- Email: [email protected]
Novel headspace gas chromatographic method for determination of oxalate in oxygen delignification liquor
Journal of Chromatography A, 1122 (2006) 209–214
Novel headspace gas chromatographic method for determination of oxalate in oxygen delignification liquor X.-S. Chai a,b,∗ , J. Samp b , H.N. Song a , H.X. Zhu a a
b
School of Light Industrial and Food Engineering, Guangxi University, Nanning, Guangxi 530004, China Institute of Paper Science and Technology, Georgia Institute of Technology, 500 10th Street, N.W., Atlanta, GA 30332, USA Received 22 February 2006; received in revised form 11 April 2006; accepted 11 April 2006 Available online 9 May 2006
Abstract A novel headspace gas chromatographic (HS-GC) method is demonstrated for an indirect determination of oxalate in oxygen delignification liquors. A small volume (50–100 L) of liquor sample is introduced into a sampling vial that contains 1.0 mL of 2 mol/L sulfuric acid. After removal of carbon dioxide (generated from carbonate in the acidic medium) by heating, the sample was mixed with a 0.5 mL of 0.02 mol/L potassium permanganate solution in a closed testing vial. At an elevated temperature (70 ◦ C), the oxalate in the sample is rapidly converted to carbon dioxide by reacting with permanganate. The carbon dioxide in the headspace can be measured by gas chromatography with a thermal conductive detector. Using a multiple headspace extraction (MHE) measurement technique, the kinetics of formation of the carbon dioxide from the other organic species in the sample can be determined, and thus a correction can be made for minimizing the interferences. The present method is simple, accurate and can be easily automated. © 2006 Elsevier B.V. All rights reserved. Keywords: Headspace; Gas chromatography; Multiple headspace extraction; Carbon dioxide; Oxalate; Oxygen delignification liquor
1. Introduction Sodium carbonate and sodium sulfate in black liquor are major species that cause fouling problem during black liquor evaporation [1,2]. With progressive water removal from the black liquor by evaporation, the solubility limits of these inorganic salts (mainly sodium carbonate and sodium sulfate) are exceeded. They deposit on heat transfer surfaces, thus reducing the efficiency of the evaporators. Fouling problem can also be caused by the precipitation of sodium oxalate [3]. Great amounts of oxalate ions are formed in the oxygen delignification process for further removing the residual lignin from pulps before the bleaching. In mill operation, the oxygen delignification spent liquor is used for washing the pulp from the digester and then fed to evaporator. Therefore, the analysis of oxalate in the oxygen delignification liquors is important for the process study to clarify how the process parameters such as temperature and alkalinity affect the formation of oxalate ions in the different
∗
Corresponding author. Tel.: +1 404 894 9992; fax: +1 404 894 4778. E-mail address: [email protected] (X.-S. Chai).
0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.04.042
processes, which can be a help to minimize the fouling problem in the evaporators. Traditionally, oxalate can be quantified by a redox titration method using permanganate in an acidic medium, which is based on the following reaction, 2MnO4 − + 5C2 O4 2− + 16H+ = 2Mn2+ + 10CO2 + 8H2 O (I) At an elevated temperature, the Reaction (I) can be accomplished rapidly. The titration endpoint is determined according to the lasting, distinct reddish color of permanganate added, i.e., the oxalate in the solution has been completely consumed by permanganate. However, this method cannot be applied to a sample containing the other reduction species, e.g., organic compounds, because these species may also react with permanganate to form carbon dioxide, although the reaction rates are slow. There is a significant amount of organic compounds – mainly lignins – in the pulping and oxygen delignification streams. The complicated liquor matrix makes it impossible to quantify oxalate using the above titration method. Compared with many other industries, especially the pharmaceutical industry, activities in the development of testing method or innovation in the pulp and paper industry based on
210
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
modern instrumentations are limited. In 1980s, capillary ion analysis and ion chromatographic technique have been introduced to the paper industry for the determination of many anions including oxalate in various mill streams [4–8]. In the TAPPI Test Methods [9], the measurement procedures for many anions existing in the pulping and bleaching liquors using ion chromatography have been standardized on the basis of the these literatures [5–8], which is also one of the few methods using advanced analytical instrument found in the TAPPI Test Methods. Attributed to the column capability in anion separation, the ions can be separated from each other and measured by UV or electrolytic conductivity detector. However, besides the relative complicated sample pretreatment procedures, the main problem in the ion chromatographic method is the poor measurement precision. As reported [9], the repeatability errors range from 10 to 30% for the inorganic anions and 15% for the organic acids (including oxalate). Furthermore, if the amount of sulfate ion or the other organic acids in the sample liquor is high, it will lead to significant errors in oxalate determination because the peaks of the respective species are very close in the chromatogram. In our previous work, we have reported a novel headspace gas chromatographic (GC) method for determination of carbonate in black liquor (i.e., pulping spent liquor) [10]. It is based on a phase conversion reaction (PCR) to convert carbonate into carbon dioxide in an acidic medium. The carbon dioxide released to headspace is determined by GC with a thermal conductivity detector. In a separate paper [11], we developed a multiple headspace extraction (MHE) GC technique for determination of the kinetics of formation of methanol in black liquors, which is a slow reaction process. As carbon dioxide formation from Reaction (I) is fast and the formation from the other organic species is slow, we believe that an analysis of the oxalate content in mill streams can be realized by a PCR combined MHE-GC technique, in which the carbon dioxide from the interference species can be corrected.
Eq. (2), At = Aoxalate + Aorg (1 − exp(−kt)) = Aoxalate +
t
At (2)
0
The GC response is directly proportional to the amount present in the headspace, i.e., At = Kmt
(3)
According to Eq. (2), the time-dependent carbon dioxide formation from the other organic species can be expressed as, At =
t
At = Aorg (1 − exp(−kt))
(4)
0
According to Eq. (4), a linear relationship can be derived between ln(Aorg,t ) and t as follows, At = At − At−t = Aorg exp[−k(t − t)][1 − exp(−kt)] (5) Therefore, for constant t, ln(Aorg,t ) = a − kt
(6)
where a = ln(Aorg [1 − exp(−kt)]) + kt. A MHE-GC method is based on repeatedly extracting an equilibrated headspace from a sample vial and the relationship between amount of species in headspace and its extraction number can be described as [12,13], ln(An ) = ln(A0 ) − bn
(7)
with t = jn − 1
(8)
According to Eq. (7), the relationship of areas in the successive runs can be written as,
2. Methodology
An = An−1 eb
When adding a sample that contains both oxalate and other organic species into a permanganate-acidic solution at an elevated temperature, oxalate is completely converted to carbon dioxide within a short time (i.e., before the first headspace measurement) according to Reaction (I). However, it is a slow process for the carbon dioxide formation from the other organic species, which is an exponential decay function, i.e., morg [1 − exp(−kt)]. The carbon dioxide formation kinetics in such a reaction system can be described as,
Eq. (7) is not valid in a system involving chemical reaction [11], e.g., carbon dioxide formation from the other organic species. When using a headspace extraction number to replace the time, Eq. (2) can be rewritten as follows, i.e.,
mt = moxalate + morg [1 − exp(−kt)]=moxalate +
t
An = Aoxalate +
(9)
n
Aorg,n
At the first headspace GC measurement, A1 = Aoxalate + Aorg,1
mt
(1)
0
All symbols are defined in Section 6. At an elevated temperature, the carbon dioxide generated from the reactions is completely released to the headspace, which can be determined by headspace GC and described by
(10)
1
(11)
According to Eq. (9), at the second headspace GC measurement, A2 = A1 eb + Aorg,2
(12)
At nth headspace GC measurement, An = An−1 eb + Aorg,n
(13)
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
211
Thus, we can calculate peak area at any nth (n = 2, 3, . . ., n) headspace GC measurement, i.e., Aorg,n = An − An−1 eb
(14)
By conducting a MHE-GC measurement on a single sample vial, we can obtain A1 , A2 , . . ., An , thus, we can calculate the Aorg,2 , Aorg,3 , . . ., Aorg,n according to Eq. (14). With Eq. (8), we can convert Aorg,n to Aorg,tn . By plotting log(Aorg,t ) versus t (where t is from t2 to tn ), the intercept (a) and slope (k) in Eq. (6) can be obtained. From this equation, we can calculate the carbon dioxide formed by the other organic species at the first (t = t1 ) headspace GC measurement, i.e., Aorg,t1 . Then, we can obtain Aoxalate according to Eq. (11). Finally, we can calculate the amount of oxalate in the sample, i.e., Coxalate =
Aoxalate KVs
(15)
3. Experimental 3.1. Chemicals All chemicals used in the experiment were from commercial sources. The standard oxalate solution (0.250 mol/L) was prepared using sodium oxalate. An oxygen delignification liquor sample was withdrawn at the end of an oxygen delignification process. A synthetic oxygen delignification spent liquor was prepared by dissolving alkaline lignin in a sodium hydroxide (2 g/L) medium, in which the lignin content is 4 g/L. On the basis of this liquor, a set of solutions containing different amount of oxalate (from 4 to 20 mmol/L) was prepared.
Fig. 1. MHE-GC measurement on (a) methanol and formic acid and (b) oxalate and mill solution.
vial is immediately put into headspace sampler for automatic MHE-GC measurement. 4. Results and discussions
3.2. Apparatus and operation All measurements were carried out using an HP-7694 Automatic Headspace Sampler and Model HP-6890 capillary gas chromatograph equipped with a thermal conductivity detector (Hewlett-Packard, now Agilent Technologies, Palo Alto, CA, USA). GC conditions were as follows: capillary column with an I.D. = 0.53 mm and a length of 30 m (model GS-Q, J&W Scientific, Folsom, CA, USA) at 30 ◦ C; carrier gas helium flow rate of 3.1 mL/min. Headspace Sampler operating conditions were as follows: oven temperature of 70 ◦ C; vial pressurized by helium and pressurization time of 0.2 min; sample-loop fill time of 0.2 min; loop equilibration time of 0.05 min; loop fill time of 0.2 min; the vial equilibration time is 2 min and MHE recycle time is 3 min. 3.3. Sample preparation and measurement procedures Two milliliters 2 mol/L sulfuric acid solution and 0.5 mL 0.02 mol/L permanganate solution in a headspace test vial were added and sealed with a septum. A 50 L of sample liquor was taken in a microsyringe and injected into the sealed vial. After quickly shaking the vial for mixing the sample and reactants, the
4.1. Kinetics of formation of carbon dioxide from interfering species There are many organic species – mainly dissolved lignin – in mill streams that can react with permanganate; however, the carbon dioxide formation is caused by the low molecular weight species, such as methanol, which is found in both alkaline pulping and oxygen delignification spent liquors. Fig. 1a and b show the results of MHE-GC on the methanol, formic acid, oxalate solutions, and mill oxygen delignification liquor, respectively. For the oxalate solution, as carbon dioxide is generated rapidly at the given experimental conditions, the carbon dioxide in the headspace decreases exponentially with the headspace extraction number. For the methanol and formic acid solution, owing to a slow reaction with permanganate, the carbon dioxide in the headspace first increases and then decreases. It has been noticed that the slope of the exponential decrease for the methanol solution is the same as the oxalate solution after the third headspace extraction, indicating that no more carbon dioxide was formed after that time. In the selected mill liquor, there are many organic species besides oxalate. The carbon dioxide in the headspace at first headspace GC measurement is contributed
212
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
the interference species on the basis of the first two headspace measurements, i.e., Aorg,t1 =
t1 Aorg,t2 t2 − t1
(16)
In this way, it significantly reduces the experiment time. 4.3. Inorganic interferences The present method is based on measuring carbon dioxide released from the reaction between oxalate and permanganate according to Reaction (I). In an acidic medium, carbonate in the sample can be rapidly converted to carbon dioxide, which affects the oxalate determination. Thus, the carbonate effect must be taken into account. In this work, a blank test run, i.e., the sample added in an acidic solution without permanganate with the same headspace GC conditions, was conducted. Thus, the carbon dioxide effect caused by carbonate acidification can be subtracted. 4.4. MHE cycle time
Fig. 2. Kinetics of carbon dioxide formation from (a) methanol-permanganate reaction and (b) sample permanganate reaction.
In the present headspace sampler, the GC sampling is based on a pressuring mode, i.e., by injecting the helium (pressuring gas) into the vial to create a high vial pressure. In this way, the headspace sample can be released from the vial to the GC sample loop through venting owing to the pressure difference. However, each action needs a certain time for the control valve to open and shut. This does not allow for very short MHE cycle times to be used in the MHE-GC measurement. In this work, we set the equilibration and MHE cycle times at 2 and 3 min, respectively.
not only by a complete Reaction (I) from the oxalate, but also by the other organic species in the sample. The decrease of the headspace carbon dioxide in these MHE-GC measurements is slower than that of oxalate solution, indicating that a complete reaction has not been achieved yet. For the methanol and the oxygen delignification liquor sample, the formation of carbon dioxide at each headspace measurements as shown in Fig. 1, i.e., the Aorg,n , can be calculated using Eq. (9). The slope (b) in the equation can be obtained by the MHE-GC measurements for a pure oxalate solution testing. Aorg,n is further changed to Aorg,tn with Eq. (8) for plotting Aorg,t1 versus the reaction time t to obtain the kinetic formation curves for carbon dioxide, as shown in Fig. 2a and b, respectively. We can see that the reaction for methanol is relatively fast, but it is slow for the interference species in the oxygen delignification liquor at the given experimental conditions.
4.5. Sulfuric acid concentration
4.2. Organic interference correction
4.6. Temperature effect
As shown in Fig. 2a, the carbon dioxide formation rates from the interference species are relatively fast in the beginning of the reaction, in which the formation rates can be regarded as the same during the reaction within 5 min. Thus, we can use Eq. (14) to calculate the carbon dioxide formation caused by
A higher temperature is desired for achieving a rapid and complete reaction between the oxalate and permanganate. However, it was noticed that the higher temperature leads to a higher water vapor pressure, which may affect the separation performance for the given GC column. In this work, we chose 70 ◦ C as
In Reaction (I), the acid concentration is very important for the reaction rate. For a rapid reaction, a high concentration of sulfuric acid is essential. However, higher acid concentration can also promote carbon dioxide formation from the interference species. Although such effect can be corrected by the present method, it will lead to a larger measurement error when the oxalate content in the sample is relatively low. Conversely, if the reaction with the interference species is too fast, i.e., it has been completed before the second headspace measurement, the relationship described in Eq. (16) is not valid. In this work, a mild concentration (2 mol/L) of sulfuric acid was chosen. As shown in Fig. 1, a complete reaction of oxalate is still ensured at the first headspace measurement at the given experimental conditions.
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
213
with five data points (n = 5) and an R2 value of 0.9979; for the standard oxalate solution, the equation is
Table 1 Repeatability of the method Sample
A1
A2
A2
A1
Aoxalate + Acarbonate a
Aoxalate
1 2 3 4 5
67.9 67.7 69.8 66.3 66.9
102.4 103.3 105.8 100.5 101.9
36.3 37.4 37.9 36.0 36.8
24.2 24.9 25.2 24.0 24.5
43.7 42.8 44.6 42.3 42.4
20.6 19.7 21.5 19.2 19.3
2.3
5.0
RSD (%) a
The average GC signal from blank tests (for carbonate) is 23.1 (with a RSD of 3.5%).
the reaction temperature, in which a complete reaction of oxalate can be achieved within 2 min (at the first headspace measurement) as shown in Fig. 1. 4.7. Method precision and method validation The repeatability of the method was determined using a 50 L of a mill oxygen delignification liquor sample, in which the oxalate content is 12.5 mmol/L measured by the present method. Table 1 lists the GC measurement signals (A1 and A2 ) and the data (A2 , A1 and Aoxalate ) were calculated according to Eqs. (12), (16) and (11), respectively. The result shows that the average relative standard deviation (RSD) from five measurements is less than 5%, which is much better than that reported in the TAPPI Test Methods [9]. Since the sample liquor can directly be analyzed by the present method without any pretreatment such as dilution, and the retention time for carbon dioxide in each GC run is relatively short, it is much convenient and efficient than the current ion chromatographic method in oxalate quantification. The present method is also validated by a set of oxygen delignification liquors with the standard oxalate addition (from 0 to 0.040 mol/L). These samples were measured by the present method and results are shown in Fig. 3, in which a standard calibration curve of oxalate is also included. For the sample with standard oxalate addition, the equation is y = 21.52(±1.17) + 1782(±47.6)x
(17)
Fig. 3. Response curves of the sample liquors with standard oxalate addition and the standard oxalate solutions.
y = −1.38(±0.29) + 1764(±12.1)x
(18)
with five data points (n = 5) and an R2 value of 0.9999. The limit of quantitation in the present method is 2 mmol/L. It can also be seen that there is no significant difference (at a 95% confidence level) for slopes of these two calibration curves, thus, the effect of sample matrix is not significant in the present method. The maximum relative difference between the oxalate standard added and measured by MHE-GC and calculated by Eq. (17) is 7.1% when the sample contains 10.0 mmol/L of oxalate, indicating that the present method is justifiable. 5. Conclusions We have demonstrated a novel HS-GC technique for the determination of oxalate content in mill liquors, based on a rapid formation of carbon dioxide from the reaction between oxalate and permanganate. Using a multiple headspace extraction (MHE) measurement technique, the kinetics of formation of the carbon dioxide from the other organic species can be determined, so that the effect of the interference species on the oxalate measurement can be corrected. In addition, this method is automated, leading to higher accuracy and simplicity in measurement. 6. Nomenclature a At b j k K moxalate morg
intercept of Eq. (6) GC peak area at the reaction time t slope of Eq. (7) headspace extraction cycle time slope of Eq. (6) GC response factor mass of carbon dioxide converted from oxalate mass of carbon dioxide converted from the other organic species mt mass of the carbon dioxide at the reaction time t n headspace extraction number t reaction time t1 time at the first headspace measurement t2 time at the second headspace measurement tn time at the nth headspace measurement Vs sample volume added into the testing vial Aorg,1 carbon dioxide generated from the other organic species during a time period from zero to t1 Aorg,2 carbon dioxide generated from the other organic species between first and second headspace measurement during a time period from t1 − t2 Aorg,n carbon dioxide generated from the other organic species between the (n − 1)th and nth headspace measurement during a time period of tn−1 − tn t time interval between headspace extractions t=tn m integrated mass of carbon dioxide from the other t 0 organic species, which is = mt1 + mt2 + · · · + mtn
214
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
Acknowledgements This work was partly supported by the Guangxi University Key Program for Science and Technology Research (2005 ZD01), Guangxi University Research Foundation (X061021), and National Natural Science Foundation of China (20367001). References [1] T.M. Grace, Survey of Evaporator Scaling in the Alkaline Pulp Industry, The Institute of Paper Chemistry, Appleton, WI, 1975. [2] W. Schmidl, W. Frederick, 1998 International Chemical Recovery Conference Proceedings, 1998, p. 367.
[3] P. Ulmgren, R. Radestrom, Nord. Pulp Pap. Res. J. 15 (2000) 128. [4] J.P. Romano, D.R. Salomom, Tappi Proceedings, 1992 Pulping Conference, 1992, p. 303. [5] J. Krishnagopalan, M. Hill, A.L. Fricke, Tappi J. 68 (1985) 108. [6] D.B. Easty, M.L. Borchardt, A.A. Webb, Pap. Puu 67 (1985) 501. [7] D.B. Easty, J.E. Johnson, Tappi J. 70 (1987) 109. [8] D.B. Easty, J.E. Johnson, A.A. Webb, Pap. Puu 68 (1986) 415. [9] TAPPI Test Method, Tappi Press, Atlanta, GA, 1987. [10] X.-S. Chai, Q. Luo, J.Y. Zhu, J. Chromatogr. A 909 (2001) 249. [11] X.-S. Chai, Q. Luo, J.Y. Zhu, J. Chromatogr. A 946 (2002) 177. [12] B. Kolb, L.S. Ettre, Static Headspace-Gas Chromatography – Theory and Practice, Wiley-VCH, New York, 1997. [13] X.-S. Chai, J.Y. Zhu, Anal. Chem. 70 (1998) 3481.
Novel headspace gas chromatographic method for determination of oxalate in oxygen delignification liquor X.-S. Chai a,b,∗ , J. Samp b , H.N. Song a , H.X. Zhu a a
b
School of Light Industrial and Food Engineering, Guangxi University, Nanning, Guangxi 530004, China Institute of Paper Science and Technology, Georgia Institute of Technology, 500 10th Street, N.W., Atlanta, GA 30332, USA Received 22 February 2006; received in revised form 11 April 2006; accepted 11 April 2006 Available online 9 May 2006
Abstract A novel headspace gas chromatographic (HS-GC) method is demonstrated for an indirect determination of oxalate in oxygen delignification liquors. A small volume (50–100 L) of liquor sample is introduced into a sampling vial that contains 1.0 mL of 2 mol/L sulfuric acid. After removal of carbon dioxide (generated from carbonate in the acidic medium) by heating, the sample was mixed with a 0.5 mL of 0.02 mol/L potassium permanganate solution in a closed testing vial. At an elevated temperature (70 ◦ C), the oxalate in the sample is rapidly converted to carbon dioxide by reacting with permanganate. The carbon dioxide in the headspace can be measured by gas chromatography with a thermal conductive detector. Using a multiple headspace extraction (MHE) measurement technique, the kinetics of formation of the carbon dioxide from the other organic species in the sample can be determined, and thus a correction can be made for minimizing the interferences. The present method is simple, accurate and can be easily automated. © 2006 Elsevier B.V. All rights reserved. Keywords: Headspace; Gas chromatography; Multiple headspace extraction; Carbon dioxide; Oxalate; Oxygen delignification liquor
1. Introduction Sodium carbonate and sodium sulfate in black liquor are major species that cause fouling problem during black liquor evaporation [1,2]. With progressive water removal from the black liquor by evaporation, the solubility limits of these inorganic salts (mainly sodium carbonate and sodium sulfate) are exceeded. They deposit on heat transfer surfaces, thus reducing the efficiency of the evaporators. Fouling problem can also be caused by the precipitation of sodium oxalate [3]. Great amounts of oxalate ions are formed in the oxygen delignification process for further removing the residual lignin from pulps before the bleaching. In mill operation, the oxygen delignification spent liquor is used for washing the pulp from the digester and then fed to evaporator. Therefore, the analysis of oxalate in the oxygen delignification liquors is important for the process study to clarify how the process parameters such as temperature and alkalinity affect the formation of oxalate ions in the different
∗
Corresponding author. Tel.: +1 404 894 9992; fax: +1 404 894 4778. E-mail address: [email protected] (X.-S. Chai).
0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.04.042
processes, which can be a help to minimize the fouling problem in the evaporators. Traditionally, oxalate can be quantified by a redox titration method using permanganate in an acidic medium, which is based on the following reaction, 2MnO4 − + 5C2 O4 2− + 16H+ = 2Mn2+ + 10CO2 + 8H2 O (I) At an elevated temperature, the Reaction (I) can be accomplished rapidly. The titration endpoint is determined according to the lasting, distinct reddish color of permanganate added, i.e., the oxalate in the solution has been completely consumed by permanganate. However, this method cannot be applied to a sample containing the other reduction species, e.g., organic compounds, because these species may also react with permanganate to form carbon dioxide, although the reaction rates are slow. There is a significant amount of organic compounds – mainly lignins – in the pulping and oxygen delignification streams. The complicated liquor matrix makes it impossible to quantify oxalate using the above titration method. Compared with many other industries, especially the pharmaceutical industry, activities in the development of testing method or innovation in the pulp and paper industry based on
210
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
modern instrumentations are limited. In 1980s, capillary ion analysis and ion chromatographic technique have been introduced to the paper industry for the determination of many anions including oxalate in various mill streams [4–8]. In the TAPPI Test Methods [9], the measurement procedures for many anions existing in the pulping and bleaching liquors using ion chromatography have been standardized on the basis of the these literatures [5–8], which is also one of the few methods using advanced analytical instrument found in the TAPPI Test Methods. Attributed to the column capability in anion separation, the ions can be separated from each other and measured by UV or electrolytic conductivity detector. However, besides the relative complicated sample pretreatment procedures, the main problem in the ion chromatographic method is the poor measurement precision. As reported [9], the repeatability errors range from 10 to 30% for the inorganic anions and 15% for the organic acids (including oxalate). Furthermore, if the amount of sulfate ion or the other organic acids in the sample liquor is high, it will lead to significant errors in oxalate determination because the peaks of the respective species are very close in the chromatogram. In our previous work, we have reported a novel headspace gas chromatographic (GC) method for determination of carbonate in black liquor (i.e., pulping spent liquor) [10]. It is based on a phase conversion reaction (PCR) to convert carbonate into carbon dioxide in an acidic medium. The carbon dioxide released to headspace is determined by GC with a thermal conductivity detector. In a separate paper [11], we developed a multiple headspace extraction (MHE) GC technique for determination of the kinetics of formation of methanol in black liquors, which is a slow reaction process. As carbon dioxide formation from Reaction (I) is fast and the formation from the other organic species is slow, we believe that an analysis of the oxalate content in mill streams can be realized by a PCR combined MHE-GC technique, in which the carbon dioxide from the interference species can be corrected.
Eq. (2), At = Aoxalate + Aorg (1 − exp(−kt)) = Aoxalate +
t
At (2)
0
The GC response is directly proportional to the amount present in the headspace, i.e., At = Kmt
(3)
According to Eq. (2), the time-dependent carbon dioxide formation from the other organic species can be expressed as, At =
t
At = Aorg (1 − exp(−kt))
(4)
0
According to Eq. (4), a linear relationship can be derived between ln(Aorg,t ) and t as follows, At = At − At−t = Aorg exp[−k(t − t)][1 − exp(−kt)] (5) Therefore, for constant t, ln(Aorg,t ) = a − kt
(6)
where a = ln(Aorg [1 − exp(−kt)]) + kt. A MHE-GC method is based on repeatedly extracting an equilibrated headspace from a sample vial and the relationship between amount of species in headspace and its extraction number can be described as [12,13], ln(An ) = ln(A0 ) − bn
(7)
with t = jn − 1
(8)
According to Eq. (7), the relationship of areas in the successive runs can be written as,
2. Methodology
An = An−1 eb
When adding a sample that contains both oxalate and other organic species into a permanganate-acidic solution at an elevated temperature, oxalate is completely converted to carbon dioxide within a short time (i.e., before the first headspace measurement) according to Reaction (I). However, it is a slow process for the carbon dioxide formation from the other organic species, which is an exponential decay function, i.e., morg [1 − exp(−kt)]. The carbon dioxide formation kinetics in such a reaction system can be described as,
Eq. (7) is not valid in a system involving chemical reaction [11], e.g., carbon dioxide formation from the other organic species. When using a headspace extraction number to replace the time, Eq. (2) can be rewritten as follows, i.e.,
mt = moxalate + morg [1 − exp(−kt)]=moxalate +
t
An = Aoxalate +
(9)
n
Aorg,n
At the first headspace GC measurement, A1 = Aoxalate + Aorg,1
mt
(1)
0
All symbols are defined in Section 6. At an elevated temperature, the carbon dioxide generated from the reactions is completely released to the headspace, which can be determined by headspace GC and described by
(10)
1
(11)
According to Eq. (9), at the second headspace GC measurement, A2 = A1 eb + Aorg,2
(12)
At nth headspace GC measurement, An = An−1 eb + Aorg,n
(13)
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
211
Thus, we can calculate peak area at any nth (n = 2, 3, . . ., n) headspace GC measurement, i.e., Aorg,n = An − An−1 eb
(14)
By conducting a MHE-GC measurement on a single sample vial, we can obtain A1 , A2 , . . ., An , thus, we can calculate the Aorg,2 , Aorg,3 , . . ., Aorg,n according to Eq. (14). With Eq. (8), we can convert Aorg,n to Aorg,tn . By plotting log(Aorg,t ) versus t (where t is from t2 to tn ), the intercept (a) and slope (k) in Eq. (6) can be obtained. From this equation, we can calculate the carbon dioxide formed by the other organic species at the first (t = t1 ) headspace GC measurement, i.e., Aorg,t1 . Then, we can obtain Aoxalate according to Eq. (11). Finally, we can calculate the amount of oxalate in the sample, i.e., Coxalate =
Aoxalate KVs
(15)
3. Experimental 3.1. Chemicals All chemicals used in the experiment were from commercial sources. The standard oxalate solution (0.250 mol/L) was prepared using sodium oxalate. An oxygen delignification liquor sample was withdrawn at the end of an oxygen delignification process. A synthetic oxygen delignification spent liquor was prepared by dissolving alkaline lignin in a sodium hydroxide (2 g/L) medium, in which the lignin content is 4 g/L. On the basis of this liquor, a set of solutions containing different amount of oxalate (from 4 to 20 mmol/L) was prepared.
Fig. 1. MHE-GC measurement on (a) methanol and formic acid and (b) oxalate and mill solution.
vial is immediately put into headspace sampler for automatic MHE-GC measurement. 4. Results and discussions
3.2. Apparatus and operation All measurements were carried out using an HP-7694 Automatic Headspace Sampler and Model HP-6890 capillary gas chromatograph equipped with a thermal conductivity detector (Hewlett-Packard, now Agilent Technologies, Palo Alto, CA, USA). GC conditions were as follows: capillary column with an I.D. = 0.53 mm and a length of 30 m (model GS-Q, J&W Scientific, Folsom, CA, USA) at 30 ◦ C; carrier gas helium flow rate of 3.1 mL/min. Headspace Sampler operating conditions were as follows: oven temperature of 70 ◦ C; vial pressurized by helium and pressurization time of 0.2 min; sample-loop fill time of 0.2 min; loop equilibration time of 0.05 min; loop fill time of 0.2 min; the vial equilibration time is 2 min and MHE recycle time is 3 min. 3.3. Sample preparation and measurement procedures Two milliliters 2 mol/L sulfuric acid solution and 0.5 mL 0.02 mol/L permanganate solution in a headspace test vial were added and sealed with a septum. A 50 L of sample liquor was taken in a microsyringe and injected into the sealed vial. After quickly shaking the vial for mixing the sample and reactants, the
4.1. Kinetics of formation of carbon dioxide from interfering species There are many organic species – mainly dissolved lignin – in mill streams that can react with permanganate; however, the carbon dioxide formation is caused by the low molecular weight species, such as methanol, which is found in both alkaline pulping and oxygen delignification spent liquors. Fig. 1a and b show the results of MHE-GC on the methanol, formic acid, oxalate solutions, and mill oxygen delignification liquor, respectively. For the oxalate solution, as carbon dioxide is generated rapidly at the given experimental conditions, the carbon dioxide in the headspace decreases exponentially with the headspace extraction number. For the methanol and formic acid solution, owing to a slow reaction with permanganate, the carbon dioxide in the headspace first increases and then decreases. It has been noticed that the slope of the exponential decrease for the methanol solution is the same as the oxalate solution after the third headspace extraction, indicating that no more carbon dioxide was formed after that time. In the selected mill liquor, there are many organic species besides oxalate. The carbon dioxide in the headspace at first headspace GC measurement is contributed
212
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
the interference species on the basis of the first two headspace measurements, i.e., Aorg,t1 =
t1 Aorg,t2 t2 − t1
(16)
In this way, it significantly reduces the experiment time. 4.3. Inorganic interferences The present method is based on measuring carbon dioxide released from the reaction between oxalate and permanganate according to Reaction (I). In an acidic medium, carbonate in the sample can be rapidly converted to carbon dioxide, which affects the oxalate determination. Thus, the carbonate effect must be taken into account. In this work, a blank test run, i.e., the sample added in an acidic solution without permanganate with the same headspace GC conditions, was conducted. Thus, the carbon dioxide effect caused by carbonate acidification can be subtracted. 4.4. MHE cycle time
Fig. 2. Kinetics of carbon dioxide formation from (a) methanol-permanganate reaction and (b) sample permanganate reaction.
In the present headspace sampler, the GC sampling is based on a pressuring mode, i.e., by injecting the helium (pressuring gas) into the vial to create a high vial pressure. In this way, the headspace sample can be released from the vial to the GC sample loop through venting owing to the pressure difference. However, each action needs a certain time for the control valve to open and shut. This does not allow for very short MHE cycle times to be used in the MHE-GC measurement. In this work, we set the equilibration and MHE cycle times at 2 and 3 min, respectively.
not only by a complete Reaction (I) from the oxalate, but also by the other organic species in the sample. The decrease of the headspace carbon dioxide in these MHE-GC measurements is slower than that of oxalate solution, indicating that a complete reaction has not been achieved yet. For the methanol and the oxygen delignification liquor sample, the formation of carbon dioxide at each headspace measurements as shown in Fig. 1, i.e., the Aorg,n , can be calculated using Eq. (9). The slope (b) in the equation can be obtained by the MHE-GC measurements for a pure oxalate solution testing. Aorg,n is further changed to Aorg,tn with Eq. (8) for plotting Aorg,t1 versus the reaction time t to obtain the kinetic formation curves for carbon dioxide, as shown in Fig. 2a and b, respectively. We can see that the reaction for methanol is relatively fast, but it is slow for the interference species in the oxygen delignification liquor at the given experimental conditions.
4.5. Sulfuric acid concentration
4.2. Organic interference correction
4.6. Temperature effect
As shown in Fig. 2a, the carbon dioxide formation rates from the interference species are relatively fast in the beginning of the reaction, in which the formation rates can be regarded as the same during the reaction within 5 min. Thus, we can use Eq. (14) to calculate the carbon dioxide formation caused by
A higher temperature is desired for achieving a rapid and complete reaction between the oxalate and permanganate. However, it was noticed that the higher temperature leads to a higher water vapor pressure, which may affect the separation performance for the given GC column. In this work, we chose 70 ◦ C as
In Reaction (I), the acid concentration is very important for the reaction rate. For a rapid reaction, a high concentration of sulfuric acid is essential. However, higher acid concentration can also promote carbon dioxide formation from the interference species. Although such effect can be corrected by the present method, it will lead to a larger measurement error when the oxalate content in the sample is relatively low. Conversely, if the reaction with the interference species is too fast, i.e., it has been completed before the second headspace measurement, the relationship described in Eq. (16) is not valid. In this work, a mild concentration (2 mol/L) of sulfuric acid was chosen. As shown in Fig. 1, a complete reaction of oxalate is still ensured at the first headspace measurement at the given experimental conditions.
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
213
with five data points (n = 5) and an R2 value of 0.9979; for the standard oxalate solution, the equation is
Table 1 Repeatability of the method Sample
A1
A2
A2
A1
Aoxalate + Acarbonate a
Aoxalate
1 2 3 4 5
67.9 67.7 69.8 66.3 66.9
102.4 103.3 105.8 100.5 101.9
36.3 37.4 37.9 36.0 36.8
24.2 24.9 25.2 24.0 24.5
43.7 42.8 44.6 42.3 42.4
20.6 19.7 21.5 19.2 19.3
2.3
5.0
RSD (%) a
The average GC signal from blank tests (for carbonate) is 23.1 (with a RSD of 3.5%).
the reaction temperature, in which a complete reaction of oxalate can be achieved within 2 min (at the first headspace measurement) as shown in Fig. 1. 4.7. Method precision and method validation The repeatability of the method was determined using a 50 L of a mill oxygen delignification liquor sample, in which the oxalate content is 12.5 mmol/L measured by the present method. Table 1 lists the GC measurement signals (A1 and A2 ) and the data (A2 , A1 and Aoxalate ) were calculated according to Eqs. (12), (16) and (11), respectively. The result shows that the average relative standard deviation (RSD) from five measurements is less than 5%, which is much better than that reported in the TAPPI Test Methods [9]. Since the sample liquor can directly be analyzed by the present method without any pretreatment such as dilution, and the retention time for carbon dioxide in each GC run is relatively short, it is much convenient and efficient than the current ion chromatographic method in oxalate quantification. The present method is also validated by a set of oxygen delignification liquors with the standard oxalate addition (from 0 to 0.040 mol/L). These samples were measured by the present method and results are shown in Fig. 3, in which a standard calibration curve of oxalate is also included. For the sample with standard oxalate addition, the equation is y = 21.52(±1.17) + 1782(±47.6)x
(17)
Fig. 3. Response curves of the sample liquors with standard oxalate addition and the standard oxalate solutions.
y = −1.38(±0.29) + 1764(±12.1)x
(18)
with five data points (n = 5) and an R2 value of 0.9999. The limit of quantitation in the present method is 2 mmol/L. It can also be seen that there is no significant difference (at a 95% confidence level) for slopes of these two calibration curves, thus, the effect of sample matrix is not significant in the present method. The maximum relative difference between the oxalate standard added and measured by MHE-GC and calculated by Eq. (17) is 7.1% when the sample contains 10.0 mmol/L of oxalate, indicating that the present method is justifiable. 5. Conclusions We have demonstrated a novel HS-GC technique for the determination of oxalate content in mill liquors, based on a rapid formation of carbon dioxide from the reaction between oxalate and permanganate. Using a multiple headspace extraction (MHE) measurement technique, the kinetics of formation of the carbon dioxide from the other organic species can be determined, so that the effect of the interference species on the oxalate measurement can be corrected. In addition, this method is automated, leading to higher accuracy and simplicity in measurement. 6. Nomenclature a At b j k K moxalate morg
intercept of Eq. (6) GC peak area at the reaction time t slope of Eq. (7) headspace extraction cycle time slope of Eq. (6) GC response factor mass of carbon dioxide converted from oxalate mass of carbon dioxide converted from the other organic species mt mass of the carbon dioxide at the reaction time t n headspace extraction number t reaction time t1 time at the first headspace measurement t2 time at the second headspace measurement tn time at the nth headspace measurement Vs sample volume added into the testing vial Aorg,1 carbon dioxide generated from the other organic species during a time period from zero to t1 Aorg,2 carbon dioxide generated from the other organic species between first and second headspace measurement during a time period from t1 − t2 Aorg,n carbon dioxide generated from the other organic species between the (n − 1)th and nth headspace measurement during a time period of tn−1 − tn t time interval between headspace extractions t=tn m integrated mass of carbon dioxide from the other t 0 organic species, which is = mt1 + mt2 + · · · + mtn
214
X.-S. Chai et al. / J. Chromatogr. A 1122 (2006) 209–214
Acknowledgements This work was partly supported by the Guangxi University Key Program for Science and Technology Research (2005 ZD01), Guangxi University Research Foundation (X061021), and National Natural Science Foundation of China (20367001). References [1] T.M. Grace, Survey of Evaporator Scaling in the Alkaline Pulp Industry, The Institute of Paper Chemistry, Appleton, WI, 1975. [2] W. Schmidl, W. Frederick, 1998 International Chemical Recovery Conference Proceedings, 1998, p. 367.
[3] P. Ulmgren, R. Radestrom, Nord. Pulp Pap. Res. J. 15 (2000) 128. [4] J.P. Romano, D.R. Salomom, Tappi Proceedings, 1992 Pulping Conference, 1992, p. 303. [5] J. Krishnagopalan, M. Hill, A.L. Fricke, Tappi J. 68 (1985) 108. [6] D.B. Easty, M.L. Borchardt, A.A. Webb, Pap. Puu 67 (1985) 501. [7] D.B. Easty, J.E. Johnson, Tappi J. 70 (1987) 109. [8] D.B. Easty, J.E. Johnson, A.A. Webb, Pap. Puu 68 (1986) 415. [9] TAPPI Test Method, Tappi Press, Atlanta, GA, 1987. [10] X.-S. Chai, Q. Luo, J.Y. Zhu, J. Chromatogr. A 909 (2001) 249. [11] X.-S. Chai, Q. Luo, J.Y. Zhu, J. Chromatogr. A 946 (2002) 177. [12] B. Kolb, L.S. Ettre, Static Headspace-Gas Chromatography – Theory and Practice, Wiley-VCH, New York, 1997. [13] X.-S. Chai, J.Y. Zhu, Anal. Chem. 70 (1998) 3481.