Photochemical complexation titrations with fluorimetric end-points

Photochemical complexation titrations with fluorimetric end-points

Analytica Chimica Acta I91 Elsevier Publishing Company, Amsterdam Printed in The Netherlands PHOTOCHEMICAL C O M P L E X A T I O N T I T R A T I O N...

498KB Sizes 3 Downloads 45 Views

Analytica Chimica Acta I91

Elsevier Publishing Company, Amsterdam Printed in The Netherlands


J O H N L. B E C K * , ]'. M. F I T Z G E R A L D * * A N D J O H N A. B I S H O P * * *

Department of Chemistry, Seton Hall University, South Orange, N.J. 07079 (U.S.A.) (Received

February 7th, 197 o)

In previous publicatioI~s work on photochemical titrations 1-4 a a d on fluorimetric titrations 5-v has been reported. The present paper discusses a combination of the two methods to permit titrations in which the titrating ion for a fluorimetric displacement reaction is generated in a photochemical decomposition. Both the photolysis and fluorescence excitation are accomplished with a single u.v. source1, a. ]~XPERIMENTAL

Reagents Reagent-grade chemicals were used without further purificatiort with tile exception of potassium trisoxalatocobalt (III) and sodium perchlorate. Deionizedwater was used throughout the study. The cobalt(III) oxalate complex was prepared as described b y BOOTR8 and purified". The ultraviolet spectrum of the final product agreed with the values found b y COP~STAKE AND UR110. The air-dried solid was stored in brown bottles under refrigeration, since the compound can decompose by both photochemical reaction and a thermal reaction 10. Stock solutions of Mg(II), Cd(II), zn(II) and cu(II), all Io-2 M, were prepared from their sulfates and were standardized b y titration with standard EDTA. Solutions of 8-hydroxyquinoline-5-sulfordc acid (Aldrich Chemical Company) were prepared b y weight as needed, at a concentration of 4" lo-a M. In preparing this solution, just enough sodium hydroxide was used to cause complete dissolution, and the final pH was adjusted to 8- 9 . Sodium hydrogen carbonate buffer (o.5 M ) w a s prepared b y weight and adjusted to pK 8.2 with sodium hydroxide. A 0.5 M phosphate buffer was prepared b y dissolving sodium monohydrogeu phosphate in water and adjusting the pl~ with sodium hydroxide. In order to study the effect of electrolytes on this titration, I M potassium chloride was prepared (by weight), and I M sodium perchlorate was prepared b y allowing stoichiometric amounts of sodium carbonate and hydrochloric acid to react in water. The final pH was adjusted to 8. 4 with sodium hydroxide. * P r e s e n t a d d r e s s : Merck I n s t i t u t e for T h e r a p e u t i c Research, R a h w a y , N . J . 07016. ** P r e s e n t a d d r e s s : D e p a r t m e n t of C h e m i s t r y , U n i v e r s i t y of H o u s t o n , H o u s t o n , T e x a s 77004; to w h o m c o r r e s p o n d e n c e s h o u l d be addressed. *** P r e s e n t a d d r e s s : D e p a r t m e n t of C h e m i s t r y a n d Chemical E n g i n e e r i n g , N e w a r k College of E n g i n e e r i n g , N e w a r k , N . J . 07102.

Anal. Chim. Acta, 51 (197o) I 9 1 - 1 9 8



A~aratus The apparatus used was based on the single source design of S,'VItTH1,a with small modifications, and is shown schematically in Fig. i. A medium-pressure mercury are (Hanovia Utility Quartz Lamp, type 30520, Englehard Hanovia, Inc., Newark, N. J.) served as the source of radiation for photolysis and also as the source of fluorescence excitation radiation. Voltage to the arc was regulated to help stabilize the intensity2, ~,9 (Model IoooS, Sorensen and Co., Inc., Norwalk, Conn.). The electrically operated shutter (Harvard Electric Shutter, Burke and James, Inc., Chicago, Ill.) L



i i L._


i i _1



' C


30 mY20

FZ !


10 0















Fig. I. Schematic design of a p p a r a t u s : (A) m e r c u r y vapor a r c , (L) lamp housing, (S) electrically operated slautter, (F1) u.v. b a n d p a s s filter, photolysis and fluorescence excitation; (F~) interference filter, fluorescence emission selection; (O) light aperture; (B) light baffle; (C) titration vessel; (P.C.) photovoltaic cells; (by1) p a t h of incident radiation; (hv2) p a t h of e m i t t e d fluorescent light; (M) magnetic stirrer. Fig. 2. Typical fluorescence titration curve: m V = r e l a t i v e fluorescence i n t e n s i t y ; time = time from start of photolysis; E P = e x t r a p o l a t e d end-point.

was mounted directly on tile lamp housing. Four selenium photocells (Type B-2M, International Rectifier, E1 Segundo, Calif.) were connected irt series, to a recorder (Model SRL, Sargeant & Co., Chicago, Ill.) operated in a linear-millivolt mode. The shutter and the chart drive were coupled so that the shutter opened when the time drive started; a chart speed of I in/min was used. Dark current was zeroed with an external bucking current between1 the photocells and the recorder input. In the excitation radiation path, a u.v.-transmitting-visible-absorbing filter ( # 7-54, Corning Glass Works, Coming, N.Y.) was combined with a heat-absorbing filter ( # 654o-ooi, Englehard Hanovia, Newark, N.J.) to prevent cracking of the u.v. filter. A 5oo-nm, 2-in2 interference filter was used in position 2 (Fig. I) to select the radiation for fluorescence measurement. A~¢al. C h i m . A c t a , 51 (197o) 191-198



Procedure for titration

Prepare a fresh solution of M(HQSOa)2 (8-hydroxyquinoline-5-sulfordc acid anion will be shortened to HQSOa for the remainder of the paper) from the stock solution of the desired metal sulfate and a freshly prepared solution of HQSOa, to a final concentration of I . io-a M. Prepare a cobalt(III) oxalate solutioi1 (5" IO-2 M) b y weight daily. Best results are obtained when the solution is stored in an ice-bath to minimize thermal decomposition10. To a 25o-ml graduated beaker add 2o ml of o.5 M phosphate buffer (pH 9) followed b y an aliquot containing I-15 /,moles of the complex to be determined. Finally, add 4 ml of the 5 "Io-~ M Co(III)-oxalate complex and dilute to a 2oo-ml m a r k on the beaker. After adding a stirring bar, place the vessel itl the apparatus, start stirring, and throw the switch which starts the recorder drive as the shutter opeHs a. A plot of mV vs. time of photolysis is obtained (Fig. 2). By using different amounts of M(HQSOa)2 it is possible to prepar.e a standardization curve in which the extrapolated time to reach the extinction of fluorescence (EP, Fig. 2) is plotted agairlst/*moles of complex takeH (Fig. 3).


75 z













/ // ..T"



/{,,I I 260 280 300

320 340 360 380 4;0


Fig. 3. Calibration titration plots for fluorescent complexes: ( • )Zn(HQSOa)2 (least-squares line 1-15 /zmole); (A)Cd(I~QSOs)2 (least-squares line, 2-15 ~mole); (Ill)Mg(HQSO3)2 (least-squares line, 1-8 #mole). Concentration of p h o t o g e n e r a t o r Co(III) o x a l a t e = IO -a M ; pH range 8-9.5. Fig. 4. Electronic absorption spectra: ( ) Co(III) oxalate, i . i o - a M , pI~ 8.8 ; (. . . . . . ) Mg(HQSO~)o,, 4 . i o - 5 M , pi~ 9; (....... ) Cd(HQSO3),2, 5 . i o - a M , p ~ 8. 5. All spectra for I-cm p a t h lengths. (]) Mercury emission wavelengths. Intensities relative to t h a t at 366 n m (corrected for detector r e sponse).

A metal ion which forms a nolffluorescent species can be determilled b y usillg an excess of a fluorescent complex, followed b y hack-titration of theexcess. Copper(TI) was determined in this way b y the use of excess of Cd(HQS03)~, the excess being titrated b y the photogeneration of cobalt(II). The calibration curve for the copper(II) titration was linear over the range 1-13.5/,mole of copper(II) for a photogeneration time range of 5-1 min when 15/,mole of Cd(HQSO3)2 were used under the conditions discussed above. Anal. Chim. Acta, 51 (197 o) 191-198




T h e c o m b i n a t i o n of a photochemical decomposition a n d fluorescent e n d - p o i n t is made possible b y the very favorable absorption characteristics of ultraviolet a n d visible light in t h e systems ii~volved (Fig. 4). The absorption of ultraviolet light b y the three fluorescent species is low at wavelengths causing decomposition of the cobalt oxalate, while the cobalt complex t r a n s m i t s relatively well at 366 rim, which is the wavelength causing fluorescence of the M(HQSOa)2 complexes. DREW 9 found t h a t incomplete a b s o r p t i o n of a strong m e r c u r y emission line at 366 n m results i n v a r i a t i o n s in the rate of the cobalt oxalate decomposition, a n d he was forced to use filters to remove this line from t h e photolytic p a t h (or alternatively, concentrated solutions4). I n the present s t u d y the M(HQSO3)2 complexes acted as " i n t e r n a l filters ''4 to remove this interference. T h e fluorescent complexes used all have b r o a d - b a n d emission spectra with m a x i m a b e t w e e n 495 a n d 515 n m 6. There is no m e r c u r y line near 5oo rim, and, i n addition, the cobalt oxalate complex has a m i n i m u m i n its a b s o r p t i o n spectrum at this wavelength 10. Consequently, a 5oo-Ilm interference filter was used t o select emitted fluorescent light. Typical d a t a for the photochemical t i t r a t i o n of one of the fluorescent complexes, Cd(HQSOa)2, are shown i n Table I. T h e complex was t i t r a t e d over a seven-fold range. A least-squares fit of d a t a yielded a slope of o.373 m i n / # m o l e with a n intercept of - o . 2 6 1 min. T h e relative s t a n d a r d d e v i a t i o n for three d e t e r m i n a t i o n s at the 8-#mole level was 1 . 3 4 ° . TABLE





Cd(HQSOa) ~ taken (#mole)

Photolysis time to e~,d-poim obsd. (rain)

Photolysis time to end-point calcd, a (mi~*)

Error b (%)

2.02 3.04 4.05 5.06 6.07 8.1o lO.12 12.14 15.18

0.48 0.93 1.31 1.62 1.92 2.74 3.51 4.20 5.48

0.49 0.87 1.25 1.63 2.00 2.76 3.51 4.27 5.4°

-+ + ---

2.04 6.45 4.58 0.62 4.17 0.73 o.oo -- 1.67 -1- 1.46

Calculated from least-squares fit of data: time to end-point = -- o.261 rain + o.373 min/#mole. b Difference between observed and calculated values. I t was n o t possible to t i t r a t e mixtures. W h e n one a t t e m p t s t h e t i t r a t i o n of a m i x t u r e of m e t a l ion a n d excess HQSOa, the fllitial species in solution are the fluorescent M(HQSOa)2 complex a n d free HQSOa. One m i g h t expect to o b t a i n the t y p e of t i t r a t i o n curve i n which there is no change in tile level of fluore~a'cence u n t i l the first break at which time all of the excess HQSOa is titrated. T h e n one m i g h t expect the fluorescence i n t e n s i t y to decrease to zero as the M(HQSOa)2 complex is t i t r a t e d a n d A n a l . Chim. Acts, 51 (197o) I91-I98



the second b r e a k is reached. The actual t i t r a t i o n curve o b t a i n e d from such a m i x t u r e has no initial flat p o r t i o n b u t begins to decrease immediately, slowly at first a n d then more r a p i d l y with a good deal of r o u n d i n g . The reasons for this behavior m a y be found in the spectral properties of the various species in the solution. HQSOa a n d Co(HQSOa)~ as well as M(HQSO3)~ absorb f a i r l y s t r o n g l y at 366 rim. The absorbance of the Co(HQSO3)2 is stronger t h a n t h a t of free HQSO3 at this wavelength. Since 366 rim is the excitation w a v e l e n g t h used to cause the M(HQSOs)2 complexes to fluoresce, one can see t h a t a n e t increase i n q u e n c h i n g results as Co(HQSOs)2 is produced during the t i t r a t i o n of t h e excess of HQSOs a n d t h u s a decrease in the fluorescence i n t e n s i t y is observed d u r i n g the first p a r t of the t i t r a t i o n . Quenching also occurs d u r i n g the second part of the t i t r a t i o n as Co(HQSOs)2 is produced i n a d d i t i o n to t h e decrease of fluorescence as t h e M(HQSOs)2 complex is t i t r a t e d ; hence, the increasing rate of decreasing fluorescence. This stepwise n a t u r e of the formation a n d displacement of the various complexes with HQSOs results i n a t i t r a t i o n curve which is unusable. Several t i t r a t i o n s were performed on the various fluorescent complexes with complete s h u t d o w n of e q u i p m e n t b e t w e e n runs. The results of these t i t r a t i o n s are condensed i n T a b l e II. Five t i t r a t i o n r u n s for the three complexes are listed, all r u n on different days. These r u n s were made d u r i n g some very hot weather i n a l a b o r a t o r y w i t h o u t air-conditioning. I t was n o t e d t h a t if a series of t i t r a t i o n s were performed TABLE



Metal ion complexes

Range (#mole)

No. of detns.

Slope~ (min/#mole)

Intemept S (mi~)

Difference b (%)

Rel. st. dev. (%)

Cd Cd Zn Mg Mg

4-12 2-15 1-15 I- 8 2- 8

9 9 7 15 18

o.355 o.373 0.430 o.31o o.4o2 o

o.o89 -- o.261 o.154 -- o.o6o -- o.337

-2.48 o-96 ---

1.34° --I.O9a 1.3oa

Least-squares fit of data: time to end-point = min+min/#mole. b Difference between observed and calculated times averaged over all points determined. o Relative standard deviation for triplicate determinations at the 8-#mole level. a Relative standard deviation for triplicate determinations at the 4-/~mole level. e Potassium trisoxalatocobalt(III) cone. = 5o% that used in other sets. late i n the evening, using a cobalt(III) oxalate solution which h a d been prepared i n the morning, the least-squares intercepts of the calibration plots became increasingly more negative. This is due to t h e t h e r m a l decomposition of the reagent solution with the resulting formation of cobalt (II) i m p u r i t y ", 10. This problem can be c i r c u m v e n t e d i n either of two ways. Since the c o b a l t ( I I I ) oxalate dissolves in aqueous solution immediately, it could be added as the solid if a c o n v e n i e n t l y reproducible w a y to accomplish this is available. Alternatively, the freshly prepared c o b a l t ( I I I ) oxalate solution can be kept in art ice-bath. At this t e m p e r a t u r e n o significant change in the intercepts of the calibration plots is n o t e d over a 24-h period. Since an a r b i t r a r y millivolt scale can be set on the recorder, the shapes of all of the t i t r a t i o n curves are q u a l i t a t i v e l y alike (Fig. 2) a n d can be m a d e as steep or as shallow as one whishes b y choosing the proper full-span setting. The place at which some difference does arise i n t h e curves is t h e end-point region (EP ; Fig. 2). Mg(HQAnal. Chim. Acta, 51 (197o) 191-198



SO3)2 exhibits the least amount of "rounding" in the end-point region, with Cd(HQSO3)2 next arid Zn(HQSO~)2 exhibiting the greatest rour~ding. This behavior falls in line nicely with the relative strengths of the complexes being titrated compared to the strength of the complex being formed in the titration 6, i.e. Co(HQSOa)2. Tile relative strengths of the complexes are in the order Co, Zn, Cd, Mg and therefore tile rounding of the end-point should be in the order Zn, Cd, Mg, which is, ill fact, observed. I t can be seen in Fig. 3 t h a t the calibration plots for the complexes do not have the same slopes nor do they have zero intercepts. Theoretically, the titration rate depends on the rate of ptlotolysis atone 4 so that all the slopes in Fig. 3 should be identical. However, complications have been observed in previous photochemical methods. DREW 9 and SMITll AND FITZGERALD 1, in their titrations, attributed their non-zero intercepts to traces of cobalt (II) in the cobalt (III)-oxalate reagent. As has been stated above, extreme care must be taken to prevent thermal degradation of the reagent yielding cobalt(II). DREW 9 arid SMITll AND FITZGERALD][ attributed the non-stoichiometric relationship of some of their titrations to induced reactions but this is unlikely in the present set of metal-ion chelate titrations. A more reasonable explanation is that the titrations are stoichiometrie but the very similar, though not identical, shape of the electronic absorption spectra of the various complexes affects the slope of the calibration curve. The monocomplex and the dicomplex have different absorption spectra and therefore the overall spectrum changes somewhat during the titration. The relative changes in the solution as a whole depend also upon the small differences in molar absorptivity at 366 nm for a given complex, compared to the absorptivity of the Co(HQSO~)~ which is being produced. These spectral changes result in the "internal filter effect ''4 for both the photolytic radiation and the fluorescence excitation radiation. In addition, the "window" for fluorescence emission is subtly altered during tile course of the titration as the concentrations of the various absorbers change dramatically. The situation is obviously quite involved but qualitatively the cobalt complex exhibits the most intense absorption of 366 nm at the pl~ used in the titration and is followed closely b y complexes of zinc, cadmium and magnesium, in that order. This is also the order of the relative slopes of the titration curve, i.e., Zn, Cd, Mg. I t should also be noted that the quantum yield for cobalt(II) production is changed slightly by heavy metal ions such as mercury(II) and copper(II)lO. It is unlikely, however, t h a t the micromolar amounts of metal ions taken for titration could account for the differences ill slopes of calibration curves. With the Cd(HQSO3)2 system, a study was made on the effect of tile buffer used. Cd(HQSOs),o (5 ml) was titrated b y the standard procedure with sodium acetate, sodium hydrogen carbonate or sodium phosphate buffers. Similar titration curves were obtained with all three buffers. I t has been previously reported 5 that the fluorescence for the complexes used decreases as the pll is lowered, owing to incomplete formation of the M(HQSO3)2 complex, tile rate of decrease with drop in pll depending on the strength of the complex. At very basic pK values the fluorescence again falls off, probably due to formation of hydroxides, pK values between 8-9. 5 gave the best results (Table II). COPESTAKE AND UR110 found that neutral electrolytes had no effect on the photoreduction of cobalt(III) oxalate s0. The electrolyte effect was studied b y using the standard procedure with the addition of neutral electrolyte, with titrations of Anal. Chim. Acta, 51 (197o) 191-198





Cd(HQS0~)2. P o t a s s i u m chloride at 0.25 M a n d 0.5 M caused 11o change in the shape of the titratiort curve, b u t s o d i u m perchlorate at the same c o n c e n t r a t i o n s curved the t i t r a t i o n plot, a n d shortened the e x t r a p o l a t e d times to end-points. However, at 0.25 M sodium perchlorate concentratiort the time was shorteHed o n l y 5%, so t h a t small cortcerttratiorts of perchlorate m i g h t be tolerated. Art examinatiort of the s t a b i l i t y cortstants of the complexes of 8 - h y d r o x y q u i nolirte-5-sulfonic acid with b i v a l e n t cations 6 indicated t h a t it should be possible to determine copper(II) b y m e a n s of excess of a fluorescent complex which is displaced b y copper (e.g., Cd(HQSOa)2) arid b a c k - t i t r a t i n g the excess. The results of such tit r a t i o n s shown ill Table I I I i n d i c a t e t h a t such is the case. These results refer to a calibration plot composed from the least-sRuares fit of the data. I t should be possible to a p p l y this b a c k - t i t r a t i o n m e t h o d to other cations which are more s t r o n g l y complexed t h a n cobalt(II), b u t which do rtot form fluorescing complexes w i t h 8-hydroxyquinoline-5-sulfortic acid. On the other h a n d , small differences irt slopes of calibration curves m a y well be observed w i t h different metal-ions used. This would again be due to the change in c o n c e n t r a t i o n of the various absorbing species d u r i n g the course of the t i t r a t i o n . TABLE












COBALT(III) OXALATE Copper(f I) (l~mole)

Time to end-point obsd. (rain)

Time to end-point calcd, a (mi~¢)

Difference b (%)

5.o3 4.7° 4.23 3.4° 3.38 2.75 2.66 2.85 2.82

5.02 4.71 4.09 3.44 3.44 2.82 2.82 2.82 2.82

+ o.16 -- o.23 + 3.45 -- 1.23 -- 1.81 -- 2.45 -- 5.64 + 1.o9




12.1 13.5 (4.3 p.p.m).

1.62 1.23

1.6o 1.34

I.O (0.32 p.p.m.) 2.0 4.0 6.0 6.0 8.o 8.0 8.0 8.0




-t- 1.22 -- 8.50

a Calculated from least-squares fit of data: time to end point = 5-33 min -- o.311 min/#mole. b Difference between observed and calculated values. Relative standard deviation for four determinations at the 8-#mole level = 3.04% . SUMMARY A fluorimetric e n d - p o i n t m e t h o d has been developed for the photochemical decomposition of cobalt(III) oxalate to produce cobalt(II) ion as t i t r a n t . The m e t h o d is used i n t i t r a t i n g the fluorescent complexes formed between 8-hydroxyquinoline-5sulfonic acid artd Mg(II), Cd(II) a n d zn(II) b y displacement of the m e t a l ion b y cobalt(II), which forms a non-fluorescent complex. A b a c k - t i t r a t i o n of art excess of c a d m i u m ( I I ) complex is used for the determirtatiort of copper(II) which also forms a non-fluorescent complex with t h e ligand. A single ultraviolet source is used for photoAmd. Chim. Acta, 51 (197o) 191-198



lysis altd fluorescellce excitation. The method is made possible by favorable u.v. absorption spectra of the various complexes involved.

Uiie mdthode X point final fluorimdtrique est mise au poi1it; elle est basde sur la d6compositio1i photochimique de l'oxalate de cobalt(III) eli io1i cobalt(II) utilisd comme titrant. Elle permet le titrage de complexes fluorescents form,s entre l'acide bydroxy-8-qui1ioldine sulfonique-5 et, Mg(II), Cd(II) et Zn(II), par [email protected] du mdtal par le cobalt(II), formant uI1 complexe non fluoresce1it. Une titration eli retour par un exc~s de complexe de cadmium permet de doser le cuivre, formant dgalemeat u1i complexe 1io1i fluorescennt avec le ligalld. Uiie simple source ultra-violette est utilisde pour la photolyse et l'excitatioI1 de fluorescence. ZUSAMMENFASSUNG

Es ist eine fluorimetrische Endpunktsmethode ftir die photochemische Zersetzung vo1i Kobalt(III)-oxalat zum als Titrant verwendeter~ Kobalt(II)-io1i entwickelt wordenn. Bei der Methode werde1i die zwischen 8-Hydroxychinoli1i-5-sulfonsaure und Mg(II), Cd(II) u1id ZH(II) gebildeten fluoresziere1iden Komplexe titriert, indem das Metallion durch Kobalt(II) ersetzt wird, das ei1ien 1iichtfluoreszierendenKomplex ergibt. Die Riicktitration eines Uberschusses an Cadmium(II)-Komplex wird ffir die Bestimmung vort Kupfer(II) be1iutzt, das ebe1ifalls mit dem Ligander~ ei1ie1i 1iichtfluoresziere1iden Komplex bfldet. Ffir die Photolyse und die Fluoreszellzanregung wird eine ei1izige Ultraviolett-Strahlenquelle verwe1idet. Die Methode wird durch die giinstigen u.v.-Absorptio1isspektren der verschiedenen Komplexe ermSglicht. REFERENCES

I E. J. SMITH AND J. M. FITZGERALD, Abstracts, 153rd National Meeting A'CS, M i a m i Beach, Fla., April ~965, No. 134o; Abstracts, z54th National Meeting ACS, Chicago, Ill., Sept. z967, No. 1346 . 2 H. D. DREW AND J. M. FITZGERALD, Alcal. Chem., 38 (1966) 778. 3 R. J. LUKASlEWICZ AND J. M. FITZGERALD, Anal. Lttrs., I (1968) 455. 4 J- M. FITZCERALD, R. J. LUI;ASlEWICZ AND H. D. DREW, Anal. Lttrs., I (1967) 173. 5 J. A. 131SHOP, Anal. Chim. Acts, 29 (1963) 172, 178. 6 J. A. 13ISHOP, Anal. Chim. Acts, 35 (1966) 224. 7 J- A. ]3ISHOP, Anal. Chim. Acts, 39 (1967) 189; Anal. Lilts., 2 (I969) I I I . 8 H. S. BOOTH, Inorganic Syr~thesis, Vol. I, McGraw-Hill, 1939, P. 37. 9 H. D. DREW, Ph.D. Thesis, Seton HalI University, S o u t h Orange, N. J., 1967, p. 35. IO T. B. COPESTAKE AND N. URI, Proc. Roy. Soc. (London), Set. A , 228 (1955) 232. Anal. C/tim. Acts,

51 (197o) I9~-198