- Email: [email protected]
Mechanism of porous core electroosmotic pump flow injection system and its application to determination of chromium(VI) in waste-water
Talanta 51 (2000) 667 – 675 www.elsevier.com/locate/talanta
Mechanism of porous core electroosmotic pump flow injection system and its application to determination of chromium(VI) in waste-water Wu-er Gan a, Li Yang a, You-zhao He a,*, Rong-hui Zeng a, M. Luisa Cervera b, Miguel de la Guardia b b
a Department of Chemistry, Uni6ersity of Science and Technology of China, Hefei, Anhui 230026, PR China Department of Analytical Chemistry, Faculty of Chemistry, Uni6ersity of Valencia, 46100 Burjassot, Valencia, Spain
Received 29 June 1999; received in revised form 2 November 1999; accepted 2 November 1999
Abstract An electroosmotic pump flow injection system is introduced in this paper. According to electroosmotic theory, the pump’s properties were described. A large flow range (ml min − 1 – ml min − 1), moderate carrier pressure ( \ 0.15 MPa), reduced performance voltage ( B500 V) and stable flow rate (RSD B 3.0% in 4 h) are the main properties of the pump. NH4OH (0.35 mM) was used as carrier for improving the pump’s flow stability. The electroosmotic efficiency of the pump’s medium, porous core, can be recovered and regenerated. A sandwich zone was used for sample and reagent introduction in order to adapt to the pump performance. Flow injection-spectrophotometry was employed for the determination of Cr(VI) in waste-water, based on the formation of the complex with 1,5-diphenylcarbazide and absorbance measurement at 540 nm. Within the calibration range of 0 – 7.0 mg l − 1 of Cr(VI), the RSD was 0.4% (n= 5). The recovery of 0.70 mg l − 1 Cr(VI) added to the waste-water sample was 94.5 92.0% (n= 5). © 2000 Elsevier Science B.V. All rights reserved. Keywords: Electroosmotic pump; Flow injection analysis; Chromium(VI)
1. Introduction Among available delivery equipment for flow injection analysis (FIA), the multiroller peristaltic pump, which can accommodate several channels and obtain different volumetric flow rate, is the most common means of propelling carrier and * Corresponding author. Fax: +86-551-3631760. E-mail address: [email protected] (Y.-z. He)
reagent solutions. The major drawbacks of this pump are its pulsed flow, especially at low flow rate, and its flow reproducibility problem due to its initial roller position or tubing’s working time. Harrison and Manz and co-workers [1–7] developed several analytical micro-systems in planar microchips. Dasgupta and Liu [8,9] established a FIA micro-system with a capillary electroosmotic pump. All these analytical systems adopted electroosmotic flow (EOF) in driving sample and
0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 9 9 ) 0 0 3 2 0 - 3
668
W.-e. Gan et al. / Talanta 51 (2000) 667–675
reagent solutions. The electroosmotic effect can control the flow rate smoothly and change the flow direction easily in the flow analysis systems. As the EOF is generated in a single capillary of which the inner diameter is tens of micrometers, the EOF rate and working pressure appear to be very low, while its working voltage is rather high. The FIA system with a single electroosmotic pump was successfully employed in the determination of fluoride in tap water [10]. The aim of this paper was to improve the FIA system by employing two home-made electroosmotic pumps [11], to investigate the specific properties of the pump’s FIA system and to apply that in the determination of chromium(VI) in waste-water.
2. Experimental
2.1. Reagents All reagents were of analytical reagent grade. Ultrapure water was double distilled and obtained from a quartz distiller. The chromium(VI) stock solution was 1.000 g l − 1 Cr(VI) and was prepared by dissolving K2Cr2O7 (Chemical Factory, Beijing, China) in distilled water.
1,5-Diphenylcarbazide (Chem. Reagent Co., Shanghai, China) was used as the chromogenic reagent for Cr(VI) determination. 1,5-Diphenylcarbazide solution (0.5 g l − 1) was prepared by dissolving 0.125 g in 25 ml of acetone. After the solution became clear, it was diluted to 250 ml with distilled water and kept in a refrigerator. Phosphoric acid (Chemical Factory, Beijing, China) was selected as the acid and masking reagent. Its concentration, both in sample and calibration solutions, was 0.10 M.
2.2. Instrumentation The electroosmotic pump as shown in Fig. 1 consisted of a pump chamber, a piece of porous core (Pc) in the middle of the chamber and two electrode cavities (Ec) on both sides of the porous core. The porous core was made from boric glass powder by high temperature sintering. Its dimensions were 35 mm in diameter, 13 mm thick, and its pore size was 2–5 mm i.d. The electrode cavity with small holes or slits on its cylinder side was composed of a 0.3-mm platinum electrode (El) fixed on its top and a tubular cellulose acetate membrane, 0.2 mm (Chemical School, Beijing, China), fixed on its inner wall to isolate tiny bubbles and reduce the pH influence resulting
Fig. 1. Scheme of porous core electroosmotic pump [11]. Pc, porous core; El, electrode; Ec, electrode chamber; F, EOF flow; M, membrane.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
669
Fig. 2. Schematic diagram of manifold in electroosmotic pump FIA system.
from the electrolysis reaction. The working voltage was provided by a DY-74 II electrophoresis power supply from Electronic Anal. Instr. Factory (Nantong, Jiangsu, China). The flow direction of the pump can be controlled by changing the polarity of the power supply, and its flow rate can be adjusted by varying the voltage of the power supply. A schematic diagram of the determination flow injection system is shown in Fig. 2. It consisted of two porous core electroosmotic pumps, one eightway valve from Zhaofa Autom. Anal. Inst. (Shenyang, Liaoning, China) and one quartz flow cell, 10-mm optical path and 2 mm i.d., from Zhaofa Autom. Anal. Inst. (Shenyang, Liaoning, China), with which a 721 spectrophotometer (Anal. Instr. Factory, Shanghai, China) was employed as an on-line detector (D) for Cr(VI) determination at 540 nm. Polytetrafluoroethylene (PTFE; 0.5 mm i.d.) was used for the FI conduits. Two reagent loops (RL1, RL2), one sample loop (SL) and one knotted reaction coil (RC) were 60, 60, 20 and 70 cm, respectively. In Fig. 2, R represents 0.5 g l − 1 1,5-diphenylcarbazide solution and S represents the sample or calibration solutions. The sample zone was sand-
wiched with two reagent zones and the Cr(VI) complex was formed by dispersion between the sample and reagents. After passing through the reaction coil, the complex was on-line determined by the spectrophotometer at 540 nm.
2.3. Procedure First, the valve was turned to the filling position. The flow rate of pump 1 (P1) was about 0.9 ml min − 1, corresponding to +100 V, and that of pump 2 ( P2), 2.8 ml min − 1, corresponding to − 300 V. Pump 1 drove the carrier to the reaction coil and detector. Pump 2 drew the sample and reagent solutions into one sample and two reagent loops. A stock tube (ST), 1 m length and 2.0 mm i.d., was installed between the loops and pump 2 to prevent the solutions from drawing into the pump. During the second step, the value changed its position to make the zones in the order of carrier–reagent–sample–reagent–carrier. Pump 1 had the same flow rate and direction as the first step, and propelled the sandwich zone to the reaction coil and detector. Pump 2 kept its flow rate, but changed its flow direction to wash the stock tube and conduits with the carrier.
670
W.-e. Gan et al. / Talanta 51 (2000) 667–675
3. Results and discussion
3.1. Selection of carrier solution The velocity of EOF generated in the electroosmotic pump depends on the electrodynamic potential, namely, zeta potential and the electric field strength, Veo =ozE/h
(1)
where Veo is the velocity of EOF, z is the zeta potential, E is electric field strength, and o and h are dielectric constant and viscosity coefficient of carrier. According to Stern theory for double electric layer of surface, the electric potential out of the Stern plane can be in conformity with Gouy– Chapman diffuse theory. The z potential is approximately equal to Stern plane potential (Cd), which is relative to the surface charge density of the Stern plane and the thickness of the double electric layer [12]. When the carrier possesses a single pair of ions sd = 2okT sinh[zeCd/(2kT)]/(zek − 1)
(2)
where sd is the surface charge density of the Stern plane, k is the Boltzmann constant, T is absolute temperature, z is ionic valence, e is electron charge, Cd is Stern plane potential and k − 1 is thickness of the double electric layer, which is inversely proportional to the ionic valence and the square root of the ionic concentration [12]. k − 1 = (okT/2n0z 2e 2)1/2
(3)
where n0 is the ionic concentration of carrier. If z potential is approximately equal to the Stern plane potential and Stern plane potential is low enough, Eq. (2) can be simplified as Eq. (4). sd = oCd/k − 1 :oz/k − 1
(4)
So a higher EOF is observed when a neutral carrier solution has lower ionic strength. Water can be the best candidate in the pump carrier selection. Our experimental results also showed the pump flow rate reduced gradually when pure water was used as the carrier. On the porous
core surface, its charge is negative in water because of silicate ionization. Hydrated hydrogen ions produced by the electrolysis near the anode will reduce the negative charge of the surface and the surface charge density during their migration. So the EOF is pH-dependent and its flow rate increases with pH [8]. In order to obtain a stable EOF rate and high EOF efficiency, the selection of an appropriate carrier solution is an important procedure. Since the suitable concentration of carrier solution can be very low (mM) and some basic buffer systems, such as phosphate and tetraborate buffer, etc. could not improve their efficiency or stability, several basic electrolyte solutions were investigated in this paper. The EOF rate and efficiency of carrier solutions at different concentration levels, from 0.1 to 1000 mM, are listed in Table 1. When the concentrations of carrier solutions increase, both the ionic strength and the pH of carrier solutions increase. In Table 1, we find that the EOF rate is almost invariant at low concentration, but the EOF rates increase at higher concentrations. As we have discussed above, the ionic strength and pH of carrier solutions have the reverse effect on EOF rate. It indicates that the ionic strength and pH have almost an equal effect on EOF rate at low concentration and the effect of pH on EOF rate is greater than ionic strength at higher concentration. But the EOF efficiency and stability will become poor at higher concentration. The EOF efficiency (R) is defined as the ratio of the EOF rate to the pump working current R= Q/I
(5)
where Q is EOF rate and I is working current. Based on the results, we chose NH4OH as the pump carrier because its flow rate and current were more stable than the other high efficiency solutions, among which 1.0 mM NaOH and Na2B4O7 both had a stability decrease as time went on. Fig. 3 shows the effect of NH4OH concentration, from 1.0 to 1000 mM, on the EOF rate and efficiency. The highest electroosmotic efficiency was observed at 0.35 mM NH4OH.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
3.2. Flow rate and efficiency
671
With 0.35 mM NH4OH, the flow rate exhibits an approximately linear relationship from 50 to 500 V. So the flow rate can easily be modified with the working voltage. For instance, the flow rates
The EOF rate is dependent on the electric field strength or the applied voltage as shown in Fig. 4.
Table 1 Effect of electrolytes and their concentration on electroosmotic current, flow rate and efficiency of electroosmotic pumpa Solution
NaOH
NH4OH
Na2SiO3
Na2B4O7
K2HPO4
a
Parameter
I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I)
Concentration ( mM) 0
0.1
1.0
10
100
1000
1.0 0.6 0.6 1.1 0.7 0.6 1.1 0.6 0.6 1.1 0.6 0.6 1.2 0.7 0.6
1.2 0.5 0.4 – – – 1.1 0.6 0.6 1.2 0.7 0.6 1.4 0.7 0.5
1.3 0.5 0.4 1.2 0.7 0.6 1.2 0.6 0.5 1.2 0.7 0.6 1.7 0.6 0.3
1.3 0.6 0.5 1.3 0.7 0.6 1.3 0.7 0.6 1.2 0.7 0.6 1.9 0.7 0.4
1.4 1.0 0.7 1.3 0.9 0.7 1.5 1.2 0.8 1.5 1.0 0.7 2.5 1.3 0.5
4.3 7.0 1.6 4.2 5.2 1.2 8.5 7.0 0.8 2.5 3.3 1.3 15.5 6.2 0.4
The working voltage was 500 V in all cases.
Fig. 3. Effect of NH4OH concentration on EOF rate and efficiency. The working voltage was 500 V.
672
W.-e. Gan et al. / Talanta 51 (2000) 667–675
Fig. 4. Effect of voltage on EOF rate and efficiency. The carrier solution was 0.35 mM NH4OH.
varied from 0.25 to 3.0 ml min − 1 when the electroosmotic voltage was adjusted from 50 to 500 V. The flow rate can also be raised by using a larger cross section of the porous core than that employed here or a higher applied voltage than 500 V. Fig. 4 indicates the effect of the working voltage on the EOF efficiency. The increase in current, which is much faster than EOF rate, is the nonlinear reason for the EOF efficiency. From our experimental results, we find the best electroosmotic efficiency was obtained at about 200 V. The EOF efficiency is dependent on the condition of the porous surface. We found that the efficiency decreased slightly along with the working time while the flow rate was practically constant over 4 h. In order to recover the efficiency, the porous core can be washed with 0.1 M NaOH and distilled water by a suction pump weekly. If the porous core is contaminated by reagent ions, the regeneration procedure should be carried out. The porous core is washed with 0.1 M Na2CO3, 1.0 M HCl and 0.1 M NaOH in turn, and with ultrapure water after each solution.
3.3. Pump pressure For the FIA system, a suitable carrier pressure is necessary, so that the flow rate variation caused by any small change in back-pressure will be negligible. According to our experimental results as shown in Fig. 5, the electroosmotic pressure was relative to the pore diameter. The higher
Fig. 5. Effect of pore size on pressure. The working voltage was 400 V.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
673
Fig. 6. Effect of electroosmotic voltage on pump pressure. Other parameters were the same as in Fig. 4. Table 2 Effect of back-pressure on flow ratea Back-pressure (MPa)
Flow rate (ml min−1) a
0
0.01
0.02
0.03
0.04
2.85
2.85
2.80
2.75
2.60
The working voltage was 500 V and the carrier solution was 0.35 mM NH4OH.
pressure was obtained with the core that has the smaller pore diameter. The electroosmotic effect is generated by the migration of counter ions in the diffuse layer with an external electric field. If the thickness of the double electric layer stays the same, a larger relative electroosmotic area will be provided by a smaller pore size. As the higher pressure is obtained from the core with smaller pore, the length of the electroosmotic medium can be reduced to 13 mm. On the condition of the same electric field strength, the working voltage can become lower and safer. The effect of the pump working voltage on its carrier pressure is shown in Fig. 6. The pressure could be up to 0.15 MPa as the voltage was 500 V. The pump backpressure influence on its flow rate is listed in Table 2. The results indicate that the flow rate reduced
by 3.5% when the back-pressure increased to 30 kPa.
3.4. Parameters of flow injection system The effect of the reaction coil length on the absorbance of chromium(VI) complex is shown in Fig. 7. By increasing the reaction coil length from 30 to 70 cm, the absorbance of chromium(VI) complex also increased. But, when the length of the coil was longer than 70 cm, the absorbance tended to an invariable absorbance. Throughout our determination, 70 cm knotted coil was chosen for the color development reaction. Sample loops of 11, 18, 26, 52, 124 and 300 cm were studied, and the half volume of the FIA system was about 60 ml for 30 cm length. A 20-cm sample loop was
674
W.-e. Gan et al. / Talanta 51 (2000) 667–675
finally selected in the determination. For the injection of sandwich zone, when the length of one reagent loop was longer than three times the sample loop, a high and fixed absorbance signal was observed in the determination. Here, two 60-cm reagent loops were adopted, so the function of sandwich reagent zones was almost the same as that of reagent carrier.
3.5. Determination of chromium(VI) in waste-water The mechanism of the chromogenic reaction between chromium(VI) and 1,5-diphenylcarbazide in acid medium has been described in the literature. The complex has its peak absorbance at 540 nm. In this paper, the chromium(VI) complex was determined by the electroosmotic pump flow injection system as mentioned above. The calibration range for Cr(VI) was 0 – 7.0 mg l − 1 with a RSD of about 0.4% (n = 5). The linear regression equation for the calibration range was A= 0.106 [Cr(VI)]+ 0.001 with a correlation coefficient of 0.9999. The concentration of chromium(VI) in a waste-water sample was 1.1290.02 mg l − 1 (n= 5). The recovery of 0.70 mg l − 1 chromium(VI)
added to the waste-water sample was 94.59 2.0% (n= 5).
4. Conclusions The FIA system with two home-made electroosmotic pumps provides a series of advantages such as a reduced working voltage (B 500 V), low electric power (maximum power consumption 5 W), moderate performance pressure (\0.15 MPa), large flow range (10 ml min − 1 –5.0 ml min − 1) and stable flow rate (RSDB 3.0%, 4 h), especially using diluted aqueous ammonia as the carrier solution. The pump also drives the carrier flow steadily and will not generate pulsed flow even at a very low flow rate. In addition, the pump’s structure and its operation are quite simple and convenient. The porous core, an electroosmotic medium, can be replaced and regenerated. As reagents cannot pass through the pump, a sandwich zone was adopted in the FIA system, and the reagent consumption and discharge was also reduced. Aqueous ammonia (0.35 mM) has higher electroosmotic efficiency than other electrolytic solutions, and was used as the
Fig. 7. Effect of reaction coil length on absorbance of Cr(VI) complex. The concentration of Cr(VI) was 2.1 mg l − 1, and the carrier flow rate was 0.9 ml min − 1, 0.35 mM NH4OH.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
pump carrier. So the sample and reagent solutions are first drawn into the sample and reagent loops, and then the sandwich zone is injected to pass through the reaction coil and detector. As a result, this FIA system will reduce reagent consumption and discharge further. The characteristics of the electroosmotic pump can be improved further. Investigations on materials of electroosmotic medium, carrier additives and organic solvent carriers will also be topics of further research.
[3] [4] [5]
[6] [7] [8] [9] [10]
References [1] D.J. Harrison, A. Manz, Z.H. Fan, H. Luedi, H.M. Widmer, Anal. Chem. 64 (1992) 11926. [2] A. Manz, D.J. Harrison, E.M.J. Verpoorte, J.C. Fet-
.
[11] [12]
675
tinger, A. Paulus, H. Luedi, H.M. Widmer, J. Chromatogr. 593 (1992) 253. D.J. Harrison, K. Fluri, K. Seller, Z.H. Fan, C.S. Effenhauser, A. Manz, Science 261 (1993) 895. K. Seller, D.J. Harrison, A. Manz, Anal. Chem. 65 (1993) 1481. A. Manz, E. Verpoorte, C.S. Effenhauser, N. Burggraf, D.E. Raymond, A.J. Harrison, H.M. Widmer, J. High Resolut. Chromatogr. 16 (1993) 433. Z.H. Fan, D.J. Harrison, Anal. Chem. 66 (1994) 177. K. Seller, Z.H. Fan, K. Fluri, D.J. Harrison, Anal. Chem. 66 (1994) 3485. P.K. Dasgupta, S.R. Liu, Anal. Chem. 66 (1994) 1792. S.R. Liu, P.K. Dasgupta, Talanta 41 (1994) 1903. Y.Z. He, W.E. Gan, M. Zhang, M.Z. Zheng, R.H. Zeng, G. Jin, Fenxi Huaxue 26 (1998) 125. Y.Z. He, W.E. Gan, Chinese Patent ZL 97212126.9, 1998. T.R. Gu, Y. Zhu, W.L. Li, J.M. Ma, L.R. Dai, H.M. Chen, Surface Chemistry, Science Press, Beijing, 1994, p. 1344.
Mechanism of porous core electroosmotic pump flow injection system and its application to determination of chromium(VI) in waste-water Wu-er Gan a, Li Yang a, You-zhao He a,*, Rong-hui Zeng a, M. Luisa Cervera b, Miguel de la Guardia b b
a Department of Chemistry, Uni6ersity of Science and Technology of China, Hefei, Anhui 230026, PR China Department of Analytical Chemistry, Faculty of Chemistry, Uni6ersity of Valencia, 46100 Burjassot, Valencia, Spain
Received 29 June 1999; received in revised form 2 November 1999; accepted 2 November 1999
Abstract An electroosmotic pump flow injection system is introduced in this paper. According to electroosmotic theory, the pump’s properties were described. A large flow range (ml min − 1 – ml min − 1), moderate carrier pressure ( \ 0.15 MPa), reduced performance voltage ( B500 V) and stable flow rate (RSD B 3.0% in 4 h) are the main properties of the pump. NH4OH (0.35 mM) was used as carrier for improving the pump’s flow stability. The electroosmotic efficiency of the pump’s medium, porous core, can be recovered and regenerated. A sandwich zone was used for sample and reagent introduction in order to adapt to the pump performance. Flow injection-spectrophotometry was employed for the determination of Cr(VI) in waste-water, based on the formation of the complex with 1,5-diphenylcarbazide and absorbance measurement at 540 nm. Within the calibration range of 0 – 7.0 mg l − 1 of Cr(VI), the RSD was 0.4% (n= 5). The recovery of 0.70 mg l − 1 Cr(VI) added to the waste-water sample was 94.5 92.0% (n= 5). © 2000 Elsevier Science B.V. All rights reserved. Keywords: Electroosmotic pump; Flow injection analysis; Chromium(VI)
1. Introduction Among available delivery equipment for flow injection analysis (FIA), the multiroller peristaltic pump, which can accommodate several channels and obtain different volumetric flow rate, is the most common means of propelling carrier and * Corresponding author. Fax: +86-551-3631760. E-mail address: [email protected] (Y.-z. He)
reagent solutions. The major drawbacks of this pump are its pulsed flow, especially at low flow rate, and its flow reproducibility problem due to its initial roller position or tubing’s working time. Harrison and Manz and co-workers [1–7] developed several analytical micro-systems in planar microchips. Dasgupta and Liu [8,9] established a FIA micro-system with a capillary electroosmotic pump. All these analytical systems adopted electroosmotic flow (EOF) in driving sample and
0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 9 9 ) 0 0 3 2 0 - 3
668
W.-e. Gan et al. / Talanta 51 (2000) 667–675
reagent solutions. The electroosmotic effect can control the flow rate smoothly and change the flow direction easily in the flow analysis systems. As the EOF is generated in a single capillary of which the inner diameter is tens of micrometers, the EOF rate and working pressure appear to be very low, while its working voltage is rather high. The FIA system with a single electroosmotic pump was successfully employed in the determination of fluoride in tap water [10]. The aim of this paper was to improve the FIA system by employing two home-made electroosmotic pumps [11], to investigate the specific properties of the pump’s FIA system and to apply that in the determination of chromium(VI) in waste-water.
2. Experimental
2.1. Reagents All reagents were of analytical reagent grade. Ultrapure water was double distilled and obtained from a quartz distiller. The chromium(VI) stock solution was 1.000 g l − 1 Cr(VI) and was prepared by dissolving K2Cr2O7 (Chemical Factory, Beijing, China) in distilled water.
1,5-Diphenylcarbazide (Chem. Reagent Co., Shanghai, China) was used as the chromogenic reagent for Cr(VI) determination. 1,5-Diphenylcarbazide solution (0.5 g l − 1) was prepared by dissolving 0.125 g in 25 ml of acetone. After the solution became clear, it was diluted to 250 ml with distilled water and kept in a refrigerator. Phosphoric acid (Chemical Factory, Beijing, China) was selected as the acid and masking reagent. Its concentration, both in sample and calibration solutions, was 0.10 M.
2.2. Instrumentation The electroosmotic pump as shown in Fig. 1 consisted of a pump chamber, a piece of porous core (Pc) in the middle of the chamber and two electrode cavities (Ec) on both sides of the porous core. The porous core was made from boric glass powder by high temperature sintering. Its dimensions were 35 mm in diameter, 13 mm thick, and its pore size was 2–5 mm i.d. The electrode cavity with small holes or slits on its cylinder side was composed of a 0.3-mm platinum electrode (El) fixed on its top and a tubular cellulose acetate membrane, 0.2 mm (Chemical School, Beijing, China), fixed on its inner wall to isolate tiny bubbles and reduce the pH influence resulting
Fig. 1. Scheme of porous core electroosmotic pump [11]. Pc, porous core; El, electrode; Ec, electrode chamber; F, EOF flow; M, membrane.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
669
Fig. 2. Schematic diagram of manifold in electroosmotic pump FIA system.
from the electrolysis reaction. The working voltage was provided by a DY-74 II electrophoresis power supply from Electronic Anal. Instr. Factory (Nantong, Jiangsu, China). The flow direction of the pump can be controlled by changing the polarity of the power supply, and its flow rate can be adjusted by varying the voltage of the power supply. A schematic diagram of the determination flow injection system is shown in Fig. 2. It consisted of two porous core electroosmotic pumps, one eightway valve from Zhaofa Autom. Anal. Inst. (Shenyang, Liaoning, China) and one quartz flow cell, 10-mm optical path and 2 mm i.d., from Zhaofa Autom. Anal. Inst. (Shenyang, Liaoning, China), with which a 721 spectrophotometer (Anal. Instr. Factory, Shanghai, China) was employed as an on-line detector (D) for Cr(VI) determination at 540 nm. Polytetrafluoroethylene (PTFE; 0.5 mm i.d.) was used for the FI conduits. Two reagent loops (RL1, RL2), one sample loop (SL) and one knotted reaction coil (RC) were 60, 60, 20 and 70 cm, respectively. In Fig. 2, R represents 0.5 g l − 1 1,5-diphenylcarbazide solution and S represents the sample or calibration solutions. The sample zone was sand-
wiched with two reagent zones and the Cr(VI) complex was formed by dispersion between the sample and reagents. After passing through the reaction coil, the complex was on-line determined by the spectrophotometer at 540 nm.
2.3. Procedure First, the valve was turned to the filling position. The flow rate of pump 1 (P1) was about 0.9 ml min − 1, corresponding to +100 V, and that of pump 2 ( P2), 2.8 ml min − 1, corresponding to − 300 V. Pump 1 drove the carrier to the reaction coil and detector. Pump 2 drew the sample and reagent solutions into one sample and two reagent loops. A stock tube (ST), 1 m length and 2.0 mm i.d., was installed between the loops and pump 2 to prevent the solutions from drawing into the pump. During the second step, the value changed its position to make the zones in the order of carrier–reagent–sample–reagent–carrier. Pump 1 had the same flow rate and direction as the first step, and propelled the sandwich zone to the reaction coil and detector. Pump 2 kept its flow rate, but changed its flow direction to wash the stock tube and conduits with the carrier.
670
W.-e. Gan et al. / Talanta 51 (2000) 667–675
3. Results and discussion
3.1. Selection of carrier solution The velocity of EOF generated in the electroosmotic pump depends on the electrodynamic potential, namely, zeta potential and the electric field strength, Veo =ozE/h
(1)
where Veo is the velocity of EOF, z is the zeta potential, E is electric field strength, and o and h are dielectric constant and viscosity coefficient of carrier. According to Stern theory for double electric layer of surface, the electric potential out of the Stern plane can be in conformity with Gouy– Chapman diffuse theory. The z potential is approximately equal to Stern plane potential (Cd), which is relative to the surface charge density of the Stern plane and the thickness of the double electric layer [12]. When the carrier possesses a single pair of ions sd = 2okT sinh[zeCd/(2kT)]/(zek − 1)
(2)
where sd is the surface charge density of the Stern plane, k is the Boltzmann constant, T is absolute temperature, z is ionic valence, e is electron charge, Cd is Stern plane potential and k − 1 is thickness of the double electric layer, which is inversely proportional to the ionic valence and the square root of the ionic concentration [12]. k − 1 = (okT/2n0z 2e 2)1/2
(3)
where n0 is the ionic concentration of carrier. If z potential is approximately equal to the Stern plane potential and Stern plane potential is low enough, Eq. (2) can be simplified as Eq. (4). sd = oCd/k − 1 :oz/k − 1
(4)
So a higher EOF is observed when a neutral carrier solution has lower ionic strength. Water can be the best candidate in the pump carrier selection. Our experimental results also showed the pump flow rate reduced gradually when pure water was used as the carrier. On the porous
core surface, its charge is negative in water because of silicate ionization. Hydrated hydrogen ions produced by the electrolysis near the anode will reduce the negative charge of the surface and the surface charge density during their migration. So the EOF is pH-dependent and its flow rate increases with pH [8]. In order to obtain a stable EOF rate and high EOF efficiency, the selection of an appropriate carrier solution is an important procedure. Since the suitable concentration of carrier solution can be very low (mM) and some basic buffer systems, such as phosphate and tetraborate buffer, etc. could not improve their efficiency or stability, several basic electrolyte solutions were investigated in this paper. The EOF rate and efficiency of carrier solutions at different concentration levels, from 0.1 to 1000 mM, are listed in Table 1. When the concentrations of carrier solutions increase, both the ionic strength and the pH of carrier solutions increase. In Table 1, we find that the EOF rate is almost invariant at low concentration, but the EOF rates increase at higher concentrations. As we have discussed above, the ionic strength and pH of carrier solutions have the reverse effect on EOF rate. It indicates that the ionic strength and pH have almost an equal effect on EOF rate at low concentration and the effect of pH on EOF rate is greater than ionic strength at higher concentration. But the EOF efficiency and stability will become poor at higher concentration. The EOF efficiency (R) is defined as the ratio of the EOF rate to the pump working current R= Q/I
(5)
where Q is EOF rate and I is working current. Based on the results, we chose NH4OH as the pump carrier because its flow rate and current were more stable than the other high efficiency solutions, among which 1.0 mM NaOH and Na2B4O7 both had a stability decrease as time went on. Fig. 3 shows the effect of NH4OH concentration, from 1.0 to 1000 mM, on the EOF rate and efficiency. The highest electroosmotic efficiency was observed at 0.35 mM NH4OH.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
3.2. Flow rate and efficiency
671
With 0.35 mM NH4OH, the flow rate exhibits an approximately linear relationship from 50 to 500 V. So the flow rate can easily be modified with the working voltage. For instance, the flow rates
The EOF rate is dependent on the electric field strength or the applied voltage as shown in Fig. 4.
Table 1 Effect of electrolytes and their concentration on electroosmotic current, flow rate and efficiency of electroosmotic pumpa Solution
NaOH
NH4OH
Na2SiO3
Na2B4O7
K2HPO4
a
Parameter
I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I) I (mA) F (ml min−1) R (F/I)
Concentration ( mM) 0
0.1
1.0
10
100
1000
1.0 0.6 0.6 1.1 0.7 0.6 1.1 0.6 0.6 1.1 0.6 0.6 1.2 0.7 0.6
1.2 0.5 0.4 – – – 1.1 0.6 0.6 1.2 0.7 0.6 1.4 0.7 0.5
1.3 0.5 0.4 1.2 0.7 0.6 1.2 0.6 0.5 1.2 0.7 0.6 1.7 0.6 0.3
1.3 0.6 0.5 1.3 0.7 0.6 1.3 0.7 0.6 1.2 0.7 0.6 1.9 0.7 0.4
1.4 1.0 0.7 1.3 0.9 0.7 1.5 1.2 0.8 1.5 1.0 0.7 2.5 1.3 0.5
4.3 7.0 1.6 4.2 5.2 1.2 8.5 7.0 0.8 2.5 3.3 1.3 15.5 6.2 0.4
The working voltage was 500 V in all cases.
Fig. 3. Effect of NH4OH concentration on EOF rate and efficiency. The working voltage was 500 V.
672
W.-e. Gan et al. / Talanta 51 (2000) 667–675
Fig. 4. Effect of voltage on EOF rate and efficiency. The carrier solution was 0.35 mM NH4OH.
varied from 0.25 to 3.0 ml min − 1 when the electroosmotic voltage was adjusted from 50 to 500 V. The flow rate can also be raised by using a larger cross section of the porous core than that employed here or a higher applied voltage than 500 V. Fig. 4 indicates the effect of the working voltage on the EOF efficiency. The increase in current, which is much faster than EOF rate, is the nonlinear reason for the EOF efficiency. From our experimental results, we find the best electroosmotic efficiency was obtained at about 200 V. The EOF efficiency is dependent on the condition of the porous surface. We found that the efficiency decreased slightly along with the working time while the flow rate was practically constant over 4 h. In order to recover the efficiency, the porous core can be washed with 0.1 M NaOH and distilled water by a suction pump weekly. If the porous core is contaminated by reagent ions, the regeneration procedure should be carried out. The porous core is washed with 0.1 M Na2CO3, 1.0 M HCl and 0.1 M NaOH in turn, and with ultrapure water after each solution.
3.3. Pump pressure For the FIA system, a suitable carrier pressure is necessary, so that the flow rate variation caused by any small change in back-pressure will be negligible. According to our experimental results as shown in Fig. 5, the electroosmotic pressure was relative to the pore diameter. The higher
Fig. 5. Effect of pore size on pressure. The working voltage was 400 V.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
673
Fig. 6. Effect of electroosmotic voltage on pump pressure. Other parameters were the same as in Fig. 4. Table 2 Effect of back-pressure on flow ratea Back-pressure (MPa)
Flow rate (ml min−1) a
0
0.01
0.02
0.03
0.04
2.85
2.85
2.80
2.75
2.60
The working voltage was 500 V and the carrier solution was 0.35 mM NH4OH.
pressure was obtained with the core that has the smaller pore diameter. The electroosmotic effect is generated by the migration of counter ions in the diffuse layer with an external electric field. If the thickness of the double electric layer stays the same, a larger relative electroosmotic area will be provided by a smaller pore size. As the higher pressure is obtained from the core with smaller pore, the length of the electroosmotic medium can be reduced to 13 mm. On the condition of the same electric field strength, the working voltage can become lower and safer. The effect of the pump working voltage on its carrier pressure is shown in Fig. 6. The pressure could be up to 0.15 MPa as the voltage was 500 V. The pump backpressure influence on its flow rate is listed in Table 2. The results indicate that the flow rate reduced
by 3.5% when the back-pressure increased to 30 kPa.
3.4. Parameters of flow injection system The effect of the reaction coil length on the absorbance of chromium(VI) complex is shown in Fig. 7. By increasing the reaction coil length from 30 to 70 cm, the absorbance of chromium(VI) complex also increased. But, when the length of the coil was longer than 70 cm, the absorbance tended to an invariable absorbance. Throughout our determination, 70 cm knotted coil was chosen for the color development reaction. Sample loops of 11, 18, 26, 52, 124 and 300 cm were studied, and the half volume of the FIA system was about 60 ml for 30 cm length. A 20-cm sample loop was
674
W.-e. Gan et al. / Talanta 51 (2000) 667–675
finally selected in the determination. For the injection of sandwich zone, when the length of one reagent loop was longer than three times the sample loop, a high and fixed absorbance signal was observed in the determination. Here, two 60-cm reagent loops were adopted, so the function of sandwich reagent zones was almost the same as that of reagent carrier.
3.5. Determination of chromium(VI) in waste-water The mechanism of the chromogenic reaction between chromium(VI) and 1,5-diphenylcarbazide in acid medium has been described in the literature. The complex has its peak absorbance at 540 nm. In this paper, the chromium(VI) complex was determined by the electroosmotic pump flow injection system as mentioned above. The calibration range for Cr(VI) was 0 – 7.0 mg l − 1 with a RSD of about 0.4% (n = 5). The linear regression equation for the calibration range was A= 0.106 [Cr(VI)]+ 0.001 with a correlation coefficient of 0.9999. The concentration of chromium(VI) in a waste-water sample was 1.1290.02 mg l − 1 (n= 5). The recovery of 0.70 mg l − 1 chromium(VI)
added to the waste-water sample was 94.59 2.0% (n= 5).
4. Conclusions The FIA system with two home-made electroosmotic pumps provides a series of advantages such as a reduced working voltage (B 500 V), low electric power (maximum power consumption 5 W), moderate performance pressure (\0.15 MPa), large flow range (10 ml min − 1 –5.0 ml min − 1) and stable flow rate (RSDB 3.0%, 4 h), especially using diluted aqueous ammonia as the carrier solution. The pump also drives the carrier flow steadily and will not generate pulsed flow even at a very low flow rate. In addition, the pump’s structure and its operation are quite simple and convenient. The porous core, an electroosmotic medium, can be replaced and regenerated. As reagents cannot pass through the pump, a sandwich zone was adopted in the FIA system, and the reagent consumption and discharge was also reduced. Aqueous ammonia (0.35 mM) has higher electroosmotic efficiency than other electrolytic solutions, and was used as the
Fig. 7. Effect of reaction coil length on absorbance of Cr(VI) complex. The concentration of Cr(VI) was 2.1 mg l − 1, and the carrier flow rate was 0.9 ml min − 1, 0.35 mM NH4OH.
W.-e. Gan et al. / Talanta 51 (2000) 667–675
pump carrier. So the sample and reagent solutions are first drawn into the sample and reagent loops, and then the sandwich zone is injected to pass through the reaction coil and detector. As a result, this FIA system will reduce reagent consumption and discharge further. The characteristics of the electroosmotic pump can be improved further. Investigations on materials of electroosmotic medium, carrier additives and organic solvent carriers will also be topics of further research.
[3] [4] [5]
[6] [7] [8] [9] [10]
References [1] D.J. Harrison, A. Manz, Z.H. Fan, H. Luedi, H.M. Widmer, Anal. Chem. 64 (1992) 11926. [2] A. Manz, D.J. Harrison, E.M.J. Verpoorte, J.C. Fet-
.
[11] [12]
675
tinger, A. Paulus, H. Luedi, H.M. Widmer, J. Chromatogr. 593 (1992) 253. D.J. Harrison, K. Fluri, K. Seller, Z.H. Fan, C.S. Effenhauser, A. Manz, Science 261 (1993) 895. K. Seller, D.J. Harrison, A. Manz, Anal. Chem. 65 (1993) 1481. A. Manz, E. Verpoorte, C.S. Effenhauser, N. Burggraf, D.E. Raymond, A.J. Harrison, H.M. Widmer, J. High Resolut. Chromatogr. 16 (1993) 433. Z.H. Fan, D.J. Harrison, Anal. Chem. 66 (1994) 177. K. Seller, Z.H. Fan, K. Fluri, D.J. Harrison, Anal. Chem. 66 (1994) 3485. P.K. Dasgupta, S.R. Liu, Anal. Chem. 66 (1994) 1792. S.R. Liu, P.K. Dasgupta, Talanta 41 (1994) 1903. Y.Z. He, W.E. Gan, M. Zhang, M.Z. Zheng, R.H. Zeng, G. Jin, Fenxi Huaxue 26 (1998) 125. Y.Z. He, W.E. Gan, Chinese Patent ZL 97212126.9, 1998. T.R. Gu, Y. Zhu, W.L. Li, J.M. Ma, L.R. Dai, H.M. Chen, Surface Chemistry, Science Press, Beijing, 1994, p. 1344.