C. M. Carter and M. D. Nunes

Research and Development, Alcoa of Australia Limited

P.O. Box 161, Kwinana, Western Australia 6167

M. V. Baker and E. L. Ghisalberti

Department of Chemistry, The University of Western Australia

Stirling Highway, Nedlands, Western Australia 6009


At Alcoa of Australia red mud is disposed of at residue lakes, and the entrained process liquor is diluted as part of the washing process. The diluted liquor is held in lakes that are exposed to solar radiation for some months before the liquor is returned to the Bayer circuit through the red mud and sand washers. It is known that this liquor has a different organic composition to the process liquor. Since it is likely that photochemistry is at least partly responsible for this difference, the role of photochemically initiated reactions on model Bayer liquor organics has been studied.

5-Hydroxyisophthalic acid and other hydroxy aromatic carboxylic acids which model some Bayer liquor organics were each irradiated in a solution of sodium hydroxide with a UV source and the degradation was monitored by UV-Visible spectroscopy. Over twenty of the degradation products of 5-hydroxyisophthalic acid were identified using GC-MS. The effect of the sparge gas on the formation of the degradation products oxalate, malonate and succinate, and the total organic carbon concentration was investigated. The effect of the hydroxyl and carboxylic acid functionalities on the reactivity of these aromatic acids was also determined.

On the basis of this work and work by other authors a mechanism has been proposed for the degradation of 5-hydroxyisophthalic acid in alkaline solution by solar radiation. This mechanism can be applied to the other hydroxy aromatic carboxylic acids investigated in this work.


C. M. Carter, M. D. Nunes, M. V. Baker and E. L. Ghisalberti


    In the Alcoa WA refineries red mud and sand is separated from pregnant Bayer liquor in clarification and is deposited in large residue lakes. The mud and sand is washed with a counter current stream of dilute liquor (referred to as lake water) from these lakes, and as a result a portion of the lake water enters the refinery. It is known that, after normalisation to the same total alkalinity (TA), lake water has a higher concentration of oxalate and a lower absorbance (liquor colour) and concentration of total organic carbon (TOC) than the plant liquor (Nunes et al, 1996; Nunes et al, 1997). It is believed that many of the organic species present in the residue lake liquor are degraded by UV radiation to lower molecular weight compounds during exposure to solar radiation. Since many of the smaller molecular weight organics are implicated in poorer gibbsite yield, quality and colour, a better understanding of the photochemistry of model Bayer compounds was sought.

    Bayer liquor is known to contain a number of hydroxybenzoic acids (Niemala, 1993). Given the propensity of phenols to undergo photo-oxidation and photodegradation it seemed worthwhile to investigate the photochemistry of these compounds. Phototransformation of phenols has been shown to involve phenoxyl radicals, and it is known that pulse radiolysis (Neta and Fessenden, 1974) and photolysis (Karakyriakos et al, 1998) of hydroxybenzoic acids in alkaline solutions produce radical species.

    In this report the effects of UV/Visible radiation on some hydroxybenzoic acids known to be present in Bayer liquor are described. The compounds chosen were 5-hydroxyisophthalic acid (5-HIA), 2-hydroxybenzoic acid (2-HBA), 3-hydroxybenzoic acid (3-HBA), 4-hydroxybenzoic acid (4-HBA) and 3,4-dihydroxybenzoic acid (3,4-DHB). For comparison purposes, isophthalic acid (IPA) and benzoic acid (BA) were also included in the study.

    Irradiation experiments were monitored by UV-Visible spectroscopy, and the concentrations of oxalate, malonate, succinate and TOC were measured as a function of irradiation time. GC-MS was used to identify other degradation products. The effect of varying the substitution pattern on the aromatic ring on the rates of formation of photodegradation products is discussed. A detailed mechanism for the formation of a number of products formed from 5-HIA is proposed.

  2. The compounds used were obtained from either Aldrich or Sigma. The test compounds (0.4gL-1) were dissolved in aqueous NaOH (total caustic (TC) 17gL-1) and sparged with oxygen before (20 min) and during irradiation. Each irradiation was carried out in a 6L reaction vessel (Figure 2-1) with a 700W medium pressure gallium iodide doped mercury arc lamp, which emits radiation of varying intensities between 230 and 580nm. A peristaltic pump was used to circulate the liquor through a 1cm quartz cell and UV spectra were collected (200nm to 800nm) every 120s by a HP 8453 UV-Visible diode array spectrophotometer.


    Figure 2-1

    Diagrammatic representation of the apparatus used for irradiation of hydroxy aromatic carboxylic acid solutions

    Aliquots were taken every 30 minutes and quantitatively analysed for oxalate, malonate and succinate using GC, and TOC using a Dohrmann DC80 Carbon Analyser. Selected samples were also analysed using gas chromatography - mass spectrometry (GC-MS) for identification of some of the other lower molecular weight organics.

  3. results AND discussion
    1. Absorbance Data
    2. The colourless solution of 5-HIA (0.4 gL-1) became deep red only a few minutes after irradiation commenced. During the first hour the colour intensified but after five hours the solution was again colourless (Figures 3-1 and 3-2). Since the radicals formed are relatively shortlived (Karakyriakos et al, 1998) they were not directly responsible for the colour. However it is likely that the highly coloured species was a quinone which are known to be produced by direct excitation of phenol derivatives in aerated aqueous solutions (Mazellier et al, 1998). Samples taken during the irradiation and kept in the dark gradually became orange/brown indicating that this coloured intermediate(s) was unstable.


      Figure 3-1

      The absorbance spectra of 5-HIA prior to irradiation, and after 0.5, 7.5 and 15 hours of irradiation

      Figure 3-2

      Kinetic trace of 5-hydroxyisophthalic acid solution. The absolute absorbance at 400nm, 450nm, 500nm, 550nm and 600nm for the solution is given as a function of time, irradiation commencing at 0.5 hours

      Solutions of all the test compounds, except for 3,4-DHB, were initially colourless, and absorbed in the visible region only when irradiated. Of the mono-hydroxybenzoic acids, 3-HBA gave the most intensely coloured solution (orange) and 4-HBA the least. This absorbance increased to a maximum, and then more slowly decreased until each solution was colourless.

      Initially the solution of 3,4-DHB was yellow, but turned brown after an hour of sparging with oxygen indicating its instability under oxygenated alkaline conditions. On irradiation, the solution gradually became paler until it was colourless.


Oxalate, Malonate, Succinate and TOC Results

      Aliquots taken during the irradiation of 5-HIA clearly showed that oxalate, malonate and succinate were produced (Figure 3-3). The final concentrations of these compounds represent a total of 18% of the initial carbon content of 5-HIA, with 12.3% being degraded to oxalate, 3.8% to malonate and 1.8% to succinate. Though these values are heavily dependent on irradiation time they do indicate that these compounds are significant degradation products.

      Figure 3-3

      Oxalate, malonate and succinate concentrations in the 0.4gL-1 5-HIA solution as a function of irradiation time

      Much of the original organic carbon (55.2% of TOC) was lost, probably as CO2, by the end of the irradiation (Figure 3-4) (see Section 3.7, Scheme 3-2). Thus, after irradiation of 5-HIA for 15 hours 73.1% of the original TOC can be accounted for as one of oxalate, malonate, succinate or carbon dioxide.

      Figure 3-4

      TOC concentration in the 0.4gL-1 5-HIA solution as a function of irradiation time

      3.3 The Effect of Solar Radiation on a Solution of 5-HIA

      The degradation products oxalate, malonate and succinate were formed, though more slowly, when a tray of 5-HIA in 17gL-1 NaOH was left exposed to solar radiation. After 15 days 8% had been degraded to oxalate, 2% to malonate and 2% to succinate, whilst 17% of the TOC had been destroyed (data not normalised for evaporation). The solution became red soon after exposure to the solar radiation.

      3.4 Importance of the Hydroxyl Functionality on the Oxalate, Malonate, Succinate and TOC Concentrations

      Figure 3-5

      Absorbance spectra of 0.04gL-1 2-HBA, 3-HBA, 4-HBA and 5-HIA in NaOH (TC 1.7gL-1) solution

      Solutions of the non-hydroxylated aromatics isophthalic acid (IPA) and benzoic acid (BA) were irradiated to determine if the hydroxyl functionality influences the degradation of 5-HIA. The rates of formation of oxalate, malonate and succinate were significantly slower, suggesting that the hydroxyl functionality activates the benzene ring toward photochemical degradation.

      Another aspect of the investigation was to determine whether the mono-hydroxybenzoic acid isomers exhibited varying reactivity given their different absorbance spectra (Figure 3-5). The rate of formation of oxalate and malonate varied between each isomer (Figure 3-6), indicating that the position of the hydroxyl substituent is important. However, succinate formation and TOC destruction were similar.

      Figure 3-6

      Oxalate concentration as a function of irradiation time for 2-, 3- and 4-hydroxybenzoic acid

      3,4-DHB underwent degradation in the oxygenated alkaline medium without irradiation, indicating a high level of reactivity under these conditions. Oxalate, malonate and succinate were formed in the 1 hours before irradiation.

      The initial TOC concentration in the 3,4-DHB solution was measured as 0.179gL-1, but was prepared at 0.218gL-1, which indicates that decarboxylation was taking place before the sample could be analysed. Based on the actual starting concentration, 42.7% of the original carbon had been lost by the end of the irradiation and 17.9% had been lost from the initial sample before analysis.

Effect of Sparge Gas on the Oxalate, Malonate, Succinate and TOC Concentrations

      The effect of sparge gas on the degradation of 5-HIA was investigated by irradiating three systems: the first was purged with 100% oxygen, the second with synthetic air (20% O2, 80% N2), and the last with 100% nitrogen. Whilst the sample sparged with pure oxygen became dark red soon after irradiation commenced, the sample purged with air, when irradiated, became orange and the nitrogen sparged solution became yellow. The latter solution was irradiated for over a week since the colour changes were slow.

      The results of the oxalate, malonate, succinate and TOC analyses showed a significant dependence on the sparge gas (Figure 3-7). It is apparent that in the absence of oxygen the oxalate is formed to a lesser extent, and at a vastly slower rate than in the presence of oxygen. However, though malonate is also formed more slowly in the absence of oxygen, eventually a concentration comparable with that of the oxygenated system was obtained.

      Figure 3-7

      The effect of changing the sparge gas on the production of a) oxalate and b) succinate in solutions of 0.4gL-1 5-HIA

      It appears that succinate accumulation is more favourable under deoxygenated conditions. The rate of succinate formation, though, is far slower in the absence of oxygen. The rate of TOC destruction, too, is heavily influenced by the presence of oxygen, with a much slower destruction rate observed for the nitrogenated system.

      These results suggest that the degradation pathway(s) that result in oxalate formation are heavily dependent on the presence of oxygen, whilst the formation of succinate is greatest in an oxygen deficient environment.

Elucidation of Degradation Products

A GC-MS analysis of the 5-hyroxyisophthalic acid solutions containing the irradiated species indicated the presence of many different hydroxy acids and polycarboxylic acids (selected compounds listed in Table 3-1). These compounds were identified primarily by interpretation of the mass spectra and comparison with entries in a database (Niemala, 1992). In many cases the relative elution time was considered when the identification was tentative. The degradation products identified for the other starting materials were very similar.

Table 3-1

Irradiation degradation products of 5-HIA under various conditions identified by GC-MS


Sparged with O2 Sparged with N2 Sparged with air

malonic acid (1)

succinic acid (2)

tartronic acid (3)


ketomalonic acid (4)


malic acid (5)

oxosuccinic acid (6)


ethenetricarboxylic acid (7)


1-hydroxy-1,1,2-ethanetricarboxylic acid (8)

2-hydroxy-1,1,2-ethanetricarboxylic acid (9)


The presence of a compound is signified by . This list is not exhaustive since many minor GC peaks were not identified. Each solution contained 0.4gL-1 5-HIA solution at a TC 17gL-1.

A quantitative analysis based on the known initial concentration was performed for 5-HIA (Figure 3-8). It is interesting to note that half of the 5-HIA was destroyed within half an hour, and none could be detected after 2 hours (Figure 3-8). The rate of destruction was first order, with k = 0.0291 min-1.

Solutions of the three isomeric hydroxy acids subjected to radiation were analysed using the same technique. They were destroyed at a slower rate than 5-HIA, each still within the GC-MS detection limits after 4 hours of irradiation. This technique showed that the order of reactivity for these acids is 2-HBA> 3-HBA > 4-HBA when irradiated with the full output of the lamp, as previously deduced on the basis of the oxalate data. The slower destruction of these acids compared with 5-HIA indicates that the extra carboxylic acid functionality leads to a faster photochemical degradation.

Figure 3-8

5-HIA concentration (ppm) as a function of irradiation time

3.7 Suggested Mechanisms of Degradation

Each of the hydroxylated benzoic acids form similar degradation products possibly reflecting a similar degradation process. The mechanism outlined is specific for 5-HIA but it is applicable to any of the hydroxybenzoic acids.

It has been shown that solutions of 5-HIA subjected to UV/Visible radiation (Karakyriakos et al, 1998) leads to the formation of the radicals shown in Figure 3-9.


5-hydroxyisophthalic dicarboxy phenoxyl dicarboxy-o-benzo- dicarboxy-p-benzo-
acid (5-HIA) radical (DPR) semiquinone (DOB) semiquinone (DPB)

Figure 3-9

Structure of 5-HIA and the radicals formed from its photo-oxidation

Karakyriakos found that there were no radicals observed in the absence of radiation, and that radical production was greatest at a wavelength of 330 5nm for DPB. Hence light with a wavelength of around 330nm is responsible for initiating the photo-oxidation of 5-HIA, and since sunlight contains radiation of this wavelength it is probable that such radicals can be generated at the residue lakes.

The formation of the dicarboxy-o-benzosemiquinone (DOB) species and the corresponding quinone from the primary radical (DPR) is illustrated in Scheme 3-1.

In the proposed mechanism for the formation of the radicals the initial step in the degradation of 5-HIA is unimolecular, which is consistent with the rate data obtained. The mechanism proposed involves a resonance stabilised positively charged species that then reacts with hydroxide to produce the other radicals observed (Scheme 3-1) (cf mechanism by Neta and Fessenden, 1974). Photo-oxidation of any of DOB, DPB and DOXB could result in quinone formation. This would account for the observed colour of the 5-HIA solution upon irradiation.

Scheme 3-1

Mechanism for the formation of DPR, DOB and the corresponding quinone

Oxidative cleavage between the two oxo groups on the quinone formed from DOB to form 1,3-butadiene-1,2,4-tricarboxylic acid would precede further oxidation at either one of the double bonds (Scheme 3-2). Oxidative cleavage and decarboxylation or addition of hydroxide accounts for the formation of ethenetricarboxylic acid (7), ketomalonic acid (4), both 1- and 2-hydroxy-1,1,2-ethanetricarboxylic acid (8 & 9), oxosuccinic acid (6), tartronic acid (3), malonic acid (1) and oxalic acid. Decarboxylation of either of the ethanetricarboxylic acids would then account for the formation of malic acid (5).


Scheme 3-2

Postulated photochemical degradation pathway for 5-HIA. For the sake of clarity, the carboxylic acid forms are shown rather than the carboxylate species. Line through a double bond indicates oxidative cleavage

Succinic acid (2) was observed as a degradation product in the oxidation of phenol by oxygen (Devlin and Harris, 1984) and is a product of a dark reaction. A probable reaction sequence for succinate involves a b -vinylogous decarboxylation of intermediate B (Scheme 3-3) (cf Gierer and Imsgard, 1977).

Scheme 3-3

Formation of succinate via b -vinylogous decarboxylation

  1. conclusions

The irradiation of organic species in a strongly alkaline medium had not been studied before. Irradiation of 5-hydroxyisophthalic acid and other hydroxybenzoic acids gave rise to highly coloured intermediates, and destruction of TOC and formation of oxalate, malonate and succinate were observed for each compound. These compounds were formed when 5-hydroxyisophthalic acid solution was exposed to sunlight, confirming that solar radiation is energetic enough for these reactions to occur.

Several experimental parameters were varied in the study of the degradation products formed by the irradiation of 5-HIA. A clear dependence on O2 was observed. The formation of succinate was shown to be more favourable in an oxygen deficient environment. However, rates of formation were greatest for oxalate, malonate and succinate in an oxygenated system.

The rates of degradation depended on the number of carboxylic acid and hydroxyl functionalities. The presence of hydroxyl groups activates the aromatic acid to photooxidation. An increase in the number of carboxylic acid groups also results in faster photochemical degradation.

Degradation of the different isomers of hydroxylated benzoic acid varied according to the position of the hydroxyl functionality. The order 2-HBA > 3-HBA > 4-HBA was found when the solutions were irradiated with the full spectrum of the lamp.

GC-MS analysis of the compounds arising from irradiation allowed the identification of several degradation products. Each degradation product was known to be a Bayer liquor organic. Thus the results of this study provide important information regarding likely processes taking place at the residue lakes.


The financial support of this work by the Department of Commerce and Trade in the form of a Neville Stanley Bursary and the Materials Institute of Western Australia is gratefully acknowledged, as is the support of Alcoa of Australia Limited and the University of Western Australia.


Devlin, H. R. and Harris, I. J. (1984). Ind. Eng. Chem. Fundam. pp.23, 390.

Gierer, J. and Imsgard, F. (1977). Acta Chem. Scand. B 31, 537-9.

Karakyriakos, E., Myers, L. P., Ghisalberti, E. L. and McKinley, A. J. (1998). Unpublished results.

Mazellier, P., Sarakha, M., Rossi, A. and Bolte, M. (1998). J. Photochem. Photobiol. A: Chem. pp.115, 117-21.

Neta, P. and Fessenden, R. W. (1974). J. Phys. Chem. pp 78, 523.

Niemala, K. (1992) The computer generated database for the analysis of low MW organic degradation products of humic materials.

Niemala, K. (1993) Determination of Carboxylic Acids in Bayer Liquor Part XVIII. A Revised, Annotated List of All Compounds Identified or Partially Characterised in Various Bayer Process Samples. Alcoa Internal Report.

Nunes, M. D., Shaw, M. and Dye, S. T. (1996). Unpublished experimental results from a comparison of lake water and refinery liquor samples.

Nunes, M. D., Binet, D. J. and Grocott, S. C. (1997). A Comparison of Lake Water and Refinery Liquor Impurity Levels. Evidence for a Lake TOC Effect. Alcoa Internal Report.