventer, 2002b C Aylmore, 2001 D Rett e ta l , 1983 E Langhans et al, 1992 F Kerley and Bernard, 1981 G Cao eta l , 1992 H Yen eta l , 1999 Appendix 3 : Determination of Sulfur Oxyanions by Ion Chromatography (modif ied from a method by Lakef ie ld R e s e a r c h Ltd) Analysis of thiosulfate. trithionate and tetrathionate Reagent Preparation Cal ibrat ion s tandards Prepare 1000 mg/l thiosulfate solution by d issolv ing the required amount of sod ium thiosulfate anhydrous in ultrapure de- ion ised water. P repare 1000 mg/l trithionate solut ion by d issolv ing the required amount of sod ium trithionate in ultrapure de- ion ised water. P repare 1000 mg/l tetrathionate solut ion by d issolv ing the required amount of sod ium tetrathionate in ultrapure de- ion ised water. F r o m these s tandards, prepare cal ibrat ion s tandards containing 2, 10, 15 and 20 mg/l thiosulfate, trithionate and tetrathionate. Cal ibrat ion s tandards should be prepared immediately prior to use . Eluant - S o d i u m perchlorate (0.038 M) with 10 % methanol D isso lve 10.675 g N a C I 0 4 . H 2 0 ( H P L C grade) in water and add 200 ml methanol ( H P L C grade). Dilute to 2 litres using ultrapure de- ion ised water. Filter the solution through a 0.22 u.m filter. Procedure Set up the ion chromatograph accord ing to the manufacturers instructions, using an O m n i P a c P A X 100 guard co lumn and an O m n i P a c P A X 100 analyt ical co lumn. P u m p the eluant through the sys tem at 1 ml/min. Monitor the abso rbance at 205 nm using a UV-v is ib le absorpt ion detector. A l low the instrument to equil ibrate s o that the abso rbance measured is constant (typically 2 hours). Inject the 2 mg/l cal ibration standard solut ion, fol lowed by the 10 mg/ l , 15 mg/l and 20 mg/l s tandard solut ion. 205 C h e c k that the cal ibrat ion curve is l inear and then proceed to ana lyse the samp le solut ions, which must be diluted to within the calibration range immediately prior to ana lys is . Typ ica l retention t imes are about 2 minutes for thiosulfate, 6 minutes for trithionate and 10 minutes for tetrathionate, depend ing on the condit ion of the co lumns. Analysis of sulfate Reagent Preparation Cal ibrat ion s tandards Prepare 1000 mg/l sulfate solut ion by d issolv ing the required amount of sod ium sulfate in ultrapure de- ion ised water. F rom this s tandard solut ion, prepare calibration s tandards containing 2, 10, 15 and 20 mg/l sulfate. Eluant - S o d i u m carbonate (1.8 mM) - sod ium bicarbonate (1.7 mM) with 10 % methanol D isso lve 0.382 g N a 2 C 0 3 and 0.286 g N a H C 0 3 in water and add 200 ml methanol ( H P L C grade). Dilute to 2 litres using ultrapure de- ion ised water. Filter the solut ion through a 0.22 u,m filter. Procedure Set up the ion chromatograph accord ing to the manufacturers instructions, using an l o n P a c A G 4 A - S C guard co lumn and an l o n P a c A S 4 A - S C analyt ical co lumn. P u m p the eluant through the sys tem at 2 ml/min. Se t the suppressor current at 27 mA. Monitor the conductivity using a conductivity detector. A l low the instrument to equil ibrate s o that the conductivity measu red is constant (typically 2 hours). Inject the 2 mg/l cal ibration s tandard solut ion, fol lowed by the 10 mg/ l , 15 mg/l and 20 mg/l s tandard solut ion. 206 C h e c k that the cal ibrat ion curve is l inear and then proceed to ana lyse the samp le solut ions, which must be diluted to within the calibration range immediately prior to ana lys is . Typ ica l retention t imes are about 8 minutes for sulfate, depend ing on the condit ion of the co lumns . 207 Appendix 4 - Chemical Impurity Analysis Impurity Analysis for a solution of Ammonium Bicarbonate Solution of 37.6 g/l N H 4 H C 0 3 Element Concentration (mg/l) A l <0.2 S b 0 .3 A s 0 .3 B a <0.01 Bi <0.1 C d <0.01 C a 0.1 C r <0.01 C o <0.01 C u <0.01 Fe <0.03 La <0.05 Pb <0.05 Mg <0.1 Mn <0.01 Hg <0.05 Mo <0.02 Ni 2.28 P <0.1 K <2 S c 0 .02 A g <0.02 N a 7 S r <0.01 Tl <0.2 Ti <0.1 W <0.1 V <0.01 Zn <0.01 Zr <0.01 208 Impurity Analysis for a solution of Ammonium Hydroxide Solut ion of 0.5 M N H 3 Element Concentrat ion (mg/l) A l <0.2 S b 0.1 A s 1.2 B a <0.01 Bi <0.1 C d <0.01 C a 1 C r <0.01 C o 0.06 C u 0.23 F e <0.03 L a 0.09 P b <0.05 Mg 2.3 Mn <0.01 Hg <0.05 Mo <0.02 Ni 1.54 P <0.1 K <2 S c <0.01 A g <0.02 N a 48 S r <0.01 Tl <0.2 Ti <0.1 W <0.1 V <0.01 Z n <0.01 Zr <0.01 209 Impurity Analysis for a solution of Sodium Thiosulfate Solution of 1.5 g/l N a 2 S 2 0 3 Element Concentration (mg/l) Al <0.2 Sb <0.1 As <0.2 Ba <0.01 Bi <0.1 Cd <0.01 Ca 0.5 Cr <0.01 Co 0.09 Cu <0.01 Fe <0.03 La 0.22 Pb 0.31 Mg 1.2 Mn <0.01 Hg <0.05 Mo <0.02 Ni <0.02 P 0.9 K 20 Sc <0.01 Ag <0.02 Na 529 Sr <0.01 Tl <0.2 Ti <0.1 W <0.1 V <0.01 Zn <0.01 Zr <0.01 210 Impurity Analysis for a solution of Sodium Trithionate Solution of 1.3 g/l Na 2 S 2 0 3 Batch 4 Element Concentration (mg/l) Al <0.2 Sb <0.1 As <0.2 Ba <0.01 Bi <0.1 Cd <0.01 Ca 0.8 Cr <0.01 Co 0.11 Cu <0.01 Fe <0.03 La 0.18 Pb 0.34 Mg 1 Mn <0.01 Hg <0.05 Mo <0.02 Ni <0.02 P <0.1 K 20 Sc <0.01 Ag O.02 Na 316 Sr <0.01 Tl <0.2 Ti <0.1 W <0.1 V <0.01 Zn <0.01 Zr 0.04 211