JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JULY 1994 VOL. 9 779 On-line Flow Injection Cobalt-Ammonium Pyrrolidin-I = yldithioformate Coprecipitation for Preconcentration of Trace Amounts of Metals in Waters with Simultaneous Determination by Inductively Coupled Plasma Atomic Emission Spectrometry Zhixia Zhuang Xiaoru Wang* Pengyuan Yang Chenlong Yang and Benli Huang Department of Chemistry Xiamen University Xiamen China The technique of on-line flow injection (FI) cobalt-ammonium pyrrolidin-1 -yldithioformate (Co-APDC) copre- cipitation for the preconcentration of trace amounts of the heavy metals Cd Cu Fe Ni Pb and Zn in rain water samples with simultaneous determination by inductively coupled plasma atomic emission spectrometry (KP-AES) has been developed. A precipitate collector system consisting of a poly(tetrafluoroethy1ene) (PTFE) membrane on a polypropylene support filtering device combined with a 1.5 m reaction coil was selected.An inorganic solution of concentrated nitric acid and hydrogen peroxide was applied as the dissolution reagent. The technique is characterized by high retention efficiency (which ranged from 77 to 99% for the six elements of interest) good enrichment factors (ranging from 10 to 50 for 100 s preconcentration depending on the elements studied) and satisfactory accuracy and precision (recoveries from two standard additions to a rain sample ranged from 92 to 104O/0 with relative standard deviations ranging from 1.9 to 5%). The sample throughput is 20 per hour. Keywords On-line coprecipitation; cobalt-ammonium pyrrolidin- 1 -yldithioformate coprecipitation ; heavy metals in rainwater; flow injection inductively coupled plasma atomic emission spectrometry The flow injection (FI) technique has been widely applied to the on-line separation and preconcentration of trace and ultra- trace amounts of heavy metals in various water samples to improve the sensitivity of the determination and to eliminate interferences from the sample matrix and some co-existing elements.Solvent extraction including liquid-liquid' and sor- bent extraction,2 chelating ion-exchange columns3 and hydride generation4 are common techniques used for FI on-line separ- ation and preconcentration. On-line precipitation-dissolution in non-segmented continu- ous-flow systems in combination with flame atomic absorption spectrometry (FAAS) has been studied extensively by Valcarcel and gal leg^.^ The technique has been used for the preconcen- tration of lead in water samples by on-line preconcentration of the hydroxide with ammonium solution followed by dissolu- tion with nitric acid and detection by FAAS.6 A preconcen- tration factor of 700 was achieved under their operating conditions.The preconcentration of Cu Ca and Co in silicate samples with this technique has also been Although coprecipitation is a very traditional chemical separation and preconcentration method it is not widely used in combination with the on-line FI technique. Recently on-line FI with hexamethylene ammonium hexamethylene dithio- carbamate coprecipitation combined with FAAS was estab- lished by Fang et al.," in which the coprecipitation of lead in the presence of high concentrations of iron was performed with the advantages of low sample consumption high operating efficiency and high tolerance of the iron matrix.The technique developed by Fang used a knotted reactor as precipi- tate collector. The knotted reactor promoted radial mixing of sample and reagent providing reproducible conditions for the precipitation. In the present work on-line FI Co-ammonium pyrrolidin- l-yldithioformate (APDC) coprecipitation for trace amounts of elements in water samples and simultaneous multi-element determination with inductively coupled plasma atomic emis- sion spectrometry (ICP-AES) is developed. Although the con- ventional batch method of Co-APDC coprecipitation has the advantages of high retention efficiency large enrichment fac- tors suitable pH range and the ability for multi-element * To whom correspondence should be addressed. preconcentration it has not been applied to on-line FI because of the technical difficulties involved in the on-line system.In this study the FI manifold for the Co-APDC coprecipi- tation-dissolution has been established. The effects of the concentration of coprecipitation reagents the pore size of the filter membrane and the selection and flow rate of the dissolu- tion reagent were optimized. Several on-line FI coprecipi- tation-dissolution systems including a poly(tetrafluor0- ethylene) (PTFE) membrane on a polypropylene support filtering device a PTFE membrane on a polypropylene sup- port filtering device combined with a reaction coil and a reaction coil combined with a filter device filled with tiny PTFE chips have been studied and compared.The technique developed was applied to the analysis of rain water samples. Experimental Apparatus A Baird (Bedford MA USA) PS-4 multi-channel ICP atomic emission spectrometer was used for simultaneous multi-element determinations. The ICP-AES operating conditions are sum- marized in Table 1. A transient signal collection and data processing software developed in this laboratory was used to collect and handle the FI signals. One hundred data points were collected for each injection with 0.8s integration per Table 1 ICP-AES operating conditions R.f. power/kW Frequency of r.f. generator/MHz Coolant gas flow rate/l min-' Plasma gas flow rate/l min-' Carrier gas flow rate/l min-' Observation height/mm Integration time/s Data collection points Torch Nebulizer Wavelength/nm 1.2 21 10 1 .o 1 .o 15 (above load coil) 0.8 100 Regular Fassel type High salt (Babington) Cd 226.5 Cu 324.1 Fe 259.9 Ni 231.6 Pb 220.3 Zn 213.8 Co 238.8780 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JULY 1994 VOL.9 point. The peak height measurement mode was used while the background was corrected with a fixed slit. A ZL 2000 FI processor (Zhaofa Institute of Automated Analysis Shengyang China) with two peristaltic pumps and a 16-port valve was applied to the on-line preconcentration and separation. Coprecipitation-dissolution Manifold and Operation Procedure The manifold of the on-line FI Co-APDC coprecipitation- dissolution system is illustrated in Fig.1. The operation sequence is summarized in Table 2 and can be interpreted as follows. In the first step coprecipitation was carried out in which the sample containing 20 pg ml-' of Co" was pumped at a flow rate of 3.3ml min-' and merged downstream with 2% APDC solution at a flow rate of 0.4 ml min-'. The precipitate was collected on the filter which was cleaned with isobutyl methyl ketone (IBMK) solution stored in the loop from the previous cycle prior to the filtration. In the second step Pump 1 was stopped whereupon the coprecipitation process was terminated. Pump 2 carried a 2% APDC solution to clean off the residues remaining on the surfaces of the filter and the reaction coil. Meanwhile the dissolution loop was filled with HN03-H202 (1 + 1) solution.In the next step Pump 2 was actuated and the valve was returned from position A to the position B (See Table 2). The doubly de-ionized water was sucked up with Pump 3 attached to the ICP atomic emission P I n 2% APDC Sample Air H2O Eluent spectrometer which carried the dissolution reagent required to dissolve the precipitate collected on the filter into the nebulizer of the ICP. During the collection of signals the IBMK solution was pumped into the loop by Pump 2 for 10 s. The procedure was then repeated. A commercial PTFE filtering device was used as the precipi- tate collector. The device has a dead volume of around 200 1-11 and 40 pl after filling with PTFE chips (3 mm in length 1.5 mm in width and 0.1 mm in thickness).Poly(tetrafluoroethy1ene) membranes with pore sizes of 0.2 0.45 and 1.0pm on a polypropylene support (Gore and Associates Membrane Products Elkton MD USA) were tested for the filtration. The membrane area of the filtering device was 0.078 cm2 in the present work. Reagent and Sample Preparation Analytical reagent grade pure chemicals were used throughout the experiments. Stock solutions of 1000pgml-' for each element of interest were used for the preparation of standard solutions. All standard solutions for calibration and sample analysis were adjusted to a pH of about 3 with dilute nitric acid and ammonium solution and contained 20 pg ml-' of Co as the coprecipitate. Isobutyl methyl ketone was selected as a cleaning reagent for the residues remaining on the surface of the coil and the filtering device from the previous sample.A saturated solution of APDC (2% m/v) was prepared with Milli-Q water (Millipore Bedford MA USA). It was then P I n 2% APDC Sample Air Eluent H2O H2O Step 1 Step 2 P I Step 3 2% APDC Sample Air H2O Eluent H20 m Step 4 I-& 2%APDC Sample Off P2 Air Eluent H2O H2O Fig. 1 and W = waste (see text for details) Manifold of the FI on-line Co-APDC coprecipitation-dissolution system; A and B refer to position of valves P1 P2 and P3 are pumps,JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JULY 1994 VOL. 9 78 1 Table 2 Operation sequence for FI on-line coprecipitation ICP-AES Step Function Time/s Pump Flowrate/ml min-' Chemicals Valve position 1 Filter clean up 120 coprecipitation and dissolution reagent filled solution filled up signal collection 2 Rinse of precipitate 30 3 Dissolution IBMK 10 4 Dissolution and 100 *The flow rate of pump 3 is 2.5 ml min- 2 3 3 3.3 0.4 4.0 1.2 0.4 - Sample A APDC APDC A HN03-Hz02 IBMK B B - purified by extraction with an equal volume of IBMK.Doubly de-ionized water was further purified with a Milli-Q water purification device. Rain water samples were collected from Xiamen Island and were also adjusted to a pH of about 3. Results and Discussion Design of the On-line Co-APDC Coprecipitation-Dissolution Preconcentration System Precipitate collector The on-line FI precipitate collector is an important part of the system. Fang" proposed a knotted reactor as a precipitate collector for the determination of lead in biological samples with FAAS.An organic solvent was used as the dissolution reagent and an enrichment factor of 20 was obtained. Several other precipitate collector systems have been studied in the present work to accommodate the technique of on-line copre- cipitate-dissolution with ICP-AES with a low power r.f. gener- ator which cannot maintain a stable plasma discharge with the introduction of an organic solvent. The experimental results obtained are compared in Table 3 and interpreted as follows. (i)The PTFE membrane on a polypropylene support filtering device. The retention efficiencies of most elements obtained with this system are satisfactory except for Fe and Zn. (ii)The combination of a knotted coil (1.5 m long x 0.5 mm i.d.) and a filter filled with PTFE chips. In order to further improve the retention efficiencies of some elements via an increase of precipitate reaction time and reaction surface area a 1.5 m reaction coil was combined with the filter device used in (i) but filled with PTFE chips instead of a PTFE membrane.The experimental results indicate that although the retention efficiencies for iron and zinc were greatly improved that of lead was very low. This might be attributed to the low dissolution efficiency of lead under these operating conditions. (iii)A PTFE membrane on a polypropylene support filtering device combined with a knotted 1.5 m reaction coil. With this precipitate collector system all elements examined showed satisfactory retention efficiencies. Therefore this system was used in the present work. The contribution of the filtering device to dispersion was measured by the method described by Ruzicka and Hansen,13 and was 2.98.The retention efficiency (RE) was examined using the follow- Table 3 Comparison of retention efficiency (%) Element Cd co c u Fe Ni Pb Zn Filter only 98 16 99 13 97 94 35 Coil + PTFE chips 79 92 100 97 96 2 94 Coil + Filter 95 93 99 77 95 93 88 ing procedure 25 ml of 0.08 pg ml-' of a multi-element stan- dard solution was preconcentrated with three precipitate collectors under the optimized conditions; the filtrates were collected; and both standard solution and filtrate were analysed directly with ICP-AES. The retention efficiency was calcu- lated as ) x 100% [Element] in filtrate [Element] in standard solution The summary of standard solution REs using three types of the precipitate collectors is given in the Table 3.Dissolution system Isobutyl methyl ketone has been successfully used to dissolve Co,Ni-APDC coprecipitates in a batch method.12 However when IBMK is introduced into the ICP-AES system the plasma is susceptible to extinguishment with a low power r.f. generator. Owing to the high dissolution efficiency of organic solvents for the Co-APDC coprecipitate several other organic solvents (ketone methanol dimethylformamide) were tested. However none of the methods are successful under the current instrumentation conditions. Therefore an inorganic dissolution system with HN0,-H202 has been considered in this work. With the preliminary test it was found that the solution was a very strong dissolution reagent for on-line FI Co-APDC coprecipitation and did not cause any problems with the ICP discharge.Optimization of the mixing ratio of HNO3-HzO2 was performed and the results are illustrated in Fig. 2 which indicates that the best dissolution would be expected with a mixing ratio of l t l corresponding to 7 moll-' HN03 and 19% m/v H202. The dissolution efficiency with the system selected was examined. A 15 min preconcentration of a 0.4 pg ml-' multi- 1400 - 1200 4- .- C 3 >. F .= 1000 e - > v) C 4- ._ 800 - 600 0 1 1.2 1 l 2 1 1 :o HNO, H,O Fig. 2 Effect of HNO H202 ratio on intensity782 - c JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JULY 1994 VOL. 9 1600 - In c C 3 .- ; 1200 c .- e - .g 800 In C m C c. - 400 0 2 4 6 8 Dissolution reagent flow rate/ml min-' Fig.3 Effect of dissolution reagent flow rate on intensity - 1 I .! c I / - 400 ' I I I I I 0 30 60 90 Preconcentration time/s Fig. 4 Effect of preconcentration time on intensity for zinc element standard solution was carried out and the coprecipitate formed was dissolved on-line with 300 pl of HN03-H202. The dissolution process was repeated three times. The solutions dissolved were nebulized and measured by ICP-AES. The intensities of all the elements studied in the third dissolution were identical to that of a blank solution. Therefore the dissolution efficiency was calculated as the amount of the element found from the first dissolution over the sum of the amounts from the first and the second dissolution without considering the third dissolution.The experimental results indicate that the dissolution efficiencies are over 97% for all elements studied. The effect of the flow rate of the dissolving solution ranging from 1 to 8 mlmin-' has been examined. The results are shown in Fig. 3. The maximum peak heights were obtained with a flow rate of 2.5 ml min-' which was subsequently used throughout the experiment. Optimization of On-line Co-APDC Coprecipitation Capacity of precipitate collector The capacity of the precipitate collector with the membrane filtering device was examined using 0.4 pg ml-' of a multi- element standard solution at flow rates of 3.3 ml min-' for the sample and 2.5 ml min-I for the dissolution reagent. Using these conditions no block from the flowing system and filtering device was observed.The capacity is calculated as concen- tration of standard solution (pg ml-') x preconcentration time (min) x flow rate of sample (ml min-') x the number of elements =0.4 x 15 x 1.5 x 6= 54 pg The concentration of Co used is 20 pg ml-' which is 50 times that of the element being studied. Hence the contribution of Co to the capacity of the system is as much as 450pg. Considering that the system used in the present work includes a 1.5 m reaction coil the actual capacity of the system is even larger than the value calculated (54 pg + 450 pg = 504 pg). Effect of Membrane Pore Size and the Life Time of the Membrane The effect of the membrane pore size on preconcentration was tested qualitatively with pore sizes of 0.2 0.45 and 1.0 pm.Initially high retention efficiency with a small pore size mem- brane and low efficiency with the large pore size membrane were expected. However no significant differences of the retention efficiencies among the three types of membrane studied were observed. This might be interpreted as indicating that the fine particles of the precipitate dissolved are mainly adsorbed on the surface of the knotted reaction coil. Only particles larger than 1.0 pm are deposited on the surface of the filtering membrane. Therefore the membrane with a pore size of less than 1.0 pm did not show a significant effect. The life-time of the membrane with 0.2 pm pore size was evaluated for almost a hundred working hours utilization. No surface distortion or unusual behaviour was observed during the period of operation.Concentration of Coprecipitation Reagent The Co-APDC complex acted as a carrier for the coprecipi- tation of the heavy metals of interest. The concentration of APDC in this work was fixed at 2% (saturated sol~tion)'~ with a flow rate of 0.4 ml min-' for reaction with the sample solutions containing Co. The effect of Co concentration on the preconcentration was examined with 0.05 pg ml- ' multi- element standard solutions containing different concentrations of Co ranging from 10 to 200 pg ml-'. Although high retention efficiency would be expected with a high concentration of Co there was difficulty associated with the dissolution in the inorganic solvent of test solutions containing 100 and 200 pg ml-I of Co. The experimental results indicate that the best signals with high and sharp peaks were obtained with the lowest concentration of Co (10 pg ml-') for Cd Cu Fe and Zn using the peak height measurement mode and there was no significant effect on Ni and Pb; 20 pg ml-' of Co was used in the present work.Sample Preconcentration Time and the ICP-AES Signal The linearity of sample preconcentration time ranging from 30 to 120s versus ICP-AES signals of the elements studied was evaluated with a multi-element standard solution of 0.08 pg ml-l. As an example the preconcentration time versus zinc signal is demonstrated in Fig. 4. For all elements studied the peak heights of the signals collected were proportional to the sample preconcentration time with correlation coefficients of 0.999 or above. Performance of On-line FI Co-APDC Coprecipita tion- Dissolution System The performance of the system was evaluated using the enrichment factor (EF) retention efficiency (RE) relative standard deviation (RSD) of the multi-element injection corre- lation coefficients of the calibration curves (r) and the detection limits (DL) which are presented in Table 4.Some typical signal profiles with and without preconcentration are com- pared in Fig. 5. Rain Water Analysis Four rain water samples collected from Xiamen Island were analysed using the proposed method. The recovery test from standard additions to one rain water sample (No. 6) was performed to verify the method. The results are summarized in Tables 5 and 6. Conclusions The batch method of Co-APDC coprecipitation for trace elements has been successfully adapted to FI on-line coprecipi- tation-dissolution with ICP-AES multi-element determination.Three different types of precipitate collector systems wereJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JULY 1994 VOL. 9 783 250 200 150 100 - v) +- .- 5 50 L. h .- G o 450 c .- v) c 0 c - - 350 250 150 1 I 1 I 1 0 50 100 550 450 350 250 150 1600 1200 800 400 0 50 100 Timeh Fig. 5 Comparison of signal profiles A with and B without preconcentration for (a) Cd; (b) Ni; (c) Cu; and (d) Zn Table 4 Characteristics of F1 on-line Co-APDC coprecipitation system Element RE*(%) EFt RSD$'(%)(n=6) r DL/pg I-' Cd 95 28 4.6 0.9980 0.7 c u 99 15 5.0 0.9984 0.2 Fe 77 7 5.0 0.998 1 0.9 Ni 95 30 1.9 0.9942 0.4 Pb 93 27 2.0 0.9898 1 Zn 88 10 4.8 0.9996 0.3 *RE = Retention efficiency.tEF = Enrichment factor. $'r = Correlation coefficient of calibration curve. Table 5 Rain water analysis; element concentrations given are the mean of triplicate analyses -~ Element concentration/pg 1- ' ~~ ~ Element No. 3 No. 6 No. 7 No. 10 Cd 9.1 8.1 7.1 8.4 cu 0.6 5.4 1.8 1.2 Fe 22.4 21.3 23.4 25.3 Ni 2.1 0.6 1.1 1.1 Pb 10 30 65 1 Zn 16.5 14.3 34.2 21.4 studied with particular attention being paid to retention efficiency. The system selected for the present work is charac- terized by high retention efficiency fairly good EF and satisfac- tory accuracy and precision for the determination of trace elements; and is suitable for simultaneous multi-element determination with ICP-AES. Future work will be focused on extension of the method to other metals and application to other sample types.This work was supported by the Chinese Natural Scientific Foundation under grant No. 292351 10-11. References Memon M. A. Zhuang Z. X. and Fang Z. L. At. Spectrosc. 1993 14 50. Ruzicka J. and Amdal A. Anal. Chim. Acta 1989 216 243. Wang X.-R. and Barnes R. M. J. Anal. At. Spectrom. 1989,4,509. Wang X.-R. and Barnes R. M. J. Anal. At. Spectrom. 1988 3 1091. Valcarcel M. and Gallego M. TrAC Trends And. Chem. 1989 8 34. Martinez-JimCnez P. Gallego M. and Valcarcel M. Analyst 1987 112 1233. Santelli R. E. Gallego M. and Valcarcel M. Anal. Chem. 1989 61 1427. Adeeyinwo C. E. and Tyson J. F. Anal. Proc. 1989 26 375. Santelli R. E. Gallego M. and Valcarcel M. J. Anal. At. Spectrom. 1989 4 547. Table 6 Recoveries from standard additions to rain sample No. 6; n = 3 Recovery Sample concentration/ Found/pg 1-' Recovery Found/pg 1 - Element PLg 1-' (added 40 pg 1-') (added 80 pg 1-l) (%I 47.2 98 86.4 98 8.1 5.4 45.8 101 88.1 103 92 93 100 104 Cd c u Fe Ni Pb Zn 14.3 54.6 101 97.8 21.3 62.6 103 95.2 0.6 40.6 100 74.9 30 67 93 110784 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JULY 1994 VOL. 9 10 Fang Z.-l. Sperling M. and Welz B. J. Anal. At. Spectrom. 11 Ruzicka J. and Hansen G. H. Flow Injection Analysis 2nd edn. T. R. P. American Chemical Society Washington D.C. 1975 1991 6 301. p. 44. Wiley New York 1988 p. 23. 12 Boyle E. A. and Edmond J. M. Anal. Chim. Acta 1977 91 189. 13 Boyle E. A. and Edmond J. M. in Analytical Methods in Oceanography Advances in Chemistry Series No. 147 ed. Gibb Paper 3 /03 95 0 A Received July 7 1993 Accepted March 3 1994