Mechanism and Rate of Butyl Phosphinate Formation from ...

1

Mechanism and Rate of Butyl Phosphinate Formation from

Reaction of Phosphinic Acid (Cyanex 272) and Tributyl

Phosphate

K.R. Barnard*

CSIRO Process Science and Engineering / Parker Centre / CSIRO Minerals Down Under National Research

Flagship, PO Box 7229, Karawara, WA 6152, Australia.

and D.W. Shiers

Abstract

An impurity species, butyl bis(2,4,4-trimethylpentyl)phosphinate (‘butyl

phosphinate’), has recently been identified in the Murrin Murrin solvent extraction (SX)

circuits. The present work established that this species is formed via direct reaction

between tributyl phosphate (TBP) and the phosphinic acid extractant found in Cyanex

272 and that the reaction is first order relative to the concentration of each reactant. The

observations are consistent with the reaction progressing via a bimolecular nucleophilic

substitution (SN2) mechanism whereby nucleophilic attack of substrate TBP by the

phosphinic acid anion occurs, resulting in cleavage of the C-O bond and ejection of

dibutyl phosphate anion.

The butyl phosphinate formation rate has been determined under synthetic extract,

strip and aqueous-free conditions, the latter at temperatures between 40 and 75°C. In

the absence of an aqueous phase, the rate coefficient was found to be 0.43 ± 0.02 M-1

* Email address: [email protected] . Phone 61-8-9334 8071. Fax 61-8-9334 8001.

2

per annum (95 % confidence interval) at 70°C with an activation energy of 122 kJ/mol.

Based on the present data and historical SX plant operating information, a model to

estimate annual butyl phosphinate generation and its build-up in the Murrin Murrin SX

circuits over the last decade was developed. The estimated accrued level of 31 g/L butyl

phosphinate by June 2009 is comparable to the measured 32-35 g/L.

Keywords: Cyanex 272; bis(2,4,4-trimethylpentyl)phosphinic acid; butyl

phosphinate; butyl bis(2,4,4-trimethylpentyl)phosphinate; tributyl phosphate; TBP;

Murrin Murrin.

1. Introduction

Cyanex 272 is a commercial solvent extraction (SX) reagent used to separate cobalt

from nickel in sulphate media (Flett, 2004). The active component is bis(2,4,4-

trimethylpentyl)phosphinic acid (‘phosphinic acid’), which is present at about 85 % in

Cyanex 272, with tris(2,4,4-trimethylpentyl)phosphine oxide (‘phosphine oxide’) being

the major impurity. This phosphinic acid is also present as the active component in the

alternative SX reagents Ionquest 290 and LIX 272.

The first known product arising from chemical degradation of Cyanex 272 has

recently been reported (Barnard, 2010). The species is the butyl ester of phosphinic

acid, namely butyl bis(2,4,4-trimethylpentyl)phosphinate (‘butyl phosphinate’, Figure

1). It was identified in organic samples from the zinc and cobalt SX circuits in Minara

Resources’ Western Australia-based Murrin Murrin (Murrin) nickel laterite operation.

The organic in these two circuits is essentially equivalent and can be interchanged. The

3

degradation product was assigned as butyl phosphinate after both its isolation and

spectroscopic characterisation from the Murrin organic, and its synthesis from the parent

phosphinic acid via a chloride intermediate.

This recently-identified species had been observed to be growing in the Murrin SX

circuits over the prior 10 years. This is undesirable from a commercial operating

perspective as besides indicating a loss of valuable extractant it could potentially have

adverse effects on the process such as increased acid (and thus aqueous) entrainment in

the organic phase.

The Murrin zinc and cobalt SX circuits in which this species formed both used a

combination of Cyanex 272 and tributyl phosphate (TBP) in Shellsol 2046 diluent. As

butyl phosphinate was not observed in several commercial Cyanex 272-containing

systems where TBP was absent, it was concluded that the impurity was most likely

formed in the plant organic by reaction (either directly or indirectly) of a component of

Cyanex 272 (phosphinic acid or phosphine oxide impurity) with TBP (Barnard, 2010).

Generation of butyl phosphinate from the phosphine oxide impurity was considered very

unlikely as it would require cleavage of a chemically stable P-C bond and direct

replacement of that with the butoxy group. A more likely path was via reaction of the

phosphinic acid hydroxy group. On this basis, two possible mechanisms were proposed,

namely direct reaction between TBP and phosphinic acid, or a two-step process

involving hydrolysis of TBP liberating a butoxy group which subsequently reacted with

the phosphinic acid.

Hydrolysis is a common degradation pathway for TBP used in the nuclear industry

(Tahraoui and Morris, 1995) and for alkyl phosphates in general (Bel’skii, 1977).

However, the butanol formed is relatively soluble in water (77 g/L), and possesses an

4

octanol on water partition coefficient of 7.9 (i.e. log Pow of 0.9, Anon., 2005). Any

butanol generated via hydrolysis of TBP would therefore be lost relatively quickly from

the organic phase by dissolution into the aqueous phase, particularly under extract

conditions. Thus for this mechanism to occur, reaction of butanol with the phosphinic

acid would need to be rapid.

The present work aimed to identify the mechanism by which butyl phosphinate forms

and assess its rate of formation under various conditions and relate this back to the

known butyl phosphinate concentration in the Murrin system and their historical

operating conditions.

2. Experimental

Two major series of experiments were performed, the first primarily under aqueous-

free conditions to determine the mechanism and rate of butyl phosphinate formation, the

second to establish the effect of various aqueous solutions and metal loading levels on

the formation rate.

2.1 Series 1 tests (primarily aqueous-free)

A series of organic solutions containing differing concentrations of phosphinic acid

and trialkyl phosphate (TBP or tris-2-ethylhexylphosphate, TEHP) in Shellsol 2046

were prepared as summarised in Table 1. Cyanex 272, used as the source of phosphinic

acid, was obtained from Murrin which in turn had purchased it from Cytec Industries

Inc. Cyanex 272 was assumed to contain 85 % phosphinic acid (782 g/L or 2.69 M) as

listed on the supplied Material Safety Data Sheet. Pure phosphinic acid used in

experiment 9 was isolated from Cyanex 272 as described previously (Barnard, 2010).

5

TBP was sourced from both Aldrich and Murrin and assumed to be 100 % pure. TEHP

and octadecane (added as an internal standard for subsequent gas chromatography

analysis) were purchased from Aldrich. Shellsol 2046 was supplied by Shell Chemical

Co. Quantification of butyl phosphinate was achieved using previously-synthesised

material (Barnard, 2010).

Relevant organic (60 mL) and the one organic/aqueous mixture (experiment 8, 30 mL

each) were placed in hexagonal glass jars (70 mm diameter x 75 mm height), covered

with a thin layer of PTFE/glass composite fabric (Fiberflon® 108.08), then sealed using

twist top lids and Teflon tape. All 70°C solutions were agitated at 200 r.p.m. using an

Innova 4000 incubator shaker maintained at 70 ± 2°C. This temperature, slightly above

that used routinely at Murrin (60°C), was employed to accelerate and thus more readily

monitor progression of the reaction. Experiments at other temperatures were performed

in duplicate in polyethylene glycol baths maintained at the appropriate temperature

(± 2°C) using IKA RCT Basic heater/stirrers with IKA ETS-D4 temperature controllers.

None of the non-70°C systems were agitated. Samples were taken intermittently

throughout the trial and stored at 4°C prior to gas chromatography (GC) analysis.

2.2 Series 2 tests (primarily aqueous-containing)

A second series of tests (Table 2) were undertaken primarily to assess the effect of

metal-free and metal-containing extract and strip aqueous solutions and the effect of

loaded metals (Co, Zn) on the rate of butyl phosphinate formation. The organic and

aqueous solutions used are listed in Table 2. Organic solutions contained 30 % v/v

Cyanex 272 (0.81 M phosphinic acid) and either 10 % v/v TBP (0.36 M) or 20 % v/v

6

butanol (Aldrich, 2.18 M). Aqueous solutions were prepared using AR grade sulfuric

acid, technical grade metal sulphate salts and de-ionised water.

In systems with no metal extraction (experiments 24-39), relevant organic (60 mL)

and organic/aqueous mixtures (A:O 40 mL:40 mL) were prepared, mixed using the

Innova shaker maintained at 70 ± 2°C, sampled intermittently and stored for subsequent

GC analysis. For systems where metal extraction occurred (experiments 40-47), the

organic and aqueous solutions (40 mL each) were initially mixed at pH 5.0 to allow

metal loading to occur. Samples of the loaded organic solutions taken throughout the

trial were stripped (A:O 3:1, 100 g/L sulfuric acid, 50°C) before storage.

The initial and final day loaded organic samples (experiments 40-47) were separated

from the corresponding aqueous phase via filtering through Whatman 1PS filter paper

before being stripped (A:O 3:1, 100 g/L sulfuric acid, 50°C, 20 minutes). The resulting

aqueous strip solution was filtered through a 0.45 µm Supor membrane filter then

assayed by inductively coupled plasma-optical emission spectroscopy (ICP-OES) for

metal ion content.

2.3 Gas chromatography

GC analyses were performed using a Varian 3800 gas chromatograph equipped with

a flame ionisation detector (FID). The chromatograph was fitted with an SGE forte BP-

1 fused silica capillary column (30 m x 0.32 mm i.d., 0.5 µm film thickness). Helium

(1.5 ml/min) was used as the carrier gas. The injection and detector ports were held at

250°C and 270°C, respectively. After sample injection at 150°C where it was held for

one minute, the oven temperature was increased to 330°C at 30°C/min. A split ratio of

50 and 1.0 µL injection volume was used.

7

Gas chromatography-mass spectrometry (GC-MS) analyses were performed using a

Saturn 2000 mass spectrometer (MS) as the detector. Acetonitrile-based chemical

ionisation (CI) and electon impact (EI) ionisation were used. Helium (1.0 ml/min) was

used as the carrier gas. The transfer line and trap temperatures were 170 and 150°C,

respectively. An AT-5MS column (Alltech, 30 m x 0.25 mm i.d., 0.25 µm film

thickness) was used. After sample injection at 100°C (injection port 270°C), the oven

temperature was increased to 340°C at 15°C/min.

To enable the ready elution of the phosphinic acid extractant, samples (100 µL) were

derivatised by dissolution in 1:1 dimethyl formamide:toluene (600 µL), adding N,O-

bis[trimethylsilyl]acetamide (BSA, 400 µL) and heating (80°C) for 0.5 hours.

Derivatisation was not required for butyl phosphinate detection.

3. Results and Discussion

3.1 Series 1 experiments: Butyl phosphinate generation at 70°C under aqueous-free

conditions

Prior to the commencement of this work, a scoping study undertaken to assess

whether butanol or TBP reacted with Cyanex 272 to generate butyl phosphinate

indicated it was the latter. Given this outcome which was subsequently confirmed-see

Section 3.2, the initial work focussed on Cyanex 272/TBP systems.

3.1.1 Effect of reagent concentration and mechanism of formation

GC traces highlighting the generation of butyl phosphinate in a 20-5-70C system (see

Table 1 for code description) with time are shown in Figure 2. The quantity of butyl

8

phosphinate (expressed as g/L) formed throughout the trial in the different systems at

70°C (experiments 1-12) is summarised in Table 3. For all tests, a linear increase in

butyl phosphinate concentration with time was observed (Figure 3). For each

experiment, the rate of butyl phosphinate formation was determined by a line of best fit

passing through the origin. The resulting annualised values (expressed as both g/L and

mM) are summarised in Table 3.

Increasing the concentration of Cyanex 272 (phosphinic acid) from 2–20 % v/v

(0.054–0.54 M) whilst keeping the TBP concentration constant at 5.0 % v/v (0.18 M,

experiments 1-9) was found to result in a proportional increase in the rate of butyl

phosphinate generation (Figure 4). Although discussed further in Sections 3.1.2, 3.1.3

and 3.1.4, the results for experiments 4-9 (all 20-5-70C experiments but under slightly

different conditions) were found to give comparable results and thus are treated as

equivalent in the present analysis. Importantly, experiment 9 shows butyl phosphinate

formation is attributable to the presence of phosphinic acid in Cyanex 272 and not the

phosphine oxide or any other impurity (Section 3.1.2).

Increasing the concentration of TBP from 1-10 % v/v (0.036 – 0.36 M) whilst

keeping the Cyanex 272 concentration constant at 20 % v/v (0.54 M phosphinic acid,

experiments 4-12) was found to result in a proportional increase in the rate of butyl

phosphinate generation (Figure 4).

These results indicate that the rate of butyl phosphinate formation primarily under

aqueous-free conditions is dependent on the concentrations of both phosphinic acid and

TBP. This relationship can be expressed as a standard rate expression;

Butyl phosphinate formation rate = k [Phosphinic Acid ]x[TBP]y (1)

9

Where k is equal to a rate coefficient that is responsive to the temperature of the

reaction.

The linear response of butyl phosphinate generation to changes in both phosphinic

acid and TBP concentrations (Figure 4) indicate that the ‘x’ and ‘y’ coefficients in

equation 1 are both equal to one. This rate expression therefore simplifies to a second

order reaction:

Rate = k [Phosphinic Acid ] [TBP] (2)

Insertion of the relevant reagent concentrations and butyl phosphinate formation rate

into equation 2 allows the k value in each of the different experiments to be calculated.

The results for experiments 1-12, which are presented in Table 3, indicate an average k

value of 0.43 M-1 per annum at 70°C with a 95 % confidence interval of 0.02 M-1 per

annum.

The fact that the reaction occurs in the absence of an aqueous phase indicates that

degradation has the potential to occur throughout all stages of a process, i.e. mixing,

settling, storage. This reaction and its possible relevance to other SX systems

containing an acidic extractant and TBP is of interest.

3.1.2 Effect of pure phosphinic acid

In addition to phosphinic acid, Cyanex 272 contains impurities such as phosphine

oxide. In order to establish whether phosphinic acid or the phosphine oxide (or any

other impurity species) was reacting with TBP, a sample of pure phosphinic acid was

isolated as described previously (Barnard, 2010). Reaction of this pure phosphinic acid

with TBP under the 20-5-70C conditions (experiment 9) yielded the expected butyl

phosphinate product at a rate consistent with the other 20-5-70C experiments using ‘as

10

supplied’ Cyanex 272. This result confirms phosphinic acid as the reactive species in

Cyanex 272 and that the presence of phosphine oxide impurity has no material effect on

the rate of reaction.

3.1.3 Effect of the presence of an aqueous acid solution

The rate of butyl phosphinate formation in the 20-5-70C system when in the presence

of a 100 g/L sulfuric acid aqueous phase was found to be 13.3 g/L per annum. This

result falls within the average rate of formation of 14.6 ± 1.5 g/L per annum (95 %

confidence interval) calculated for the five anhydrous 20-5-70C experiments. Based on

this single result there would seem to be no evidence to suggest that the presence of

such an aqueous phase affects the rate of butyl phosphinate formation. However,

additional results from subsequent experiments under both extract and strip conditions,

presented in Section 3.2, suggest otherwise.

3.1.4 Effect of different TBP sources

Of the four 20-5-70C experiments performed using Cyanex 272 under anhydrous

conditions, two were performed using Murrin TBP (experiments 6 and 7) and two were

performed using TBP sourced from Aldrich (experiments 4 and 5). The results

overlapped, confirming TBP and not an impurity specific to the Murrin-supplied reagent

is associated with the reaction.

3.1.5 Effect of a long chained alkyl phosphate

It was possible that phosphinic acid could react with trialkylphosphate species other

than TBP. If so, the reaction of TEHP (instead of TBP) with phosphinic acid might be

expected to result in an analogous 2-ethylhexyl phosphinate species. This reaction

(experiment 13) was undertaken using the same molar amount of TEHP (0.18 M) as was

11

used in the TBP-containing 20-5-70C experiments. It resulted in the generation of a

small, phosphorus-containing peak growing with time in the GC trace at a retention time

comparable to or slightly higher than that seen for the butyl phosphinate species. GC-

MS analysis and inspection of the resulting chemical ionisation mass spectrum revealed

prominent ions at m/z 468 and 396, inconsistent with the molecular weight of 2-

ethylhexyl phosphinate (402.5). As hydrolysis of TEHP to di-2-ethylhexyl phosphoric

acid (DEHPA) was possible, a reference sample of DEHPA (molecular weight 322.42)

was derivatised and analysed. The resulting product was found to possess the same

retention time and mass spectrum as the species generated in the present work. The ion

at m/z 468 is assignable to doubly-derivatised DEHPA, whilst the ion at m/z 396

corresponded to the loss of one trimethylsilyl group. The product generated in the

present work is therefore assigned as DEHPA arising from hydrolysis of TEHP. The

inability of TEHP to react with phosphinic acid like TBP does is assumed to be

attributable to its increased steric bulk. This is discussed further in Section 3.3.

3.1.6 Effect of dibutyl phosphate

Hydrolysis of TBP generates butanol and dibutyl phosphate. Although butanol does

not react with phosphinic acid (Section 3.2), another (very unlikely) possibility is

reaction between phosphinic acid and dibutyl phosphate. Two additional anhydrous

experiments were therefore undertaken by combining 20 % v/v Cyanex 272 (0.54 M)

with 7.9 % v/v dibutyl phosphate (0.36 M) and shaking for 12 days at 70°C. No butyl

phosphinate was formed.

12

3.1.7 Effect of temperature on the rate of butyl phosphinate generation

The effect of temperature (40-75°C) on the rate of butyl phosphinate formation was

assessed using an organic solution containing 20 % Cyanex 272 (0.54 M) and 5.0 %

TBP (0.18 M, experiments 4-9 and 14-23). The quantity of butyl phosphinate formed

throughout the trial at the different temperatures is summarised in Table 4, along with

the resulting annualised rate of formation (in g/L and mM) and calculated rate

coefficient, k. As expected, increased butyl phosphinate formation was observed as the

operating temperature increased.

The activation energy (Ea) for butyl phosphinate formation under anhydrous

conditions can be determined from the available data by using the Arrhenius equation:

k = A exp(-Ea/RT) (3)

where k is the rate coefficient, R is the gas constant (8.314 J K-1 mol-1), T is

temperature (in Kelvin) and A, the pre-exponential factor, is the rate constant at infinite

temperature. Determination of Ea is most readily achieved by taking the logarithm of

each side of this equation which then becomes

ln(k) = ln (A) – (Ea/RT). (4)

This format is equivalent to a straight line equation y = mx +c, where x is (1/T) and

m (the gradient) equals - Ea/R. The Arrhenius plot based on the available data gave a

very good linear fit (Figure 5). From this plot, Ea was calculated to be 122 kJ/mol and

"A” is 1.57x1018 per annum. This can be used to readily assess the k value at other

operating temperatures under anhydrous conditions.

13

3.2 Series 2 experiments: Effect of aqueous solutions and metal loading

The quantity of butyl phosphinate formed throughout the trial in the different systems

at 70°C is summarised in Table 5 along with the resulting annualised values (expressed

as both g/L and mM) and resulting rate coefficient, k.

Butyl phosphinate was not formed in any of the systems containing butanol

(experiments 24-29), confirming this alcohol is not associated with butyl phosphinate

formation.

The rate coefficients derived from the two anhydrous tests (experiments 30, 31) were

found to be comparable with the result of 0.43 ± 0.02 M-1 per annum (95 % confidence

interval) obtained previously under anhydrous conditions (Section 3.1.1), providing

confidence in the ability to compare the two series of results. Relative to the anhydrous

system, the presence of metal-free strip aqueous solution (20 g/L sulfuric acid) saw the

rate coefficient decrease to 0.33-0.35 M-1 per annum (experiments 32, 33), suggesting

dilute aqueous acid or water interferes with the reaction. The same rate coefficient was

obtained for tests done under strip conditions in the presence of 10 g/L each of Co, Ni

and Zn (experiments 34, 35), suggesting the presence of these uncomplexed metals play

no role in butyl phosphinate formation. Combined, these four results were found to

indicate a rate coefficient of 0.34 ± 0.02 M-1 per annum (95 % confidence interval) for

butyl phosphinate formation under strip conditions at 20 g/L sulfuric acid concentration.

In contrast, the single experiment performed previously in the presence of 100 g/L

sulfuric acid solution gave a higher result of 0.39 M-1 per annum (Section 3.1.3),

suggesting increased acidity leads to increased butyl phosphinate formation.

Confirmation of this possible trend would be of interest.

14

Tests done under metal-free extract conditions (experiments 36, 37) and under

extract conditions in the presence of 10 g/L Ni (experiments 38, 39) saw comparable

results obtained, with rate coefficients in the range 0.32-0.37 M-1 per annum.

Combined, these four results were found to indicate a rate coefficient of 0.34 ± 0.04 M-1

per annum (95 % confidence interval) for butyl phosphinate formation under these

extract conditions at pH 5.0. This result is equivalent to that observed under strip

conditions using 20 g/L sulfuric acid, suggesting the rate of butyl phosphinate formation

is independent of acidity in that range. This implies the reduced reaction rate seen under

extract and strip conditions relative to the anhydrous system is due to the presence of the

aqueous phase.

The two Murrin SX circuits are used to sequentially extract zinc and cobalt from the

nickel rich, sulfate-based feed solution. Complexation of cobalt with phosphinic acid at

low (0.12 M) and high (0.28 M) concentrations (experiments 40-43) saw the rate of

butyl phosphinate formation increase from the metal-free rate of 0.34 M-1 per annum to

0.37 and 0.46 M-1 per annum, respectively. A plot of these results (Figure 6) suggests a

possible relationship exists between cobalt loading and butyl phosphinate rate

coefficient, although additional data is ideally required to confirm this. The rate

coefficient changing as a function of cobalt loading could be associated with the actual

amount of cobalt loaded, changes in complex structure / stoichiometry, or the ratio of

loaded cobalt to total phosphinic acid present in the organic phase. Further investigation

would also be required to clarify this matter.

Like cobalt, complexation of zinc with phosphinic acid at low (0.15 M)

concentrations (experiments 44, 45) saw no change in the butyl phosphinate rate

coefficient relative to the metal-free aqueous system (0.33 M-1 per annum), but

15

increasing loading to 0.30 M (experiments 46, 47) saw a marked increase to 0.85 M-1

per annum.

It should be noted that all of these tests were performed under continual agitation

conditions and therefore are likely to be more representative of the rates expected under

mixing conditions. It is possible that the rate observed under non-mixing conditions

such as those prevailing in the settling stage of a commercial operation differ, with the

rates for stripped or partially loaded organic being more aligned to the results obtained

here under anhydrous conditions. This possibility ideally requires further investigation.

3.3 Mechanism of butyl phosphinate formation

The current results preclude butyl phosphinate formation via reaction of phosphinic

acid with any species arising from TBP hydrolysis. Instead, they demonstrate it is

formed via direct reaction between phosphinic acid and TBP, with the rate of formation

being first order relative to both TBP and phosphinic acid concentration and thus second

order overall.

Hydrolysis of alkyl phosphates proceeds through nucleophilic substitution arising

from attack of the P-O-C linkage by a nucleophile which can occur via two distinct

processes (Bel’skii, 1977 and references therein; Kocieński, 2005). The type 1 process

sees attack occur at the phosphorus atom leading to P-O bond cleavage, whereas the

type 2 processes progress via attack of the carbon substituent, leading to C-O bond

cleavage. Type 1 processes (P-O cleavage) can occur via three mechanisms, namely

addition-elimination, elimination-addition (e.g. hydrolysis of trialkyl phosphates under

basic conditions), and direct substitution. In all cases where such a reaction with a

trialkyl phosphate is successful, the nucleophile ends up being attached to the dialkyl

16

phosphate with ejection of an alkoxy group. Type 2 processes (O-C cleavage) can

progress via two mechanisms: a bimolecular nucleophilic substitution (or SN2) reaction,

and a unimolecular nucleophilic substitution (or SN1) reaction. In both cases, the

phosphate acts as the leaving group.

Given a pKa of 5.22 (Zhang et al, 1995), some dissociation of phosphinic acid in the

organic matrix would be expected. The generation of butyl phosphinate even under

anhydrous conditions suggests phosphinic acid or, more likely, its anion can act as a

nucleophile. As such, reaction via any of the Type 1 processes to yield butyl

phosphinate can be discounted as it would instead result in the generation of a molecule

containing a P-O-P linkage. Such P-O-P-containing species could not form from either

of the Type 2 processes where phosphate acts as the leaving group. Instead, the alkyl

group excised from the trialkyl phosphate would end up attached to the attacking

nucleophile, in this case yielding butyl phosphinate. The second order nature of butyl

phosphinate formation as determined here is consistent with this reaction being assigned

as an SN2 reaction involving C-O cleavage (Figure 7).

Increasing the steric bulk of the substrate (alkyl phosphate in this case) results in

inhibition of SN2 reactions. The outcome obtained when TBP was replaced by the more

sterically bulky tris-2-ethylhexylphosphate is consistent with this expectation.

3.4 Implications for the Murrin site

It was previously determined that butyl phosphinate was present in the Murrin plant

organic solutions at 32-35 g/L in mid 2009 (Barnard, 2010). It was of interest to

estimate the annual amount of butyl phosphinate formed and cumulative concentration

present based on the data from the present work. In order to do this, Murrin kindly

17

provided access to available relevant operating data over the past several years. The

results were then extrapolated to cover the period for which data was not readily

available. As such these calculations are only approximate.

Based on the present work, the key parameters affecting butyl phosphinate generation

are:

• Phosphinic acid (Cyanex 272) concentration.

• TBP concentration.

• Organic temperature.

• Aqueous strip acidity.

• Metal (Co, Zn) loading levels.

Based on the five years of analysed data, the following was concluded:

1. The average Cyanex 272 and TBP concentrations in the Murrin SX circuits

were about 21 % (0.54 M phosphinic acid) and 11 % (0.41 M), respectively.

2. The average organic temperature was 60°C for all of mixing, settling and

storage stages. The relatively short residence times in the storage tanks

between cycles meant the organic would not materially decrease in

temperature during that time.

3. Average zinc and cobalt loading levels and strip acidity in the relevant stages

and average residence time in the various stages of the Murrin circuit were

assessed. Rates during mixing and settling were assumed to differ as

discussed above. Specific parameters are not disclosed for confidentiality

reasons.

These conclusions were assumed to be valid over the entire life of the Murrin

operation. On the basis of points 2 and 3 above, the average rate coefficient was

18

estimated to be 0.12 M-1 per annum and thus the annual rate of butyl phosphinate

formation calculated to be about 27 mM (9.5 g/L) per annum. Based on a 95 m3 organic

inventory, this equates to 2600 mole of butyl phosphinate being generated annually at a

reagent cost of approximately A$77,000 based on A$65/L for Cyanex 272 (965 L

consumed) and A$20/kg for TBP (690 L consumed). At the time of analysis in 2009

(Barnard, 2010), there was approximately A$250,000 of Cyanex 272 and TBP

associated with butyl phosphinate in the Murrin circuits.

Given the known level of butyl phosphinate in the circuit in 2009, it was of interest to

assess whether the calculated annual rate of formation would reasonably allow the

observed circuit concentration to be reached. Such a calculation requires the ongoing

rate of loss of butyl phosphinate from the circuit to also be known.

As the amount of Cyanex 272 added to the circuit per annum to compensate for

ongoing losses was known, it was originally intended to assume the percentage of butyl

phosphinate lost per annum matched that for Cyanex 272. However, two species are

readily detected in Cyanex 272 by GC analysis, namely the phosphinic acid extractant

and the phosphine oxide impurity. Although not reported, quantitative analysis of

Cyanex 272 phosphinic acid and the phosphine oxide impurity was undertaken at the

same time the butyl phosphinate in the Murrin circuit was quantified (Barnard, 2010). It

was noted that whereas the phosphinic acid was present at levels equating to about 20 %

v/v Cyanex 272, the phosphine oxide impurity was present at levels equating to about

28 % v/v Cyanex 272. That is, the more chemically inert nature of the phosphine oxide

impurity appears to slightly decrease its rate of loss from the Murrin circuit organic

relative to the phosphinic acid. (Although beyond the scope of the present analysis, it is

worth noting that any gradual increase of this phosphine oxide species with time could

19

eventually have operating consequences in regard to increased acid entrainment, e.g.

Sarangi et al., 2006; Haghshenas et al., 2009). A salient example of this is the inability

of the phosphine oxide to form butyl phosphinate, unlike phosphinic acid. It was

therefore concluded that ongoing loss of butyl phosphinate from the Murrin circuit

would be best assessed relative to that determined for the phosphine oxide impurity in

Cyanex 272.

Based on the known average annual loss of phosphinic acid being about 40 % and the

ratio of phosphine oxide to phosphinic acid in the circuit organic being about 1.4 times

the ratio present in fresh Cyanex 272, it was estimated that phosphine oxide was being

lost from the circuit at about 28 % per annum. Combining the calculated rate of

formation with this estimated percent annual loss (both calculated using a monthly

interval, i.e. 0.79 g/L formed per month and 2.3 % lost per month) commencing from

June 1999 gave rise to an estimated value of 31 g/L butyl phosphinate in June 2009

(Table 6), comparable to the previously measured value of 32-35 g/L (Barnard, 2010).

This result suggests the developed model can reasonably explain the build-up of butyl

phosphinate in the Murrin plant. Unfortunately, no additional (historic) Murrin plant

samples were available for analysis to validate the present model.

This model is sensitive to operating temperature. Increasing the average operating

temperature from 60°C to 62°C increases the rate of formation to 12.4 g/L per annum,

resulting in a calculated accrued butyl phosphinate level of 41 g/L by June 2009.

Conversely, decreasing the operating temperature to 58°C decreases the rate of

formation to 7.3 g/L per annum and calculated accrued level of 24 g/L.

It should be noted that this analysis using the phosphine oxide species as a reference

material was possible due to only Cyanex 272 being added to the circuit over the

20

relevant timeframe. Substitution to an alternative reagent with a different phosphine

oxide concentration such as Ionquest 290 (phosphine oxide impurity < 5 % cf. about

15 % in Cyanex 272) during the timeframe examined would have made analysis based

on phosphine oxide concentration problematic.

4. Summary and conclusion

The impurity species butyl bis(2,4,4-trimethylpentyl)phosphinate (‘butyl

phosphinate’) has recently been identified in a commercial SX plant using Cyanex 272

and TBP. The present work has established that this species is formed via direct

reaction between phosphinic acid (the extractant in Cyanex 272) and TBP, with the rate

of formation being first order relative to both TBP and phosphinic acid concentration.

The observations are consistent with nucleophilic attack of TBP by the phosphinic acid

anion resulting in cleavage of the C-O bond and ejection of dibutyl phosphate anion via

a regular SN2 mechanism. Phosphinic acid was not observed to react with the more

sterically bulky phosphate species tris-2-ethylhexyl phosphate, an outcome consistent

with the proposed mechanism.

In the absence of an aqueous phase, the rate coefficient was found to be 0.43 ± 0.02

M-1 per annum (95 % confidence interval) at 70°C with an activation energy of

122 kJ/mol. The reaction occurring in the absence of an aqueous phase indicates that

degradation has the potential to occur throughout all stages of a process, i.e. mixing,

settling, storage. Agitation in the presence of an aqueous solution between pH 5.0 and

20 g/L sulfuric acid reduced the reaction rate relative to the anhydrous system by about

21

20 % to 0.34 ± 0.02 M-1 per annum (95 % confidence interval) at 70°C, increasing to

0.39 M-1 per annum with 100 g/L sulfuric acid.

Complexation of 0.28 M cobalt saw an increase in the rate of formation to 0.46 M-1

per annum. Complexation of 0.15 M Zn had no effect on the rate, although increasing

loading to 0.30 M saw the rate increase markedly to 0.86 M-1 per annum. That the rate

coefficient appears to change as a function of metal loading could be associated with the

amount of metal loaded, complex structure / stoichiometry, or the ratio of loaded metal

to total phosphinic acid present in the organic phase. Further investigation would be

required to clarify this matter.

Based on the current results and historical SX plant operating information, the rate of

butyl phosphinate formation at the Murrin plant was calculated to be about 27 mM (9.5

g/L) per annum, equating to 2600 mole generated annually. Along with the estimated

annual rates of loss, this data was used to estimate butyl phosphinate build-up in the

Murrin circuit over the last decade. The estimated value of 31 g/L butyl phosphinate by

June 2009 is comparable to the measured 32-35 g/L.

Acknowledgements

Mr. John O’Callaghan and Dr. Sian Miller of Minara Resources are thanked for

providing information pertaining to the operation of the Murrin Murrin solvent

extraction circuits. The support of the Minerals Down Under National Flagship and

Parker CRC for Integrated Hydrometallurgy Solutions (established and supported under

the Australian Government’s Cooperative Research Centres Program) are gratefully

acknowledged.

22

References

Anon., 2005. 1-butanol MSDS as found on International Programme of Chemical

Safety, http://www.inchem.org/documents/icsc/icsc/eics0111.htm, accessed 30 Sept.,

2010.

Barnard, K.R., 2010. Identification and characterisation of a Cyanex 272 degradation

product formed in the Murrin Murrin solvent extraction circuit, Hydrometallurgy, 103,

190-195.

Bel’skii, V.E., 1977. Kinetics of the hydrolysis of phosphate esters, Russ. Chem.

Rev., 46, 828-841.

Flett, D.S., 2004. Cobalt-Nickel separation in hydrometallurgy: A review, Chemistry

for Sustainable Development, 12, 81-91.

Haghshenas, D.F., Darvishi, D., Rafieipour, H., Alamdari, E.K., Salardini, A.A.,

2009. A comparison between TEHA and Cyanex 923 on the separation and the

recovery of sulfuric acid from aqueous solutions, Hydrometallurgy, 97, 173-179.

Kocieński, P.J., 2005. “Protecting Groups, 3rd Edition”. Druckhaus Thomas

Muntzer, Bad Langensalza, Germany, 454-457.

Sarangi, K., Padhan, E., Sarma, P.V.R.B., Park, K.H., Das, R.P., 2006.

Removal/recovery of hydrochloric acid using Alamine 336, Aliquat 336, TBP and

Cyanex 923, Hydrometallurgy, 84, 125-129.

Tahraoui, A., Morris, J.H., 1995. Decomposition of solvent-extraction media during

nuclear reprocessing - literature review, Sep. Sci. Technol., 30, 2603-2630.

Zhang, P., Inoue, K., 1995. Recovery of metal values from spent

hydrodesulfurization catalysts by liquid-liquid extraction, Energy and Fuels, 9, 231-239.

23

Table 1 Summary of conditions used in the predominantly aqueous-free experiments.

Table 2 Summary of conditions used in the primarily aqueous-containing experiments.

Table 3 Concentration (g/L) of butyl phosphinate observed with time in the 70°C systems

along with annualised rate of formation and calculated rate coefficient.

Table 4 Concentration (g/L) of butyl phosphinate observed with time in the 20 % Cyanex

272 / 5.0 % TBP systems at different temperatures along with the annualised rate of formation

and calculated rate coefficient.

Table 5 Concentration (g/L) of butyl phosphinate observed with time in the various

systems at 70°C along with annualised formation rate and calculated rate coefficient.

Table 6 Estimated butyl phosphinate concentration (g/L) in the Murrin circuit over time.

24

Table 1 Summary of conditions used in the predominantly aqueous-free experiments.

Expt. Experiment Temp. NotesNo. Code* (M) % (v/v) (M) % (v/v) (°C)1 2-5-70C 0.054 2 0.18 5.0 702 5-5-70C 0.14 5 0.18 5.0 703 10-5-70C 0.27 10 0.18 5.0 70

4, 5 20-5-70C 0.54 20 0.18 5.0 706, 7 20-5-70C 0.54 20 0.18 5.0 70 Murrin TBP used

8 20-5-70C-Acid 0.54 20 0.18 5.0 70 Aqueous present9 20-5-70C-Pure 0.54 20 0.18 5.0 70 Pure phos. acid

10 20-10-70C 0.54 20 0.36 10.0 7011 20-2.5-70C 0.54 20 0.091 2.5 7012 20-1.0-70C 0.54 20 0.036 1.0 7013 20-9-70C-TEHP 0.54 20 0.18 5.0 70 TEHP used

14, 15 20-5-60C 0.54 20 0.18 5.0 6016, 17 20-5-50C 0.54 20 0.18 5.0 5018, 19 20-5-40C 0.54 20 0.18 5.0 4020, 21 20-5-55C 0.54 20 0.18 5.0 5522, 23 20-5-75C 0.54 20 0.18 5.0 75

TBPPhosphinic acid

* Code is Cyanex 272 concentration (% v/v)-TBP concentration (% v/v)-Temperature (°C)-unique identifier, e.g. 'acid' is sample mixed with sulfuric acid, 'pure' used pure phosphinic acid and 'TEHP' used tris-2-ethyl hexyl phosphate instead of TBP. For instance 20-5-70C refers to an experiment using 20 % Cyanex 272, 5 % TBP at 70°C.

25

Table 2 Summary of conditions used in the primarily aqueous-containing experiments.

Expt. Organic* Condition Aqueous phase No.

24, 25 B Anhydrous nil26, 27 B Strip 20 g/L H2SO4

28, 29 B Extract pH 5.0, 10 g/L Na2SO4

30, 31 T Anhydrous nil32, 33 T Strip 20 g/L H2SO4

34, 35 T Strip 20 g/L H2SO4 plus 10 g/L each Co, Ni, Zn36, 37 T Extract pH 5.0, 10 g/L Na2SO4

38, 39 T Extract pH 5.0, 10 g/L Na2SO4 plus 10 g/L Ni40, 41 T Extract pH 5.0, 10 g/L Na2SO4 plus 8.8 g/L Co42, 43 T Extract pH 5.0, 10 g/L Na2SO4 plus 17.7 g/L Co44, 45 T Extract pH 5.0, 10 g/L Na2SO4 plus 9.8 g/L Zn46, 47 T Extract pH 5.0, 10 g/L Na2SO4 plus 19.5 g/L Zn

* Contains 30% v/v Cyanex 272 and either 10% v/v TBP (T) or 20% v/v n-butanol (B).

26

Table 3 Concentration (g/L) of butyl phosphinate observed with time in the 70°C systems

along with annualised rate of formation and calculated rate coefficient.

Expt. Experiment Rate coeff.No. Code 10 20 34 45 55 67 (g/L p.a) (mM p.a.) (k, M-1 p.a.)1 2-5-70C 0.06 0.10 0.16 0.17 0.21 0.26 1.5 4.2 0.432 5-5-70C 0.11 0.22 0.44 0.50 0.59 0.70 4.0 11.5 0.473 10-5-70C 0.27 0.50 0.87 1.10 1.25 1.46 8.4 24.3 0.504 20-5-70C (a) 0.47 0.87 1.51 2.01 2.15 2.67 15.0 43.4 0.445 20-5-70C (b) 0.36 0.76 1.34 1.71 2.02 2.43 13.6 39.1 0.406 20-5-70C (c) 0.37 0.76 1.43 1.80 2.04 2.38 13.7 39.5 0.407 20-5-70C (d) 0.50 1.00 1.72 2.10 2.50 2.80 16.4 47.3 0.488 20-5-70C-Acid 0.38 0.77 1.34 1.69 1.98 2.35 13.3 38.4 0.399 20-5-70C-Pure 0.58 1.04 1.42 1.73 2.11 2.45 14.1 40.8 0.42

10 20-10-70C 0.78 1.47 2.52 3.28 3.92 5.18 27.2 78.4 0.4011 20-2.5-70C 0.20 0.43 0.76 0.97 1.17 1.40 7.8 22.4 0.4612 20-1.0-70C 0.06 0.15 0.26 0.35 0.41 0.52 2.8 8.1 0.42

Formation rateDay

27

Table 4 Concentration (g/L) of butyl phosphinate observed with time in the 20 % Cyanex

272 / 5.0 % TBP systems at different temperatures along with the annualised rate of formation

and calculated rate coefficient.

Expt. Experiment Rate coeff.No. Code 10 20 34 45 55 112 (g/L p.a) (mM p.a.) (k, M-1 p.a.)14 20-5-60C (a) 0.10 0.17 0.36 0.51 0.69 1.36 4.4 12.6 0.1315 20-5-60C (b) 0.07 0.18 0.36 0.50 0.66 1.41 4.4 12.8 0.1316 20-5-50C (a) 0.02 0.05 0.08 0.10 0.13 0.21 0.75 2.2 0.0217 20-5-50C (b) 0.03 0.05 0.09 0.13 0.15 0.23 0.84 2.4 0.0218 20-5-40C (a) 0.02 0.02 0.04 0.04 0.05 0.08 0.29 0.8 0.0119 20-5-40C (b) 0.02 0.03 0.04 0.05 0.05 0.09 0.33 0.9 0.01

7 16 25 31 4920 20-5-55C (a) 0.05 0.11 0.17 0.21 0.23 2.0 5.9 0.0621 20-5-55C (b) 0.08 0.12 0.16 0.20 0.21 2.0 5.7 0.0622 20-5-75C (a) 0.42 1.25 2.16 2.93 32.4 93.7 0.9523 20-5-75C (b) 0.43 1.17 2.15 2.97 32.4 93.6 0.95

Formation rateDay

28

Table 5 Concentration (g/L) of butyl phosphinate observed with time in the various

systems at 70°C along with annualised formation rate and calculated rate coefficient.

Expt Rate coeff.No. 7 16 25 31 g/L p.a. mM p.a. k, M-1 p.a.24 0.0 0.0 0.0 0.0 0.0 0.00 0.0025 0.0 0.0 0.0 0.0 0.0 0.00 0.0026 0.0 0.0 0.0 0.0 0.0 0.00 0.0027 0.0 0.0 0.0 0.0 0.0 0.00 0.0028 0.0 0.0 0.0 0.0 0.0 0.00 0.0029 0.0 0.0 0.0 0.0 0.0 0.00 0.0030 0.8 1.9 3.1 3.8 45.2 0.13 0.4431 0.9 2.1 3.2 4.0 47.1 0.14 0.4632 0.7 1.7 2.4 3.1 36.2 0.10 0.3533 0.6 1.5 2.3 2.8 33.3 0.10 0.3334 0.7 1.6 2.4 2.9 34.9 0.10 0.3435 0.7 1.7 2.4 2.9 35.2 0.10 0.3436 0.6 1.4 2.3 2.8 33.3 0.10 0.3337 0.7 1.3 2.3 2.8 32.7 0.09 0.3238 0.7 1.6 2.4 2.9 35.4 0.10 0.3539 0.8 1.7 2.6 3.2 37.8 0.11 0.3740 0.8 1.7 2.6 3.2 38.1 0.11 0.3741 0.8 1.8 2.6 3.2 38.4 0.11 0.3842 0.9 2.2 3.1 4.0 46.7 0.13 0.4643 0.9 2.0 3.2 4.0 47.1 0.14 0.4644 0.7 1.5 2.4 3.0 35.3 0.10 0.3545 0.6 1.4 2.2 2.6 31.0 0.09 0.3046 1.5 3.6 5.7 6.7 81.2 0.23 0.8047 1.9 4.5 6.3 7.7 93.1 0.27 0.91

Formation rateDay

29

Table 6 Estimated butyl phosphinate concentration (g/L) in the Murrin circuit over time.

k rate formed(M-1) (g/L p.a.) '00 '01 '02 '03 '04 '05 '06 '07 '08 '090.124 9.5 8 14 19 23 25 27 29 30 31 31

Year

30

Figure 1 Structure of butyl bis(2,4,4-trimethylpentyl)phosphinate.

Figure 2 GC traces highlighting the growth of the butyl phosphinate peak at 4.83 min with

time in a 20-5-70C system.

Figure 3 Butyl phosphinate generation observed over time at different phosphinic acid

concentrations (2-20 % v/v Cyanex 272; 0.054-0.54 M). TBP concentration was fixed at 5.0 %

v/v (0.18 M).

Figure 4 Butyl phosphinate generation per annum with changing phosphinic acid

concentration at a fixed TBP concentration of 0.18 M (5.0 % v/v, squares), and changing TBP

concentration at a fixed phosphinic acid concentration (0.54 M, 20 % v/v Cyanex 272,

triangles).

Figure 5 Arrhenius plot describing the relationship between the rate coefficient, k, and

temperature.

Figure 6 Plot of butyl phosphinate rate coefficient versus cobalt loading.

Figure 7 Butyl phosphinate formation via the proposed SN2 reaction mechanism.

31

P

O

(CH3)3CCH2CH(CH3)CH2

(CH3)3CCH2CH(CH3)CH2

O(CH2)3CH3

Figure 1 Structure of butyl bis(2,4,4-trimethylpentyl)phosphinate.

32

3 4 5 Minutes0

10

20

30

40

mVolts

Octadecane4.05 min Butyl phosphinate

4.83 min

Day 45

Day 0

Day 67

Phosphinic acid4.42 min

Phosphine oxide5.55 min

TBP3.48 min

Day 20

Diluent region

Figure 2 GC traces highlighting the growth of the butyl phosphinate peak at 4.83 min with

time in a 20-5-70C system.

33

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 10 20 30 40 50 60 70Time (days)

But

yl p

hosp

hina

te (g

/L) .

2-5-70C5-5-70C10-5-70C20-5-70C (a)20-5-70C (b)20-5-70C (c)20-5-70C (d)20-5-70C-Acid20-5-70C-Pure

Figure 3 Butyl phosphinate generation observed over time at different phosphinic acid

concentrations (2-20 % v/v Cyanex 272; 0.054-0.54 M). TBP concentration was fixed at 5.0 %

v/v (0.18 M).

34

y = 26.8xR2 = 0.97

y = 77.4xR2 = 0.98

0

3

6

9

12

15

18

21

24

27

30

0.00 0.10 0.20 0.30 0.40 0.50Reagent concentration (M)

But

yl p

hosp

hina

te (g

/L p

.a.)

.

TBPphosphinic acid

Figure 4 Butyl phosphinate generation per annum with changing phosphinic acid

concentration at a fixed TBP concentration of 0.18 M (5.0 % v/v, squares), and changing TBP

concentration at a fixed phosphinic acid concentration (0.54 M, 20 % v/v Cyanex 272,

triangles).

35

y = -14653x + 41.90

R2 = 0.991

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

0.0028 0.0029 0.0030 0.0031 0.00321/T (K-1)

ln(k

)

Ea = = 122 kJ/molA = 1.57 x 1018 p.a.

Figure 5 Arrhenius plot describing the relationship between the rate coefficient, k, and

temperature.

36

0.33

0.35

0.37

0.39

0.41

0.43

0.45

0.47

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28Loaded Cobalt (M)

Rat

e co

effic

ient

Figure 6 Plot of butyl phosphinate rate coefficient versus cobalt loading.

37

P O-

RR

O

SN2 POC4H9R

R

O-O

O

POBu

OBuOBuOBu

P

O

H

O

H

C3H7

C +

Figure 7 Butyl phosphinate formation via the proposed SN2 reaction mechanism.