SERS Surfaces Modified with a 4-(2- Pyridylazo) Resorcinol Disulfide Derivative: Detection of Copper, Lead, and Cadmium L. Gayle Crane, DaoXin'Wang, L. Malia Sears, Bernard Heyns, and Keith Carron 1995 Journal Article mc-95-04 In Analytical Chemistry Volume 67, No. 2 L. Gayle Crane Department of Chemistry University of Stellenbosch Stellenbosch, South Africa DaoXin Wang and L. Malia Sears Desalination System Inc. Escondido, California Bernard Heyns and Keith Carron Department of Chemistry University of Wyoming Laramie, Wyoming 1995 Reprinted from Adyticai Chemistry, 1995,34. Copyright Q 1995 by the Amencan Chemical Society and reprinted by permission of the copyright owner. SERS Surfaces Modified with a 44 2=Pyridylazo)resorcinol Disulfide Derivative: Detection of Copper, Lead, and Cadmium L. Cayle Crane,t DaoXin Wang, L Malii Sears,* Bernard Heyns, and Keith Camon* Department of Chernistfy,UniverSrty of Wyoming, hank, Wyoming 82071-3838 We have developed a surface-enhanced Raman spectm complex coatings with thiol or &sulfide groups are strongly scopic technique for the determination of W+,Cd2+, and ' anchored and do not desorb under strongly acidic or basic Cu2+ concentration using a 4(2-~&~)reso~olconditions. (PAR) coating modified with a didde. The didde High sensitivity with SERS coatings can be achieved through provided a strong anchor to a roughened silver substrate. three sources of signal enhaocement First is the inherent doe- Atomic~citywasdemonstratedbythedistinctspec- fold enhancement available through surfaceenhanced Raman bralc~thato~curredthroughthe~tera~onofW+,spectroscopy. Second, if the coating has accessible dectronic Cd2+, and Cu2+ ions with the indicator. The absorption absorptions, it is possible to tune the laser to achieve additional of these me* by the coatingwas followed with SERS and resonance Raman enhancements. Last, in solution systems the couldbefittoaFNmkinisotherm. Iangmuirianbehaoior large partition coefficients allow us to create a locaIfsr high was not observed; this is most likely due to the elm concentration of analyte at the metal surface. static repulsions as the metai cations are absorbed. The In previous metal ion studies we examined Eriochrome Black detection limits at pH 6 for PW+,Cd2+, and Cu2+ were T (Ern1and a thiol-modiiied diienU>-l%crown-6.4The EBTwas 522,50.3, and 1.49 ppb, respe&veiy. Flow qeriments not chemically bound to the substrate and, therefore, represented using an optical mM?r probe indicated instantaneous an early feasiiili@ study of the potential for SEE in detecting response to changes in metal concentration. species with no inherent viimtional spectmm (single atoms). SERS of EBT metal complexes exhiiited better sel- beWeen We have recently demonstrated the ability to detect ionic chemically similar metal ions than electronic absorption spectme species'-4 and organic compounds in aqueous solutions using copy due to the narrow banMdth of the SERS spectral feabms. chemically modified silver surfaces.S-6 The approach we have The bands that showed the strongest Raman signal were ftWures developed combiies sukedunced Raman spectroscopy 0,associated with the azo bridge of EBT. These bands sbifted in anatyte-specific fitycoatings, and aer-optic sensors. The frequency and intensity according to the size and charge of the coatings vary from simple alkanes to complex macrocycles with metal ion. Trace sensitbity was demonstrated with EBT due to the common theme of containing a thiol or &sulfide group to the strong resonance enhancement that occurs with green Ar+ anchor the coating to a silver substrate. Longchain all~4 laser excitation. disGdes and thiols have been shown to form self-assembled More recendy, a study with a thiobmodified d2xnzelkrwn-6 monolayers (SAM) at noble metal suTfaces.7J These surfaces illustrated the possibility of enhanced chemical selectivity with mimic reversed-phase E3PK coatings and provide excdent SERS coating. Crown ethers are well-known for their atonric surfaces for the detection of hydrophobic organic anajyteS The radius dependent chemical selectivity? Diaminodiil8- Same types of intermolecular interactions responsible for d- crown4 was coupled with meraptonicotinic acid to brm an assembly may not be present in some of the complex metal ion attachable SEE indicaar. It was demonstrated that chemical &elating and macrocyclic coatings that we have synthesized for selectivity could be built into the sensor with a crown ether. metal ion detection. However, it has been found that these However, the sensitivity tended to be lower than EBT due to the ~~ of resonance enhancement report, we will show + Permanent ad* Deparlment of Chem&y, 'Ihe Uaiversi~of stdka lack In this bosch. Stenmboscfi 7600, south Mica results of a surfaceenhanced resonance Raman spectroscopic + Permanent ad* Desalination System Inc. 1240 Sipson Way, Esam (SERRS) sensor with a disulfidemodified 4(2-pyridyhm)resor- dido, CA 92029. (I) CamnK T; Mullen. K L, Lanouent, k;'AnSeffbach. H. AppL Spccaos. CinOl~AR). 1991,45,420-423. This report discusses the atkiiutes of PAR disulfide (PARDS) (2) Mullen. K I; Wag, D.; Cranc L G.; Carron. K T. Ad.Ckem. 1992, a, as a bound metal ion indicator. The absorption of metal ions was 930-936. (3) Mullen. K I; Wag, D.: Huriey, L G.;Cmn. K T. Spectmc@ 1992.7 followed by correlating the intensity ratio of a Raman band (SL24-32 sensitive to metal coordination and its counterpart in the uncom- (4) Heyns, J. B. B; Sears. L U; Corcoran, R C.; Carron, K. T. AdChm. plexed indicator. This approach provides internal standard and 1994.66. 1572-1574. an (5) C~KO~.KT;EL;ImWisM. L; carron. ~~.~~vitrpnsci ~dd corrects for changes in Iaser power or the substrate. 1992.26,1950-1954. The coordination of PAR with metal ions is very pH dependent (s) MUnea K k hn.K. T. Ad. chem. 1994,66,478-#3. We obtained absorption at pH The hthenns 0 wg&g, C. k; chung. C.; Porter, M.J Ekctmad Qornr 199 1,310,335- isotherms 3-6. 359. (8) Bryant M. A; Pmberton, J. E.]. Am. chnn Soc. 1991,113.3629-3637. (9) Christen~ea1.; Eatough. D.; Izan. K Rto: 1974, 74.351-384. 360 Analytical Chemistly, Vol. 67, No. 2, January 75, 7995 0003-2700/9510367-03.00/0 0 1995 American Chemical sodety chloride was produced by the reaction of 393’dthiodip~pionk c H‘ acid and thionyl chloride. PARDS was obtained by combining PAR and 3,3.dithiodiptp pionyl chloride in a Llmix&ure of didoromethane and methanol The major product of this reaction was an asymmetric didde It is assumed that an asymmetric &sulfide is foxmed due to steric . reasons. The product was isolated using hhcbromatographp. with a mixture of acetonitrile and methanoL The structure was I PAR 1 verilied by proton and carbon-13 NMR Adytes. We chose three metals to test the PARDS for sensitivity and selectivity. These metals were chosen because of their environmental interest and their lack of complicated spe- ciation and hydrolysis in aqueous solutions. Reagent grade Pb 0 0 (NO&, Cu(N0&2.5H20, and Cd(N03)rH20 were purchased &omAl&ch and used withoutfurtherpudication. Measuremm no of pH were made with an Orion 520A pH meter. The pH Figure 1. Synthetic route forthe addition of a disulfide functionality adjustments were made with acetic acid and NaOH. Bufferswere to 4-(2-pyridylazo)resorcinol. not used to avoid complexation of metal ions with buffer components (for example, PO43- species). All solutions were made using water obtained from a Millipore system. , were fit to a Frumkin isotherm to account for coveragdependent Substrate Prepamtion. Two types of SEE substrates were electrostatic repulsion.10 In order to combine results with our usedintheexperimentsdiscussedinthispaper. Repetitivetesting accepted fonnulations, we interpreted the in the anaEytical data to determine optimal conditions for the indicators was conducted hmof conditionalfonnation constants. The conditional formation using dver foil substrates? Once these conditions were found, constants allow one coeflicients and sensitivk to calculate partition ~~~were~rmedwiththeopti~~ersubstrates.The ties at any pH. The dynamic response of the SEEindicator was foil substrates were prepared by etching 0.1 mm thick silver foil using a fibemptic flow system. It was found that the system tested (99.9%, Adrich) in a stirred solution of 30%nitric acid for 60-90 responded in less than 60 s or in this case the integration time s. These substrates are known to produce large SERS enhance- required for the spectroscopic measurement The use of an ments.5 After rinsing thoroughly with water, the foil was soaked optical fier codgumtion also demonstrated the potential of this in a 10-5 M solution of the in methanol (speCtrat grade, method for in situ monihing of groundwater metal ion contax& PARDS Baker) for 3-4 h., A methanol wash was nation necessary to remove any physisorbed PARDS. The use of an emission spectroscopy for the determination of The optical Ben were prepared as previously descriied with metal ions is not new. Fluorescencebased metal ion fiberqtic softdad 600 pm silica core mrs purchased kom General Fiber sensors have been investigated by Wangbai,u Carroll,a BrighCu Optics5 We found that the Wenhanced porysiloxane-heduo- and Iiebe~ma.n.1~Detection limits in the micromolar to parts per rine copolymer cladding and nylon buffer interfered with the trillion range were reported. However, while the use of fluores detection of metal ions. Ow interpretationof this interferenceis cence provides excellent sensitivity, it may suffer drom inadequate that a small amount of unreacted monomer in the cladding is selectivitybetween metal ions. For example, Carroll et aL showed slightly soluble in methanol and foms an impermeable film over that lifebe discrimination could be used to gain further select+ the indicator. The problem was deviated by removing the ity.a We will demonstrate that the SERRS spectrum of a bound dadding kom the first centimeter of the fiber by soaking in indicator between ions due to the sharp can distinguish metal coneentratedsulfuricacidfor15min.