ISSN:0003-2654    

DOI:10.1039/AN99116FX041

出版商:RSC    

年代:1991

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN99116FP043

出版商:RSC    

年代:1991

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN99116BP044

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

1094 ANALYST, NOVEMBER 1991, VOL. 116 Journal Style Update Spelling of Sulfur The new (1990) edition of IUPAC’s ‘Nomenclature of Inorganic Chemistry’ contains a table of IUPAC-approved names ‘for use in the English language’. These include ‘caesium’, ‘aluminium’, and ‘sulfur’ (spellings as given here). There is an increasing use of the ‘f‘ rather than the ‘ph’ spelling for sulfur in English publications, in particular the English language versions of I S 0 and European standards, and those British Standards that implement I S 0 standards verbatim. Furthermore, there is no good etymological basis for prefer- ring the ‘ph’ spelling. In view of these considerations, the Royal Society of Chemistry’s Nomenclature Committee has recently recommended that the RSC change to using the ‘f spelling in all publications. This recommendation will be implemented for the RSC’s primary journals in 1992. Alan McNaught Manager, RSC Journals Thomas Graham House, Science Park, Cambridge, UK

ISSN:0003-2654    

DOI:10.1039/AN9911601094

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1991, VOL. 116 1095 Determination of Anions by Flow Injection A Review Danhua Chen,* M. D. Luque de Castrot and Miguel Valcarcel Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, 14004 Cordoba, Spain Summary of Contents Introduction Flow Injection Configurations Used Types of Reactions Redox Reactions Complex Formation Reactions Catalysed Reactions Substitution Reactions Precipitation React ions Miscellaneous Reactions Dye formation reactions Chemiluminescence reactions Ion-pair formation reactions Aci d-b a se reactions Enzymic reactions Types of Detection Multi-determinations Physical Spectral Resolution Separation by High-performance Liquid Chromatography (Time Resolution) Chemical Discrimination Selectivity Sensitivity Analysis of Real Samples Comparison With Other Techniques Conclusions References Keywords: Review; anions; flow injection Introduction Since its inception over 15 years ago, flow injection (FI) has undergone extensive development, as reflected in the large number of papers published on this topic over this period' and in four monographs published recently.*-5 Flow injection is now well established; hence reports on theoretical aspects, new modes and revolutionary designs appear less frequently in the literature; the incipient commercialization of instrumen- tation is bound to contribute to its consolidation and new trends point to the start of a new age in which the emphasis will be placed on solving analytical problems in various fields of social interest.6 The 3000 or so papers published on FI to date cover a wide range of methods and analytes, as shown in reviews on areas such as clinical,7-9 pharmaceutical,*0.11 en~ironmental~~J3 and food analysis.14-16 Obviously, inorganic analysis has been one of the main targets of workers concerned with FI. Cationic species have been largely determined by atomic absorption spectrometry (AAS),17.18 while anions, both organic and inorganic, have been sensed by various detection techniques depending on the features of the species concerned. To date, no attempts have been made to discuss critically and system- atically the various FI methods developed for the determi- nation of anions. Only one paper has considered this subject, but the evaluation was neither systematic nor ~ritical.~9 The aim of this paper is to provide a review of the * Permanent address: Department of Chemistry, Wuhan Univer- t To whom correspondence should be addressed.sity, Wuhan, China. determination of anions by FI. As a rule, the features of anions make it difficult to find suitable reagents for developing sensitive and selective methods for these species covering the determination ranges usually required.20 For these reasons, it was decided to compile a systematic review of the state-of-the-art of this subject. For clarity, a series of tables have been compiled for the different anions summarizing the main features of each method [reaction and/or separation technique used, type of detector, determi- nation range, detection limit, relative standard deviation (RSD) and sampling frequency, and also the type of sample in which the species was determined].Commonly determined anions have been allocated an individual table, viz., phospho- rus (Table l), sulphur (Table 2), chlorine (Table 3) and fluoride (Table 4). Other anions with similar features appear in groups of two or more, viz., nitrite and nitrate (Table 5), cyanide and thiocyanate (Table 6), chromate, molybdate and vanadate (Table 7), and bromine and iodine anions (Table 8). Other anions not included in Tables 1-8 are listed in Tables 9 and 10. Data have been gathered from virtually every paper published to date (some 300 references are also listed in the tables). Different aspects of the FI configurations used are illustrated schematically in Figs. 1 4 . The type of configur- ation used in each method (modifications are denoted by a lower case letter) is also listed in the tables.The manifolds are discussed in the next section. Flow Injection Configurations Used The 14 different manifolds depicted in Figs. 1-4 represent the various designs used to develop determinative FI methods.1096 ANALYST, NOVEMBER 1991, VOL. 116 Table 1. Flow injection determination of phosphorus anions Analytical features Reaction/ separation Configuration technique* I-a I-a I-a XI I-a VI I-a I1 I-a V-b IX 111-a I-a I-a I-a I-a I-a XI I-a I-a I1 I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a XI XI V-a V-b V-a XI XI CF CF CF CF CF CF CF CF CF CF CF-LLE CF CF CF CF CF CF CF CF CF CF CF CF CF CF CF su CF CF CF R CF EC EC EC EC CF-EC CF-EC EC NCR NCR R HPLC-R-CF HPLC-R-CF R-CF R-CF HPLC-CF HPLC-CF Detection? P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P F F A A A A A cu cu PO ICP-AES P P P P P P P Range pg ml-1 5-25 5-30 - 1 .&24 1.947.5 0-10 0-2 0.1-2 1-12.5 - - 0-60 04.5% P 0.04-2.5 0.25-1.5 5-30 - 0.95-95 0.95-4.75 04.38 0-268 0-30 50-500 0 4 .8 0.95-9.5 - - - 8.95-9.5 - 0.03-0.6 04.045 - 0.09-47.5 0.1-30.0 0.95-95 0 4 . 8 0.48-95 - 1.9-950 - 19-57 mmol dm-3 0.95-95 0.95-47.5 0.95-47.5 0.95-47.5 0.95-9.5 - RSD 1.32 1 .o 2.3 (Yo 1 - - - 0.3-1.0 - 0.97 - - - - - 4 . 0 - - - - - 1.5 <1 .o t 0 . 7 0.5 - - 0.5-1.5 1 .o - - - - 0.9 <1.0 1.2 2.6 - - - - - 2.1 <1.0 <1.5 <1.0 4 . 5 - - Sampling frequency h-' 420 >200 - - - - 10 120 200 - - - 60 180 60 120 100 - - - >90 90 240 s120 40 120 300 45 30 60 20 - - - - - 70 - - 68 100 15 60 80 100 80 - - Real samples Plants - Water Water Water - - Plasma, urine Plants - Water Fertilisers Rock - - Fruit Sea-water Rock Plants Rock Water - - Water Water Powder Blood, serum - - Urine Compounds - Water Smoke Reference 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49,50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 65 67 68 * R = Redox; CF = complex formation; CT = catalytic; SU, substitution: PR = precipitation: DF = dye formation; CL = chemiluminescence; IP = ion-pair formation; AB = acid-base; EN = enzymic; EC = electrochemical; NCR = no chemical reaction; DIA = dialysis; LLE = liquid-liquid extraction; HPLC = high-performance liquid chromatography; IE = ion exchange; DIS = distillation; and GD = gas diffusion.t P = Spectrophotometry; F = spectrofluorimetry; PO = potentiometry; A = amperometry; CL = chemiluminescence; AAS = atomic absorption spectrometry; ICP-AES = inductively coupled plasma atomic emission spectrometry; S = chemical sensor; CU = coulometry; TC = thermochemical; MECA = molecular-emission cavity analysis; and MS = mass spectrometry. Tables 1-10 relate each configuration to the analyte deter- mined using that configuration and to the paper in which the determination was reported. Fig. 1 illustrates the simplest configurations used with Manifold I has been used to develop methods with (one or normal and reversed FI methods.ANALYST, NOVEMBER 1991, VOL. 116 1097 Table 2.Flow injection determination of sulphur anions Analytical features Anion sop SO& s032- sop S04’- sod?- sop sop S04’- sop SO42- sop S04’- soj2- soj2- sop sop SO& sop sop SO& sop so32- s o p SO3’- S03’- s o p so32- S2- S* - S’- S’ - S2- S2 - S2- S2- Reaction/ separation Configuration technique* I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a I-a IX I-a I-a I-a I-a I-a I-a I-a I-a I-a V-b I-b I-a I-a IX I-a I-a I-a I-a III-b IX I-a I-a I-a I-a PR PR PR PR PR PR PR PR PR PR su su DIA-PR CT su su su su NCR NCR NCR NCR R-EN R R-CF GD-R R-CL R NCR NCR PR CF CT CL R-CL DIA-EC * Abbreviations as in Table 1. I Abbreviations as in Table 1. Range1 Detection? pg ml-l P 10-100 P P P P P P P P P P P P P P P P P PO PO ICP-AES MECA A 50-200 40-1 60 100-900 1-30 5-200 1oo-1OOo 1-100 10-120 1-20 0.2-20 50-500 0.25-6.0 0-30 0.5-30 1-lo00 25-1000 0-1 .o 1 .7-83 0.8-69 - - - - P 1-100 P 8-800 P 1-20 CL 0.09-3.5 F 0.026-0.25 PO 0.1-17.8 PO 10-“10-’ mol dm-3 AAS 0-2 - A P M5.1 P 0-32 CL 0 .0 0 1 ~ . 1 6 CL 0.0140 Detection limit/ pg ml-1 10 - - - - - - - 0.45 - - - - - 0.1 - - - - - 0.0028 - 1 x 10-5 2 x 10-5 mol dm-3 0.3 mol dm-3 0.1 0.0032 - - - 0.01 0.00015 0.42 0.ooOol 0.0004 - RSD 0.85 0.95 (Yo 1 - - <1.0 - - <2.0 - 0.78 - t 2 . 0 2.0 4.1 - - - 0.94-1.2 - - 98 - 2.2 1 .5 2.0 1.0-2.0 5.0 - - - 1.2 4.0 - - 3.0 4.4 Sampling frequency1 h- 180 60 30 120 - - 30 200 30 90 40-60 40 20 20 <20 30 20-38 20 40 100 - - - <15 60 60 720 24 - - 100 210 240 240 - - Real samples Water, plants Sea-water Water, plants Water - - - - - Water Rain, snow Urine Rain Water Water Water - - - - Water - - - - - - - - - - - - - Water - Reference 69 70 71 72 73 74 75 63 76 77 78 79 80 81,82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 both broken lines for reagent streams) or without chemical reaction (a); with the use of an ion-exchange, enzyme or redox column (b); or with a mini-mixing chamber, generally used in FI titrations (c).Methods for the determination of phosphate, sulphate, sulphite, suIphide, chloride, chlorate, fluoride, nitrite, nitrate, nitritehitrate, cyanide, thiocyanate, chro- mate, molybdate, vanadate, bromide, bromate, iodide, iodate, iodidehodate, borite, permanganate, oxalate and selenate have been developed with different variants of this general scheme.Manifold 11 illustrates the simple reversed FI mode in which a reagent plug occasionally resulting from the mixing of two different streams is injected into the sample stream acting as carrier. Only phosphate has been determined with this type of configuration, which is particularly suited to the analysis of waste water and, as a rule, of abundant, inexpensive samples. The merging-zones approach (Figs. 1-3) was conceived in order to decrease reagent consumption by simultaneously injecting sample and reagent. The mixed plug can be passed (b) or not passed (a) through an ion-exchange column, as in the determination of phosphate, sulphide, chloride, nitrite, molybdate and bromide. The stopped-flow mode has been implemented in manifolds such as IV, both in the normal and reversed (in parentheses in Fig.1) FI modes. By synchronizing the injection and propulsion units, the reacting plug can be halted at the reactor to ensure optimum development of the rkaction without increasing the dispersion (non-kinetic stopped-flow mode) or at the flow cell to monitor the development of the reaction by collecting signal-time data pairs (kinetic stopped-flow mode), thereby avoiding matrix effects and interferences from reac- tions faster or slower than the analytical reaction. The determination of fluoride and the simultaneous determination of chloride and chlorate have been accomplished using this configuration. Multi-determination designs are more complicated (Fig. 2). Manifold V was designed for differential kinetic determi- nations (a) by using the same reagent which reacts at a different rate with two or more analytes (serial detectors) and (b) for simultaneous determinations based on splitting the sample into two channels, each of which is merged with a suitable reagent and then driven to different parallel detec- tors.These assemblies have been used for the determination of phosphate/ammonia and nitritehitrate. Multi-determinations requiring different reagents or reac- tion media have been carried out using configuration VI. A1098 ANALYST, NOVEMBER 1991, VOL. 116 Table 3 Flow injection determination of chlorine anions Anion C1- c1- c1- c1- c1- c1- c1- c1- c1- c1- C1- c1- c1- c1- Cl- c1- c103- c10-- c103 - c10-- c103- c10-- c103- c104- c104- Reaction/ separation Configuration technique* I-a PR I-a NCR I-a NCR IX NCR I-a NCR I-a NCR I-a NCR I-a CL I-a SU-CF I-a SU-CF I-a R III-a su I-a SU-CF I-a SU-CF I-a PR I-a R IV R IX GD-EC I-a R I-a R IX IP-LLE IX IP-LLE * Abbreviations as in Table 1.? Abbreviations as in Table 1. Analytical features Detection-t AAS PO PO PO PO PO PO PO CL P P P P P P P P P P P AAS P Range/ pg ml-1 3-100 5-5000 250-5000 0.7 1-1 78 - - 10-200 1.8-3.6 0.2-15 0.1-6.0 1.1-15.3 0.5-75 2-10 2-200 0-14 0.084-0.84 - - 0 . O M . 0 0.01-10.1 0.1-8.3 0.1-5.0 0-2.5 Detection limit/ pg ml-1 1.3 - - 3.5 - - - - 0.043 - - - - - - - 110-6 mo dm-3 0.04 (CIO-) 0.03 (C103-) - 0.07 0.036 RSD 2.0 1.7 0.5 (”/) - - 1.75 2.0 3.0 <1.0 1.0-3.0 - - - - - - (3.0 - - 94 1.2 1.2 - Sampling frequency1 h-1 50 120 120 100 60 120 30 120 90 90 15 60 - - - - - - - - 45 20 Real samples Water Water Milk Water Plants Serum - - - Water Ethanol Rain - - - - - - Water - - Chlorate Reference 105,106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127-129 Table 4 Flow injection determination of fluoride Analytical features Anion F- F- F- F- F- F- F- F- F- F- F- F- F- F- F- F- F- F- F- ~ ~~~~ Reaction/ Detection Sampling separation Range/ limit/ RSD frequency/ Real h-1 samples Reference Configuration technique* Detection? pg ml-1 pg ml-1 (Yo) I-a I-a I-a I- a I-a I-a I-a I-a I-a I-a I-a I-b I-a I-a IV I-a I-a I-a XI1 * Abbreviations as in Table 1.t Abbreviations as in Table 1. NCR NCR NCR NCR NCR NCR NCR NCR NCR NCR NCR NCR CT CF CF su CF CF CF PO PO PO PO PO PO PO PO PO PO PO PO P P P P F F-S F-S 0.02-0.2 0.019-19 0.3-10 0.019-19 0.001-1.0 0.19-1.9 - - - - 1.9-190 0.019-0.19 - 0.03-1.2 0.08-1.2 0.1-10 0.00024.02 0.5-8 0.001-0.04 - - - - 0.001 - - 0.001 0.0019 0.02 - 0.001 0.01 - - - - - - 3.0 5.0 - - 0.5-5.0 - - - - - 1.7 3.0 - - 0.2 0.2 15.0 1.0 - 60 120 60-180 360 120 40-60 80 60 24 24 60 100 60 t l 30 - - - - Rain Rain - - - - - Water, urine beverages Rain Various materials - - - - Water Water Water - - 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 selection valve allows the appropriate medium to be selected, An approach based on an internally coupled valve assembly as in the sequential determination of arsenate/arsenite, (Fig.2, manifold VIII) has been used for simultaneous arsenate/phosphate and arsenate/arsenite/phosphate.speciation using a prior redox step implemented in a solid The reversed FI mode has been used for the sequential reactor located in the secondary valve. determination of anions such as nitritekyanide, nitritelsul- Configurations involving on-line separation units are de- phide and cyanide/sulphide by sequentially injecting the picted in Fig. 3. required selective reagent into the sample stream. The typical location of a gas-phase separator, dialyser orANALYST, NOVEMBER 1991, VOL. 116 1099 I C C - P Fig. 1 Simple manifolds used in the determination of anions. I , Normal FI configuration with (broken lines) or without chemical reaction (a), with an on-line ion-exchange, enzyme or redox column (b) or with a mini-mixing chamber (c). 11, reversed FI manifold with injection of one or two mixed reagents.111, Merging-zones approach without (a) or with (b) an ion-exchange column. IV, Stopped-flow arrangement using a timer to synchronize injection with the halting of the flow at a pre-set time in the normal or reversed (in parentheses) FI mode. P, Peristaltic pump: C. carrier; R, reagent; S, sample; IV, injection valve; D, detector: and W. waste liquid-liquid extraction unit (between the injection and detection systems) as shown in manifold IX has been used for the separation/dilution of phosphate in plasma and urine, for the separation of sulphate, sulphite, sulphide, chlorine, cyanide, cyanidekhiocyanate and carbon dioxide by gas diffusion and perchlorate, nitrate and nitritehitrate by liquid- liquid extraction with ion-pair formation.Two different streams emerge from the separation unit in all instances, one of which is driven to waste while the other reaches the flow cell for measurement of the analyte. One application involves placing the separation unit in the loop of the injection valve (manifold X) to accomplish sample preconcentration in addition to separation, as in the determi- nation of thiocyanate after liquid-liquid extraction. This type of unit is also used in its typical location for on-line ion-exchange processes, and always when a chromatographic or a precipitation-separation-preconcentration step is required (configuration XI). This approach has been used for the determination of phosphate and polyphosphates by high-performance liquid chromatography (HPLC) , and for chloride, bromide and iodide. U VI P P I wc U VII P Vlll P I I Fig.2 Manifolds used for the multi-determination of anions. V, Differential kinetic or non-kinetic determination with serial (a) or parallel (b) detectors, respectively. VI, Using a switching valve to select the appropriate reagent or medium. VII, Using the reversed mode, with sequential injection of a selective reagent for each analyte. VIII, Speciation with a prior redox step. SV, Selecting valve; and D1 and DZ, detectors. Other symbols as in Fig. 1 Fig. 4 shows some special designs such as that used to implement integrated reaction-detection (XII) , and those used to obtain multi-peak recordings (XI11 and XIV). Chemical flow cell sensors have been used in manifolds similar to manifold XII, including a suitable support (gener- ally an ion-exchange resin) in the flow cell.In this way, it is possible to retain one of the components of the analytical reaction temporarily or permanently at the detection point, thereby increasing sensitivity and selectivity as an in situ concentration step is performed during the monitoring pro- cess. The determination of fluoride by transient retention of a ternary complex in the flow cell, and that of cyanide by retention of a fluorescent product resulting from its reaction with pyridoxal 5-phosphate, have been carried out with this type of sensor.1100 ANALYST, NOVEMBER 1991, VOL. 116 Table 5 Flow injection determination of nitrite and nitratc Analytical features Reaction] separation Configuration technique* 111-a I-a I-a I-a I-a I-a I1 I-a I-a I-a I-a I-a I-a I-a I-a IX I-a I-a I-a I-a I-a I-b I-b I-b I-a I-a I-a I-a I-a I-a I-a I-a I-a I-b I-b I-a I-a IX I-b I-a V-b V-b I-a VIII - I-a I-a I-a I-b 1-a IX R R R R R R R R R R R R R EC R-EC GD-NCR EC R-EC EC R R R R R R R NCR su NCR R R-EC R-EC NCR NCR NCR NCR NCR IP-LLE R R-CL R R R R R R R R EC NCWEC IP-LLE * Abbreviations as in Table 1.t Abbreviations as in Table 1. P P P P P P P P P P P P P A A PO A A A F F P P P P P P P P P A A PO PO PO PO PO AAS F CL P P P P P P PIA PIA A POIA AAS Rangel Detectionl' yg m1-I 0.025-0.5 0.002-60 0.2-2 0-0.1 0.05-1.3 0.1-5 - - 0.0054). 03 1-10 10-1000 0.0032-0.064 0.4646 0.0092-0.2 15 0.23-46 0-30 0.046-9.2 2.3-230 2.3-33 0-0.014 0-10 0-2.5 0.05-0.3 0-12.4 1-10.0 0.089-2.2 - 1.24-8.7 0.2-20 0-22 0.62-6.2 31-310 6.2-3 10 - 8 x 10-5- 10-1 mol dm-3 1-5 3.1-310 0.13-2.2 0-12.4 - 0.1~0.5 1-5 0.0092-0.092 0.12-1.2 0.1-0.5 1-5 0.025-16 0.1-16 0.1-1 .o 0.5-5.0 5 x 10-5- 5 x 10-3 rnol dm-3 0-7.5 x 10-5 mol dm-3 - - - 1-10 0.2-2 Detection limit1 pg ml-1 - - - - - - - - 0.0082 - - - - 0.07 0.0015 0.003 - - - 0.oooO46 0.002 0.0062 0.0015 - - 0.089 - - - - - 6.2 - - 1 x 10-5 rnol dm-3 - - - 0.00062 0.001 - 0.0023 0.0062 - - 0.001 0.01 - 5 x 10-6 rnol dm-3 mol dm-3 5 x 10-5 0.027 - - RSD 0.5 0.14.4 1.5 1.67 2.2 S1.3 (Yo 1 - - - 0.7-1.8 - 0.7 - - - - - - 4 . 0 0.4 1.7-2.8 <1.0 >1 .o 0.4 3.0 3.0 3.0 - - - - 2.64.8 1-2 0.8 - 2.4 1-2 0.8 0.5 1.5 1 .o 0.85 1.96 1.1 0.9 1.0 0.7 - - - - - - 2.5 4.7 3.7 Sampling frequency1 h-' 70 300 - - 50 60 130 30 120 50 - - - - 100 30 - - - 45 30 75 40 40 - - 120 18-35 100 - - - 90 125 120 100 100-200 20 - - 90 30 72 65 60 - - - - - 35 * 5 Real samples Waste water - - - Water Water Water - - - - Water - - - - - - - Sea-water Saliva Water Water Rain Water Rain - - - Plants - - Soil, water, fertilizers Waste water, Soil - Water Foodstuffs Water - - - - Water - Water, soil - - - Soil - - Reference 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 19 1 192 193 194 195 196 197 198 199ANALYST, NOVEMBER 1991, VOL.116 1101 Table 6 Flow injection determination of cyanide and thiocyanate Analytical features Anion CN- CN- CN- CN- CN- CN CN- CN- CN - CN- CN CN- CN CN- CN CN - CN- CN- CN- CN- CN CN - SCN ~ SCN - CN-/ SCN ~ CN-/ SCN CN-/ SCN- Configuration v11 I-a/II IX IX IX I-a I-a I-a I-a I-a XI I I-a I-a I-a IX I-a I-a I-a I-a I-a I-a I-a I-a X IX I-a * Abbreviations as in Table 1 t Abbreviations as in Table 1 Reaction/ separation technique” R-DF R-DF GD-DF GD-DF GD-DF CT CL CL CL CT CT NCR NCR NCR NCR NCR EC EC EC EC DIS-EC R-S U EC IP-LLE GD-DF GD-DF NCR Detection i P P P P P P CL CL CL F F-S PO PO PO PO PO A A A A A AAS A P P P PD Range/ pg ml-1 0.3-5.0 0.1-1 .0 0.4-5.0 0.1-10.0 - - 0.00005- 0.026 - - 0.1-20 0.05-3 .0 20- 1000 5-50 0.3-100 0.2626 0.02626 0.26-26 0-10 0.4-40 0.65-26 - 2.6-5.2 0.0 12-0.23 0.026-2.6 0.58-5.8 0.058-5.8 0.4-2 0-6 0.2626 2.32-1 74 The open-closed system (manifold XIII) provides a means of obtaining enhanced information by using a single conven- tional detector yielding multi-peak recordings.298J99 The system was used for studying chromium speciation. A multi-peak recording can also be obtained by using a simple FI manifold with a programmed pump to change the flow direction iteratively and hence to achieve the repeated passage of the sample plug to the detection point.288.3(H) One such system was used for the kinetic discrimination of perchlorate and iodide based on the different rate of extrac- tion of their ion pairs with hexacyanoferrate(Ii1) into an organic phase injected into the sample stream.Types of Reactions The distribution of reaction/separation techniques is sum- marized in Fig. 5 , where ‘separation’ includes dialysis, liquid-liquid extraction, HPLC, ion-exchange and distillation; ‘no chemical reaction’ denotes determinations based on features of the analyte, e .g . , absorbance, fluorescence and redox properties. The different types of chemical reactions used for the determination of anions by FI are discussed below. Redox Reactions A spectrofluorimetric method for the determination of phosphate has been described based on the oxidation of thiamine to thiochrome by a heteropolyacid.5’ Phosphinate and phosphonate were determined with sodium hydrogen sulphite as oxidant,65.66 and hypophosphate was determined with potassium iodate in sulphuric acid.28 Methods for the Detection limit/ pg ml- - - - 0.006 0.006 0.2 0.00005 0.00025 0.05 - - - 0.08 - - - - 0.001 - - - - - - 0.0078 0.012 0.05 0.08 - RSD 0.66 <0.8 0.8 1.4 1.4 2.1 3.3 2.4 <2.0 I .39 1.2 2.5 2.0 5 .o <1.0 (%) - - - - - - <2.0 3.@4.0 1 .o 1.5 - - - Sampling Frequency/ h-1 28 20 10 40 - - 360 700 360 10 60 40 20 - - - - 60-70 100 60 40-50 240 18 10 - - - Real samples - - - Waste water Waste water - - Plants, water - - - - - Waste water - - - Water Waste water - - - Urine - - - - Reference 200 20 1 202 203 204 205 206.207 208,209 210 211 2 12 213 214 2 I5 216 217 218 219 220 22 I 222 223 224 225 226 227 228 determination of sulphite have been reported based on the reduction of sulphite by phosphonate,93 the oxidation of an organic disulphide to a coloured thio anion at pH 6.0,9* the decolorization of Malachite Green94 and the reaction of sulphur dioxide with N-acridinylmaleimide in a water-N,N- dimethylformamide medium to form a fluorescent product .9h Most of the methods for the determination of nitrite are based on a diazotization-coupling reaction.Nitrite is diazotised with sulphanilamide and the product coupled with N-( l-naphthy1)- ethylenediammonium dichloride to form a highly coloured azo dye which is measured at 520 nm;339-156.170.1*9-192 a similar reaction with other reagents has also been developed. 1577158.193 Nitrate was determined similarly after on-line reduction to nitrite in a copper-coated cadmium column or with a dissolved reactant.171-171 Potassium bromate has been shown to be a selective colour-forming reagent for halide anions. 117 The sequential iodimetric determination of chlorate, and chlorite-chlorate, has been described by Gordon and co- workers.122-125 The reaction between iodate and iodide with the release of iodine was used for the spectrophotometric determination of both species.26hJ75 Bromide was determined on the basis of the bromination of Phenol Red to Bromo- phenol Blue after oxidation of bromide to bromine with chloramine-T.260,*61 Periodate was determined by measuring the red colour in the reaction with salicylaldehyde guanyl- hydrazone in a basic medium.274 A hydride generation reaction was used for the determination of ar~enic2~33294 and selenium.’94 Complex Formation Reactions The Molybdenum Blue method for the determination of phosphate has frequently been used in this con-ANALYST, NOVEMBER 1991, VOL.116 - 1102 IX S P , XI P n I C . Separ- IV I ation .0 1 unit Fig. 3 Flow injectiodon-line separation assemblies. IX, Conven- tional location of separation units involving two emerging streams as in gas-diffusion, dialysis and liquid-liquid extraction processes. X, Special location of the separation unit for preconcentration in addition to separation (dialysis, gas diffusion, ion exchange or liquid-liquid extraction). XI, Conventional location of separation units involving a single emerging stream ( e . g . , in precipitation and ion-exchange processes). Symbols as in Fig. 1 text.21-41,49-51,54,55,58,64-68 Phosphate was first reacted with molybdenum(v1) to form the yellow 12-molybdophosphoric acid, which was then reduced to the blue heteropoly phospho- molybdenum. The formation of a heteropoly acid was also the basis for the determination of silicate and arsenate.24.26 Phosphate also forms a yellow heteropoly acid complex with vanadomolybdic acid.42-45 The reaction is very fast, although it is not as sensitive as that of the Molybdenum Blue method.The Methylene Blue method was applied to the determination of sulphide in industrial waters.101 The spectrophotometric determination of fluoride with lanthanum(rIr)/alizarin com- plexone has been reported; the slow reaction rate was overcome by either increasing the temperature to accelerate the reaction143 or by using the stopped-flow-reagent injection method.14 Spectrofluorimetric methods for the determina- tion of fluoride have been described, based on enhancement of the fluorescence of the ternary complex formed between fluoride, zirconium and Calcein Blue.146-148 Boron was determined spectrofluorimetrically on the basis of a ternary complex formed with 2,6-dihydroxybenzoic acid and Crystal Violet .278 Most of the spectrophotometric methods reported for the determination of chromium(v1) are based on its reaction with 1,5-diphenylcarbazide.229-237 The thiocyanate R Xlll PI c P7 R XIV f=l IP S U P W t Fig. 4 Special manifolds for integration of retention and detection using a chemical sensor (XII), and for the iterative passage of the sample plug through a conventional detector in an open-closed system (XIII) or the iterative change of the flow direction with a programm- able pump (XIV). IP, Injected miscible or immiscible phase. Other symobls as in Fig. 1 Redox 22% ectrochemical Complex formation 21 % 5% No chemical reaction Other reactions 15% 9% Fig.5 Frequency of use of the different types of reactions used for the determination of anions (separation techniques are also included) method was adapted to the determination of molybdenum, in which molybdenum(v1) was first reduced to molybdenum(v), and the molybdenum(v)-thiocyanate complex formed was measured.24Q-242 The colour forming reagent 4-(2-pyridylazo)- resorcinol was used for the spectrophotometric determination of vanadium.254.255ANALYST, NOVEMBER 1991, VOL. 116 Table 7 Flow injection determination of chromate, molybdate and vanadate 1103 Analytical features Configuration I-a I-a I-a I-a I-a 111-a XI11 I-a I-a IX I-a I-a IX 111-a - I-a I-a I-a I-a I-a I-a I-a I-b I-b I-a I-a I-a I-b I-b Reaction/ separation technique* CF CF CF CF CF CF CF CF CF NCR CF CF IP-LLE LLE-CF IE-CF CT CT CT CT CF CF CT CT IE-CT CT CF CF IE IE * Abbreviations as in Table 1.1- Abbreviations as in Table 1. Detection? P P P P P P P P P P ICP-AES P P P P P P A PO PO AAS P P P P P P AAS ICP-AES Rangel pgml-* 0.1-20 0.540 0.1-20 0-2.0 0.1-1.2 0.2-10.0 W . 0 0-20 0-1 .o 0.14.0% M O O 0.05-1 .O 0.00015- 0.006 0.001-1.0 0.00 14.04 0.02-1.0 0.17-30 - - 0-0.02 - 0.054.5 0.01-1 .0 0.00004- 0.002 0. oooO5- 0.002 6.6-26.4 0.5-3 .0 0.2-1.0 0.01-0.05 Detection limit/ pg ml-1 - - - - 0.018 - - - - 0.44 0.0002 - - - 0.00015 0.007 0.0012 0.17 0.24 - - - - - 0.oooo2 - - - - RSD 1.3 31.0 (% 1 - - 0.7-3.0 - - 0.8 2.5 0.73 1.1 <2.0 2.3 - - 0.83-1.9 <2.0 1.3 - - - 0.37-0.55 - - 0.5 - - 3.0 <5.0 Sampling frequency1 h-1 70 180 120 - - - - 40 120 24 270 40 30 3 4 90 40 1 00 60 - - - 60-70 12 1.5 - - - - 17 Real samples Soil - - - - - - - - - - Steel Steel Plants water Soil Plants Soil Sea-water - - - - Rain - - - - - Reference 229 230 23 1 232 233 234 235 236 237 238 239 240 24 1 242 243 244 245 246 247 248 249 250 25 1 252 253 254 255 256 257 Catalysed Reactions Sulphate was determined on the basis of its catalytic action on the reaction between zirconium and Methylthymol Blue.81.s2 The determination of sulphide has been described, based on its catalytic action on the iodine-azide reaction.102 Iodide catalyses the redox reaction between arsenic(rr1) and ce- rium(rv), and this is the basis for the thermochemical,267 spectrofluorimetric268 and spectrophotometric205 determina- tion of iodide.Thiocyanate was determined by Tanaka et al.205 on the basis of its catalysis of the oxidation of arsenic(rr1) by cerium(1v). A spectrofluorimetric method for the determina- tion of cyanide with pyridoxal and pyridoxal5-phosphate has been described;21l7212 the reaction is catalytic in nature. Trace amounts of bromide were determined spectrophotometrically by the catalytic action of bromide on the oxidation of Pyrocatechol Violet with hydrogen peroxide in an acidic medium.262 A method for the determination of molybdenum has been reported, based on its catalytic action on the oxidation of iodide by hydrogen peroxide in an acidic medium.244-247 Trace amounts of vanadium in water were determined by the catalytic effect of vanadium on the oxidation of some organic compounds by bromate.*5(&253 The catalytic effect of selenium on the reduction of Tolidine Blue by sulphide was also used for the catalytic determination of this element.295 Substitution Reactions Yoza et aZ.47 proposed a method for the determination of orthophosphate, diphosphate and triphosphate based on the substitution reaction between phosphates and coloured metal complexes to form a colourless M-P, complex.Sulphate was determined by its competitive reaction with barium chelates such as those of Methylthymol Blue,72,76?83,84, dimethylsul- phonazo-III,8S786 nitchromazo78 and 3-(2-carboxyphenylazo)- 6-(2-sulphophenylazo)chromotropic acid.79 The spectropho- tometnc detection of the iron( 111)-thiocyanate complex was used for the determination of nitrate based on the displace- ment of thiocyanate from a strongly basic ion-exchange resin.176 Fluoride was reacted with zirconium-Alizarin Red S to form a colourless zirconium fluoride product, the resulting absorbance being measured at 520 nm.145 A method for the determination of chloride with mercury(l1) thiocyanate and iron(rI1) as reagents has been reported;11s3116 the method involves measurement of the absorbance at 480 nm of the coloured iron(Ir1)-thiocyanate complex formed. Another method for the determination of chloride has been described, based on its solid-liquid reaction with mercury(I1) chloranilate to yield coloured chloranilate ions. 118 The spectrofluorimetric determination of oxalate in urine has been described, based on the quenching effect of oxalate on the highly fluorescent complex formed between zirconyl ions and flavonol in sulphuric acid.292 Precipitation Reactions Sulphate has frequently been determined by measuring the turbidity of a barium sulphate suspension spectrophotometric- ally63,69-77>8* the method was first reported by Krug et al.69 One problem with this method is that the precipitate formed tends to settle in the reactor and flow cell, thus reducing the precision and leading to clogging of the manifold.Alkaline1104 ANALYST, NOVEMBER 1991, VOL. 116 Table 8 Flow injection determination of bromine and iodine anions Analytical features Reaction/ separation Anion Configuration technique* Br- Br- Br- Br- Br- Br- Br03- I- I- I- I- I- I- I- I- I- 103- 1 0 4 - I-no3- I-a III-a I-b I-b I-a I-a I-a I-a I-a I-a I-a I-a I-a - I-a I-a I-a I-a I-a Cl--Br--1- XI NCR R-CT R R CT R R R R CT CT CT EC EC NCR NCR EC R R IE-NCR * Abbreviations as in Table 1.Abbreviations as in Table 1. Detection Range/ limit/ Detection? pg ml-l pg ml-l PO CL P P P P PO P P TC F P A A PO PO A P P PO 1-5000 - 1-lo00 O.oooO625 0.16-2.4 - 1-10 - 0.01-0.6 - 0.13-0.65 0.1-10 0.05-15 0.00127-1.27 0.1-1 .o 5.1-51 5 x 10-10- mol dm-3 0.254-58.8 0.1 mol dm-3 - 1 x 10-3 5 x 10-5- 0.0127-127 5-50 2 x 10-3- 1 x 10-5- 8 x 10-3 mol dm-3 6 x 10-5 rnol dm-3 20-5000 0.01 15 nmol dm-3 0.1 0.05 0.00125 3 mol dm-3 0.1 5 x 10-10 mol dm-3 - - - - - 0.35 - 5 RSD 1.6 Gl.0 0.2-2.5 (Yo ) - 1.4-2.5 - - - 1.0 0.3-1.0 - - 0.6-3.0 1 .o - - 1.0 1 .o - 1.5-2.1 Sampling frequency1 h-1 88 113 80 120 45 5 - - 52 30 20 - - - - 60 - 20 100 - Real samples Soil Sea-water Sea-water chloride Water Water - - - - - - - - - Salt - - - - - Reference 258 259 260 261 262 263 264 265 266 267 268 205 269 270 27 1 272 273 274 275 228,276 Table 9 Flow injection determination of borate and hydroxide Analytical features Reaction/ separation Anion Configuration technique* H2B03- I-a CF H2BO3- I-a CF HzBO3- I-c AB H2BO3- I-a DF H2B03- 1-a DF H2BO3- 1-a DF H2BO3- I-a DF OH- I-a AB OH- I-c AB OH- I-c AB * Abbreviations as in Table 1.t Abbreviations as in Table 1. Detection? F P P P P P P PO PO P Range/ pg ml- 0.02-0.5 0.0005-5.0 2.5 x 10-4- 5 x 10-2 rnol dm-3 0.140 0-3.0 0.254.0 0.2-2.5 0.01-0.1 rnol dm-3 0.01-1 0.01-1 mol dm-3 Detection limit/ pgml-l 0.0002 - - - - 0.02 - - - - RSD 4 .0 G0.4 (Yo 1 - <1.0 0.8 1.0-1.4 2.6% 4.0 0.8 c1.0 Sampling frequency/ h-l - - >60 60 60 120 25 180 - - \ Real samples Reference Sea-water 277 - 278 - 279 - 280 28 1 - 282 - 283 - 284 - - 285 - 286,287 ethylenediaminetetraacetate was introduced into the system in different ways7@-73,76 in order to dissolve the residual precipitate and hence flush the system. Sulphide in the sample solution was precipitated with cadmium ion, the precipitate being freely driven to the detector while the excess of cadmium was retained in the ion-exchange column and detected by AAS.99 Chloride was injected into a carrier containing silver nitrate and the precipitate retained on a stainless-steel filter. In this way, chloride was detected by AAS at a sampling frequency of 50 h-1105,1“ and by spectrophotometry based on the turbidity of silver chloride.121 Miscellaneous Reactions These include dye formation, chemiluminescence, ion-pair formation, acid-base and enzymic reactions. Dye formation reactions Cyanide has been determined by using chloramine-T; the CNCl produced forms a red-blue dye on addition of pyridine- barbituric acid. The method was first adapted to FI by Rios et al.2007201 The method can also be applied to thiocyanate.227 The reaction of cyanide with isonicotinic acid and 3-methyl-l-ANALYST, NOVEMBER 1991, VOL. 116 Table 10 Flow injection determination of other anions 1105 Analytical features Anion Mn04- Mn04- Salicylate c 2 0 4 2 - c 2 0 4 2 - ~ s 0 ~ 3 - ~ s 0 ~ 3 - As02- AsOd3- Se04*- Se042- Se042- Se03*- co32- Configuration I-a IX IX I-a I-a VI VI - IX IX I-a I-a I-a IX Reaction/ separation technique* NCR IP-LLE DIA NCR su CF CF R R GD-R CT CT CT GD-AB * Abbreviations as in Table 1.7 Abbreviations as in Table 1. Detectiont P P PO PO F P P AAS MS MS P P P P Range1 pg ml- * 5 x 10-2 mol dm-3 0-25 mol dm-3 0-100 0.88-31.7 0.86-2 1.4 1.3 x 10-4- 10-4-10-2 1.1-27.8 - 5-1 100 5440 2.5-25 1-12 1-12 0-72 Detection limit/ pg ml-1 - 0.58 - - 0.53 - - 0.0001 - - 1.0 - - 0.3 RSD (To) =1.0 - - t l . O 7.0 0.95 0.40 1.5 - - - 0.5 0.5 0.8 Sampling frequency1 h- - 24 - - 60 36 36 220 - - 35 90 90 15 Real samples - - Serum - Urine Water Water Soil and plants - - - - - - Reference 288 289 290 29 1 292 26 26 293 294 294 295 296 296 297 phenyl-2-pyrazolin-5-one has been exploited in FI.The unstable red intermediate was monitored spectrophotometric- ally at 548 nm.2033204 Azomethine-H was used as a dye formation reagent for the determination of boron.28+283 Chemiluminescence reactions The chemiluminescence detection of sulphite was achieved by measuring the radiation emitted by sulphite on oxidation by permanganate in an acidic solution, the signal being greatly increased by the addition of a sensitizer.95 A chemilumines- cence system was developed by Qian et a1.103 for the determination of sulphide; the weak light emission resulting from the reaction of 7,7,8,8-tetracyanoquinodimethane with sulphite was effectively enhanced by the use of Rhodamine B as a sensitizer in an organic reaction medium. The oxidation of sulphide by sodium hypochlorite in the presence of fluorescein is another alternative to the chemiluminescent detection of sulphide, the maximum intensity of the transient signal being largely determined by the rate of mixing of sample and reagent.104 An NO, chemiluminescent analyser was used for the sensitive determination of nitrate.188 Yamada and co- workers206210 developed several chemiluminescence systems for the determination of cyanide , the weak chemilumines- cence induced by injection of alkaline cyanide into the reagent stream being effectively sensitized by the presence of uranine in the organic medium.A synthetic anionic luminol derivative was bound to a strong anion-exchange resin for the chemilu- minescence determination of anions.152 Ion-pair formation reactions This type of reaction is closely related to liquid-liquid extraction and has been studied extensively in this laboratory.Nitrate forms ion pairs with copper(i)-organic chelate com- plexes, the atomic absorption of copper from the organic phase being proportional to the analyte concentration. 186,199 Anions have been determined on the basis of the formation of ion pairs with ferroin; the ion pairs were subsequently trans- ferred into the organic phase and monitored spectrophoto- metrically at 570 nm (23 anions were tested225). The indirect atomic absorption spectrometric determination of perchlorate has been described, based on the liquid-liquid extraction of copper( 1)/6-methylpicolinaldehyde azine/perchlorate .I26 Per- chlorate was also measured as the Brilliant Green-perchlorate complex at 640 nm after continuous extraction into the organic phase (benzene).127-129 Acid-base reactions The concentration of hydroxide ion has been determined by FI titration284-287 based on an acid-base reaction. Detection was carried out spectrophotometrically with phenolphthalein as indicator2869287 or by means of a pH electrode.284~285 Enzymic reactions An enzymic method was developed by Masoom and Town- shendgl for the determination of sulphite. On passing through the immobilized enzyme column, sulphite was oxidized to sulphate, with the release of hydrogen peroxide which was measured amperometrically . Types of Detection The distribution of the different detection techniques used for the determination of anions by FI is shown in Fig.6. Spectrophotometry is the most widely used detection technique because of its simplicity, the availability of instru- ments and the rapid response of the detectors, the variety of colour-forming reactions available and the adequate sensitiv- ity achieved in most instances. It was preferentially used in the determination of phosphorus anions,21-5* 764-682301 sul- phate,63,69-86 nitrate and nitrite,149.161 and has been applied to virtually all of the anions listed in the tables. Potentiometry is second to spectrophotometry in terms of frequency of use. The chief advantage of this technique is its instrumental simplicity, provided a direct response to the analyte can be obtained. There are some exceptions when the detector responds to the analyte reaction product.The slow dynamic response of electrodes in the flow system is the main shortcoming of this technique, particularly at low concentra- tions, although some improvements have been made by polishing the electrode surface133,136 or by adding an appro- priate concentration of analyte to the carrier solution.134 Potentiometry has been widely used in the determination ofANALYST, NOVEMBER 1991, VOL. 116 1106 70 60 50 E 2 40 a 30 c Z 20 10 0 Fig. 6 P PO A F CL AAS Others Detection technique Freauencv of use of the different detection techniaues (a6breviatio;s as Ln tables) fluoride,13@-141 and in the determination of phosphate,61 sulphate,87,88 sulphide97.98 nitrite,l@ nitrate,182-185 cyan- molybdenum(v1) ,247,248 salicylate290 and oxalate.291 Amperometry has been used in the determination of phosphate,54-58 sulphite ,91 sulphide ,100 nitrite ,162,163,165-167 nitrate,179J80 cyanide,218.222 thiocyanate,224 iodide,269,270 iodate273 and molybdenum(v1) .246 Spectrofluorimetry is usually more sensitive than spectro- photometry.The lack of spectrofluorimetric determinations of anions (only 5%) is probably due to the paucity of reaction systems, some of which require the application of indirect methods. Spectrofluorimetry has been used for the determina- tion of phosphate,52753 nitrite ,168,169 nitrate ,187 cyanide ,2113212 fluoride ,146,148 iodide ,268 boron277 and oxalate .292 Chemiluminescence methods have a reputation for high sensitivity and wide calibration ranges; however, the lack of chemiluminescence reactions and their poor selectivity have limited their application.Some improvements have been achieved in the determination of anions by FI, and this has been applied to sulphite,95 sulphide,103~1~ nitrate,188 cyan- ide ,206-210 chloride114 and bromide.259 Atomic absorption spectrometry used to develop an indirect method for the determination of anions, which usually entails a liquid-liquid extraction, substitution or precipitation process. It has been used for the determination of sulphide,99 nitrate ,186,199 nitrite ,199 cyanide ,223 chloride ,105,106 perchlor- ate,126 molybdenum(v1) ,249 vanadium(v)256 and arsenic(v) .293 Other detection techniques used in the determination of anions by FI include inductively coupled plasma atomic emission spectrometry,62.89,239,257 coulometry,59,60 molecular- emission cavity analysis ,% mass spectrometry ,294 thermo- chemical methods267 and those based on chemical sen- ide,213-217 ha~ide~,lO7-113,258,264,271,272,276 hydroxide ion,284,285 sors - 147,148,212 Multi-determinations Multi-determination is one of the more interesting trends in analytical chemistry on account of advantages such as a high analytical rate, efficiency and economy.302.303 About 10% of the papers examined reported the multi-determination of anions by FI.They can be divided into three groups. Physical Spectral Resolution Configuration V (Fig. 2) was adapted to split the sample plug into two (or more) streams merging with different reagent streams for spectrophotometric determinations such as that of phosphate by the Phosphomolybdenum Blue method and nitrate (or nitrite) by the Indophenol Blue method30.33 and sulphate by the turbidimetric method with barium(i1).Mer- cury(~~) thiocyanate and iron(ri1) were used as reagents for chloride ,63 while the diazotization-coupling reaction was used for nitrite, nitrate being reduced on-line to nitrite with a copper-coated cadmium c01urnn.120~150,170,189-192 The Phos- phomolybdenum Blue method was applied to phosphate, while phosphonate was oxidized to phosphate with sulphite, after which they were detected by two parallel66 or series65 detectors. Pyridine-barbituric acid and chloramine-T were used as reagents for cyanide, while the diazotization-coupling reaction was used for the determination of nitrite by reversed FI.2(W Chloride and nitrate were determined potentiometric- ally113 and simultaneously with phosphate by the Phospho- molybdenum Blue method by means of a spectrophotometric determination.191 Separation by High-performance Liquid Chromatography (Time Resolution) Related species contained in a sample have previously been separated by HPLC (configuration VIII). Flow injection post-column detection applications reported in this context include the spectrophotometric determination of silicate, phosphate and arsenic,24*38 inorganic polyphosphate~,67~68 phosphinate, phosphonate and orthophosphate by the Molyb- denum Blue method,@ and the potentiometric determination of chloride, bromide and iodide .276 Chemical Discrimination Multicomponent mixtures have been differentiated by means of their different chemical properties.Hence, nitrate and chloride were determined simultaneously by displacement of thiocyanate from an ion-exchange column, followed by spectrophotometric detection of the iron(rr1)-thiocyanate complex; chloride was removed by previously passing the sample stream through a column containing a cation-exchange resin in the silver form.176 Cyanide and thiocyanate were determined by the pyridine-barbituric acid method, cyanide being masked by using nickel as a complex-forming agent.227 Thiocyanate and iodide were determined spectrophotometric- ally as catalysts for the redox reaction between cerium(rv) and arsenic(rI1); their simultaneous determination is based on the selective inactivation of the catalytic activity of thiocyanate.205 The sequential spectrophotometric determination of mixtures of arsenite-arsenate, arsenate-phosphate and arsenite- arsenate-phosphate using the Molybdenum Blue heteropoly acid method was accomplished by using a selecting valve to provide an appropriate medium for the indicator reaction to develop with one, two or three species in the mixture.26 Mixtures of chloride and iodide-were determined by indirect AAS.Total chloride and iodide were determined by the decrease in silver absorption, on precipitation with silver nitrate, after which chloride was determined selectively by dissolving the silver chloride precipitate in ammonia solu- tion.3" The speciation of both oxidation states of selenium (selenite and selenate) has been accomplished by Linares et al.296 by exploiting the catalytic action of selenium(1v) on the reaction between chlorate and hydrazine in a hydrochloric acid medium, the coloured oxidation product of o-toluidine being measured spectrophotometrically .Selenium(v1) did not interfere with the determination. The total selenium content was quantified after reduction of selenium(v1) to selenium(rv) with hot concentrated hydrochloric acid. Mixtures with ratios between 1 : 4 and 4 : 1 were determined with errors of less than 3 and 8%, respectively. Selectivity Interferences are not uncommon in the determination of inorganic anions owing to their similar chemical physico- chemical features. Hence, direct detection or the use of derivatization reactions does not lead to particularly selective analytical methods.ANALYST, NOVEMBER 1991, VOL.116 1107 Flow injection can advantageously replace manual ana- lytical methodologies because of its tolerance to matrix effects. The enhanced selectivity achieved can be the result of physico-chemical kinetics,305 material separation or signal discrimination,306 or a combination of two or all of these effects. The multi-determinations described above are the ultimate goal in terms of selectivity. Kinetics and selectivity are very closely related.307 The intrinsic nature of FI makes it more tolerant to foreign species than its manual counterparts, as no unwanted reactions develop to a significant extent during the short residence time typical of FI. The FI spectrophotometric determination of cyanide with several reagents2~204,226,*27 is a representative example.On the other hand, the implementation of stopped- flow modes in FI configurations (manifolds I and IV) avoids the occurrence of parasitic signals from the sample matrix because of the relative rather than the absolute character of the measurements. Hence, the natural colour of wine does not interfere when sulphite23 is determined spectrophotometric- ally from reaction-rate measurements. The ready implementation of non-chromatographic con- tinuous separation techniques in FI configurations (Fig. 3) also enhances selectivity indirectly through the isolation of the analyte anions from the sample matrix.308 It is interesting to note that the kinetics of these processes can also contribute to selectivity.The selectivity factor (ratio of the tolerated level for the FI-extraction method to that of its manual counter- part) is very high for various anionic species in the determi- nation of cationic chelates and atomic absorption measure- ment of the metal content in the organic phase after liquid-liquid extraction.309 One of the most salient advantages of fast detectors (e.g., rapid-scan voltammetric and spectrophotometric diode-array detectors) is the ease of implementation of multi-determi- nation procedures based on signal discrimination. Although interesting developments have been reported in this respect ,31@312 none were concerned with the determination of anions. Sensitivity As a rule, FI methods are less sensitive than their manual and air-segmented counterparts because no chemical equilibrium is attained and because of the dispersion of the sample plug in the carrier(s). Nevertheless, the great versatility of this technique allows sensitivity to be enhanced by introducing appropriate modifications into FI manifolds such as reversed FI, preconcentration steps and multi-peak recordings.In reversed FI modes (configuration I1 in Fig. 1 and VII in Fig. 2), the sensitivity is clearly increased compared with the conventional alternative because the amount of sample in the reaction zone increases with increasing dispersion, albeit at the expense of the large sample volume required. This approach is particularly useful for implementing on-line monitoring of anions in waters and industrial effluents. 146 The use of non-chromatographic continuous separation techniques in FI configurations (Fig.3) allows the implemen- tation of preconcentration approaches based on the use of a large sample volume, retention of the analyte anion in one phase and subsequent elution with a small volume and rapid transport to the detector. The preconcentration factor thus achieved ranges from 2 to 10000. The use of derivatizing chemical reactions to facilitate the retentiodelution process is of particular relevance in this context. The use of flow-through chemical sensors in FI manifolds based on the integration of retention (reaction) and detection by using appropriate supports in the flow cell is the most recent approach in this context .3*3 These systems allow the development of kinetic 'in situ' concentration methods which circumvent one of the most serious drawbacks of preconcentration methodologies, namely, the need for a large sample volume (manifold XIII, Fig.4). It is also possible to enhance sensitivity in FI through the use of multi-peak recording alternatives, exploiting fast detector capabilities and the repetitive passage of the reacting plug through the flow cell, by using an open-closed assembly or the iterative change of the flow direction (configurations XI11 and XIV in Fig. 4). Rather than a single FI peak, a multi-peak recording is obtained from a single injection. By summing one of the parameters (e.g., peak area, peak height) of two or more peaks, the analyte concentration range can be auto- matically amplified by several orders of magnitude.Such is the situation with the FI determination of nitrite by using a traditional spectrophotometric reaction. In the conventional mode, the determination limit is 0.2 pg ml-1, whereas if an amplification method is applied with the aid of a diode-array detector, the determination limit is 0.002 yg ml-1.314 Analysis of Real Samples Of the 300 papers examined, only 109 dealt with real samples, distributed as follows: water, 64 (including waste water, 6; sea-water, 6; rainwater, 4; and melted snow, 1); plants, 11; soils, 9; urine, 6; plasma and serum, 5; chemicals, 5; materials, 3; rocks, 3; fertilizers, 2; foodstuffs, 1; fruit, 1; beverages, 1; milk, 1; saliva, 1; smoke, 1; air, 1; and salt, 1. The over-all number exceeds 129 as some papers dealt with more than one type of sample.The lack of papers on sample analysis is obviously due to the interference of foreign species contained in the sample matrix. However, some improvements have been achieved. Hence, a column packed with cation-exchange resin was incorporated in an FI system to remove interfering metal ions present in the sample for the spectrophotometric determination of sul- phate.79,83,86 A dialysis or gas-diffusion unit is usually used to trap the analyte from the sample matrix, as in the turbidi- metric determination of urinary inorganic sulphate,gO and the potentiometric determination of anionic cyanide202-204.226 and thiocyanate .226 Soil samples are usually pre-treated manually to obtain the liquid extracts; however, we have developed an automated method for the direct introduction of soil samples into an FI system.The weighed sample was placed in the sample cell and inserted into the manifold. It was leached for 30 s by circulating 0.1 mol dm-3 hydrochloric acid through the cell at 80 "C under ultrasonic irradiation. The extraction was then injected into the carrier stream as the valve was switched. The over-all sampling rate achieved was 25 samples h-1.283 Bergamin et a1.241 developed an on-line electrolytic alloy dissolution method for the determination of molybdenum. A direct current pulse was applied to the polished steel sample and the dissolved sample was driven to the detector by the carrier stream. Comparison With Other Techniques The paucity of selective, sensitive methods for the determi- nation of anions made the inception of ion chromatography a revolution within the analytical chemistry of these species and led to the development of more complex and expensive, albeit also clearly more sensitive and selective, methods than their earlier counterparts.Flow injection and HPLC have been compared.315 In spite of the similarities derived from their hydrodynamic character, they differ in major respects such as the high separation potential of HPLC, which is of particular relevance to ions on account of their intrinsic analytical features and permits the determination of from 2 to more than 20 anions in the same sample. On the other hand, FI is more useful than HPLC for the determination of one or two anions in a large number of samples in a straightforward, automated and fast way, and this is particularly appealing to control laboratories.The implementation of continuous separation techniques such as liquid-liquid extraction, distillation, ion1108 ANALYST, NOVEMBER 1991, VOL. 116 exchange, precipitation, dialysis and gas diffusion in FI configurations allows the two important properties of sensitiv- ity and selectivity to be enhanced in an indirect fashion. These advantages have increased the potential of FI even further with respect to ion chromatogTaphy and other analytical techniques according to Braun and Zsindely.316 Conclusions Any significant practical contribution to the analytical chem- istry of inorganic anions is, in principle, of some interest on account of the well-known difficulties encountered in develop- ing suitable analytical methods for these species.This review was compiled with the aim of promoting the use of FI by routine control laboratories by transferring the analytical methods reported in the literature to the daily work of the analytical chemist involved in monitoring anionic species in a variety of matrices encountered in environmental, phar- maceutical, food and clinical chemistry, among others. Flow injection methods are straightforward, fast and can easily be automated; in principle, they only require the purchase of items such as peristaltic pumps, flow cells, Teflon tubing, connectors and an autosampler with computer control, as they are compatible with most of the instruments typically used in laboratories (spectrophotometers, spectrofluorimeters, potentiometers). The increasing commercialization of FI analysers should also contribute to the gradual adoption of FI by control laboratories.Compared with other analytical alternatives, FI allows the development of more sensitive, selective, precise and rapid methodologies which can be applied to a wider range of situations and standardized much more readily. We should not wish to leave the reader with the impression that the determination of anions by FI is already on the decline from the point of view of basic research. There are still a variety of unexplored aspects which offer interested workers promising prospects. Hence, some of the issues dealt with in this review ( e . g . , the use of fast detectors and the iterative passage of the sample or sampleheagent plug through the detector) have not yet been exploited to their full potential.Miniaturization and the increasing efficiency of non-chroma- tographic continuous analytical separation techniques will hopefully permit the development of methods for the auto- mated determination of anions that are clearly superior to those used at present.308 The direct introduction of solid samples into continuous automated systems as a means of avoiding tedious preliminary manual stages (e.g., dissolution, digestion) might lead to interesting prospects for the determination of anions in a large variety of samples judging by the promising results obtained to date using this method.283 The Comision Interministerial de Ciencia y Tecnologia (CICyT) is thanked for financial support.References Rfiiitka, J., and Hansen, E. H., Anal. Chim. Acta, 1986,179.1. Valcarcel, M., and Luque de Castro, M. D., Flow injection Analysis: Principles and Applications, Ellis Horwood, Chiches- ter, 1987. RfiiiEka, J., and Hansen, E. H., Flow Injection Analysis, Wiley, New York, 2nd edn., 1988. Karlberg, B., and Pacey, G. E., Flow Injection Analysis. A Practical Guide, Elsevier, Amsterdam, 1989. Burguera, J. L., Flow Injection Atomic Spectroscopy, Marcel Dekker, New York, 1989. Valcarcel, M., and Luque de Castro. M. D., Fresenius J. Anal. Chem., 1990,337,662. Rocks, B., and Riley, C., Clin. Chem., 1982, 28, 409. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 Riley, C., Rocks, B., and Sherwood, R.A., Talanta, 1984,31, 879. Linares, P., Luque de Castro, M. D., and Valcarcel, M., Rev. Anal. Chem., 1985,8, 229. Rios, A., Luque de Castro, M. D., and Valcarcel, M., J. Pharm. Biomed. Anal., 1985, 3 , 105. Luque de Castro, M. D., and Valcarcel, M., J. Pharm. Biomed. Anal., 1989, 7, 1291. Lazaro, F.. Luque de Castro, M. D., and Valcarcel, M., Analusis, 1985, 13. 147. Luque de Castro, M. D., and Valcarcel, M., Int. J . Environ. Anal. Chem., 1990, 38, 171. Fang, Z., Anal. Chim. Acta, 1986, 180,8. Fang, Z., Xu, S., Wang, X., and Zhang, S., Anal. Chim. Acta, 1986, 179, 325. Osborne, B. G., and Tyson, J. F., Int. J . Food Sci. Technol., 1988, 23, 541. Gallego, M.. Luque de Castro, M. D., and Valcarcel, M., At. Spectrosc., 1985, 6, 16.Tyson, J. F., Adeeyinwo, C. E., and Bysouth. S. R., J . Anal. At. Spectrom., 1989, 4, 191. van Staden, J. F.. Water SA, 1987, 13, 197. Williams, W. J., Handbook of Anion Determination, Butter- worth, London, 1979. RGiiEka, J., and Stewart. J. W. B., Anal. Chim. Acta, 1975,79, 79. RfiiiEka, J.. and Hansen, E. H., Anal. Chim. Acta, 1975, 78, 145. Hansen, E. H., and RfiiiEka. J., Anal. Chim. Acta, 1976, 87, 353. Narusawa, Y., Anal. Chim. Acta, 1988, 204, 53. Schulze, G., and Thiele, A., Fresenius Z. Anal. Chem., 1988, 329, 71 1. Linares, P., Luque de Castro, M. D., and Valcarcel, M., Anal. Chem., 1986,58, 120. Korenaga, T.. and Okada, K., Bunseki Kagaku, 1984,33,683. Worsfold, P. J., and Clinch, J. R., Anal. Chim. Acta, 1987,197, 43. Fang, Z., Sun, L., Gao, Z., Zhu, Y., Wang, X., and Li, L., Turang Tongbao, 1982,4,40.Stewart, J. W. B., and RfiiiEka, J., Anal. Chim. Acta, 1976,82, 137. Ogata, K., Taguchi, K., and Imanari, T., Bunseki Kagaku, 1982, 31, 641. Bergamin Fo, H., Zagatto, E. A. G., Krug, F. J., and Reis, B. F., Anal. Chim. Acfa, 1978, 101, 17. Reis, B. F., Zagatto, E. A. G., Jacintho, A. O., Krug, F. J., and Bergamin Fo, H., Anal. Chim. Acta, 1980, 119,305. Janse, T. A. H. M., van der Wiel, P. F. A., and Kateman, G., Anal. Chim. Acta, 1983, 155, 89. Jacintho, A. O., Kronka, E. A. M., Zagatto, E. A. G., Arruda, M. A. Z., and Ferreira, J. R., J. Flow Injection Anal., 1989, 6, 19. Hansen, E. H., Krug, F. J., Ghose, A. K., and RiiiCka, J., Analyst, 1977, 102,714. Israel, Y., and Barnes, R. M., Analyst, 1989, 114, 843.Narusawa, Y., and Hashimoto, T., Chem. Lett., 1987, 1367. Scholz, F., Henrion, G.. and Weinberg, K., 2. Chem., 1986,26, 298. List, D., Ruwisch, I., and Langhans, P., Fluess. Obst., 1986,53. 10. Johnson, K. S., and Petty, R. L., Anal. Chem., 1982,54, 1185. Basson, W. D., van Staden, J. F., and Cattin, P. M., Fresenius 2. Anal. Chem., 1981, 307, 373. Royset, O., Anal. Chim. Acta. 1985, 178, 217. Wang, S., Mei, K.. and Luo, Y., Lihua Jianyan, Huaxue Fence. 1988, 24, 305. van der Berg, J. H. M.. Deedler, R. S., and Egberink, H. G. M., Anal. Chim. Acta, 1980, 114, 91. Motomizu, S., Wakimoto, T., and TBei. K., Talanta, 1983,30, 333. Yoza, N., Kurokawa, Y., Hirai, Y., and Ohashi, S., Anal. Chirn. Acta, 1980, 121, 281. Engelhardt, H., and Klinkner, R., Fresenius 2. Anal.Chem., 1984, 319, 277. Hirai, Y., Yoza, N., and Ohashi, S., Anal. Chim. Acta, 1980, 115, 269. Hirai, Y., Yoza, N., and Ohashi. S., Chem. Lett., 1980, 499. Hirai, Y., Yoza, N.. and Ohashi. S., Bunseki Kagaku, 1981.30, 465.ANALYST, NOVEMBER 1991, VOL. 116 1109 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 7s 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 9s 96 97 Linares. P., Luque de Castro. M. D., and Valcarcel. M., Talanta. 1986, 33, 889. Motomizu, S., Mikasa, H., Oshima, M.. andTBei, K., Bunseki Kagaku, 1984.33, 116. Fogg, A. G., Cripps, G. C., and Birch, B. J.. Analyst, 1983,108, 1485. Fogg, A. G.. and Bsebsu, N. K., Analyst, 1982, 107,566. Fogg, A. G . , and Bsebsu, N. K.. Analyst, 1981, 106, 1288. Ikariyama, Y., and Heineman, W.R., Anal. Chern., 1986, 58, 1803. Harden, S. M., and Nonidez, W. K., Anal. Chern., 1984, 56, 2218. Hori. T., and Fujinaga, T., Talanta, 1983.30, 925. Fujinaga, T., Okazaki. S . , and Hori. T., Bunseki Kagaku, 1980, 29, 367. Alexander, P. W., and Koopetngarm, J., Anal. Chirn. Acta. 1987, 197, 353. Jacintho, A. O., Zagatto, E. A. G . , Bergamin F", H., Krug, F. J.. Reis, B. F., Bruns, R. E., and Kowalski, B., Anal. Chirn. Acta, 1981, 130, 243. Basson, W. D.. and van Staden, J. F.. Water Res.. 1981,15,333. Hirai, Y., Yoza, N., and Ohashi, S . , J . Chrornatogr., 1981.206, 501. Baba, Y., Yoza, N., and Ohashi, S . , J . Chrornatogr., 1985,318, 319. Baba, Y., Yoza, N . , and Ohashi, S., J . Chromatogr., 1984,295, 153. Yoza, N., Sagara. Y., Morioka. H., Handa, T., Hirano, H., Baba, Y..and Ohashi, S . , J . Flow Injection Anal., 1986, 3, 37. Brazell. R . S . , Holmberg, R. W., and Noneyhan. J . H., J . Chrornatogr., 1984, 290, 163. Krug, F. J., Bergamin Fo, H., Zagatto, E. A. G . , and Storgaard Jorgensen, S . , Analyst, 1977, 102, 503. van Staden. J . F.. Fresenius Z . Anal. Chern.. 1982, 310, 239. Baban, S.. Beetlestone, D.. Betteridge, D.. and Sweet, P., Anal. Chirn. Acta, 1980. 114, 319. Moller, J . , and Winter, B., Fresenius Z . A n d . Chern.. 1985, 320, 4S1. Krug, F. J . , Zagatto, E. A. G., Reis, B. F., Bahia F", O., and Jacintho, A. 0.. Anal. Chirn. Acta, 1983, 145, 179. van Staden, J . F.. Water SA, 1986. 12, 43. Toei, J . . Analyst, 1987. 112, 1067. Marsden, A. B., and Tyson, J. F., Anal. Proc., 1989, 26, 157.Xia, B.. and Zhang, L., Fenxi Huaxue, 1987. 15, 465. Kuznetsur, V . , and Hangos, M., Magy. Kern. Foly., 1982, 88, 569. Eremina, I . D., Shpigun, L. K.. and Zolotov, Yu. A., Zh. Anal. Khirn., 1987, 42, 1631. van Staden, J . F., and Bassen, W. D.. Lab. Pract., 1980, 29, 1279. Jones, E. A . , Anal. Chirn. Acta, 1984, 156, 313. Jones. E. A.. MINTEK. 1983, M111, 32. Madsen. B. C., and Murphy, R. J., Anal. Chern.. 1981, 53, 1924. Koch, B., FreJenius Z. Anal. Chern.. 1988,329, 707. Kondo, 0.. Miyata, H., and TBci, K., Anal. Chirn. Acta, 1982, 134, 353. Nakashima, S . . Yagi, M., Zenki, M., Doi. M., and TBei, K., Fresenius Z . Anal. Chem., 1984, 317, 29. Coetzee. J . F., and Gardner, C. W., Jr.. Anal. Chern., 1986.58, 608. Shpigun, L. K.. Eremina, I . D., and Zolotov, Yu.A., Zh. Anal. Khirn.. 1986. 41. 1557. Cox, A. G., McLeod. C. W., Miles, D. L., and Cook, J. M., J . Anal. At. Spectrorn., 1987. 2. 553. Burgucra, J . L., and Burguera. M.. Anal. Chirn. Acta, 1984, 157. 177. Ma?oom, M.. and Townshend, A., Anal. Chirn. Acta, 1986, 179, 399. Jenke. D. R., Anal. Chirn. Acta. 1989, 220, 281. Hirai. Y., Yoza. N . , and Ohashi. S., Bunseki Kagaku, 1982,31, 681. Sullivan. J. J.. Hollingworth, T. A., Wekell, M. M., Newton, R. T.. and Larose, J . E., J . Assoc. Off. Anal. Chern., 1986.69. 542. Yamada, M., Nakada, T.. and Suzuki, S., Anal. Chirn. Acra, 1983, 147, 401. Desgupta, P. K . . and Yang. H. C.. Anal. Chern., 1986.58.2839. Glaister. M. G.. Moody. G. J., andThomas, J . D. R., Analyst. 1985, 110. 113. 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 I42 Duffield, E.J., Moody, G . J., and Thomas, J. D. R., Anal. Proc., 1980, 17, 533. Petersson, B. A., Fang, Z., RGZiEka, J . , and Hansen, E. H . , Anal. Chirn. Acta, 1986, 184, 165. Milosavljevic, E. B., Solujic, L., Hendrix, J . L., and Nelson, J. H., Anal. Chern., 1988, 60, 2791. Legett, D. J., Chen, N. H., and Mahadevappa, D . S., Anal. Chirn. Acta, 1981, 128, 163. Kamson, 0. F., Anal. Chim. Acta, 1988, 211,299. Qian, X., Guo, Y., Yamada. M., Kobayashi, E . , and Suzuki. S . , Talanta, 1989, 36, 505. Burguera, J . L., and Townshend, A., Anal. Chirn. Acta. 1980, 114, 209. Martinez-Jimenez, P., Gallego, M., and Valcarcel.M., Anal. Chern., 1987, 59, 69. Valcarcel, M., and Gallego, M., TrAC, Trends Anal. Chern. (Pers. Ed.), 1989, 8, 34. van Staden, J. F., Anal. Chirn. Acra, 1986, 179, 407. van Staden. J. F., Anal. Lett., 1986, 19, 1407. Trojanowicz, M., and Matuszewski, W., Anal. Chirn. Acra, 1983, 151, 77. Hauser. P. C., Tan, S. S., Cardwell, T. J., Cattrall, R . W., and Hamilton, 1. C., Analyst, 1988, 113, 15.51. Alegret, S., Alonso, J., Bartroli, J., Garcia-Raurich, J.. and Martinez-Fabregas, R., J . Pharrn. Biorned. Anal., 1988,6,749. Krawczyinski vel Krawczyk, T., Szostek, B., and Trojanowicz, M., Talanta, 1989, 36, 811. Cardwell, T. J . , Cattrall, R. W.. and Hauser, P. C., Anal. Chirn. Acta, 1988, 214, 359. Cooper, M. M., and Spurlin, S. R., Anal. Lett., 1986, 19,2221.RGiiEka, J., Stewart, J . W. B., and Zagatto. E. A. G., Anal. Chirn. Acra. 1976, 81, 387. Krug, F. J.. Pessenda, L. C. R., Zagatto, €. A. G., Jacintho, A. 0.. and Reis. B. F., Anal. Chirn. Acta, 1981, 130,409. Miyazaki, M., Okubo, N., Hayakawa, K., and Umeda, T., Chern. Pharrn. Bull.. 1984,32. 3702. Eremina, I . D., Shpigun, L. K., and Zolotov, Yu. A . , Zh. Anal. Khirn., 1989, 44, 399. Thijssen, P. C., Prop, L. T. M., and Kateman, G., Anal. Chirn. Acta, 1985, 174, 27. Frenzel, W., Fresenius Z . Anal. Chern., 1988, 329, 660. Zaitsu, T.. Maehara, M.. and TBei. K., Bunseki Kagaku, 1984, 33. 149. Miller, K. G., Pacey, G . E.. and Gordon, G., Anal. Chern., 1985, 57, 734. Pacey. G. E.. Hollowell, D. A.. Muller, K. G., Straka. M. R., and Gordon, G., Anal. Chirn.Acta, 1986, 179, 259. Gordon, G.. Yoshino, K., Themelis. D. G., Wood, D., and Pacey, G. E., Anal. Chim. Acta, 1989, 224, 383. Themelis, D. G., Wood, D., and Gordon, G., Anal. Chirn. Acfa. 1989, 225. 437. Gallego, M., and Valcarcel, M., Anal. Chirn. Acta, 198.5, 169, 161. Burns, D. T., Chimpalee, N., and Harriott. M., Anal. Proc., 1989, 26, 4. Burns, D. T., Chimpalee, N., and Harriott. M., Anal. Chirn. Acta. 1989, 217, 177. Harriott, M.. and Burns, D. T., Anal. Proc., 1989. 26, 315. Slanina, J., Lingerak, W. A.. and Bakker, F., Anal. Chim. Acta, 1980, 117, 91. Toth, K . . Fucsko, J . , Lindner. E . , Feher, Zs., and Pungor, E . , Anal. Chirn. Acta, 1986, 179, 359. van der Winkel, P., and Backer. G., Anal. Chirn. Acta. 1983, 145, 207. van Dort, W. J . , and van Eerd, E.J . J . M., Anal. Chirn. Acta, 1983, 155. 21. Frenzel, W.. and Bratter, P.. Anal. Chirn. Actu, 1986,185,127. Frcnzel, W.. and Bratter, P.. Anal. Chirn. Acra, 1986, 187, 1. Davey. D. E., Mulcahy, D. E., and O'Connell. G . R., Arral. Letr., 1986, 19, 1387. Frenzel, W., and Bratter, P., Anal. Chirn. Acta. 1986, 188, 151. Fucsko, J., Toth, K., and Pungor, E., Anal. Chirn. Acta, 1987, 194. 163. Cui, H., and Fang, Z., Fenxi Huuxue, 1987. 15, 825. Han, W., and Fan, L., Fenxi Huaxue, 1986, 14, 387. Okabayashi, Y., Hikawa. M., Nakagawa, T., Tanaka, H., and Chikuma, M., Analyst, 1989, 114, 1267. Toda, K., Sanemasa, I., and Deguchi, T., Bunseki Kagaku, 1985.34. 31.1110 ANALYST, NOVEMBER 1991, VOL. 116 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 Wada, H., Mori, H., and Nakagawa, G., Anal.Chim. Acta, 1985, 172,297. Leon-Gonzalez, M. E., Santos-Delgado, M. J., and Polo-Diez, L. M., Anal. Chim. Acta, 1989,219, 329. Cardwell, T. J., Cattrall, R. W., and Mitri, M., Anal. Chim. Acta, 1988, 214,433. Chen, D., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1990, 230, 137. Russell, D. A., and Narayanaswamy, R., Anal. Chim. Acta, 1989, 220, 75. Chen, D., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1990, 234, 345. Zagatto, E. A. G., Jacintho, A. O., Mortatti, J., and Bergamin FO, H., Anal. Chim. Acta, 1980, 120, 399. Lerigue, D., and Ughetto, L., Eau Ind. Nuisances, 1983,75,64.LAzaro. F., Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1986, 179,279. Jones, E. A., MINTEK, 1985, M200,65. van Staden, J. F., Joubert, A. E., and van Vliet, H. R., Fresenius Z. Anal. Chem., 1986, 325, 150. Yuan, Y., Fenxi Huaxue, 1987, 15,454. Caiiete, F., Rios, A., Luque de Castro, M. D., and Valcarcel, M., Analyst, 1988, 113, 739. Bergamin FO, H., Reis, B. F., and Zagatto, E. A. G., Anal. Chim. Acta, 1978,97,427. Nakashima, S . , Yagi, M., Zenki, M., Takahashi, A., and TGei, K., Anal. Chim. Acta, 1983, 155,263. Verma, K. K., and Stewart, K. K., Anal. Chim. Acta, 1988,214, 207. Lynch, T. P., Analyst, 1988, 113, 1597. Motomizu, S., Rui, S. C., Oshima, M., and TGei, K., Analyst, 1987,112, 1261. Kok, W.Th., Thuy, D. T., Nghi, T.V., and den Boef, G., Anal. Chim. Acta, 1987, 200, 533. Cox, J. A., and Kulkarni, K. R., Analyst, 1986, 111, 1219. Trojanek, A., and Bruckenstein, S., Anal. Chem., 1986, 58, 866. Martin, G. B., and Meyerhoff, M. E., Anal. Chim. Acta, 1986, 186, 71. Chamsi, A. Y., and Fogg, A. G., Analyst, 1988, 113, 1723. Fogg, A. G., Bsebsu, N. K., and Abdalla, M. A., Analyst, 1982, 107, 1040. Fogg, A. G., and Bsebsu, N. K., Analyst, 1984, 109, 19. Motomizu, S., Mikasa, H., and T6ei, K., Talanta, 1986,33,729. Ohta, T., Arai, Y., and Takitani, S., Anal. Sci., 1987, 3, 549. Johnson, K. S., and Petty, R. L., Limnol. Oceanogr., 1983,28, 1260. Nakashima, S., Yagi, M., Zenki, M., Takahashi, A., and TGei, K., Fresenius, Z. Anal. Chem., 1984, 319, 500. Rui, S. C., Motomizu, S., and TGei, K., Bunseki Kagaku, 1987, 36,207.Madsen, B. C., Anal. Chim. Acta, 1981, 124,437. Al-Wehaid, A., and Townshend, A., Anal. Chim. Acta, 1986, 186,289. Slanina, J., Bakker, F., Bruyn-Hes, A. G. M., and Mols, J. J., Fresenius 2. Anal. Chem., 1978, 289, 38. Faizullah, A. T., and Townshend, A., Anal. Chim. Acta, 1986, 179, 233. Slanina, J., Bakker, F., Bruyn-Hes, A., and Mols, J. J., Anal. Chim. Acta, 1980, 113, 331. GinC, M. F., Reis, B. F., Zagatto, E. A. G., Krug, F. J., and Jacintho, A. O., Anal. Chim. Acta, 1983, 155, 131. Fogg, A. G., Chamsi, A. Y., and Abdalla, M. A., Analyst, 1983, 108,464. Fogg, A. G., Scullion, S. P., and Edmonds, T. E., Analyst, 1989, 114,579. Hansen, E. H., Ghose, A. K., and RfigiEka, J., Analyst, 1977, 102, 705. RfiiiEka, J., Hansen, E.H., and Zagatto, E. A., Anal. Chim. Acta, 1977,88, 1. Alegret, S., Alonso, J., Bartroli, J., and Paulis, J. M., Anal. Chim. Acta, 1984, 164, 147. Cammann, K., Fresenius 2. Anal. Chem., 1988,329, 691. Hansen, E. H., RfiEiEka, J., and Ghose, A. K., Soil Nitrogen Fert. Pollut., Proc. Rep. Res. Coord. Meet., 1980, 77. Silva, M., Gallego, M., and Valcarcel, M., Anal. Chim. Acta, 1986, 179, 341. Motomizu, S., Mikasa, H., and TGei, K., Anal. Chim. Acta, 1987, 193, 343. Yoshizumi, K., and Aoki, K., Anal. Chem., 1985,57, 737. 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 2 10 21 1 212 213 214 215 216 217 218 219 220 22 1 222 223 224 225 226 227 228 229 230 23 1 232 233 Gin& M. F., Bergamin Fo, H., Zagatto, E. A. G., and Reis, B.F., Anal. Chim. Acta, 1980, 114, 191. Anderson, L., Anal. Chim. Acta, 1979, 110, 123. van Staden, J. F., Anal. Chim. Acta, 1982, 138,403. Bermudez, B., Rios, A., Luque de Castro, M. D., and Valcarcel, M., Talanta, 1988, 35, 810. Xu, S., and Fang, Z., Fenxi Huaxue, 1983, 11,93. Schothorst, R. C., Reijn, J. M., Poppe, H., and den Boef, G., Anal. Chim. Acza, 1983, 145, 197. Schothorst, R. C., and den Boef, G., Anal. Chim. Acta, 1985, 175, 305. Schothorst, R. C., van Veen, J. J. F., and den Boef, G., Anal. Chim. Acta, 1984, 161, 27. Hulanicki, A., Matuszewski, W., and Trojanowicz, M., Anal. Chim. Acta, 1987, 194, 119. Almhofer, N., and Frenzel, W., Fresenius Z. Anal. Chem., 1988,330,494. Gallego, M., Silva, M., and Valcarcel, M., Fresenius Z. Anal. Chem., 1986,323, 50.Rios, A., Luque de Castro, M. D., and Valcarcel, M., Analyst, ~ 1984, 109, 1487. Rios, A., Luque de Castro, M. D., and Valcarcel, M., Talanta, 1984,31, 673. Figuerola, E., Florido, A., Aguilar, M., and de Pablo, J., Fresenius Z. Anal. Chem., 1988, 331, 620. Zhu, Z., and Fang, Z., Anal. Chim. Acta, 1987, 198, 25. Zhu, Z., and Fang, Z., Kexue Tongbao (Engl. Transl.), 1986, 31, 1728. Tanaka, A., Miyazaki, M., and Deguchi, T., Anal. Lett., Part A , 1985, 18, 695. Ishii, M., Yamada, M., and Suzuki, S., Anal. Lett., 1986, 19, 1591. Ishii, M., Yamada, M., and Suzuki, S., Bunseki Kagaku, 1986, 35,955. Wu, X., Yamada, M., Hobo, T., and Suzuki, S., Anal. Chem., 1989, 61, 1505. Ishii, M.. Yamada, M., and Suzuki. S., Bunseki Kagaku, 1987, 36,316. Ishii, M., Yamada, M., and Suzuki, S., Bunseki Kagaku. 1986, 35, 542.Linares, P., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1984, 161, 257. Chen, D., Luque de Castro, M. D., and Valcarcel, M., Talanta, 1990. 37, 1049. Pohlandt-Watson, C., MINTEK, 1986. M275, 7. Petak, P., and Stulik, K., Anal. Chim. Acta, 1986, 185, 171. Okumoto, C., Nagashima, M., Mizoiri, S., Kazama, M., and Akiyama, K., Eisei Kagaku, 1984,30,7. Figuerola, E., Florido, A., Aguilar, M., and de Pablo, J., Anal. Chim. Acta, 1988, 215, 283. Muller, H., in 3rd Symposium on Ion-selective Electrodes Matrafured, ed. Pungor, E., Elsevier, Amsterdam, 1980, p. 279. Fogg, A. G., and Alonso, R. M., Analyst, 1987, 112, 1071. Ma, H., Jin, L., and Yan, H., Kexue Tongbao (Chin. Ed.), 1982,27, 1145. Pihlar, B., Kosta, L., and Hristovski, B., Talanta, 1979,26,805.Polta, J. A., and Johnson, D. C., Anal. Chem., 1985,597,1373. Pihlar, B., and Kosta, L., Anal. Chim. Acta, 1980, 114,275. Haj-Hussein, A. T., Christian, G. D., and RfiiiEka, J., Anal. Chem., 1986,58, 38. Cox, J. A., Gray, T., and Kulkarni, K. R., Anal. Chem., 1988, 60, 1710. Caiiete, F., Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1989, 224. 169. Tanaka, A., Mashiba, K., and Deguchi, T., Anal. Chim. Acta, 1988,214, 259. Sweileh, J. A., Anal. Chim. Acta, 1989, 220, 64. Alexander, P. W., Haddad, P. R., and Trojanowicz, M., Anal. Chem.. 1984, 56. 2417. Storgaard Jgrgensen, S., and Regitano, M. A. B., Analyst. 1980, 105, 292. Bubnis, B. P., Straka, M. R., and Pacey, G. E., Talanta, 1983, 30,841.Lynch, T. P., Kernoghan, N. J., and Wilson, J. N., Analyst, 1984, 109, 839. de Andrade, J. C., Rocha, J. C., and Baccan, N., Analyst, 1984, 109, 645. de Andrade, J. C., Rocha, J. C., and Baccan, N., Analyst, 1985, 110, 197.ANALYST, NOVEMBER 1991, VOL. 116 1111 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 25 1 252 253 254 255 256 257 258 259 260 26 1 262 263 264 265 266 267 268 269 270 27 1 272 273 274 275 276 Ruz, J . , Torres, A.. Rios, A., Luque de Castro, M. D., and Valcarcel. M., J . Autom. Chem., 1986, 8, 70. Ruz, J . , Rios, A., Luque de Castro, M. D., and Valcarcel, M., Talanta, 1986, 33, 199. Ruz. J . , Rios, A., Luque de Castro, M. D . , and Valcarcel, M., Anal. Chim. Acta, 1986, 186. 139. de Andrade, J. C., Rocha, J .C., Pasquini, C., and Baccan, N., Analyst, 1983, 108, 621. Burns, D. T., Chimpalee, N., and Harriot, M., Anal. Chim. Acta, 1989, 225. 241. Cox, A. G., Cook, I. G., and McLeod, C. W., Analyst. 1985, 110, 331. Krug, F. J., Bahia Fo, D., and Zagatto, E. A. G., Anal. Chim. Acta, 1984, 161, 245. Bergamin Fo, H., Krug, F. J . , Reis, B. F., Nobrega, J . A., Mesquita, M.; and Souza, I. G . , Anal. Chim. Acta. 1988, 214, 397. Bergamin Fo, H., Medeiros. J . X., Reis, B. F., and Zagatto, E. A. G., Anal. Chim. Acta, 1978, 101, 9. Yoshimura, K., Matsuoka, S., and Waki, H., Anal. Chim. Acta, 1989, 225, 313. Fang, Z., and Xu, S., Anal. Chim. Acta, 1983, 145, 143. Pessenda, L. C. R . , Jacintho, A. 0.. and Zagatto, E. A. G., Anal. Chim. Acta, 1988, 214, 239. Trojanowicz, M., Hulanicki, A., Matuszewski, W., Palys, M., Fuksiewicz, A., Hulanicka, T., Raszewski, S., Szyller, J., and Augustyniak, W., Anal.Chim. Acta, 1986, 188, 165. Muller, H., and Muller, V., Anal. Chim. Acta, 1986, 180, 30. Hua, C., Jagner. D., and Renman, L., Anal. Chim. Acta, 1987, 192, 103. Tyson, J . F., and Sarkissian, L. L., Anal. Proc.. 1985, 22, 19. Hwang, J. M., Tsung, J . C., and Chen, Y. M., J . Chin. Chem. SOC. (Taipei). 1985, 32, 405. Fukasawa, T., Kawakubo, S . , and Unno, A., Anal. Chim. Acta, 1986. 193, 269. Fukasawa, T., Kawakubo, S., Okabe, T., and Mizulke, A., Bunseki Kagaku, 1984,33, 609. Nakano, S . , Tago, M., and Kawashima, T., Anal. Sci., 1989,5, 69. Betteridge, D., and Fields, B . , Anal. Chem., 1978, 50, 654. Rios, A . , Luque de Castro, M. D., and Valcarcel, M., Anal.Chem., 1986,58, 663. Patel, B., Haswell, S. J., and Grzeskowiak, R., 1. Anal. At. Spectrom., 1989,4, 195. Hirata, S., Umezaki, Y . . and Ikeda, M., Anal. Chem., 1986,58, 2602. van Staden, J. F., Analyst, 1987, 112, 595. Shakir. I . M. A., and Faizullah, A. T., Analyst, 1989, 114, 951. Anfalt, T., and Twengstrom, S . , Anal. Chim. Acta, 1986, 179, 453. Anagnostopoulou, P. I . , and Koupparis, M. A., Anal. Chem., 1986, 58, 322. Yonehara, N., Akaike, S . , Sakamoto, H., and Kamada, M., Anal. Sci., 1988, 4,273. Carlsson, A., Lundstrom, U., and Olin, A.. Talanta, 1987, 34, 615. Ohura, H., Imato, T., Yamasaki, S . , and Ishibashi, N., Bunseki Kagaku, 1986,35,349. Fujiwava, K., and Fuwa, K., Anal. Chem., 1985, 57, 1012. Al-Wehaid, A., and Townshend, A., Anal.Chim. Acta, 1987, 198, 45. Elvecrog. J . M., and Carr, P. W., Anal. Chim. Acta, 1980, 121, 135. Tanaka, A., Obata. K.. and Deguchi, T., Anal. Sci., 1986, 2. 197. Wasberg, M., and Ivaska, A., Anal. Chim. Acta, 1986,179,433. Pratt, R., Jr., and Johnson, D. C., Anal. Chim. Acta, 1983, 148, 87. Ilcheva. L., Cammann, K., and Georgieva, T., Analusis, 1988, 16, 101. Alegret, S., Florido, A., Lima, J . L. F. C., and Machado, A. A. S. C., Talanta, 1989,36. 825. Abdalla, M. A., and Al-Swaidan, H. M.. Analyst, 1989, 114, 583. Berzas Nevado, J. J., and Valiente Gonzalez, P., Analyst, 1989, 114, 989. Kamson, 0. F., Anal. Chim. Acta, 1986, 179,475. van Staden, J . F., Anal. Chim. Acta, 1989. 219, 55. 277 278 279 280 28 1 282 283 284 285 286 287 288 289 290 29 1 292 293 294 295 296 297 298 299 300 30 1 302 303 304 30s 306 307 308 309 310 311 312 313 314 315 316 Motomizu, S . , Oshima, M., and TGei, K., Bunseki Kagaku, 1983,32, 458. Lopez-Garcia, I . , Sanchez-Pedreiio, C., Hernandez-Cordoba. M.. and Garcia-Lorente, A., Analusis, 1988, 16, 196. Anagnostopoulou, P. I., and Koupparis, M. A., J . Pharm. Sci., 1985, 74,886. Krug, F. J . , Mortatti, J . , Pessenda, L. C. R., Zagatto, E. A. G . , and Bergamin Fo, H., Anal. Chim. Acta. 1981, 125, 29. Arruda, M. A . Z., and Zagatto, E. A. G., Anal. Chim. Acfa, 1987, 199, 137. de Andrade, J . C., Ferreira, M., Baccan, N., and Bataglia, 0. C., Analyst, 1988, 113, 289. Chen, D., Lazaro, F., Luque de Castro, M. D., and Valcarcel. M., Anal. Chim. Acta, 1989, 226, 221. Astrom. O., Anal. Chim. Acta, 1979, 105, 67. Simpson, S. F.. and Holler, F. J . , Anal. Chem., 1982, 54, 43. Stewart, K. K., and Rosenfeld, A. G., J . Autom. Chem., 1981, 3, 30. Toei, J . , Saigusa, Y., and Baba, N., J . Flow Injection Anal., 1986, 3, 27. Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chem., 1988, 60, 1540. Burns, D. T., Chimpalee, N., and McKillen, G . M., Anal. Chim. Acta, 1989, 217, 183. Chang, Q., and Meyerhoff, M. E., Anal. Chim. Acta, 1986,186, 81. Fogg, A. G., Alonso, R. M., and Fernandez-Arciniega, M. A., Analyst, 1986, 111, 249. Gaetani, E., Laureri, C . F., Vitto, M., Borghi, L., Elia, G. F., and Novarini, A., Clin. Chim. Acta, 1986, 156, 71. Wang, X., and Fang. Z . , Fenxi Huaxue, 1988, 16, 912. Canham, J . S . , and Pacey, G. F . . Arlol. Lett., 1988, 21, 1619. Martinez-Lozano, C., PCrez-Ruiz, T., Tomas, V . , and Abellan, C., Analyst, 1989, 114, 715. Linares, P., Luque, de Castro, M. D., and Valcarcel, M., Analyst, 1986, 111, 1405. Kuwaki, T., TGei, K . , Akiba, M., Oshima, M., and Motomizu, S., Bunseki Kagaku, 1987,36, T132. Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chem., 1985, 57, 1803. Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1986, 179,463. Caiiete, F., Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chem., 1988, 60, 2354. Yamane, T., and Kamijyo, M., Bunseki Kagaku, 1989,38,46. Lazaro Boza, F., Luque de Castro, M. D., and Valcarcel Cases, M., Analyst, 1984, 109, 333. Luque de Castro, M. D., and Valcarcel, M., TrAC, Trends Anal. Chem. (Pers. E d . ) , 1986, 5 , 71. Martinez-Jimenez, P., Gallego. M., and Valcarcel. M., Anal. Chim. Acta, 1987, 193, 127. Valcarcel, M., Analyst, 1987, 112, 729. Smith. F., and Chang, C., The Practice of Ion Chromatography. Wiley, New York, 1983. Fritz. J . , and Gjerde, P., Ion Chromatography, Hiithig, Heidel berg, 1982. Valcarcel, M., and Luque de Castro, M. D., Non-Chromato- graphic Continuous Separation Techniques, Royal Society of Chemistry, Cambridge, in the press. Gallego, M.. Silva, M., and Valcarcel, M., Anal. Chem., 1986, 58, 2265. Caiiete, F., Rios, A., Luque de Castro. M. D., and Valcarcel, M., Anal. Chim. Acta, 1988, 214, 375. Caiiete, F., Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1988, 211,287. Bermudez, B., Lazaro. F., Luque de Castro, M. D., and Valcarcel. M., Analyst, 1987, 112, 535. Valcarcel, M., and Luque de Castro, M. D . , Analyst, 1990,115, 699. Lazaro, F., Rios, A., Luque de Castro, M. D., and Valcarcel, M., Anal. Chim. Acta, 1986, 179, 279. RGiiEka, J., and Christian, G. D., Analyst, 1990, 115, 475. Braun, T., and Zsindely, S . , TrA C , Trends Anal. Chem. (Pers. Ed.), 1990, 9, 144. Paper OlO3Y02 K Received August 29th, 1990 Accepted June 6th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911601095

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1991, VOL. 116 1113 Estimation of Effective Diffusion Coefficients of Model Solutes Through Gastric Mucus: Assessment of a Diffusion Chamber Technique Based on Spectrophotometric Analysis M. A. Desai and P. Vadgama Department of Medicine (Clinical Biochemistry), University of Manchester, Hope Hospital, Salford M68HD, UK A diffusion chamber technique based on spectrophotometric analysis t o determine effective diffusion coefficients for solutes with various relative molecular mass ( Mr) values and properties, passing through native mucus gel, is reported. For all solutes studied, a reduction in effective diffusion coefficients is observed with a retardation of solute flux by a factor of at least two. For the solutes investigated (with Mr values ranging from 126-14400 u), no consistent effect of solutes of low Mr was evident with regard t o the retarding effect of mucus; however, at high Mr values (>4000 u) the retardation was greatly enhanced.A possible relationship between charged solutes of low Mr and the degree of retardation was observed, which possibly suggests the presence of ionic interactions of the solutes with the largely negatively charged mucus gel. The results provide further evidence for the suggestion that mucus is acting more than simply as a gel support for an unstirred water layer and is perhaps a more potent diffusion barrier t o specific solutes. Keywords: Diffusion coefficient; gastric mucus; diffusion chamber; steady state Mucus forms a continuous viscoelastic gel layer that covers many epithelial surfaces, including those of the gastrointesti- nal tract, the upper respiratory tract and parts of the genitourinary tract.In particular, it constitutes an extracel- Mar barrier to the diffusion of nutrients and therapeutic agents to the absorptive epithelial surface. The precise thickness of the gel has been a matter of controversy depending on the techniques used.1.2 However, the thickness of the gel has been reported by some to be between 100 and 600 pm in the gut3.4 and its key gel forming constituent has been considered to be a 2 x 106 u glycoprotein subunit5 with a high (70% m/m) carbohydrate component,6 mainly present as oligosaccharides, confined to certain regions along the linear protein core. The glycoprotein is thought to be responsible for the physical and structural properties of mucus.' Macro- molecules such as secretory IgA, lysozyme and lactoferrin may also contribute to the protective function of mucus.3 Although rheological properties are important for the action of mucus as a lubricant, mixing barrier and particle trap, and have been the focus of much attention,%lO less information is available about mucus as a diffusion barrier. Again, some controversy exists as to whether mucus acts as a diffusion barrier by stabilizing a surface water layer,ll-12 or whether it could present an additional, more potent, and perhaps more selective barrier to the transfer of nutrients of both low and high relative molecular mass ( M , ) and drugs within the small intestine.13 Peppas et al.l4 have developed a theory for solute diffusion in intestinal mucus; this takes into account the concentration of the constituent glycoprotein, the size of the diffusing species and the density of the macromolecular cross-links.However, direct practical observation is necessary concerning the influence of mucus upon the diffusion of specific solutes. There are indications that the diffusion of certain species such as the ergo1 alkaloids,*5 aminoglycosides,I6 fatty acids17 and some antibioticsI*,lg is significantly retarded. Here, an estimation of the effective diffusion coefficients ( D e ) for a series of solutes of various M , and properties passing through a native mucus gel is reported, in an attempt to examine the precise retarding effect of mucus on the access of such species to the intestinal epithelium. Theory The diffusion coefficients of compounds with various M , values were determined using steady-state analysis.2° Mem- branes we1.e held between two well mixed chambers of concentrations c1 and c2 in the component whose diffusion coefficient was to be measured.Assuming that the mass transfer resistance of the film between the bulk fluids in the two chambers and the membrane was negligible, the transient diffusion process inside one volume element within the membrane is governed by the partial differential equation dc d2c dt dx2 -=D- where c is the concentration in the membranes, D is the diffusion coefficient, t is time, and x is the distance subject to the boundary conditions' c = c l at x = O c = c 2 at x = l (2) At steady state, eqn.(1) becomes d2c O=D- d x 2 (3) Applying the boundary conditions [eqn. (2)J to eqn. (3) gives the steady-state concentration profile Y A. c = c1 + (c2 - c1)- (4) 1 The flux at steady state is given by where A is the surface area of the membrane. The slope of the straight line in the graph of Q (total amount of solute transferred through the membrane) versus t is dQldt at steady state. Therefore, Experimental Reagents Phenolphthalein diphosphate tetrasodium salt, phenolph- thalein, 5-hydroxytryptamine, 1,3,5-trihydroxybenzene (phloroglucinol), 5-hydroxy-~-tryptophan, b-nicotinamide adenine dinucleotide (NAD) and lysozyme (muramidase E.C.1114 ANALYST, NOVEMBER 1991, VOL. 116 3.2.1.17) were purchased from Sigma (Poole, Dorset, UK). Ribonucleic acid from T.utilis (RNA), glycyrrhizic acid and cyanocobalamine (vitamin BI2) were purchased from Fluka (Glossop, Derbyshire, UK). Track-etched porous poly- carbonate membranes (0.1 pm pore size) were obtained from Poretics (Livermore, CA, USA). The remaining chemicals including the buffer components were purchased from Merck (Poole, Dorset, UK). An isotonic phosphate buffer was used consisting of 2.44 g of NaH2P04, 7.5 g of Na2HP04, 3 g of NaCl and 0.6 g of ethylenediaminetetraacetic acid in 11, pH 7.4. Porcine gastric mucus was obtained from an abattoir using animals immediately after slaughter. The stomach was split open, the luminal surface washed with water and the gel collected by scraping the intact mucosal surface with a glass slide. The mucus was frozen at -20°C until required for diffusion studies.Validation experiments on acid diffusion through fresh and frozen porcine gastric mucus have given similar responses over the entire pH spectrum;21 freezing and thawing of porcine mucus has also been demonstrated to have no apparent effect on permeability.17 Diffusion Experiment Diffusion experiments carried out at room temperature (21 k 2 "C) were performed in a Perspex diffusion chamber consist- ing of two half-cells. Each half-cell contained a round chamber of 170ml volume, held together with screw connections (Fig. 1). Two polycarbonate membranes (50 mm; 0.1 pm pore size) were placed on either side of a 200pm nylon netting spacer, and the laminate held between two stainless-steel discs. The chambers were secured with O-rings placed in position to provide a water-tight connection.The nylon spacer provided an unstirred aqueous layer for diffusion. For diffusion through mucus, thawed porcine mucus was layered on the nylon spacer before placing the upper polycarbonate membrane in position. Both chambers were filled with isotonic buffer and allowed to equilibrate for 1 h. The chambers were then emptied, one chamber was filled with buffer alone and the other with a 1 mg ml-* solution of the solute in buffer. Both chambers were filled simultaneously and rapid stirring was instituted; 1OOpl volumes of samples were taken from the buffer chamber at fixed time intervals. Initial and subsequent fixed time concentrations were deter- mined by measuring the absorbance of the solutions at 280 nm, except for phenolphthalein which was measured at 550 nm.Results and Discussion The graphs of Q (amount of solute transferred) versus time for two selected compounds 5-hydroxytryptamine and lysozyme are shown in Fig. 2(a) and (b), respectively, in the presence and absence of mucus. The plots show that the relationship Q versus t is linear and therefore the application of eqn. (6) is Motor drive A I) Fig. 1 Schematic representation of the diffusion chamber assembly valid for the operating conditions and duration of the experiment, allowing calculation of D, values. Furthermore, linear regression analysis gave regression coefficients of >0.99 [Fig. 2(a) and (b)] for both aqueous and mucus layers. The effect of mucus is to reduce Q substantially over a given time.Table 1 shows the effect of including mucus within the 200 pm diffusion gap created by the nylon spacer. For all solutes studied, a reduction in D, values is observed with a retardation of solute flux by a factor of at least two. Over the range of solutes investigated with M, values between 126 and 14400u, no consistent effect of solutes of low M, is evident with regard to the retarding effect of mucus. However, at very high M, values (>4000u) the retardation due to mucus is greatly increased as shown by a plot of log M, versus the ratio A diffusion chamber technique (IUPAC22) based on spectrophotometric analysis for the estimation of D, values of solutes that absorb in the ultraviolet region and which are therefore easier to monitor is described here.Gastrointestinal of De(aqueous) : De(mucus) (Fig. 3)- 30 (4 25 20 15 10 5 € 0 cn a 40 35 30 25 20 15 10 5 0 25 50 75 100 125 150 175 200 Time/min Fig. 2 Plots of Q (amount of solute transferred) versus time in aqueous and mucus layers. (a) 5-Hydroxytry tamine, A, aqueous (y = 0.37857 + 0.13167x, r = 0.998); B, mucus & = -9.2857 x 10-2x, r = 0.998); and (b) lysozyme, A, aqueous (y 1.3740 + 0.26857x, r = 0.996); B, mucus 0, = 24250 + 1.517 x lO-*x, r = 0.957) Table 1 Effective diffusion coefficients of selected solutes through aqueous and native porcine mucus layers Solute Phloroglucinol 5-Hydroxy-~- tryptophan Phenolphthalein 5-H ydroxytryptamine Phenolphthalein diphos- NAD Glycyrrhizic acid Cy anocobalamine RNA Lysoz yme phate OJlO-7 cm* s-1 De(aqueous): M , Aqueous Mucus De~,u,,s~ 126 78 24 3.3 220 68 14 4.9 318 83 18 4.6 387 63 14 4.5 566 49 33 1.5 663 11 1.7 6.5 840 67 27 2.5 1355 26 10 2.6 4000-8000 160 9 17.8 14400 120 4.5 26.7ANALYST, NOVEMBER 1991, VOL.116 1115 ,- 30 I i 30 I 2 Fig. 3 Plot of log M, versus ratio of De(aqueous):De(mucus) (values obtained from Table 1) mucus forms a tightly bound gel that is thought to restrict the diffusion of protein molecules.6 Estimates of D, values given in Table 1 clearly demonstrate that although species of low M , are retarded by a factor of 2-5, which is consistent with recently reported values,23 they nonetheless diffuse more freely than species of high M,. Diffusion of hydrogen ions through mucus has been demonstrated to be retarded approxi- mately 5-fold when compared with diffusion through an unstirred layer of similar thickness .24 This study has found anomalously high diffusion resistance for NAD (Fig.3) and to a lesser extent for 5-hydroxy-~- tryptophan, phenolphthalein and 5-hydroxytryptamine through mucus (Table l), raising the possibility of some specific binding of these solutes by mucus; some antibiotics bearing nitrogen groups have been shown to be selectively retarded, possibly on the basis of ionic interactions and charge differences.16 Mucus gel has also been shown to retard diffusion of ionic species such as K+ ions.lO This may be explained by the fact that mucus has a net surplus of negative charges due to the presence of charged amino, carboxyl, sulphate and neuraminic acid groups on mucus fibrils, thus providing a highly electrified surface.The charged glyco- protein is capable of retarding diffusion of other charged species by the Donnan exclusion phenomenon .25 The significant retardation of charged solutes such as NAD, 5-hydroxy-~-tryptophan and 5hydroxytryptamine in this study may therefore be attributed to the interaction of these species with charged mucus, whereas neutral species such as phloroglucinol , phenolphthalein diphosphate and cyanoco- balamine are less retarded (Table 1 and Fig. 3). The reduction of net charge by treatment with a species such as N-acetyl- neuraminidaselo or N-acetylcysteine, sodium taurodeoxy- cholate and acetylsalicylic acid24 has been reported to de- crease the effectiveness of the mucus as an ionic diffusion barrier.It can be seen from Table 1 that solutes as large as RNA (4-8000 u) and lysozyme (14 400 u), although significantly retarded, can still diffuse through the tightly formed mucus matrix and that defined diffusion coefficients can be measured as for small solutes. However, a network of channels due to the ordered arrangement of glycoprotein molecules in mucus has been proposed,26 and this could permit macromolecular transfer. Indeed, even multilayer liposomes (-200 nm) intro- duced into the lumen of the small intestine have been shown to be able to penetrate the mucus layer.27 It may be possible therefore that the existence of such channels in porcine gastric mucus can account for the finite diffusion of RNA and lysozyme. It is important to point out that lysozyme did not appear to have any proteolytic effect on mucus under the conditions used in this study, as confirmed by the straight line graph of Q versus t [Fig.2(b)], which does not show any abnormality that would have occurred in the presence of any decomposition of mucus. I .- t E 0 2 000 4 000 6 000 8000 10 000 Time/s Fig. 4 Plots of amount of 1*5I-lysozyme transferred through an aqueous layer versus time at stirring speeds of A, 200 rev min-l 0 = 105.60 + 0.29160x, r = 0.987); and B, 1400 rev min-' 0, = 536.73 + 2.7042x, r = 0.996) It should also be noted that the diffusion coefficient values for the solutes reported here are effective diffusion coeffi- cients obtained strictly under the experimental conditions described here and assumptions made in the calculation of D,.They may therefore show a deviation from literature values normally obtained under ideal diffusion conditions. However, the main purpose of this paper is to assess the technique and compare mass transfer of solutes of various M , values under identical experimental conditions. The high estimates of D, values obtained in this study may be due to the fact that very rapid stirring (1400 rev min-I) of solutions was instituted to ensure that the barrier to diffusion presented by the stagnant boundary layer was minimized, and, therefore, D, values reflected only the barrier presented by the membrane/gel/water barriers. As a consequence, the model here assumes that the barrier to diffusion presented by such a stagnant boundary layer is negligible. However, under such rapid stirring conditions, some turbulence in the non- stirring aqueous diffusion compartment in the control experi- ments was unavoidable, and may have given rise to higher D, values as a result of convection (Fig.4). Lucas28 has also reported diffusion coefficients of sodium and hydrogen ions to be of one order of magnitude higher than free solution values when similar stirred compartmental systems were used, a similar explanation being provided for this observation. This problem could have been avoided by use of smaller pore polycarbonate membranes, but then the diffusion barrier presented by the membranes themselves would have been significant, particularly for macromolecular diffusion. The highly retarded diffusion of RNA and lysozyme, however, emphasizes the role of mucus as a barrier to molecules such as final nutrient peptide or saccharide products which are required to diffuse to their hydrolase or transport sites on the epithelial membrane.The intestinal mucus coat has been shown by some workers to be an important diffusion barrier for such nutrients and other oligomers that require to be digested, transported or bound to receptor sites on the outer intestinal membrane. 13 The high diffusional resistance demonstrated for mucus here is therefore of relevance to the absorption of a range of therapeutic and other exogenous compounds to which the small intestine is exposed. The ratios of aqueous to mucus D, in this study clearly show a significant dependence on M,, and therefore provide further evidence for the suggestion that mucus is acting as more than simply a gel support for an unstirred water layer.It is not clear from this general study what the precise relationship of M , may be to mucus resistance as the solutes examined varied in charge, shape and class of compound. Future work will determine diffusional resistance using model diffusants of consistent shape and charge. Also, comparison of diffusion coefficients over an extended range of M , values remains to be carried out.1116 ANALYST, NOVEMBER 1991, VOL. 116 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 References Kerss, S., Allen, A., and Garner, A., Clin. Sci., 1982, 63. 187. Bickel, M., and Kauffman, G. L., Gastroenterology, 1981, 80, 770. Munster, D. J., Roberton, A. M., and Bagshaw, P. F., N . Z .Med. J . , 1989, 102,607. van Hoogdalem, E. J., De Boeboer, A. G., and Brimer, D. D., Pharmacol. Ther., 1989, 44, 407. Carlstedt. I., and Sheehan, J. K., Biochern. SOC. Trans., 1984. 12, 615. Allen, A., Br. Med. Bull., 1978, 34, 28. Neutra, M. R., and Forstner, J . F., in Physiology of the Gastrointestinal Tract, ed. Johnson, L. R., Raven Press, New York, 2nd edn., 1987, vol. 2, p. 975. Crowther, R. S., Marriott, C., and James, S. L., Biorheology, 1984, 21,253. Allen, A., in Physiology of the Gastrointestinal Tract, ed. Johnson, L. R., Raven Press, New York, 1981, vol. 1, p. 617. Lee, S. P., and Nicholls, J. F., Biorheology, 1987, 24, 565. Morris, G. P., Gastroenterol. Clin. Biol., 1985, 9, 106. De Simone, J. A., Science, 1982, 220, 221. Smithson, K. W., Millar, D. B., Jacobs, L. R., and Gray, G., Science, 1981, 214, 1241. Peppas, N. A., Hansen, P. J.. and Buri, P. A., Int. J . Pharrn., 1985, 20, 107. Nimmerfall, F. N., and Rosenthaler, J., Biochem. Biophys. Res. Cornmun., 1980, 94, 960. 16 17 18 19 20 21 22 23 24 25 26 27 28 Niibuchi, J . J . , Aramaki, Y., and Tsuchiya, S., Int. J . Pharm., 1986, 30, 181. Smith, G. W., Wiggins, P. M., Lee, S. P., andTasman-Jones, C.. Clin. Sci., 1986, 70, 271. Cheema, M. S., Rassing, J. E., and Marriott, C., J . Pharm. Pharmacol. Suppl., 1986, 38, 53. Kearney, P., and Marriott, C., Int. J. Pharm.. 1987, 38, 211. Hanoun, B. J. M., and Stephanopoulos, G., Biotechnol. Bioeng., 1986, 28, 829. Nicholas, C. V.. Desai, M., Vadgama, P., McDonnell, M. B., and Lucas, S., Analysr, 1991, 116,463. IUPAC, Pure Appl. Chem.. 1979, 51, 1575. Desai, M. A., Nicholas. C. V., and Vadgama, P., J . Pharrn. Pharmacol., 1991,43, 124. Turner, N. C., Martin, G. P., and Marriott, C.. J. Pharm. Pharrnacol.. 1985, 37, 776. Bokris, J. O’M., and Reddy, A. K. N., in Modern Electro- chemistry, eds. Bokris, J. O’M., and Reddy. A. K. N., Plenum Press, New York, 1973. vol. 1, p. 623. Lazarev, P. I., Dokl. Akad. Nauk SSSR, 1986, 286, 761. Brodskii, R. A., Gal’perin, Yu. M., Lazarev, P. I.. Nadkochii, V., and Popov, G. A., Dokl. Akad. Nauk SSSR, 1983,273,464. Lucas, M. L., Dig. Dis. Sci.. 1984, 29, 336. Paper 1 I02 71 8 B Received June 7th, 1991 Accepted June 20th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911601113

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1991, VOL. 116 1117 Development of a Micro-electrochemical Flow Cell Using Carbon or Gold Fibres for Voltammetric and Amperometric Analyses Chi Hua, Kamal A. Sagar, Kieran McLaughlin, Manuel Jorge, Mary P. Meaney and Malcolm R. Smyth* School of Chemical Sciences, Dublin City University, Dublin 9, Ireland The development of a micro-electrochemical flow cell using either carbon or gold fibre electrodes is described. The cell was tested using anodic and adsorptive stripping voltammetric determination of selenium(iv) and chromium(vi), respectively, and for the amperometric detection of salbutamol and hydroquinone after high-performance liquid chromatography. The cell is simple, easy t o prepare and is compatible with many electroanalytical systems. The cost of the flow cell is considerably less than conventional electrochemical detectors equipped with glassy carbon macro-electrodes.Keywords: Carbon and gold fibre; flow cell; amperometric and voltammetric detection; selenium( ~ v ) and chromium(vi); salbutamol The past decade has seen the widespread and increased use of microelectrodes for electrochemical investigations. This im- petus has been driven by the unique properties exhibited by such electrodes; namely, the reduced iR (current x resistance) losses and capacitative charging currents and the greatly increased rates of mass transfer. Carbon fibre electrodes have been developed for poten- tiometric and amperometric analyses in flowing systems. I-5 Several flow cells have been reported for use with high- performance liquid chromatography (HPLC) using carbon fibres.All are based on the positioning of a fibre in the outlet of a chromatographic column . 6 - 1 1 However, a considerable amount of skill is required to fabricate these cells and the resulting signals are easily affected by any change in the position of the fibre in relation to the outlet of the column. In a previous paper, a flexible and simple carbon fibre flow cell was reported.5 This cell was constructed by fixing a carbon fibre into a polyethylene tube, which served as the working electrode. A silver wire coated with silver phosphate and a carbon fibre electrode were mounted downstream of the working electrode to serve as the reference and counter electrodes, respectively. However, in this system, it was found to be necessary to add some phosphate to the mobile phase because a silver-silver phosphate reference electrode was used in the eluent stream.In addition, any impurities in the eluent might contaminate the reference electrode. This paper describes a modified flow cell, with an independent reference electrode and a stainless-steel counter electrode, which are more simple and reliable to operate compared with the previous design. This cell has been tested for anodic and adsorptive stripping voltammetric analyses in addition to amperometric detection in HPLC. Experimental Preparation of Carbon Fibre Flow Cells Carbon fibre flow electrodes were prepared as described previously.5 A gold fibre (25 ym in diameter) was used instead of the carbon fibre when anodic stripping voltammetric determination was performed for selenium(1v).The reference electrode was prepared by inserting a silver wire, coated with silver chloride or silver phosphate, into a polyethylene tube (15 mm long, 1 mm i.d. and 2 mm 0.d.). A porous ceramic rod (2 mm long, 1 mm in diameter) was fitted to one end of the polyethylene tube, which was then gently heated with an electric iron so that the melted polyethylene tube closed the gap between the tube and the porous ceramic rod. The * To whom correspondence should be addrcsscd. reference electrode tube was then filled with an internal reference solution containing 1 mol dm-3 potassium chloride or phosphoric acid. Finally, this end was also closed by gentle heating. A stainless-steel tube (20 mm long, 1.5 mm 0.d.and 0.5mm i.d.) was used as a counter electrode. The working electrode, reference electrode and counter electrode were mounted in a T-tube as shown in Fig. 1. Coating AgCl or Ag3P04 onto a Silver Wire Silver chloride or Ag3P04 was coated onto a silver wire (20 mm long, 0.1 mm in diameter), by connecting the wire to the anode of a 1.5 V battery and connecting a platinum electrode to the cathode of the battery for 2 min in a solution of 1 mol dm-3 hydrochloric acid or 1 mol dm-3 phosphoric acid as the electrolyte. Instrumentation and Reagents For anodic stripping voltammetric determination of selen- ium(iv), an EG & G Princeton Applied Research Model 264A polarographic analyser was used. For the adsorptive stripping voltammetric determination of chromium(vI), a Metrohm Model 626 polarographic analyser was used.An Ag-AgC1 reference electrode was used for the voltammetric analyses. The HPLC-amperometric detection analysis was performed using a Waters Model 510 pump in association with a reversed-phase octadecylsilane (ODS) bonded column (250 x 5mm i.d.; Supelco). The column outlet was connected directly to the fibre flow cell. Either a Metrohm 641 VA detector or a Hewlett-Packard 1049A programmable electro- chemical detector was used in connection with the fibre cell. An Ag3P04 reference electrode was used in the analyses by Flow -I/* w I \I// I I 1 Fig. 1 Structure of the carbon fibre micro-flow cell: 1, silver wire coated with silver chloride; 2, reference electrode body; 3, internal reference solution; 4, T-tube; 5 , ceramic rod; 6, stainless-steel counter electrode; and 7, fibre flow electrode1118 ANALYST, NOVEMBER 1991, VOL.116 HPLC. Carbon fibres, 14.5pm in diameter, were obtained from the Kureha Chemical Company. The gold fibres, 25 pm in diameter, were purchased from Goodfellow Metals. Salbu- tamol was purchased from Sigma. All of the other reagents were of analytical-reagent grade. De-ionized water was obtained by passing distilled water through a Millipore Milli-Q water purification system. Analytical Procedures Adsorptive stripping voltammetric analysis Adsorptive stripping voltammetric analyses were performed using the carbon fibre flow electrode as the working electrode. The working electrode was plated with mercury by passing an aqueous solution containing 200 mg 1-1 of mercury(I1) nitrate and 0.1 rnol dm-3 hydrochloric acid through the cell at a flow rate of 1 ml min-1 for 60 s with an applied potential of - 1.8 V on the working electrode.The sample solution containing 1 x 10-8 mol dm-3 chromium(vI), 0.008 mol dm-3 diethylene- triaminepentaacetic acid (DTPA), 1.04 mmol dm-3 sodium nitrate and 0.08moldm-3 sodium acetate was adjusted to pH 6.2 and then allowed to pass through the flow cell at a flow rate of 1 ml min-1 for 60 s with an adsorptive potential of -0.80 V on the working electrode. After stopping the pump for 30 s, the differential pulse stripping voltammetric curve for chromium(vI) was obtained by scanning the working electrode potential up to - 1.5 V at a rate of 10 mV s-1, the pulse height was 50 mV and pulse duration 0.5 s.The peak potential of chromium was located at - 1.22 V. Anodic stripping voltammetric determination of selenium(1v) In the determination of selenium(iv), a gold fibre electrode was used as the working electrode. In the deposition process the sample solution, containing 50 pg 1-1 of selenium(1v) and 0.1 rnol dm-3 perchloric acid, was passed through the flow cell at a flow rate of 1 ml min-1 for 20 s with a potential of -0.2 V on the working electrode. The differential pulse stripping analysis was performed by scanning the working electrode potential anodically up to + 1.2 V at a rate of 10 mV s-1, with a pulse height of 25 mV and a pulse duration of 1 s. Amperometric detection in HPLC Salbutamol and hydroquinone were detected at a carbon fibre working electrode in the flow cell, held at +1.5 V.The compounds were eluted with a mobile phase containing methanol and water (70 + 30), 0.02 rnol dm-3 phosphate, 0.017 rnol dm-3 sodium dodecyl sulphate and 0.0027 mol dm-3 diethylamine (adjusted to pH6.0) at a flow rate of 0.9 ml min-1. For the hydrodynamic studies a mobile phase consisting of methanol, water, and orthophosphoric acid (90 + 10 + 0.1) was used. Results and Discussion Adsorptive Stripping Voltammetric Determination of Chromium(v1) Cathodic stripping voltammetry has previously been used for the measurement of total chromium in natural water.12 This analysis is based upon the fact that Cr'" forms a complex with the ligand dtpa [log equilibrium constant (K = 15.3)]. Under appropriate conditions, this complex is adsorbed onto a mercury surface and a cathodic scan reveals a peak at -1.22 V, corresponding to the reduction of CrI'l-dtpa to Cr".Chromium(v1) is reduced electrochemically to Cr"' at poten- tials more negative than -0.05 V. A catalytic effect caused by the oxidation of Cr" to Crllf by nitrate ions gives rise to a signal enhancement. Constant current stripping analysis has been used for trace metals and organic compounds using carbon fibre flow electrodes.2-4 However, application of the voltammetric 0.9 1.1 1.3 1.5 -EN versus Ag-AgCI Fig. 2 Adsorptive stripping voltammetric curve of chromium(v1) obtained by analysing a sample solution containing 1 X 10-8 mol dm-3 chromium(v1) in an electrolyte consisting of 0.008 mol dm-3 DTPA, 1.04 mol dm-3 sodium nitrate and 0.08 mol dm-3 sodium acetate, adjusted to pH 6.2. The background is shown by the broken line stripping mode incorporated with carbon fibre flow electrodes has not been reported.Because the voltammetric mode is widely used in unstirred solutions, efforts have been made to apply this mode of detection with the fibre flow cell. Chromium(v1) has been determined by adsorptive stripping voltammetry using a mercury electrode in a batch system.12 The response of chromium(v1) in the carbon fibre flow cell was therefore tested. A 1 x 10-gmoldm-3 solution of chro- mium(v1) yielded a well-defined adsorptive stripping peak as shown in Fig. 2. In adsorptive stripping voltammetric analysis, it is usually necessary to plate mercury onto the carbon electrode surface before adsorbing the organic or metal complexes.A medium exchange mechanism is usually necessary in order to introduce solutions into the flow cell for mercury plating, sample adsorption, stripping and electrode cleaning. Medium exchange is facilitated in the flow cell without the trouble of introducing air bubbles into the system, or affecting the electrode surface, which can occur when using a batch system approach. Sometimes contact with air will change the behav- iour of the electrode surface and cause irreproducibility of the response. Application of the carbon fibre flow cell is also convenient for computerized automated analysis. When a multiple valve system is used, the whole analytical procedure, including the introduction of different solutions into the flow cell, the application of different potential waveforms to the working electrode for electrode pre-treatment, signal sampling and electrode cleaning can be controlled using computer-driven instrumentation.Anodic Stripping Voltammetric Determination of Selenium(1v) The determination of selenium in various matrices has been carried out using both cathodic13.14 and anodicl5.16 stripping voltammetric methods of analysis. The cathodic , stripping voltammetric determination of selenium(1v) is based upon the accumulation of mercury selenide, HgSe, on the surface of a mercury electrode. Selenium(1v) has also been determined by anodic stripping voltammetric procedures using a gold disc electrode in a batch system. This procedure was adapted for operation in the micro-flow cell system.The anodic stripping voltammogram for selenium(rv), obtained using a gold fibre electrode incorporated into the flow cell, is shown in Fig. 3. These data further confirm the possibility of applying the voltammetric mode in this flow system. Passivation is a common problemANALYST, NOVEMBER 1991, VOL. 116 1119 0.7 0.9 E N versus Ag-AgCI Fig. 3 Anodic stripping voltammetric curve of selenium(iv) obtained by analysing a sample solution containing 50 yg I- * of selenium in an electrolyte of 0.1 rnol dm-3 perchloric acid. The background is shown by the broken line t .Id C 2 3 u 5 nA A 0 5 10 Ti rn e/m i n Fig. 4 HPLC trace of A, 5ng of hydroquinone; and B, 50ng of salbutamol, obtained by injection of the mixture using amperometric detection.The carbon fibre working electrode was held at 1.5 V versus Ag-Ag3P04. The compounds were separated on a CIS column and eluted with a mobile phase containing methanol-water (70 + 30). 0.02 mol dm-3 phosphate, 0.017 rnol dm-3 sodium dodecyl sulphate and 0.0027 rnol dm-3 diethylamine adjusted to pH 6.0, at a flow rate of 0.9 ml min-1 when a gold electrode is used, caused by the formation of gold oxides on the electrode surface. This problem can be solved by applying an anodic potential of +2.0V for 10s to the gold electrode, with a solution consisting of 6 mol dm-3 nitric acid and 2mol dm-3 sulphuric acid passing through the cell, followed by applying +0.9 V at the working electrode for 2 s, with 5 rnol dm-3 hydrochloric acid passing through the cell, as reported for the constant current stripping determination of arsenic17 using gold fibre electrodes.Amperametric Detection in HPLC A clear and well-defined chromatogram for hydroquinone and salbutamol was obtained by injecting a mixture containing 5 ng of hydroquinone and 50 ng of salbutamol into the system 0 5 10 Time/min I I -1.0 0 +1.0 +2.0 E N versus Ag-Ag3P04 Fig. 5 ( a ) HPLC trace and (b) linear sweep voltammogram of hydroquinone and background. A 5 pg amount of hydroquinone was injected onto the HPLC column. The compound was eluted with a mobile phase consisting of methanol-water-orthophosphoric acid (90 + 10 + O.l), at a flow rate of 1.2 ml min-1. The working electrode potential was held at + 1.5 V for amperometric detection. When the hydroquinone peak appeared, the voltammogram was recorded by stopping the pump and scanning the working electrode potential from -1.0 to +2.0 V at a rate of 10 mV min-1 used for HPLC, as shown in Fig.4. As the flow cell is small in size and is easy to handle, it is simple to connect the cell to any liquid chromatographic system. The impurities in the mobile phase do not contaminate the reference electrode because it is isolated from the eluent by the porous ceramic rod. At the working electrode, the hydroquinone is oxidized to its quinone form and the salbutamol also possibly to its quinone form. In addition to obtaining a conventional chromatogram, it is also possible to obtain a linear sweep voltammogram for the component eluted from the column by stopping the flow and scanning the potential of the working electrode.The voltammogram obtained in this way for hydroquinone is shown in Fig. 5. In the future it is anticipated that this design of cell can be used to obtain hydrodynamic voltammograms without the requirement to stop the flow. If so, the capacitance current caused by scanning would be much smaller for a micro- electrode compared with an electrode with a larger surface area. 18 Conclusions The microelectrode flow cell described in this paper is suited to voltammetric and amperometric analyses. As the cell is1120 simple in structure, small in size and easy to construct, it can easily be used in many electrochemical systems. If a carbon fibre flow cell stops functioning owing to contamination or for any other reason, it is easy to exchange it for a new fibre without great expense.References Bixler, J . W., Bond, A. M., Lay, P. A., Thormann, W., van den Bosch, P. Fleischmann, M., and Pons, B. S., Anal. Chim. Acta, 1986, 187,67. Hua, C., Jagner, D., and Renman, L., Talanta, 1988, 35, 597. Hua, C., Jagner, D., and Renman, L., Anal. Chim. Acta, 1987, 197, 265. Hua, C., Jagner, D., and Renman, L., Talanta, 1988, 35, 525. Hua, C., Yunping, W., Chunglian, J., and Tonghui, Z., Anal. chim. Acta, 1990, 235, 273. ?tulik, K . , Analyst, 1989, 114, 1519. Stulik, K., Pacakova, V., and Podolak, M., J . Chromatogr., 1984, 298, 225. Kounaves, S. P., and Young, J. B., Anal. Chem., 1989, 61, 1469. 9 10 11 12 13 14 15 16 17 18 ANALYST, NOVEMBER 1991, VOL. 116 White, J. G., St. Clair, R. L., and Jorgenson, J. W., Anal. Chem., 1986,58, 293. Knecht, L. A., Guthrie, E. J., and Jorgenson, J. W., Anal. Chem., 1984,56, 479. Goto, M., and Shimada, K., Chromatographia, 1986, 21, 631. Golimowski, J., Valenta, P., and Nurnberg, H. W., Fresenius 2. Anal. Chem., 1985,323, 315. Batley, G. E., Anal. Chim. Acta, 1986, 187, 109. Adeloju, S. B., Bond, A. B., Briggs, M. H., and Hughes, H. C., Anal. Chem., 1983,552076. Posey, R. S., and Andrews, R. W., Anal. Chim. Acta, 1981, 124. 107. Andrews, R. W., and Johnson, D. C., Anal. Chem., 1975,47, 294. Hua, C., Jagner, D., and Renman, L., Anal. Chim. Acta, 1987, 201, 263. Lunte, C. E., Wheeler, J. F., and Heineman, W. R., Anal. Chim. Acta, 1987, 200, 101. Paper 1 l02350K Received May 20th, 1991 Accepted July 23rd, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911601117

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1991, VOL. 116 1121 Chemical Sensing of Amine Antioxidants in Turbine Lubricants Richard J. Price and Lionel J. Clarke Shell Research Ltd., Thornton Research Centre, P.O. Box 1, Chester CHI 3SH, UK A chemically modified microelectrode has been developed for the determination of amine antioxidants in turbine lubricants. The electrochemical probe is coated with a thin film of conducting poly(ethy1ene oxide), which, when placed into a lubricant environment, results in the extraction of the electroactive species into the polymer. Cyclic voltammetric experiments have been performed and an ability to measure the phenyl-a- napthylamine additive quantitatively has been demonstrated. Keywords: Chemical sensor; cyclic voltammetry; amine antioxidant; poly(ethy1ene oxide); lubricant The analysis of lubricating oils during or after use is an established approach for determining the extent to which they might have deteriorated or been contaminated during opera- tion in an engine.This is commonly carried out by removing samples for analysis. There are many tests that are routinely used, a combination of which is generally suitable for determining the condition of the lubricant, or diagnosing problems with the engine. Analytical techniques are typically employed to determine changes in physical properties such as viscosity, acidic content, colour, odour, water content and flash point.' However, from these measurements, it is not possible to determine the remaining useful lifetime of a lubricant.2 It has been demonstrated that lubricants rapidly degrade as the concentrations of antioxidant additives dec- rease,3 indicating the importance of being able to monitor additive concentration in oils as operating time progresses.Indeed, the concentration of antioxidant is used as an important indicator of the remaining useful lifetime, and a number of methods435 are presently used in its determination, including electrochemical techniques.6.7 However, with these methods the measurement cannot be made in situ without sample pre-treatment. In this study, the possibility of intro- ducing a single technique that could be used in situ, or at least could be applied easily without additional sample preparation, to used oil samples has been explored. Such a method could be very attractive in providing a rapid on-the-spot analysis of the lubricant, which has previously not been possible. The proposed approach uses an electrochemical probe for direct chemical sensing of amine antioxidants in turbine lubricants.The resistance of a typical turbine lubricant is so high that standard electrochemical experiments, which require a conductive medium, are impossible. Therefore, a micro-electrode has been chemically modified8 so that the electrochemistry of the amine antioxidant can be monitored. Experimental Probe Construction A schematic representation of the electrochemical probe that was developed for this work is shown in Fig. 1. The three- electrode arrangement was constructed by feeding three metal wires through the end of a soda-glass tube before heat-sealing.The extreme tips of the wires were then exposed by careful polishing, using emery paper followed by a succession of 1,0.3 and 0.04pm alumina polishes. The wires were positioned inside glass sleeves to ensure they were electrically insulated from each other and then connected to metal terminal caps. The Ag wire acted as a pseudo-reference, the Pt wire as a counter electrode and the Pt microdisc as a working electrode. The arrangement was then coated with a polymer film cast from a solution of poly(ethy1ene oxide) [PEO; Aldrich, average relative molecular mass ( M r ) , 600 0001 and LiCI04 in acetonitrile-methanol (9 + 1, by volume). These chemisorbed films are typically about 10 ym thick and have the composition PEOl6LiCIO4. Apparatus An EG & G Model 173 potentiostat and Model 175 programmer were used for the electrochemical experiments and current-potential data were recorded on a Gould 1604/ 104 storage oscilloscope.The sample temperature was kept constant to within 0.5 "C using a thermostatically controlled water-bath. Method Typically, 25 yl of the polymer-salt solution was dispensed onto the end of the polished electrode, and evaporation and curing of the film were allowed to take place for 10min at 50°C. The electrochemical probe was then placed in the oil sample (1-2 ml) at an elevated temperature (about 60 "C). After a set time (lOmin), a cyclic voltammogram (CV) was recorded at a sweep rate of 200 mV s-1. Results and Discussion Unused base oil with no additive species present shows a featureless CV between the cathodic and anodic limiting currents [see Fig.2(a)]. These limiting currents register the onset of the decomposition of the polymer matrix. Com- parison of this with the CV of the fully formulated oil [Fig. 2(b)] reveals three distinct features attributable to the presence of additives in the formulation. Samples of the individual additives dissolved in the base oil, giving concentra- tions equivalent to those in the formulated oil, were prepared. From the CVs, it was apparent that the peaks marked A and B in Fig. 2(b) can be assigned to the amine antioxidant, Soda g Electrically insulated with glass sleeves f il ilass- PEO- r v 4- -w c- 4- / Pt wire (0.5 mm) R wire - (50 pm) Ag wire (0.9 mm) i e w Lubricant Fig. 1 ical probe with an ionically conducting PEO film Schematic representation of the three-electrode electrochem-1122 ANALYST, NOVEMBER 1991, VOL.116 I Potential - Fig. 2 Cyclic voltammograms of (a) a base oil with no additives and (b) a fully formulated turbine lubricant. Peaks A and B can be assigned to PAN and peak C to a rust inhibitor also present in the formulated oil 2.5 I I 1 I I 2.7 2.8 2.9 3 3.1 3.2 3.3 -lK-’ 103 T Fig. 3 Plot of voltammetric peak current [peak A (see text for details)] as a function of 1/T for PAN in a turbine lubricant phenyl-a-naphthylamine (PAN), and peak C to a rust inhibitor also present in the formulated oil. The PEO film is polar in nature and has an affinity towards polar molecules in the oil, such as the amine antioxidant. Although high M, polymers, such as PEO, exhibit mechanical properties that are similar to those of a true solid, at the atomic level, local relaxation processes still provide liquid-like degrees of freedom.9 These give rise to ionic conductivity, with ion transport considered to take place by a combination of motion linked to both movement of the polymer segments and ion transition between suitable coordination sites.At elevated temperatures, the ionic conductivity is such that the electrochemistry of the preconcentrated additive species can be successfully observed at the working microelectrode. For this type of electrochemical experiment PEO is ideal because it is relatively resistant to oxidizing and reducing conditions, and gives a reasonably large electrochemical window in which to observe additive species.A sample of the formulated oil was taken, and the peak current for the oxidation of the PAN additive [Fig. 2(b), peak nA 1 h V A Potential - Fig. 4 Cyclic voltammograms of (a) PAN in base oil (polymer- coated electrode) and (b) PAN in acetonitrile-0.1 mol dm-3 Bu4NBF4 (Pt-wire electrode). For details of peaks labelled A, A‘ and B see text A] recorded as a function of the increasing temperature. The relationship between the two is represented in an Arrhenius- type plot (Fig. 3). It can be seen that the peak current increases in a similar manner to the way that ionic conductivity has been observed to increase with temperature.10 This reflects the mobility of the additive species in the PEO matrix and can be explained in terms of the ‘free-volume model’.11 As the temperature is increased, the expansivity of the polymer produces local empty spaces (free volume) into which ionic carriers or polymer segments can move.The curve in Fig.3 shows an abrupt change in slope at about 56 “C, which can be traced to a change in phase of the polymer. At room temperature, PEO has a variety of morphologies and consists of both crystalline and amorphous phases. This multi-phase behaviour strongly influences the transport of species. As the temperature is raised above the glass transition temperature ( Tg) , a rapid increase in the observed current is seen and the PEO becomes macroscopically rubbery, rather than glassy. This means that the local environment around any given polymer chain becomes liquid-like. Above the Tg, the increase in current with a rise in temperature is much less rapid.Therefore, in terms of the operating conditions for the probe, it would seem advantageous to work at temperatures above the Tg, where large currents are observed and the morphology of the PEO is not variable, thus improving the reproducibility of the data obtained. The electrochemistry of PAN in acetonitrile was also studied at a Pt wire electrode, using Bu4NBF4 as a supporting electrolyte and Ag-AgN03 (0.1 mol dm-3) as the reference. The CV obtained at a sweep rate of 200 mV s-1 is given in Fig. 4,, along with the CV of PAN in base oil recorded using the polymer-coated electrode. The peak marked A’ in the CV from the acetonitrile solvent shows that the oxidation peak, A, is part of an electrochemical couple.Additional CVs demon- strated that the peak current of A was proportional to the square root of the sweep rate (for v 4 0 0 V s-I), and that the peak potential was independent of sweep rate. It was noted that peak B was only observed if the initial sweep direction of the CV was in the anodic direction. This would indicate the existence of the following reaction scheme: R + ne- e 0 (peaks A and A’) (1) k o + x X -+ Y + ne- (peakB) (3)ANALYST, NOVEMBER 1991, VOL. 116 1123 Fig. 5 Reaction of PAN at a Pt electrode to form a cation radical 60 50 40 $ 30 a 20 10 A I I I I I I 0 0.05 0.1 0.15 0.2 0.25 PAN ("/o m/m) Fig. 6 Relation between voltammetric peak current [peak A (see text for details)] and concentration of PAN The oxidized species (0) is partially consumed in a chemical reaction that produces another species (X), the reduction of which gives rise to peak B.The values of the reaction rate (k) and sweep rate ( v ) are such that the A-A' couple is not completely reversible. The actual oxidation involved is likely to be as shown in Fig. 5.12 Although some stability is conferred to the cation radical by the phenyl and naphthyl systems, this species is sufficiently unstable to be consumed in the subse- quent chemical reaction. Peak B, compared with peak A, is much larger under the base oil-polymer conditions than it is under the acetonitrile-Pt wire electrode conditions. In addi- tion, peak A' does not feature in the solid electrolyte, indicating that the diffusion coefficients and reaction rate constants in the polymer matrix and solvent are different.A study was carried out to establish if the magnitude of the oxidation current of peak A could be related to the concentra- tion of PAN in a lubricant package. All of the CVs were recorded after a set time in order to allow the polymer- lubricant interface to equilibrate (eqn. 4) (4) The magnitude of the current is dependent on the ionic conductivity and the thickness of the polymer film, in addition to the diffusion coefficients of the electroactive species dissolved in the polymer. 13 Constant temperature was there- fore required to achieve reproducible data. A series of solutions with varying concentrations of PAN in base oil were prepared. The electrode was polished and a fresh polymer- film coated onto the surface for the analysis of each sample.The response of each new polymer-coated electrode, in a sample of the fully formualted oil, was used as a reference for the signal obtained for the antioxidant. The current observed for peak A is plotted against YO m/m PAN in Fig. 6. A reasonably linear response was obtained over the range 04.15% m/m PAN, a span that encompasses the concentra- tion of the antioxidant in the fully formulated lubricant. At higher concentrations (0.2% m/m PAN), non-linearity ocurred probably because of an uncompensated iR (current x resistance) drop or possibly even from saturation of the polymer film. The electrochemical probe has been used to monitor the depletion of the PAN additive in an oil during a turbine oil stability test14 by taking samples from the test rig at regular intervals.The error associated with this electrochemical analysis was of the order of 5%, which permitted the trend of additive consumption to be satisfactorily followed over a period of time. Towards the end of the lifetime of a turbine oil, neutral and acidic degradation products are known to build up. Samples of used oil, which were known to have these products present, were taken and their CVs recorded using the polymer-coated electrode. It was shown that these degradation products do not produce interferences at the potential for which the electroactivity of the antioxidant is observed. Conclusions An approach for establishing an ionically conducting medium into which additive species can be extracted from a poorly conducting oil environment, for quantitative analysis, has been demonstrated for the amine antioxidant PAN in a turbine lubricant.The experiment can be simplified from one of cyclic voltammetry to a simple potential step at which the additive is known to oxidize, and by monitoring the steady- state current after a set time. The electronics required for such an arrangment could conceivably be miniaturized into a simple hand-held device with some additional facility to enable constant temperature to be maintained during the analysis of oil samples. Of greater attraction is the ability to place a sensor directly into a working environment to give continuous on-line monitoring. Initial work on PEO indicates that it is physically robust and is capable of retaining reasonable levels of ionic conductivity over periods of several days.The ability to perform this type of sensing is attractive, although technically demanding, with the reliability and durability of the probe becoming crucial factors, along with the ability to calibrate and reference the recorded signals. The ability to monitor the oil condition in situ is desirable in terms of providing more accurate information on drain-time require- ments and to provide a forewarning of oil degradation brought about by unusual operating conditions. The conducting polymer-coated electrode would appear to hold some promise in achieving these goals, with respect to monitoring amine antioxidants in turbine lubricants. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 References Snook, W. A., Lubrication, 1963, 54, 97. Kauffman, R. E., and Rhine, W. E., Lubr. Eng., 1988,44,154. Ravner, H., and Wohltjen, H., Lubr. Eng., 1986,39, 701. Sniegowski, P. J., J. Chromatogr. Sci., 1977, 15, 328. Mowery, R. L., Helicopter Lubricant Sensor, Naval Research Laboratory (NRL), Washington, DC, Project No. 61-2236-0-5, June 2, 1986. Cheek, G. T., and Murray, R. L., Anal. Chem., 1989,61,1467. Kauffman, R. E., Lubr. Eng., 1988,45, 709. Murray, R. W., Ewing, A. G., and Durst, R. A., Anal. Chem., 1987, 59, 379A. Vincent, C. A., Chem. Br., 1989, April, 391. Reed, R. A., Geng, L., and Murray, R. W., J. Electroanal. Chem., 1986,208, 185. Ratner, M. A.. and Shriver. D. F., Chem. Rev., 1988, 88, 109. Nelson, R. F., in Anodic Oxidation Pathways of Aliphatic and Aromatic Nitrogen Functions, ed. Weinberg, N. L., Wiley, New York, 1974, pp. 591-780. Geng, L., Reed, R. A., Longmire, M., and Murray, R. W., J. Phys. Chem., 1987,91, 2908. ASTM Standard Method D943, 1987 Annual Book of ASTM Standards: Section 5. American Society for Testing and Materials, Philadelphia, PA, 1987. Paper 11024771 Received May 28th, 1991 Accepted July 4th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911601121

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1991, VOL. 116 1125 Examination of Ammonia-Poly( pyrrole) Interactions by Piezoelectric and Conductivity Measurements Jonathan M. Slater and Esther J. Watt Analytical Science Group, Birkbeck College, University of London, 20 Gordon Street, London WCI H OAJ, UK The conducting polymer poly(pyrrole), electrochemically prepared and doped with anions, has been found to be a responsive coating for a piezoelectric gas detection system. Polymers doped with bromide, nitrate and sulphate ions were tested. It was found that samples of ammonia gas cause a measurable frequency decrease, interpreted as adsorption by the polymer coating of the quartz crystal; the linear range was 0.051% for mixtures of the gas in nitrogen. These signals were found to correspond to simultaneous conductivity changes of a similarly prepared poly(pyrro1e) sample, showing analogies in the two sensing mechanisms.The duality of the poly(pyrro1e) response increases the possibilities of using it as a gas sensor. Keywords: Ammonia detector; piezoelectric gas detection system; gas sensor; poly(pyrrole) gas sensor Poly(pyrro1e) is a conducting polymer with conductivity ranging from 1 to 100 S cm-1 (Fig. 1).1 It may be conveniently prepared by the electrochemical oxidation of the pyrrole monomer in an electrolytic solution, onto gold or platinum electrodes .24 The reaction is initiated by the electrochemical generation of monomer radicals which combine with other units in solution to form the polymer chains, along which electrons may conduct. Excess oxidation of the polymer generates a net positive charge resulting in the uptake of counter ions from the electrolyte solution.These dopant anions render the polymer a p-type semiconductor, giving it a second mode of conduction."." However, its true structure, which is dependent on conditions such as pH, potential, solvent and dopant anion, has still not been fully charac- terized. Despite this apparent lack of characterization poly- (pyrrole) is a potentially useful material for the fabrication of sensors. It has been used as a sensing material in field effect transistor7 and ion-selective electrode8 devices and as a conducting matrix for enzyme entrapment electrodes.9 It shows interesting gas sensing possibilities which were first demonstrated by Nylander et al.10 in 1983. The polymer was prepared by chemical oxidation, the precipitated 'pyrrole black' being impregnated into filter-paper which was then shown to give a 30% change in conductivity on exposure to 1% More recently Miasik et af. I have reported a device utilizing electrochemically prepared polymer promising convenient and controllable preparation which should lead to a more stable and potentially reproducible sensor material. The mechanism of interactions was attributed to the p-type semiconducting nature of poly(pyrro1e). Exposure to elec- trophilic gases, such as NO,, tends to attract electrons out of the polymer matrix, causing an increase in conductivity, whereas nucleophilic gases, such as NH3, will have the opposite effect. This clear evidence of polymer-gas interac- tions renders poly(pyrro1e) a suitable coating for piezoelectric sensor crystals which should be a useful means of investigating such interactions.Furthermore piezoelectric crystals already NH3. H Fig. 1 NO3-, SO4'- or Br- Structure of doped poly(pyrrole), X- = doping anion, e.g., contain electrodes suitable for the electrochemical deposition of poly(pyrro1e) layers. The piezoelectric gas detection systems described to date work on the principle that a gas adsorbed onto a crystal coating changes the crystal mass resulting in a shift in its fundamental frequency. 11-12 The Sauerbrey relationship is commonly used to relate the observed frequency change to the adsorbed mass: 13 AF = -2.3 x 106 Fo2 AmlA where A m is the change in mass of the crystal (g), AF the related frequency change (Hz), A the gas-sensitive area (cm2) and Fo the initial frequency of the quartz crystal (MHz).It is well-known that moisture is a major interferent in piezoelectric measurement systems and other workers have reported methods of overcoming this problem.14 Poly- (pyrrole) readily adsorbs moisture; in fact poly(pyrro1e) tosylate is reported to be hygroscopic.15 However, this work was carried out using cylinder gases which are assumed to be of constant moisture. A possible application of poly(pyrro1e) gas sensors is in arrays and in this instance the moisture response would be treated as a mixed sensor response. Experimental Apparatus The piezoelectric crystals used were 5 MHz, AT-cut quartz crystals with gold electrodes (Webster Electronics, Ilminster, Somerset, UK).The measurement electronics were con- structed by British Gas. The sensor and reference crystals were built into matched oscillator circuits, driven by a 5 V d.c. supply, and a signal which related to the difference in frequency between the two crystals was extracted. This output could be displayed on a y-t recorder (Siemens Kompenso- graph X-T C1011) or transferred to an IBM PC via a suitable interface (Blue Chip Technology AIP-24, 12 bit analogue input card). Data logging software was written to provide a permanent record of results. A further output allowed the frequency of the sensor crystal to be monitored directly (Thandar TF-600 frequency counter). Measurements were made in two flow systems. The first, illustrated in Fig.2, is similar to those previously described.16 A continuous flow of air is pumped through activated charcoal and silica gel drying agent. The scrubbed, dry air is passed through an injection cell where the sample may be introduced. The gas system was split into two, one stream passed through a reference cell, containing an uncoated crystal, the other passed through an identical cell containing the poly(pyrro1e)1126 r Chart Double oscillator recorder Digital computer meter frequency - crystal or ANALYST, NOVEMBER 1991, VOL. 116 5 V d.c. - supply coated crystal. These cells were based on the double impinger cell design of Karmarkar and Guilbault17 wherein the gas is split into two streams impinging directly onto the two faces of the crystal. Exhaust gases were monitored by flow meters, a constant flow optimized at 40 ml min-1 being maintained.The second flow system utilized a British Gas multi-gas sensor test rig which allowed both crystals to be exposed to alternate 5min pulses of carrier gas (air mixture; Air Products) and sample gas. Similar flow conditions were used with the same double impinger flow cell. The test rig also has a facility for determining the moisture content (relative humidity) of test and carrier gases. Crystal Coating Crystal coatings were prepared by the electrochemical oxida- tion of the pyrrole monomer using a Princeton Applied Research 174A polarographic analyser in a three-cell poten- tiostatic assembly. The three-electrode cell consisted of a saturated calomel reference electrode, a glassy carbon auxiliary electrode and the gold electrodes of the piezoelectric crystal as the working electrode.The electrolysis solution was a combination of pyrrole (freshly distilled, Aldrich) and supporting electrolyte, either KBr, K2S04 (both Fisons), KCI or KN03 [both Merck (formerly BDH)]. Different polymer coatings were grown onto the gold electrodes of the crystals by sweeping aqueous solutions of the pyrrole and electrolyte between 0.0 and 0.9V. Table 1 shows the preparation conditions of different types and thicknesses of the poly- (pyrrole) coatings prepared. Poly(pyrro1e) coated crystal Uncoated crystal Perspex Perspex sample crystal reference crystal cell cell (through rubber septum) Fig. 2 Piezoelectric gas detector flow apparatus Conductivity Measurements A Degussa Dew Point Sensor E was used as the basis of a conductivity sensor.The device (4 x 6mm) consists of alumina, coated with a platinum film, into which a number of meanders have been cut to form finely spaced multi-interdigi- tated electrodes. The shape and size of the meanders are shown in Fig. 3, the total number being 24, each of dimensions 100 x 700 pm with an electrode gap of 10 pm. Poly(pyrro1e) may be deposited on the platinum surface by anodic oxidation of pyrrole producing a coating sufficiently thick to bridge the gap between the device electrodes (sensor I, Table 1). The conductivity of the polymer can therefore be measured at these points. A simple Wheatstone bridge was constructed so that changes in the polymer conductivity caused by different gases could be measured on a Thurlby 1503-HA digital multimeter and recorded on a chart recorder.Gas Samples The test gas samples were obtained from cylinders of 1% sample gas, in a nitrogen balance (Air Products). Gas samples were taken with a 10 ml glass syringe. Dilution of the gas was carried out by syringe dilution, a procedure previously described by Karmakar and Guilbault .17 Results and Discussion Poly(pyrro1e) Preparation A typical cyclic voltammogram obtained during poly(pyrro1e) preparation is shown in Fig. 4. It has similar characteristics to those previously reported for the electrochemical oxidation of pyrrole while cycling between 0.0 and 0.9 V.18 The polymer 700 pm t I I100 prn 7 Fig. 3 Geometry of the conductivity sensor (total number of meanders is 24) Table 1 Preparation conditions of poly(pyrro1e) coated onto piezoelectric crystals and the conductivity sensor Change in crystal parameters due to poly(pyrro1e) coating Sample A B C D E F G H Ill Electrolyte solution" K2S04 (0.5 rnol dm-3) K2S04 (0.5 mol dm-3) K2SO4 (0.5 mol dm-3) KBr (1 .O mol dm-3) KBr (1 .O mol dm-3) KBr (1.0 mol dm-3) K2SO4 (0.5 mol dm-3) KN03 (1 .O mol dm-3) KBr (1 .O mol dm-3) Preparation timetls 54 54 54 180 180 36 36 99 1800 Drying conditions 110 "C, 10 min 110 "C, 10 min llO"C, 10 min Ambient Ramp§ 110 "C, 10 min 110 "C, 10 min 110 "C, 10 min 110 "C, 10 min FrequencyIHz 4 322 3 967 2 754 29 861 22 347 4 396 17 782 11 742 MassSIyg 85.3 78.3 54.2 589.0 441 .o 86.7 350.0 232.0 * Solutions contain electrolyte and freshly distilled pyrrole (0.05 mol dm-3, aqueous).t Prepared by cyclic voltammetry, sweeping between 0.0 and 0.9 V, scan rate = 100 mV s-*. $ Mass calculated from the Sauerbrey equation. 5 Ramp-dried sample was heated in the oven from room temperature to 110 "C over a period of 10 min. '1[ Conductivity sensor.ANALYST, NOVEMBER 1991, VOL. 116 5 , I 1127 B CI 2 3 a 0 e a o 0.3 0.6 EN versusSCE 0.9 Fig. 4 Preparation of sensor G by the cyclic voltammetry of pyrrole (0.05 mol dm-3) in aqueous K2S04 (0.5 mol dm-3), by scanning between 0.0 and 0.9 V, scan rate = 100 mV s-*. A, First cycle; B, second cycle; and C, third and subsequent cycles 0 0 50 100 1 50 Reaction time/s Fig. 5 Preparation of poly(pyrro1e) onto piezoelectric crystals, mass of polymer coating versus reaction time, scan rate = 100mVs-1.Doping anion: 0, bromide; 'I, sulphate; and M, nitrate preparation is a three-dimensional nucleation and growth reaction; thus, it would be expected that growth is favoured on the polymer surface rather than on the bare electrode. This type of polymerization is characterized by voltammograms of increasing current which eventually reaches a steady-state limit.3 Fig. 4 exhibits the first stages of this polymerization process; further cycling, in other experiments, has shown small increases in current eventually reaching a maximum value. The growth profiles were similar in shape for the different electrolytes used. In all instances the cycling process was terminated at the more positive potential in order to ensure that the polymer was in the oxidized form.The mass of deposited poly(pyrro1e) increases with the reaction time (see Table l), Fig. 5. Samples A, B and C (sulphate-doped) were prepared under similar conditions (54 s polymerization time, equivalent to three cycles between 0.0 and 0.9V followed by a final, half-scan ending at 0.9V, 100 mV s-1 followed by drying for 10 min at 110 "C), and therefore have a comparable mass of polymer deposited, 85.3 x 10-6, 78.3 x 10-6 and 54.2 x 10-6g. Duplicate bromide- doped samples, D and E, were also prepared (180 s poly- merization time, equivalent to ten cycles between 0.0 and 0.9 V, 100 mV s-I), but while sample D was left to dry in the atmosphere sample E was ramp-dried in an oven from ambient temperature to 110 "C. The two samples have different total masses, the variation in which is attributed to the drying conditions.Thus sample E (oven dried) weighed 441 x 10-6g and sample D (air dried) weighed 589 X 10-6g, the additional mass probably being due to excess moisture trapped in the polymer structure. Further dryinglheating did not significantly change the polymer mass. Later it will be shown that this trapped moisture has a significant effect on the response of the piezoelectric system to gases. Flow Injection Experiments The sensors prepared by coating with doped poly(pyrro1e) (Table 1) were used as the gas-sensitive elements in the flow apparatus, Fig.2. A typical response to repeated 10ml 1% NH3 I 1 0.5% NH3 I 30 min Time - Fig. 6 Typical recorder trace showing the correlation of res onse of A, piezoelectric and B conductivity sensors coated with poly&yrrole) (Table 1, sensor D) to 10ml injections of NH3 into an air carrier stream 0 N -1oL \ g) -20 5 -30 - E -40 - $ -60 - k -70- c 0 $ -50- i -lo - g - 2 0 - -80 1 I 1 I I I I 0 100 200 300 400 500 600 Mass of polymer coating/vg Fig.7 Flow injection apparatus responses to 10 ml injections of 1% NH3, frequency change versus mass of polymer coating. Doping anion: a, bromide; 'I, sulphate; and ., nitrate injections of 1% ammonia in nitrogen can be seen in Fig. 6. The observed decrease in frequency is instantaneous (within the limitations of equipment response), as the gas interacts, but subsequent recovery takes approximately 10 min. The average frequency change observed for each of the crystals in responses to these injections versus the total mass of the crystal coating was plotted, Fig.7. The response increases with the mass of crystal coating; for the sulphate-doped poly(pyrro1es) (all dried at 100 "C for 10 min), this increase is almost linear, but for bromide-doped polymers, which were prepared under varying conditions, the response appears to be related to the drying conditions. Assuming that the two coatings contain the same amount of polymer and the mass difference is solely due to absorbed moisture, crystal D gives a response enhanced by the trapped water. This would be expected if the response was due to ammonia gas dissolving in the water; however, this is unlikely to result in the full recovery of the sensor. If mass was continually being added to the crystal, the baseline would drift dramatically in the direction of decreasing frequency.Additionally, the response would be expected to diminish as the surface of the poly- (pyrrole) became saturated with base, thus becoming unfav- ourable for interactions with ammonia. Moisture does, however, enhance the gas adsorption. The relative responses of the other, differently doped poly(pyrro1es) were also plotted (Fig. 7). Although individual responses vary the results are of the same magnitude. The nitrate-doped polymer gives the smallest response while the sulphate-doped polymers give the largest. Multi-gas Sensor Test Rig Experiments Typical responses to alternate pulses of air and 1% ammonia gas can be seen in Fig. 8. The response of sensor D (bromide-doped and dried at room temperature), Fig.8(a),1128 ANALYST, NOVEMBER 1991, VOL. 116 400 300 200 100 0 -100 - 200 - 300 -400 0 i p -loo c 2’ -200 S Q) CT -300 -400 0 10 20 30 1 1 I I 0 5 10 15 20 25 30 0 10 20 30 Time/min Fig. 8 Typical responses from gas test rig of ( a ) sensor D, ( b ) sensor E and (c) sensor A to alternate 5 rnin pulses of 1% NH3 and air shows an initial frequency decrease of over 300Hz on exposure to 1% NH3 for 5 min; no plateau is reached. The frequency then recovers to the baseline on subsequent exposure of the crystal to a stream of air. The following cycle of exposure to the sample gas causes a frequency decrease of less than 200Hz followed by a recovery of 400Hz. This pattern is repeated. Thus, if the frequency change is due to change in the mass of the sensor, after initial exposure to ammonia twice the amount of material is lost from the polymer surface as is originally adsorbed.This could simply be moisture lost from the polymer in the dry stream of air, or moisture depletion due to reaction with ammonia. However, this would be expected to result in a more rapid decrease in response as the experiment progresses due to a loss of moisture available for the ammonia reaction. Sensor E (bromide-doped and oven dried), Fig. 8(b), shows an initial frequency drop of about 250Hz but a reduced recovery, approximately 75 Hz. Subsequent exposure to ammonia causes a similar effect, in contrast to crystal D where the response diminished considerably. Also sensor E has a drifting baseline indicating increasing mass.This indicates the gas is not desorbed as efficiently as it is adsorbed; possibly Table 2 Response of three nominally identical sensors to 5 rnin pulses of ammonia gas Response? Mass of Standard polymer Frequency Weighted Mean deviation coating*/pg change/Hz response response (Yo) 55.9 99 1 985 886 882 825 56.5 838 1039 1069 1032 952 69.3 953 1068 1084 1239 1224 8 0.0142 0.0131 7 0.0141 0.0127 0.0126 0.0119 0.0180 0.01538 0.0147 0.0151 0.0156 0.0135 0.01 10 0.01284 10 0.0123 0.0125 0.0143 0.0141 * Mass calculated from the Sauerbrey equation. ? The frequency change caused by 5 rnin exposure to 1% ammonia. The weighted response is obtained by dividing the frequency change caused by the ammonia by that caused by the poly(pyrro1e) loading on the crystal. remaining water in the dried polymer is more strongly bound or deeply imbedded than in the ‘wet’ polymer.The response of sensor A (sulphate-doped and oven dried) to an analogous experiment was observed, Fig. 8(c). The results are similar to those for sensor E: a large initial decrease in frequency (100 Hz) followed by a recovery of approxi- mately 40Hz. Subsequently the size of the response is matched by the recovery. The sulphate response is marked by the virtual plateauing of the signal after 5min of exposure. This implies that all reaction sites have been filled. Recovery is very rapid suggesting that a response due to reaction with water vapour is not occurring. This pattern of response is most promising for an effective gas sensor. As shown in Fig. 8, the response to the first and subsequent exposures to ammonia was found to be different. The initial response was enhanced with the sulphate-doped material being most sensitive and the nitrate-doped material least sensitive.However, for subsequent measurements the responses were reduced and the order of sensitivity changed. Crystal D (air dried) gave a lower response than crystal E (oven dried). It appears that the wetter crystal has been desensitized by exposure to the ammonia, suggesting that there is a difference in the interaction between ammonia and either the water on the wet polymer or more strongly or deeply bound moisture on the dried polymer. It would appear that increasing the thickness of the poly(pyrro1e) coating will enhance the sensitivity of the sensor.There is however a limit to the mass of coating a crystal will allow, as large mass loadings will stop its steady oscillation.19 Additionally there is speculation as to the validity of the Sauerbrey equation at high crystal coatings. Table 2 shows the reproducibility of three nominally identical sensors to 5 min pulses of ammonia gas. Conductivity Measurements There is a strong correlation between the responses of the piezoelectric and conductivity sensors, Fig. 6. The bromide- doped poly(pyrro1e) coated conductivity sensor (Table 1, sensor I) was positioned downstream of the piezoelectric sensor, being impacted by the carrier gas and injected sample gas as they left the piezoelectric sensor cell. The initial resistance of the poly(pyrro1e) was 5800 kQ. A 10 ml injectionANALYST, NOVEMBER 1991, VOL.116 0 1 1 I 1 1 I I 1129 of 1% NH3 caused a resistance increase of 890 kQ, agreeing with results observed by Nylander et al. 10 They explained the change as a movement of electrons in the poly(pyrro1e) structure. Our piezoelectric experiments indicate that a response is caused by the reaction between ammonia gas and water; it is also expected that this would increase the poly(pyrro1e) conductivity. The implication is that two mechanisms may cause conductivity changes; however, the similarity between the results obtained for piezoelectric and conductivity measurements (Fig. 6) suggests that the sensing mechanisms are linked to those reported in the literature, with electrophilic gases giving increases in conductivity while nucleophilic gases give decreases in conductivity.Measurements with the piezoelectric sensors showed that of the gases investigated (NH3, CO, NO,, CH4 and C02) in the range 0.05-1%, only NH3 gave an initial sharp frequency decrease and could easily be discriminated. All other gases were characterized by an initial frequency increase. The shapes of these responses suggest a biphasic response mechan- ism and are reported in detail elsewhere.19Jo Calibration Graphs The Sauerbrey equation relates the mass adsorbed on the crystal surface to the frequency change. Beitnes and Schroder20 have questioned the validity of the Sauerbrey equation, at trace concentrations of gas, by showing that a plot of log frequency change versus log concentration should be a straight line with a slope of unity.Although our measurements are at higher concentrations a similar method of looking at the validity of the poly(pyrro1e) coated piezoelectric crystal response to ammonia was employed. A log-log plot of the NH3 response from 0.05 to 1% gas has a slope of 0.9 (Fig. 9), within the acceptable limits for such measurements. The exact mcchanism of interactions between the ammonia gas and the poly(pyrro1e) coated crystal is not clear although the response profiles, shown in Fig. 8, suggest a sorption related rather than diffusion related reaction. These profiles are in agreement with those of Bartlett and Ling-Chung21 who have used partial least squares curve fitting to indicate the profiles fit a sorption model better than either of two diffusion models considered.However, it is important to note that other schemes can yield a similar functional form and further work is required to elucidate the mechanism of response. The system responds linearly to gas concentration in the range 0.05-1% (Fig. 9). The conductivity mode is less sensitive to the gas concentration but can discriminate between different gases. Current work into evaluating a combined sensor may over- come some of the limitations experienced with separate devices. 2.5 r 1 Selectivity Poly(pyrro1e) is known to change in conductivity on exposure to a range of different gases, the exact response depending critically on the method of polymer preparation, particularly on the dopant ion.1.10 Other workers have reported the conductivity change of differently doped polymers to vapours and gases such as methanol and ethano1,21922 which cause a conductivity decrease, and N02,23 PC13 and S02,24 which give increases in conductivity.Our investigations have included NO2 and methanol for which conductivity increases and decreases, respectively, were also obtained; the response profiles are analogous to those for ammonia (Fig. 8). Conclusion This study has shown the potential of the conducting polymer poly(pyrro1e) as a gas sensor operating in a dual mode. Measuring mass and conductivity changes simultaneously is an important step in overcoming the ever present problem of moisture interference in the piezoelectric response. Other workers have employed several methods of overcoming this problem.14 Rigorous steps have been taken to dry gas samples.For example, tubing of the cation-exchange polymer Nafion (Perma Pure Products, Toms River, NJ, USA), a hygroscopic material which allows the transfer of water molecules through its walls, will facilitate the exchange of moisture from the gas travelling through the tube, to a desiccating agent packed around the tube. A second sensor may be used to obtain a measure of the moisture content which is then subtracted. The conductivity measurements used in this work are unaffected by moisture; thus the dual operation reported here provides a method of characterizing and compensating for the humidity response. The sensors reported have been used for periods of up to 1 month with no noticeable loss of sensitivity. Long-term stability trials are underway and will be reported elsewhere.25 Although some workers have reported that ammonia causes irreversible decreases in conductivity,26 we have not observed this effect and this is consistent with results reported by Miasik et al.1 and Nylander et al. 10 The authors thank British Gas for the funding and support of this work and in particular B. Price and D. Byrne of the British Gas Midlands Research Station for the development of the double crystal oscillator and the data acquisition software. We are grateful to Dr. M. J. Freeman, also of the Midlands Research Station, for assisting in the gas sensor test rig experiments and T. Cross for helpful and stimulating dis- cussions. 1 2 3 4 5 6 7 8 9 10 References Miasik, J. J., Hooper, A., and Tofield, B.C., J. Chem. SOC., Faraday Trans. I , 1986,82, 1117. Diaz, A. F., Castillo, J. I., Logan, J. A., and Lee, W., J. Electroanal. Chem., 1981, 129, 115. Asavapiriyanont, S., Chandler, G. K., and Gunawardena, G. A., J. Electroanal. Chem., 1984, 177, 229. Diaz, A. F., Kanazawa, K. K., and Gardini, G. P., J. Chem. SOC., Chem. Commun., 1979,635. Diaz, A. F., and Castillo, J. I., J. Chem. Soc., Chem. Commun., 1980, 397. Kanazawa, K. K., Diaz, A. F., Geiss, R. H., Gill, W. D., Kwak, J. F., Logan, J. A., Talbot, J. F., and Street, G. B., J. Chem. SOC., Chem. Commun., 1979, 854. Josowicz, M., and Janata, J., Anal. Chem., 1986, 58, 514. Dong, S., Sun, Z., and Lu, Z., J. Chem. SOC., Chem. Commun., 1988, 993. Faulds, N. C., and Lowe, C. R., J. Chem. SOC., Faraday Trans. I , 1986, 82, 1259. Nylander, C. Armgarth, M., and Lundstrom, I., Proceedings of the International Meeting on Chemical Sensors, Fukuoka, 1983, eds. Seiyama, T., Fueki, K., Shiokawa, J., and Suzuki, S., Elsevier, Amsterdam, 1983, pp. 203-207.1130 ANALYST, NOVEMBER 1991, VOL. 116 11 12 13 14 15 16 17 18 19 Alder, J. F., and McCallum, J. J., Analyst, 1983, 108, 1169. McCallum, J. J., Analyst, 1989, 114, 1173. Chemical Sensors, ed. Edmonds, T. E., Blackie, Glasgow, 1988, Ho, M. H., Guilbault, G. G., and Reitz, B., Anal. Chem., 1980, 52, 1489. Advances in Electrochemical Science and Engineering, eds. Gerischer, H., and Tobias, W., VCH, Weinheim, 1990, Lai, C. S. I., Moody, G. J., and Thomas, J. D. R., Analyst, 1986, 111, 511. Karmarkar, K. H., and Guilbault, G. G., Anal. Chim. Acta, 1974, 71, 419. Slater, J. M., and Watt, E. J., Anal. Proc., 1989, 26, 397. Lu, C., J. Vac. Sci. Technol., 1975, 12, 578. pp. 295-317. pp. 1-74. 20 Beitnes, H., and SchrGder, K., Anal. Chim. Actu, 1984,158,57. 21 Bartlett, P. N., and Ling-Chung, S . K., Sens. Actuators, 1989, 19, 141. 22 Gardner, J. W., Bartlett, P. N., Dodd, G. H., and Shurmer, H. V., NATO ASZ Ser., 1990, H39, 131. 23 Hanawa, T., Kunabata, S., and Yoneyama, H., J. Chem. SOC., Faraday Trans. I , 1988, 84, 1587. 24 Hanawa, T., and Yoneyama, H., Bull. Chem. SOC. Jpn., 1989, 62, 1710. 25 Slater, J. M., and Watt, E. J., unpublished results. 26 Gustafsson, G., and Lundstrom, I., Synth. Met., 1987,21,203. Paper 1 l00872B Received February 22nd, 1991 Accepted July 2nd, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911601125

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1991, VOL. 116 1131 Potentiometric Determination of Proton Activities in Solutions Containing Hydrofluoric Acid Using Thermally Oxidized Iridium Electrodes Michael L. Hitchman and Subramaniam Ramanathan" Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgo w GI IXL, UK A robust sensor based on iridium oxide is shown to be suitable for use in determining proton activities in solutions containing hydrofluoric acid. Fabricated by a thermal growth process, the low-impedance sensor possesses analytical utility for regulating acidity levels in etching media and pickling-baths. The thermally oxidized electrodes function satisfactorily in hydrofluoric acid solutions with concentrations up to 28 mol dm-3 and show responses comparable to electrodes prepared by the more involved method of reactive sputtering .Keywords: Iridium oxide sensor; h ydrofluoric acid monitoring; thermally oxidized electrode; direct current reactive sputtered electrode Common formulations used for etching in the microelectron- ics industry are based on hydrofluoric acid.132 For example, silicon dioxide and silicate glasses are readily attacked by HF at room temperature, whereas both single crystal and poly- crystalline silicon are wet-etched in mixtures of nitric and hydrofluoric acid. In both instances in order to control the rate of etching it is necessary to regulate the acidity of the media by the addition of buffering agents; NH4F is often used for SiOZ etching and CH3COOH for Si etching. Hydrofluoric acid based pickling- and etching-baths used in the metallurgical industries are similarly dependent on pH control for their optimum functioning3 Thus, there is a need to monitor the proton activity in both etching media and pickling-baths containing HF in order to ensure good process control.The glass electrode is, of course, the most commonly used pH probe, but it is not suitable for use in solutions containing HF because of the etching of the silica-based glass by HF. It has been found, for example, that a glass electrode does not give reliable results when the concentration of HF >5 mmol dm-3.4 This has been commonly attributed to the formation of hydrofluorosilicic acid on the glass membrane, which leads to a masking of exchange sites. The hydrogen electrode can be used reliably to monitor proton activity in media containing HF,5-7 but there is, of course, the need to equilibrate the test solution with a defined partial pressure of hydrogen and this is not very practicable for industrial processes. The quinhydrone electrode has also been used in strong HF solutions4.8~9 where no other redox couple has been present.The necessity to spike the test media with quinhy- drone in this instance is also a serious disadvantage for process control, particularly in flowing solutions. The drawbacks of the glass, hydrogen and quinhydrone electrodes have led to significant interest in the use of alternative probes based on various solid-state configurations. Among the advantages attributed to such sensors are rugged- ness, low-impedance characteristics and ease of miniaturiza- tion.l0 One example of a solid-state sensor is a semiconductor electrode, which makes use of the dependence of etch-rate kinetics on proton activity in HF solutions.11-14 A potential is applied between a semiconductor and stainless-steel electrode and the resulting current flow can be correlated with the acidity, up to about 50 mmol dm-3.Beyond this level the rapid kinetics of the etching process requires sample dilution before a measurement can be made. Another solid-state sensor that * Present address: Singapore Science Centre, Science Centre Road, Singapore 2260, Singapore. has been investigated is a palladium electrode cathodically charged to form palladium hydride.15 This electrode has been used to monitor the acidity of NH4F-HF baths.The electrodes remain viable for 1-2 weeks in this environment. For low concentration fluoride baths, a tungsten oxide electrode has been shown to operate satisfactorily as a pH probe.16 For more concentrated HF solutions, an iridium oxide electrode has been found to be particularly useful.17 However, it has to be prepared by d.c. reactive sputtering, which is not a particularly straightforward or cheap fabrication technique. Therefore, the feasibility of using a simple, readily implemen- ted, thermal method of growth for preparing iridium oxide sensors for monitoring the proton activity in solutions containing HF has been investigated. Experimental The iridium oxide electrodes were made from iridium wires of length 1 cm, diameter 0.5 mm and purity 99.9% (Goodfellow Metals).An electroactive coating was formed by heating the wire, which had previously been soaked in 2 mol dm-3 NaOH, at 800 "C in a furnace for 30 min.18 This process was repeated three times in order to obtain a uniform blue-black coating. The electrode was then cooled in air and immersed in doubly distilled water for 2 d . This step ensured that the iridate coating formed by the thermal treatment was converted into iridium oxide. A small area of the oxide coating was then scraped off at one end of the wire in order to expose the base metal and onto this a length of platinum wire (Johnson Matthey) was spot welded to form an ohmic contact. This junction and most of the platinum wire was encapsulated within heat-shrink poly(tetrafluoroethy1ene) (PTFE) tubing (Farnel Electronics) such that only the iridium oxide surface was exposed.When not in use the electrode was stored in distilled water. For the measurements involving HF a simple PTFE cell was used. The PTFE lid on the cell had two holes drilled through it to accommodate the iridium oxide electrode and a salt bridge consisting of a porous PTFE diaphragm. A saturated calomel electrode (SCE) dipping into saturated KCI contained in the salt bridge provided a long term, stable reference electrode; earlier attempts to use a porous ceramic frit for the salt bridge only gave stable reference potentials over a matter of days. The volume of the sample solution in the PTFE cell was 300 ml. All of the reagents used were of AnalaR grade and the solutions were made up in water, which had been twice1132 - - - ANALYST, NOVEMBER 1991, VOL.116 710 W 690 2 670 2 v) 2 210 I+\ \h 730 W $J 170 $ 150 v) 5 190 1 - . > E 130 \ al f110 Lu" 90 c . ' ' 650 Po q-J- -1.5 -1.3 -1.1 -0.9 -0.7 - Log[ H F] Fig. 1 Variation in potential of iridium oxide electrodes with concentration of HF. A, Thermally oxidized iridium electrode (this work); and B, d.c. reactive sputtered iridium oxide electrode, ref. 17 distilled in glass. Buffers were prepared according to standard recipes. 19 A Corning 150pH-ion meter was used for the potentio- metric measurements. It had an impedance of 1012 Q and the current drawn was typically 1 PA. A voltage output from the pH meter was fed to a Keithley Model 175 digital voltmeter, which was used for logging the potentiometric data.All experiments were performed at 25 k 0.1 "C. Results and Discussion The variation in electrode potential, of a thermally oxidized iridium electrode, with HF concentration in the range 5.6 < [HF] < 28mol dm-3 is shown in Fig. 1. The upper limit represents the concentration of HF as usually supplied, i.e., about 48% m/m. The lower limit represents a typical concen- tration of HF that might be used for etching silicon dioxide and binary and ternary silicate glasses. * Initially, on the introduc- tion of the electrode into the solution the potential drift was typically 0.5-1.0 mV min-1, but after 10 min the drift had decreased by more than one order of magnitude (see below) and this allowed an effective steady-state reading to be obtained.Thus, all of the potential values plotted in Fig. 1 were obtained after a 10min stabilization period. Clearly, a thermally oxidized electrode responds to variations in the activity of the protons. The operative equilibrium, in general, for such an electrode, has been suggested to be as follows:20 211-0~ + 2H+ + 2e- Ir203 + H20 E(mV versus SCE) = 681.0 - 59.1 pH where the pH response is conferred by the acid-base properties of a mixed-valence oxide. However, as can be seen from Fig. 1, the response in HF does not show either a Nernstian slope or, indeed, linearity. The super-Nernstian behaviour ( i e . , a slope >59mV per pH unit) cannot be attributed to variations in the concentrations of H+ with HF as expected from equilibria involving protons and HF;17 e .g . , H+ + F- e HF. Neither would contributions from liquid junction potentials be expected to have such a significant influence on electrode potential. The most likely reason for the deviations from Nernstian behaviour is that in concentrated HF, the proton activity coefficient is a complex function of ionic strength. However, although the dependence of potential on the concentration of HF is non-linear, the reproducibility for separate runs made with the same electrode over several days was good. At each concentration of HF the variation in the steady-state potential measured was never greater than k6 mV and at best it was reproducible to within about +1 mV. As far as the stability of the electrode is concerned, after the initial drift, already mentioned, of =1 mV min-1 the reading Table 1 Fundamental pH electrode parameters measured for ther- mally oxidized iridium electrodes at 25 "C in standard buffer solutions (2.0 d pH S 12.0) before and after treatment in: ( a ) 5% HF + 5% HCI; and ( b ) concentrated HF Before treatment After treatment Electrode E"'ImV (aEl3pH)lmV E"'1mV (3ElapH)lmV identifier versus SCE per pH unit versus SCE per pH unit ( a ) HFl 586.8 58.3 596.3 58.1 HF2 559.8 58.1 551.2 56.8 HF3 620.5 59 .O 547.8 56.7 HF5 583.6 59.2 552.4 54.2 ( 6 ) HF4 680.8 59.1 621.4 54.9 was stable to better than 0.1 mV min-1, and so it would be possible to monitor solutions of HF continuously over several hours without the need for re-calibration.Drift in the electrode potential was also assessed in more detail for mixed electrolytes, and is discussed below.The variation in potential measurements made for the same concentration of HF but with different electrodes could be as great as +lo-15 mV, but this variation would be consistent over the HF concentration range shown in Fig. 1; i e . , the whole curve in the plot would be shifted by the same amount. This variation between different electrodes is not unexpected and is, in fact, commonly observed for pH glass electrodes. It arises from variations in the apparent standard potentials of electrodes.21 The advantage of solid-state electrodes, such as the iridium oxide system, over glass electrodes on this matter is that close agreement between different electrodes can be achieved by a field-induced poising technique. This technique is based on the concept of the pH response of thermally oxidized iridium electrodes being dependent on the acid-base properties of a mixed-valence oxide.** By using this concept, the intervalency transitions induced by an electric field can be used to shift the Fermi level of electrons in the d-n* band to either higher or lower electronic energy levels and so allow the enhancement of the agreement between apparent standard electrode potentials (E"') in a batch of sensors.This can be important not only when sensors are freshly made, and invariably values of E"' differ, but also when they are used in particularly aggressive environments and shifts of E" ' occur as a result of the differential attack of the mixed oxide; for example, the use of electrodes in solutions containing HF.After the experiments in HF, no sign of damage to the thermal oxide coating could be detected on microscopic examination, but changes in the fundamental electrode constants, and the Nernstian slope (3EI3pH) could be seen. Table 1 summarizes the data for three electrodes for which measurements were made in a series of standard buffer solutions over the pH range 2.0-12.0 after being exposed to a mixture of 5% HF and 5% HCI, a typical metal etching solution,3 or to concentrated HF. For each, a diminution in the Nernstian slope is seen to occur in addition to a shift in the values of E"', with a greater effect being observed after exposure of the electrodes to concen- trated HF. The shift in E"' values can, as indicated above, be understood in terms of the variations in the acid-base characteristics of a mixed-valence oxide, and this can be overcome by the application of an appropriate electric field to the electrode in order to restore the original, or for that matter, any desired value of E" '.The variations in the Nernst slope might be due to the incorporation of F- ions onto the oxide sites, possibly in the form of oxy-fluoro species. This would mean that the open circuit potential would no longer be controlled solely by a proton equilibrium but might also include a component from a parallel equilibrium involving coordinated F- ions. A mixed equilibrium involving, on average, more than one electron would then give a sub- Nernstian pH dependence. Some support for this line ofANALYST, NOVEMBER 1991, VOL.116 1133 Table 2 Differences in potential for a thermally oxidized electrode (El) and a sputtered electrode (E2) as a function of the concentration of HF -Loglo[HF] -0.80 -0.90 -1.00 -1.10 -1.20 -1.30 -1.35 -1.40 -1.45 (El -E2)/mV 534 536 542 544 545 550 545 547 551 61 0 590 W 0 2 570 E 2 , 550 E G 530 I I HF1 H F2 L L 0 20 40 60 80 100 120 140 160 180 Time/h Fig. 2 Time dependence of potential of iridium oxide electrodes in 5% HF + 5% HCl. HF1 and HF2 were previously conditioned in 5% HF + 5% HCl for 1 week whereas HF3 was a newly prepared electrode reasoning is given by the observation that electrodes eventu- ally recover a near-Nernstian slope after storage in distilled water when, presumably, the F- ions adsorbed are leached out of the oxide layer.Also plotted in Fig. 1 are the results of Lauks et aZ.17 obtained with electrodes prepared by d.c. reactive sputtering. The general shapes of the two curves are very similar. As has already been indicated the absolute potential value, at any concentration of HF for a given electrode, will largely depend on the E"' for the electrode, so the significantly different electrode potentials shown in Fig. 1 for the two electrodes are probably simply a reflection of the different E"' values. If that is so then taking the difference of the two potentials at the same concentration of HF should give a constant value. Table 2 shows that this is almost true. The mean value is 543.7 k 4.6mV at the 95% confidence level and so the two electrodes show the same dependence of potential on the concentration of HF to within a few per cent.Thus, thermally oxidized electrodes clearly show comparable behaviour to d.c. reactive sputtered electrodes and the much simpler, less involved and cheaper method of preparation of the former type of electrode is obviously a distinct advantage. In addition to concentrated HF being used in the semicon- ductor industry, mixtures of HF and HCI are, as mentioned above, used for metal etching.3 For example, a typical mixture used for aluminium etching would be 5% HF and 5% HCI. The long-term stability of electrodes in this mixture was evaluated and the results are shown in Fig. 2. The electrodes are the same as those used in the investigation of the fundamental parameters reported in Table 1.The greater drift of potential with time of electrode HF3 is consistent with the significantly greater variation observed for this electrode in its electrode parameters, particularly in the E"' arising from the shift in the acid-base properties of the mixed-oxide sensor, as discussed above. Conclusions Iridium oxide electrodes can be readily made and re-gener- ated by thermal oxidation. The reproducibility of the response and stability in HF used in typical industrial environments are sufficiently good to warrant further investigation for process analysis applications. The support of an ORS award for S. R. is gratefully acknowledged. We also acknowledge financial support for this work from Ingold Messtechnik AG, and thank Drs. R. Bucher and H. Buehler of that company for the many useful discussions.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 References Kern, W., and Deckert, C. A., in Thin Film Processes, eds. Vossen, J. L., and Kern, W., Academic Press, New York, 1978, Wolf, S., and Tauber, R. N., Silicon Processing for the VLSZ Era, Volume I-Process Technology, Lattice Press, California, 1988, ch. 15. Smithells, C. J., Metals Reference Book, Butterworth, London, 1976, pp. 310 and 1490. Warren, L. J., Anal. Chim. Acta, 1971, 53, 199. Wynne-Jones, W. F. K., and Huddleston, L. J., J. Chem. SOC., 1924, 125, 1031. Jahn-Held, W., and Jellinek, K., 2. Elektrochem., 1936, 42, 401. Broene, H. H., and de Vries, T., J. Am. Chem. SOC., 1947,69, 1644. Entwistle, J., Weedon, C., and Hayes, T., Chem. Ind. (London), 1973,9,433. Farrer. H., and Rossotti, F., J . Znorg. Nucl. Chem., 1964, 26, 1959. Ives, D. J., and Janz, G. J., Reference Electrodes, Academic Press, New York, 1969, ch. 7. Turner, D. R., Anal. Chem., 1961, 33, 959. McKaveney, J. P., Anal. Lett., 1970, 3, 17. McKaveney, J. P., and Byrnes, C. J., Anal. Chem., 1970, 42, 1023. McKaveney, J. P., and Byrnes, C. J., Anal. Chem., 1972, 44, 290. Jasinki, R., J. Electrochem. SOC., 1974, 121, 1579. Veselkov, E. A., Elektrokhimiya, 1970, 6, 1701. Lauks, I., Yuen, M. F., and Dietz, T., Sens. Actuators, 1983,4, 375. Macur, R. A., US Patent., 1 348 912, 1974. CRC Handbook of Chemistry and Physics, ed. Weast, R. C., CRC Press, Boca Raton, F1, 20th edn., 1985, p. D-149. Pourbaix, M., Atlas of Electrochemical Equilibria in Aqueous Solutions, National Association of Corrosion Engineers, Hous- ton, TX, 1974, p. 397. Hitchman, M. L., and Ramanathan, S., Electroanalysis, in the press. Hitchman, M. L., and Ramanathan, S., Talanta, in the press. ch. V-I. Paper 1 I022236 Received May 13th, 1991 Accepted June 25th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911601131

出版商:RSC    

年代:1991

数据来源: RSC