METHOD FOR QUANTITATIVE DETERMINATION OF INDIVIDUAL

POLYSULPHIDE SPECIES IN OXIDIZED WHITE LIQUORS BY

MEANS OF RAMAN SPECTROSCOPY

TECHNICAL DISCLOSURE

The present invention relates to an on-line method of monitoring sodium polysulphide production in oxidized white liquor by means of Raman spectrometry and, optionally, the control of any process which generates sodium polysulphide by the oxidation of white liquor.

DESCRIPTION OF PRIOR ART

The advantages of polysulphide pulping have been documented extensively over the last forty years [1-7]. When polysulphide is added to a conventional kraft process liquor, this pulping method represents an effective way for increasing pulp yield by 1 to 4 percent through the inhibition of alkaline peeling reactions of hemicelluloses, or for releasing less lignin at a given yield into the liquor system. In addition, liquor throughput is improved for mills with digester and/or recovery bottlenecks.

Currently available polysulphide generating technologies produce oxidised white liquors by partially oxidising white liquor with air or oxygen in the presence of a catalyst such as wet-proofed activated carbon (MOXY ^® ) [6] or manganese dioxide in the presence of lime mud (Paprilox ^® ) [3-4]. The latter process produces polysulphide liquor during recausticizing [3-4] and requires less capital expenditure than competing processes [5-7]. However, a by-product of both processes is sodium thiosulphate, a dead-load chemical which accumulates during the recovery process and which can significantly increase the corrosion rate of reactor surfaces exposed to the process liquor, even though it plays no role during pulping. Therefore, one must monitor the selectivity of the process by keeping the formation of thiosulphate and its precursors to a minimum, and thereby ensure that an adequate level of residual sodium sulphide is present in the liquor so as to maintain an adequate pulping rate in the digester. Thus, the development of a suitable on-line analytical method would ensure that mills using a polysulphide generating process get maximal benefit from its use.

In the past, a number of laboratory analytical methods have been used for the analysis of sulphide, polysulphide and other oxidised sulphur anions in oxidized white

liquors [8-10]. Generally, these methods are laborious, time-consuming and are not suitable for rapid at-line or slipstream analysis.

UV-visible spectrometry has been proposed as an alternative by several investigators [11-16]. Recently, Van Heek et al. [16] showed that the pulp yield increase obtained with polysulphide is proportional to the ratio of polysulphide measured either at 285 nm or 416 nm to the amount determined by the gravimetric method described in [8]. This ratio is termed the "UV/gravimetric ratio". The difference between the gravimetric amount and that determined by UV- Visible spectroscopy did not appear to correspond with any pulp yield increase. Hence the name, "inactive polysulphide" (p. 10 of [16]). Heat treatment of the liquor, on the other hand, produced an increase in the UV/gravimetric ratio which translated into a higher pulp yield than that obtained with a non-heat treated liquor [16].

Following recent kinetic evidence provided by Licht and Davis [13] and earlier by Giggenbach [17], the inactive polysulphide is believed to be a thiosulphate precursor with the formula S ₄ O ^2' . These authors proposed that pentasulphide (S ₅ ^2" ) (which is in equilibrium with tetrasulphide and is present in small amounts under alkaline conditions) is being subjected to non-oxidative alkaline hydrolysis through a nucleophilic attack by a hydroxide ion with expulsion of hydrosulphide [13, 17]. This compound appears to be stable within a narrow range of alkalinities and/or sulphidity. Although the UV method has some potential for on-line application, the necessity of having a small pathlength limits its application to either a rapid laboratory method or a near-process, i.e. at-line, method. Danielsson and Chai [15] proposed the use of UV attenuated total reflectance (UV-ATR) to overcome the problems associated with a small pathlength. However, at low concentrations, because of the small pathlength, the strongly overlapping bands that are seen in the UV-visible region do not enable the determination of individual polysulphide species with sufficient accuracy by monitoring the rise and fall during the process of readily identifiable absorption peaks produced by these individual species. On the other hand, such monitoring may be possible with Raman spectrometry.

Raman spectroscopy is a technique that measures the intensity of light produced by the inelastic scattering of photons originating from a monochromatic light source, such as a visible-light (785 nm) or near-infrared laser (1064 nm). Raman scattering arises when the incident-light photons collide with molecules in the sample, which subsequently scatter the light with a change in energy. The energy difference between the incident light and the Raman scattered light is equal to the energy involved in changing the molecule's vibrational state. This discrete energy shift is called the Raman shift. Raman spectra are expressed in terms of scattering intensity versus wavelength. Raman spectra generally have sharp and narrow bands that are indicative of both the types of molecular bonds present and their respective concentrations.

The Raman technique is ideal for capturing the intensity and frequency of inelastic scattering produced by the vibrations of non-polar bonds such as those of polysulphide chains. This is impossible to do with infrared spectroscopy, which only records the intensity and frequency of light absorbed by strongly polar bonds. Thus, the information which one can gather from a Raman spectrum often complements that which is obtained from an infrared spectrum. For example, water is a strong IR absorber but a weak Raman scatterer and consequently does not interfere with the corresponding Raman spectrum.

Recent technological advances in visible-light Raman spectroscopy and the introduction of fibre-optic probes have simplified the practice of Raman spectroscopy by facilitating the delivery and collection of signals. It is now possible to use visible- light excitation for samples that exhibit significant amounts of fluorescence since the Raman spectrum is more intense and the background fluorescence is significantly reduced. Also, advances in multivariate analysis have shortened the time for on-line data processing. The key principles and developments of Raman technology in the pulp and paper industry have been reviewed by Leclerc and Trung [18].

Meyer et al. [19] report that the Raman spectrum of sodium thiosulphate in aqueous solution has several peaks, namely at 339 cm ^'1 (^ ₆ ), 451 cm ^"1 (^ ₃ ), 672 cm ^"1 (p ₂ ), 1002 cm ^"1 and 1125 cm ^"1 (V _A ). They have also reported the respective Raman absorptivities of these peaks. Sodium sulphate also absorbs at 981 cm ^"1 {v{). On the

other hand, following earlier work by Janz et al.[21] on polycrystalline polysulphide compounds, El Jaroudi et al. [22-23] recently established that the main Raman peak for solid sodium trisulphide appears at 476 cm ^"1 from the laser line, and confirmed that the corresponding peak for solid sodium tetrasulphide appears at 451 cm ^"1 . However, the Raman absorptivities of these two peaks are unknown and have still not been reported, which makes the quantitative determination of individual species in aqueous solution by

Raman spectrometry virtually impossible, especially because of the numerous equilibria involved between these species and the sensitivity of the equilibria to hydroxide and sulphide concentrations. Consequently, it would not be obvious to a person ordinarily skilled in the art of such measurements to successfully perform quantitative Raman determination of individual polysulphide species in either aqueous or alkaline solutions.

In the following, it is demonstrated that the instant invention overcomes previous limitations on the quantitative determination of individual polysulphide species by Raman spectrometry.

DISCLOSURE OF THE INVENTION

The instant invention is used for the on-line analysis of liquor samples from various polysulphide-generating techniques such as, but not limited to, electrolysis, oxidation processes, with or without catalysts such as manganese, lime mud, and wet- proofed agent. The online analytical sensor can be used to determine polysulphide from various points of the generator from which the selectivity and conversion efficiency can be calculated. The Raman sensor of the instant invention overcomes the limitations of and replaces conventional analyses techniques such as gravimetric analysis, gas chromatography, and narrow-pathlength UV determination. Rapid, real-time analysis of polysulphide provides much needed information to optimally control the production of polysulphide. Real-time analysis of polysulphide liquor going to the digester allows for better control of the pulping condition of the kraft process.

In accordance with the invention, there is provided a method for measuring a sulphur compound selected from sodium tetrasulphide, sodium trisulphide or inactive polysulphide in kraft process liquor comprising: (a) withdrawing a sample of the liquor

from the process liquor, (b) subjecting the sample to a monochromatic light source of predetermined wavelength, (c) collecting a scattered Raman intensity over a predetermined wave number region so as to produce a Raman scattered peak-intensity spectrum, (d) determining a peak-intensity measurement for the sulphur compound, and (e) correlating a relationship between the peak-intensity measurement of the sulphur compound in the sample with a known concentration of the sulphur compound to determine polysulphide concentration in the sample.

In another aspect of the invention, there is provided in a process for generating polysulphide in a kraft liquor, the improvement wherein the aforementioned method of the invention is carried out, and at least one process parameter of the generation is controlled in response to the determined polysulphide concentration of the sample.

In accordance with one embodiment of the invention, there is provided a method for measuring sodium tetrasulphide in kraft process liquor comprising: (1) withdrawing a sample from the process liquors, (2) subjecting the sample to a monochromatic light source of predetermined wavelength, (3) collecting the scattered Raman intensity over a predetermined wave number region so as to produce a Raman-scattered peak-intensity spectrum, (4) recording peak-intensity measurements for tetrasulphide, (5) correlating by the relationship between the peak-intensity measurements of the sample with the concentration of sodium tetrasulphide to determine polysulphide concentration in the sample.

In another aspect, there is provided a process in which polysulphide is generated in white liquor which comprises carrying out the method set out hereinbefore on said liquor and controlling at least one parameter of the polysulphide generation is response to the determined polysulphide concentration to obtain an optimum polysulphide concentration in said liquor.

In accordance with yet another embodiment of the invention, there is provided a method for measuring sodium trisulphide in kraft process liquor comprising: (1) withdrawing a sample from the process liquor, (2) subjecting the sample to a monochromatic light source of predetermined wavelength, (3) collecting the scattered Raman intensity over a predetermined wave number region so as to produce a Raman-

scattered peak-intensity spectrum, (4) recording peak-intensity measurements for trisulphide, (5) correlating the relationship between the peak-intensity measurements of the sample with the concentration of sodium trisulphide to determine polysulphide concentration in the samples. In accordance with a further embodiment of the invention, there is provided a method for measuring inactive polysulphide in kraft process liquor comprising: (1) withdrawing a sample from the process liquor, (2) subjecting the sample to a monochromatic light source of predetermined wavelength, (3) collecting the scattered Raman intensity over a predetermined wave number region so as to produce a Raman- scattered peak-intensity spectrum, (4) recording peak-intensity measurements for inactive polysulphide, (5) correlating the relationship between the peak-intensity measurements of the sample with the concentration of inactive polysulphide species to determine polysulphide concentration in the samples.

In another aspect of the invention, there is provided a process in which polysulphide is generated in white liquor which comprises carrying out the method set out hereinbefore on said liquid and controlling at least one parameter of the polysulphide generation in response to the determined polysulphide concentration, to minimize inactive polysulphide concentration in the liquor.

In the aforementioned methods of the invention, the step (5) is suitably carried out by univariate calibration or multivariate calibration.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a diagrammatic view of a sensing apparatus according to one embodiment of the present invention, whereby the sample is withdrawn from the process.

FIG. 2 is a diagrammatic view of a sensing apparatus according to a further embodiment of the present invention, whereby the sample is not withdrawn from the process. FIG. 3 is a diagrammatic view of a sensing apparatus according to a third embodiment of the present invention.

FIG. 4 is a diagrammatic view of a sensing apparatus according to a fourth embodiment of the present invention, whereby the analysis is carried out by remote sampling.

FIG. 5 is a graph of a Raman spectrum of active polysulphide at five different concentrations.

FIG. 6A is a calibration graph of the predicted versus actual active polysulphide concentration.

FIG. 6B is a validation graph of the predicted versus actual active polysulphide concentration. FIG. 7A is a calibration graph of the predicted versus actual inactive polysulphide concentration.

FIG. 7B is a validation graph of the predicted versus actual inactive polysulphide concentration.

FIG. 8A is a calibration graph of the predicted versus actual trisulphide concentration.

FIG. 8B is a validation graph of the predicted versus actual trisulphide concentration.

FIG. 9A is a calibration graph of the predicted versus actual tetrasulphide concentration. FIG. 9B is a validation graph of the predicted versus actual tetrasulphide concentration.

DETAILED DESCRIPTION OF THE INVENTION FIG. 1 illustrates a diagrammatic view of a sensing apparatus according to one embodiment of the present invention. Referring to FIG. 1, the excitation light 10 from a monochromatic light source is delivered to a sample 12 of interest in a sample cell 14 through a fiber-optic cable 16 and a combined system excitation-collection optic 18, thereby generating the Raman-scattered light that is collected by the optic 18 and delivered to a detector/analyser 20 controlled by a computer 22. Sample 12 is delivered to cell 14 in sampling feed 13.

FIG. 2 illustrates a diagrammatic view of a sensing apparatus according to one embodiment of the present invention where sample is not withdrawn from the process but analyzed directly within process lines or reaction/mixing chamber 24 through a window 26 made of sapphire or similar material. Referring to FIG. 2, the excitation light from a monochromatic light source 10 is delivered to a sample 28 of interest within the reaction/mixing chamber 24 through the window 26, thereby generating the Raman-scattered light that is collected by the optic 18 and delivered to a detector/analyser 20 controlled by a computer 22.

FIG. 3 is a diagrammatic view of a polysulphide generator system, based on, but not limited to, the air oxidation process for polysulphide generation with or without catalyst, complete with the sensing and control apparatus as described in FIG. 1 and FIG. 2 of the present invention. Green liquor sample flows from the smelt-dissolving tank into the slaker. Lime is added to the green liquor in the slaker so as to produce sodium hydroxide and the reaction continues as it flows from the slaker in a line 31 and through the various causticizers 30, 32, and 34. Liquor samples are withdrawn from causticizer 32 and diverted to a polysulphide reactor 36 where oxygen 38 and / or air 40 is sparged into the reactor to oxidize the sulphide to form polysulphide. Generally, the polysulphide reactor 36 also incorporates a heater 42 to maintain and/or control polysulphide liquor temperature to provide optimum production temperature. The final polysulphide liquor is then pumped to the digester or first mixed with a white liquor stream prior to use as cooking chemicals.

According to FIG. 3, white liquor samples are taken at withdrawal point 44, prior to the polysulphide reactor through small bore conduit 46 into a sample multiplexer 48 and a flow through cell 50 of indefinite pathlength equipped with a sapphire or similar material for analysis. Monochromatic light source is sent by the fiber optic cable 16(FIG. 2) to the excitation-collection optic 18 (FIG. 3) and brought back to the detector/analyser 20 controlled by the computer 22. Similarly, polysulphide liquors are taken from withdrawal points 52, 54, and 56 to allow for optimization of the polysulphide produced. Analysis of liquor taken from sample withdrawal point 58 can

be used as a feed-forward input to control the proper addition of chemicals to digester 60 for effective and uniform digesting of wood chips.

Alternately, samples can be analyzed directly within the pipe or vessels provided that these sample locations are equipped with a window 26 of sapphire or of similar material, to allow for sample analysis. FIG. 4 illustrates this remote sampling implementation as applied to the electrolytic method of producing polysulphide from white liquor. The electrolytic cell consists of an anode compartment 62 and a cathode compartment 64, separated by a cation-selective membrane 66 and having a power supply 69. White liquor is fed through line 25 into the anode compartment 62 as shown by arrow X and, as current is applied to the anode electrodes 68, Na ₂ S is oxidized to form polysulphide liquor while sodium hydroxide is formed on the cathode electrode 70 and removed through line 71. Production of polysulphide concentration changes with applied current, with increasing current resulting in increasing polysulphide concentration. Sample analyses windows 26 located at the incoming white liquor view point 72 in line 25, within the anode chamber at location 74, and at the polysulphide liquor outlet 27 at location 76 can all be equipped with excitation-collection optics 18 (see FIG. 2) connected to a detector/analyser 20 controlled by the computer 22 via fiber optic cables through an optical multiplexer 78. Rapid determination of the polysulphide at these various locations will allow one to optimally control the production of polysulphide suitable for controlling the digester.

A combined system of pipes, light guides or visible fiber optic cables, and excitation-collection optics are used for remote sensing in the various locations of the polysulphide generator system thereby minimizing costs by allowing for multiple streams to be analyzed with a single Raman spectrometer.

EXPERIMENTAL

Apparatus A laboratory reactor similar to that depicted in [16] was used for generating oxidised white liquors under the same conditions as in [16]. However, a capillary tube was used instead of a sparger to deliver the air to the mixture.

Design Liquors were prepared according to a 3 x 3 full-factorial design. Nine liquors were prepared: the starting sulphide was 15, 20 and 25 g/L as S, with each sulphide level having an initial EA of 70, 85 and 100 g/L as Na ₂ O. Sodium carbonate (15 and 30 g/L as Na ₂ O were also added to the liquor containing 15 g/L as S of sodium sulphide and 85 g/L as Na ₂ O of EA, for a total of eleven liquors. Table I gives the concentration ranges for the 3 x 3 full-factorial design, as well as the sample ID. Having a three-tiered design with different levels of sodium sulphide prevents data closure. Five synthetic liquors having varying levels of total sulfur were also added so as to further open the data and to prevent unintended colinearity between sulphide, total polysulphide and thiosulphate. The liquors are identified in Table I below:

TABLE I

Synthetic white liquors samples used for the 3 x 3 full-factorial experimental design

Sample ID [Sulphide] [EA] [Carbonate] g/L as S g/L as Na ₂ O g/L as Na ₂ O

DO-1718-0min 15 70 0

DO-1719-0min 15 85 0

DO-1720-0min 15 100 0

DO-1721-0min 20 70 0

DO-1722-0min 15 85 15

DO-1723-0min 15 85 30

DO-1725-0min 20 85 0

DO-1727-0min 20 100 0

DO-1728-0min 25 70 0

DO-1729-0min 25 85 0

DO-1730-0min 25 100 0

Procedure

Aliquots were taken at 0, 15, 30, 45 and 60 minutes for the five liquors having 15 g/L sulphide as S; at 0, 20, 40, 60 and 80 minutes for the three liquors having 20 g/L sulphide as S; and at 0, 25, 50, 75 and 100 minutes for the three liquors having 25 g/L sulphide as S. Once generated, each aliquot was identified with a self-explanatory suffix appended to the original sample ID and stored under nitrogen prior to analysis. One mill liquor sample (EA = 81.25 g/L as Na ₂ O; Na ₂ S = 14.01 g/1 as S; Na ₂ CO ₃ = 14.01 g/1 as Na ₂ O) was also used for generating four oxidised white liquors. The oxidation times were from 25 to 75 minutes. Each of the four liquors was thermally treated for 17 hours at 77 ⁰ C. ABC, UV, Raman and gravimetric analysis were performed on all samples.

Reagents

Finely powdered MnO ₂ (Mallinckrodt, 99.8%) was used as the catalyst for making oxidised liquors. Reagent-grade chemicals were used for making synthetic white liquors prior to oxidation. Sodium tetrasulphide (Na ₂ S ₄ , ca. 90% purity) was obtained from Alfa Aesar (Word Hill MA) and used to make the synthetic polysulphide liquors. All liquors were stored under nitrogen prior to analysis so as to prevent oxidation.

Spectroscopic Analysis of Oxidised and Synthetic Polysulphide Liquors

A computer-controlled Chromex Inc. (Albuquerque NM) Sentinel visible- excitation Raman spectrometer was used for measuring the concentration of thiosulphate and that of individual polysulphide species. The Sentinel uses a continuous wave (cw) diode laser source with an excitation wavelength of 785 nm and a fibre-optic sample probe. Raman spectra were collected with 45 seconds signal integration and 6 co-added scans in a 180° back-scatter mode over a range of 240 to 2250 cm ^" with a resolution of 0.25 cm ^"1 . The power intensity at the sample was approximately 75 mW.

UV-visible spectra were collected over a range of 200 to 600 nm with a resolution of 1 nm with a Varian (Palo Alto CA) Cary 50 Probe. A 0.2-mm Suprasil ^® fused-silica transmission cell was used for UV analyses. Total polysulphide was

determined gravimetrically using the method of Dorris and Uloth [8]. Inactive polysulphide was taken to be the value of the difference between the amount of total polysulphide determined gravimetrically and that of active polysulphide determined at 416 nm. Thermo Galactic's (Thermo Instruments, Madison WI) Grams32™ and its

PLSIQ™ multivariate calibration software package was used for spectral processing and developing partial least-squares calibrations.

Liquor Analysis A Brinkmann 682 Titroprocessor equipped with a 665 Dosimat accessory was used for liquor analysis and determining EA, AA and TTA. The results were used to determine initial and residual sulphide in Paprilox liquors. By assuming a mass balance with the sodium sulphide measured at the start of the reaction, sodium thiosulphate concentrations were obtained by subtracting the total polysulphide determined by UV- visible spectrometry and the residual sulphide obtained by titration from the initial sulphide concentration.

The initial sulphide concentration present prior to oxidation in the eleven synthetic white liquors of the 3x3 design was first calculated from chemical weights. When time permitted, this value was also confirmed by titration. In most cases, results were within 0.5 g/L as S of the chemical-weight value. When available, the initial- sulphide values obtained by titration were used together with the total polysulphide and residual sulphide concentrations for calculating the thiosulphate concentrations.

Calibration Modelling The concentrations of effective alkali, residual sulphide, active polysulphide, inactive polysulphide and thiosulphate obtained previously were used as input variables for building the calibration. The respective concentrations of trisulphide and tetrasulphide were calculated with the use of the equilibrium constants derived by Giggenbach [11] and the activation energy for the alkaline hydrolysis of pentasulphide reported by Licht and Davis [13], as well as that for thiosulphate formation reported by

Giggenbach [17]. The calculated trisulphide and tetrasulphide values were then used for building a calibration model that could predict the concentration of these species without bias. The spectral region from 300 to 1800 cm ^"1 was used, with the exception of regions where sulphate (981 cm ^"1 ) and carbonate (1065 cm ^"1 ) absorb. Optimal calibrations were achieved with three or four principal components in most cases. About 20 samples were left out of the calibration for validation purposes. The additional set of four oxidized mill white liquor samples and four thermally treated samples generated from the mill liquor was used as an independent validation set.

EXAMPLES Example 1

The Raman spectrum of active polysulfide at five different concentrations is seen in FIG. 5. The peak at 451 cm ^"1 is mainly attributable to the absorbance of the tetrasulphide species in solution while the peak at 475 cm ^"1 is related to the trisulphide species present in solution. As expected, when tetrasulphide is dissolved, equilibrium is established whereby both tri-and tetrasulphide are present.

Example 2

Calibration and validation results for active and inactive polysulphide are shown in FIGS. 6A, 6B, 7A and 7B. Four principal components were used to describe 99 % of the measurement variance for active (FIG. 6A and 6B) and inactive (FIG. 7A and 7B) polysulphide. As seen in FIG. 6B, prediction results for active polysulphide have a slightly lower error of prediction than for the calibration (FIG. 6A), with a prediction error of less than 0.30 g/L as S.

On the other hand, prediction results for inactive polysulphide (FIG. 7B) are slightly more scattered than for the calibration (FIG. 7A). The prediction error is about 0.85 g/L as S.

Example 3

Calibration and validation results for bisulphide and tetrasulphide are shown in FIGS. 8 A, 8B, 9A and 9B, respectively. Four principal components were used for trisulphide, whereas six were needed for tetrasulphide, probably because of the strong overlap with the thiosulphate peak at 445 cm ^'1 .

As seen in FIG. 8B, the prediction error for trisulphide is very low, i.e. 0.13 g/L as S, and is of the same order of magnitude as the calibration, i.e. 0.16 g/L as S (FIG. 8A). This is probably because of the lack of overlapping bands around 475 cm ^"1 . On the other hand, the prediction error in FIG. 9B for tetrasulphide is double that of trisulphide, i.e. 0.25 g/L as S, but nevertheless slightly lower than that found for active polysulphide (0.29 g/L as S).

FIGS. 6A, 6B, 7A, 7B, 8A, 8B, 9A and 9B are further described hereinafter:

FIG. 6A is a graph of multivariate calibration regression (4 principal components obtained for active polysulphide showing the actual versus predicted active-polysulphide concentration;

FIG. 6B is a graph of multivariate validation regression obtained for active polysulphide showing the actual versus predicted active-polysulphide concentration.

FIG. 7A is a graph of multivariate calibration regression (4 principal components) obtained for inactive polysulphide showing the actual versus predicted inactive-polysulphide concentration;

FIG.7B is a validation regression obtained for inactive polysulphide showing the actual versus predicted inactive-polysulphide concentration. The actual inactive- polysulphide concentration was obtained by calculating the difference between the amount of total polysulphide determined gravimetrically and that of active polysulphide determined at 416 nm. The value of inactive polysulphide determined in this manner was further verified by measuring inactive polysulphide at 330 nm.

FIG.8 A is a multivariate calibration regression (4 principal components) obtained for trisulphide showing the actual versus predicted trisulphide concentration;

FIG.8B is a validation regression obtained for trisulphide showing the actual versus predicted trisulphide concentration.

FIG.9A is a multivariate calibration regression (6 principal components) obtained for tetrasulphide showing the actual versus predicted tetrasulphide concentration;

FIG.9B is a validation regression obtained for tetrasulphide showing the actual versus predicted tetrasulphide concentration.

REFERENCES

1. Kleppe, P.J., "Polysulfide Pulping in a Dual- Vessel Kamyr Digester", Tappi J., 58(8): 172 (1975).

2. Landmark, P.A., Kleppe, PJ. and Johnsen, K., "Cooking Liquor Oxidation and Improved Cooking Technique in Polysulfide Pulping", Tappi J. 48(5): 56 (1965).

3. Dorris, G. M., "Oxidation of White Liquor in the Presence of Lime Mud", Pulp Paper Can, 95(10): 44 (1994).

4. Dorris, G.M., "Process of Producing Kraft Pulp Liquor by the Oxidation of White Liquor in the Presence of Lime Mud", U.S. Patent No. 5,082,526 (1992).

5. Yamaguchi, A., "Operating Experiences with the MOXY Process and Quinoid Compounds", 1983 TAPPI Pulping Conference Proceedings, TAPPI Press, Atlanta, GA, vol. 2, p.544.

6. Smith, G.C., Knowles, S.E. and Green, R.P., "All its Takes is MOXY: Mead Oxidation System Generates Polysulfide Liquor", Paper Trade J. 159: 38 (1975).

7. Nakamura, M. and Ono, T., "Production of Polysulfide Using a New Catalyst", 1988 TAPPI Pulping Conference Proceedings, TAPPI Press, Atlanta, vol. 2, p.407.

8. Dorris, G.M. and Uloth, V.C., "Analysis of Oxidized White Liquors. Part I: Determination of Polysulphides by Gravimetry", J. Pulp Pap. Sci. 20(8): J211 (1994).

9. Dorris, G.M. and Uloth, V.C., "Analysis of Oxidized White Liquors. Part II: Potentiometric Titrations for the Determination of Polysulphides and Sulphoxy Anions", J. Pulp Pap. Sci. 20(9): J242 (1994).

10. Sullivan, J. and Douek, M., "Analysis of Hydroxide, Inorganic Sulphur Species and Organic Anions in Kraft Pulping Liquors by Capillary Electrophoresis", J. Chromatogr. A, 1039:215-225 (2004).

1 1. Giggenbach, W., "Optical Spectra and Equilibrium Distribution of Polysulfide Ions in Aqueous Solution at 20°", Inorg. Chem. 11(6): 1201 (1972).

12. Licht, S., Hodes, G. and Manassen, J., "Numerical Analysis of Aqueous Polysulfide Solutions and Its Application to Cadmium Chalcogenide/Polysulfide Photoelectrochemical Solar Cells", Inorg. Chem. 25(15): 2486 (1986).

13. Licht, S. and Davis, J., "Disproportionation of Aqueous Sulfur and Sulfide: Kinetics of Polysulfide Decomposition," J. Phys. Chem.B, 101 : 2540 (1997).

14. Danielsson, L.-G., Chai, X.-S., Behm, M., and Renberg, L., "UV Characterization of Sulphide-Polysulphide Solutions and Its Application for Process Monitoring in the Electrochemical Production of Polysulphides", J. Pulp Pap. Sci. 22(6): J187 (1996).

15. Danielsson, L.-G., and Chai, X.-S., "Spectroscopic Measurements in Opaque Solutions: UV- Vis Spectroscopy on Process Liquors in the Paper and Pulp Industries", Process Cont. Qual., 6:149 (1994).

16. van Heek, R.P., Dorris, G.M., Uloth, V.C., Page, N., Hu T.Q., and Leclerc, D.F., "Improved Kraft Pulp Yield by Heat Treatment of Polysulphide Liquors Generated by Oxidation", WIPO PCT Publication No. WO 02/059417 A2 (August 1, 2002).

17. Giggenbach, W.F., "Kinetics of the Polysulphide-Thiosulphate Disproportionation up to 240°", Inorg. Chem. 13(7): 1730 (1972).

18. Leclerc, D.F., and Trung, T.P., "Vibrational Spectroscopy in the Pulp and Paper Industry" in Handbook of Vibrational Spectroscopy, J.M. Chalmers and P.R. Griffiths (Eds.), John Wiley and Sons, Chichester UK, Volume 4, Part 9: Industrial Applications, pp. 2952-2976 (2002).

19. Meyer, B., Ospina, M., and Peter, L.B., "Raman Spectrometric Determination of Oxysulfur Anions in Aqueous Systems", Anal. Chem. Acta, 117, 301 (1980).

20. Janz, GJ. , Downey, J.R. Jr., Roduner E., Wasilczyk, G.J., Courts, J.W., and Eluard A., "Raman Studies of Sulfur-Containing Anions in Inorganic Polysulfides. Sodium Polysulfides", Inorg. Chem. 15(8): 1759 (1976).

21. El Jaroudi, O., Picquenard, E., Demortier, A., Lelieur, J.-P., and Corset J., "Polysulfide Anions. I. Structure and Vibrational Spectra of the S ₂ ^2- and S ₃ ^2" Anions. Influence of the Cations on Bond Length and Angle", Inorg. Chem. 38(10): 2394 (1999).

22. El Jaroudi, O., Picquenard, E., Demortier, A., Lelieur, J.-P., and Corset J., "Polysulfide Anions. II. Structure and Vibrational Spectra of the S ₄ ^2" and S ₅ ^2" Anions. Influence of the Cations on Bond Length, Valence, and Torsion Angle", Inorg. Chem. 39(12): 2593 (2000).