The Use of Solid Phase Reactors as the Source of a Catalytic Solution . The Indirect Flow-Injection Spectrophotometric Determination of Amino Acids

A determinação indiretade aminoácidos é realizada em um sistema de injeção de fluxo utilizando-se um reator de fase sólida que contém sais cúpricos, imobilizados em esferas de resina de poliéster. Uma substância farmacêutica é forçada através do reator e os íons cúpricos liberados (complexados pela substância farmacêutica) agem como catalisadores da reação subseqüente en tre Fe(III) e tiosulfato de sódio. A curva de calibração é lin ear no intervalo de concentrações 0,1-0,3 μg L para glicina, com um desvio padrão relativo de 2,3% e uma velocidade de passagem de 28 amostras por hora. Foi analisada a influência de substâncias estranhas e o método foi aplicado à determinação de glicina em duas formulações farmacêuticas diferentes.


In tro duc tion
Solid phase re ac tors are now firmly es tab lished in contin u ous-flow ap pli ca tions.Their use in FI as sem blies of fers some ad van tages over the ho mo ge neous so lu tion sys tems that have been il lus trated in sev eral pa pers 1,2,3 .In fact, FI solid phase re ac tors af ford a va ri ety of ap pli ca tions and facil i tate sam ple der i va tion.So far they have been used for the pu ri fi ca tion of both sam ple and re agent so lu tions, and for the in situ prep a ra tion of un sta ble re agents 4 .
Sev eral meth ods have been pro posed for the de ter mi nation of amino ac ids as rel e vant in di ca tors of nu tri tional require ments.The di rect titrimetric de ter mi na tion of amino ac ids in mul ti vi ta min for mu la tions has been of fi cially recom mended 5 .How ever, most of the pub lished meth ods for the de ter mi na tion of amino ac ids are based on the der i vation of the analyte, which al lows an in crease in sen si tiv ity and fa cil i tates the sep a ra tion step in real sam ples, which usu ally con tain a mix ture of sev eral amino ac ids.In this way, fluorimetric 6 , polarographic 7 , and chro mato graphic meth ods (TLC 8 , GC 9,10 and HPLC [11][12][13] ) have been proposed.
Some meth ods based on the FI meth od ol ogy have been de vel oped.Glycine and cysteine can be fluorimetrically de ter mined by means of prior oxydation with po tas sium Ar ti cle iodate in the pres ence of 2-mercaptoethanol 14 .In di rect deter mi na tions of amino ac ids in phar ma ceu ti cal for mu lations have been car ried out us ing solid-phase re ac tors 15,16 and atomic ab sorp tion for the de tec tion of the me tal lic cation lib er ated from the re ac tor.Sen sors for in dus trial process mon i tor ing us ing en zyme-cartridge flow-injection anal y sis 17,18 , electrodialysis with amino acid oxydation to en hance se lec tiv ity in FI for L-amino ac ids 19 , and bienzyme sen sors based on elec tri cally wired peroxidase 20 , have im proved the an a lyt i cal strat e gies for amino ac ids.
This pa per deals with a new strat egy for ex ploit ing the use of solid phase re ac tors for the in di rect de ter mi na tion of phar ma ceu ti cal sub stances.The pro posed man i fold includes a re ac tor con tain ing cop per car bon ate im mo bi lized on a poly mer resin.On pas sage through the re ac tor, the sam ple forms a com plex and re leases the metal ion; then the re leased cu pric ion acts as the cat a lyst in a sub se quent re ac tion be tween Fe(III) and thiosulfate ion.The prod uct of this re ac tion (fer rous ion) is spec tro pho to met ri cally mon itored (at 510 nm) by re ac tion with 1,10-phenanthroline (hence forth re ferred to as phen).

Re agents
Aque ous so lu tions of glycine (Probus) at pH 9.5 were potentiometrically ad justed by drip ping di luted NaOH (Panreac).The other aque ous so lu tions were fer ric ni trate (Panreac), so dium thiosulfate (Panreac), phen in 0.05 mol L -1 HCl (Panreac).The solid-phase re ac tor was pre pared ac cord ing to a pre vi ously pub lished pro ce dure 15 with CuCO 3 .Cu(OH) 2 .2H 2 O (Panreac), poly es ter resin so lu tion AL-100-A (Reposa) con tain ing the co balt ac ti va tor and add ing methylketone (Akzo) as the cat a lyst.All re agents used were of an a lyt i cal grade, un less oth er wise stated.

Ap pa ra tus
The pro posed FIA man i fold is de picted in Fig. 1.The as sem bly was pro vided with a Minipuls 2 peri stal tic pump (Gilson) and a 5041 in ject ing valve (Rheodyne).The complex formed was mon i tored at 510 nm by a Lambda 16 (Perkin-Elmer) spectrophotometer, pro vided with a 18 µl flow-cell (Hellma).The PTFE tubes were of 0.8 mm ID, ex cept the solid-phase re ac tor which was pre pared with 1.5 mm ID PTFE tubes.

Pro ce dures
578 µL of aque ous al ka line glycine so lu tion at pH 9.5 (potentiometrically ad justed by drip ping di luted NaOH) was in jected into the car rier so lu tion of NaOH (pH 9.5, flow-rate 2.6 mL min -1 ).The in jected so lu tion was forced through the solid-phase re ac tor (10 cm long, 1.5 mm ID, and par ti cle size 150-200 µm), and then merged with a 7.0 x 10 -4 mol L -1 Fe(III) so lu tion.The sub se quent mix ture with a 2.58 x 10 -3 mol L -1 Na 2S2O3 so lu tion pro duced Fe(II), which was spec tro pho to met ri cally de ter mined at 510 nm by means of re ac tion with a 0.5% (w/v) phen so lution.

Re sults and Dis cus sion
Fig ure 2 shows the ex per i men tal setup and so lu tions used in pre lim i nary ex per i ments with a num ber of phar maceu ti cal sub stances that re leased cu pric ion from the re actor by a dis place ment re ac tion.The cu pric ion, re leased from the solid-phase re ac tor by complexation with the phar ma ceu ti cal sub stance, acts as a cat a lyst for the re ac tion be tween Fe(III) and thiosulfate ion.The in jec tion of a phen so lu tion and absorbance read ings at 510 nm pro vided the amount of Fe(II) pres ent.Ex per i ments were car ried out by al ter nately cir cu lat ing so lu tions of the phar ma ceu ti cal under study through chan nel d at pH 8.5 and an aque ous stream at the same pH.The most fa vor able re sults were obtained with adren a line, chlorpromazine, dipyrone, pyramidon, and glycine (Ta ble 1).These phar ma ceu ti cal sub stances were used in fur ther ex per i ments where the reagent (cu pric salt) was re moved from the re ac tor in or der to as cer tain whether the pres ence of free phar ma ceu ti cal substance in the re ac tor af fects the re ac tion be tween Fe(III) and thiosulfate.These tests re vealed that adren a line provides the most prom is ing re sults (un der the work ing con ditions used, the phar ma ceu ti cal sub stance was readily ox i dized by Fe(III)).
Ex per i ments per formed with three amino ac ids provided re sults as good as those pre vi ously ob tained with adren a line.In fact, the cal i bra tion graphs for glycine, histidine, and proline were lin ear (with cor re la tion co ef ficients of ca.0.998) and had sim i lar slopes.The through put was close to 250 sam ples h -1 .
Not with stand ing the ex cel lence of these re sults, we tested al ter na tive con fig u ra tions in or der to re duce phar ma -Fig ure 1.The flow-injection as sem bly for the de ter mi na tion of amino ac ids.P: peri stal tic pump; V: in jec tion valve; L: sam ple loop; D: detec tor.
Fig ure 2. The flow-injection as sem blies from pre vi ous stud ies on the in flu ence of chem i cal and FIA pa ram e ters.ceu ti cal sub stance con sump tion.In fact, the pre vi ous assem bly re quired a high vol ume of sam ple to flush the re ac tor on switch ing be tween sam ples of dif fer ent con centra tions; oth er wise, carry-over would have been quite signif i cant.In ad di tion, the new ex per i ments were aimed at in creas ing the sen si tiv ity and de creas ing the de tec tion limit by di min ish ing the blank sig nal.
A se ries of pre lim i nary ex per i ments were car ried out in or der to char ac ter ize and con trol the in flu ence of sev eral ex per i men tal vari ables on the chem i cal sys tem.The in fluence of the re ac tant con cen tra tion was stud ied by keep ing it con stant in the cir cu lated so lu tions and chang ing the flow-rate of the streams.The flow-rates of the Fe(III) and thiosulfate streams (iden ti cal for both) were al tered and the in flu ence of the pres ence of the cat a lyst at vari able con centra tions of fer ric and thiosulfate ions was stud ied (the glycine flow-rate was kept con stant and the amino acid was re placed with wa ter to eval u ate the blank sig nal).The absorbance de creased with in creas ing flow-rate of the Fe(III) and thiosulfate streams.The high est absorbances were ob tained at low flow-rates (lower than 1 mL min -1 ) of both Fe(III), sul fate and aque ous glycine, which sug gests that the re dox re ac tion is fa vored by long res i dence times.How ever, the dif fer ences be tween the sig nals re vealed that the best re sults were achieved with in ter me di ate cat a lyst flow-rates (about 3 mL min -1 ) and low Fe(III) and thiosulfate flow-rates.In other words, a com pro mise must be made be tween res i dence and con tact times, and an ef fective cat a lyst con cen tra tion.Ob vi ously, the lon ger the res idence time, the greater the ex tent of re ac tion de vel op ment, and hence the greater the de tec tor re sponse.How ever, the ef fect of in ter fer ences may be in creased.The dif fer ences be tween the sig nals for the blank (dis tilled wa ter) and the sam ple (phar ma ceu ti cal sub stance so lu tion) were also illus tra tive of the re ac tion ki net ics in the ab sence of a cat alyst.
We stud ied the de vel op ment of the re ac tion us ing the stopped-flow tech nique.For this, we car ried out a se ries of ex per i ments chang ing the phen con cen tra tion in the presence and ab sence of the cat a lyst.The flow was stopped for 10 min and 1,10-phenanthroline con cen tra tions in the range of 0.02-0.5% (w/v) in 0.05 mol L -1 HCl were cir culated.In these ex per i ments, the re agent col umn and in jection valve in Fig. 2 were re moved from the as sem bly, and a 1,10-phenanthroline so lu tion of vari able con cen tra tion was cir cu lated through chan nel d.
There were two ex per i men tal batches: one with out a cat a lyst and one with a cat a lyst (us ing the same as sem bly but in clud ing Cu(II) con cen tra tions be tween 0 and 3.11 mg L -1 in the fer ric so lu tion).These ex per i ments al lowed us to draw the fol low ing con clu sions: ( a) the absorbance decreased mark edly with de creas ing phen con cen tra tion; ( b) the more con cen trated (0.1, 0.5%) ligand so lu tions gave much greater slopes in the absorbance vs. time plots; ( c) the absorbance in creased sharply with an in crease in the concen tra tion, both in the pres ence and ab sence of a cat a lyst (absorbance val ues in the lat ter case were quite high anyway); ( d) a Cu(II) con cen tra tion of 3.11 mg L -1 quan ti tatively con verted Fe(III) to Fe(II) ion at time zero; ( e) un der these con di tions, a Cu(II) con cen tra tion as low as 6 µg L -1 could be de tected.
The in flu ence of other ex per i men tal vari ables was studied us ing the man i fold de picted in Fig. 3. Chan nel a was used with and with out an in ter ca lated solid phase re ac tor.In the ab sence of the re ac tor, the chan nel was used to cir culate Cu(II) at a vari able con cen tra tion; in its pres ence, var ious amino ac ids at dif fer ent con cen tra tions were passed.By com par ing the sig nals ob tained in both cases, the conver sion ef fi ciency of the re ac tor and the ex tent of the decom po si tion of the Cu(II)-amino acid com plex were si mul ta neously de ter mined.
The poor per for mance of the amino ac ids in the pre vious as sem bly was the likely re sult of the low quantitativeness of the re ac tion tak ing place in the solid phase re ac tor or of the Cu(II)-amino acid com plex be ing too sta ble to al low Cu(II) to act as a cat a lyst.In or der to ascer tain the rea son, the man i fold de picted in Fig. 4 was assem bled.The Cu(II)-glycine com plex was merged with the acidic Fe(III) stream to re lease Cu(II) by dis place ment of H + ion.Two types of tran sient sig nals were ob tained by circu lat ing a Cu(II) stream through chan nel a in the ab sence of the solid phase re ac tor, and an amino acid stream in the pres ence of the re ac tor.The re sults were sim i lar to the previ ous ones; how ever, they al lowed us to as cer tain that the fact that absorbances could not be higher was not the re sult of poor mix ing (in the pre vi ous as sem bly, three streams con verged on the same point), nor of a high sta bil ity of the Cu(II)-aminoacid com plex.Rather, the sen si tiv ity was limited by the quantitativeness of the re ac tion be tween the amino acid and the im mo bi lized cu pric salt.
In creased re ac tor lengths up to 10-15 cm (Fig. 4) resulted in in creased sig nals.The ef fect of lon ger lengths was off set by the in creased sam ple dis per sion.The peak absorbances ob tained at re ac tor lengths of 10 and 15 cm were 0.138 and 0.139, re spec tively.
Placing the solid phase re ac tor in the loop of the in jection valve (Fig. 5) was in di cated in prin ci ple, since it resulted in no sam ple dis per sion dur ing the re ac tion.In ad di tion, the time the so lu tion and solid bed re mained in con tact was con trolled by switch ing the valve on and off with out the need to stop the flow, which would have required au to matic con trol and de tracted from the reproducibility, ow ing to the typ i cal flow-rate os cil la tions ob served within the first few mo ments af ter a peri stal tic pump is started.Nev er the less, the many chan nels in cluded in this man i fold caused prob lems that are usu ally ab sent from sim pler con fig u ra tions.Thus, switch ing the valve on and off caused peak dis tor tions aris ing from changes in the hy dro static pres sure; the ef fect var ied with col umn length, which could not be in creased at will.The peak for the blank stan dard could not be min i mized.Also, con di tion ing af ter the analyte con cen tra tion was changed took a fairly long time since the sys tem was not pe ri od i cally flushed with carrier.
Placing the solid phase re ac tor be tween the in jec tion valve and the merg ing point with fer ric so lu tion re sulted in a de crease in the blank tran sient sig nal; the con tin u ous flow of the car rier so lu tion through the re ac tor avoids the con dition ing or wash ing steps re quired in the other man i folds.This man i fold was tested by pre par ing and in ject ing dif ferent con cen tra tions of glycine (up to 8 µg mL -1 ; five rep licates for each point), and re sulted in a wider lin ear graph.This as sem bly was se lected for fur ther work.
The se lected man i fold (Fig. 1) was op ti mized in terms of all of the in flu en tial chem i cal and FIA vari ables involved in the pro cess us ing the univariate method.The results are sum ma rized in Ta ble 2. The op ti mi za tion se quence was the same as de scribed in the ta ble, and the exper i ments in volved in ject ing 280 µL of a 5 µg mL -1 glycine so lu tion.
The se lec tion of the op ti mum value that could be obtained for each tested pa ram e ter was the one which gave the best com pro mise of peak height-peak base width-rsd (%).
Fig ure 4. The flow-injection as sem blies from pre vi ous stud ies on the in flu ence of chem i cal and FIA pa ram e ters.
Fig ure 3. The flow-injection as sem blies from pre vi ous stud ies on the in flu ence of chem i cal and FIA pa ram e ters.Fig ure 5.The flow-injection as sem blies from pre vi ous stud ies on the in flu ence of chem i cal and FIA pa ram e ters.
In op ti miz ing the pH, tran sient sig nals were cor rected for the base line at pH 11.5, where no sig nal was ob served.Pos si bly, the lack of a re sponse at pH 11.5 was due to the in situ pre cip i ta tion of the re leased Cu(II).The op ti mal pH, 9.5, was that re sult ing in the best com pro mise be tween the max i mum absorbance and the base line at pH 8.5.
The Fe(III) con cen tra tion re sulted in a crit i cal vari able, and then it was re-optimized with the se lected FIA as sembly at three dif fer ent glycine con cen tra tions: 0.5, 1.0, and 2.0 µg mL -1 .The Fe(III) con cen tra tion range was from 2.0 x 10 -4 to 10.0 x 10 -4 mol L -1 .The op ti mum value, 7.0 x 10 -4 mol L -1 , was also rel a tive to the base line since an in crease in the Fe(III) con cen tra tion also ac cel er ated the uncatalysed re ac tion, thereby in creas ing the blank sig nal.

An a lyt i cal fig ures of merit
The cal i bra tion graph was lin ear from 0.1 to 3.0 µg mL -1 glycine, and fit ted the equa tion, A = -0.0093+ 0.3380 C, where A de notes the absorbance and C the glycine concen tra tion, in mi cro grams per mil li li ter.The cor re la tion coef fi cient was 0.9957.We ran cal i bra tion graphs for other a-amino ac ids, the equa tions for which are shown in Ta ble 3. A com par i son of the slopes of the graphs with those for pre vi ously re ported meth ods 15,16 re veals that the pro posed sys tem is 20 times more sen si tive in one case and 7 times more in the other; hence, the ap pli ca tion range is nec es sarily short in a spec tro pho to met ric tech nique with a nar row ap pli ca bil ity "win dow" such as this.
The day-to-day reproducibility was de ter mined by running five cal i bra tion graphs for glycine on dif fer ent days.The av er age of the slopes ob tained was 0.3380, and the rela tive stan dard de vi a tion (RSD) was 4.36%.The re peat ability of the pro cess was cal cu lated us ing 16 rep li cates of an in jec tion of 1 µg mL -1 glycine.The mean peak height and RSD thus ob tained were 0.3719 AU and 2.3%, re spectively.In the same ex per i ment, the through put was es timated to be 28 sam ples h -1 .
The in flu ence of for eign sub stances, some of them accom pa ny ing glycine in phar ma ceu ti cal for mu la tions, was stud ied by pre par ing so lu tions con tain ing 2.46 µg mL -1 of the amino acid and dif fer ent con cen tra tions of the po ten tial interferent.Re sults ob tained (rel a tive er ror in %) were as fol lows: 0.9 for acetyl ace tic acid at 5 µg mL -1 ; 3.8 for caffeine at 50 µg mL -1 , and 5.2 for NaCl at 50 µg mL -1 .High er rors were cal cu lated for ascor bic and cit ric acid at 5 µg mL -1 , but by pre par ing con cen tra tions sim i lar to or smaller than that of glycine (which is case of the phar ma ceu ti cal for mu la tions) the er rors were al ways smaller than 3%.
Finally, glycine was de ter mined two dif fer ent for mula tions: Okal (tab lets, con tents ac cord ing to la bel claim: acethyl sal i cylic acid 500 mg, glycine 100 mg and caf feine 30 mg) and Actilevol (am poules).A few Okal tab lets were ground in an ag ate mor tar and an aliquot of ca.0.35 g of the re sult ing pow der was ac cu rately weighed and dis solved in dis tilled wa ter.The re sult ing mix ture was fil tered and brought to 500 mL with dis tilled ter; ten mil li li ters of the re sult ing so lu tion were lev eled to 100 mL; aliquots of 5 mL were ad justed to pH 3 with 0.1 mol L -1 HCl and the acetyl sal i cylic acid was re moved by ex trac tion with di ethyl ether.The sec ond for mu la tion, Actilevol (am poules, com po sition ac cord ing to la bel claim: ascor bic acid 500 mg, hemathoporphyrin 6 mg, glycine 500 mg, con cen trated live ex tract 20 mg, id yeast 100 mg, so dium ci trate 300 mg, etha nol 0.05 g) was pro cessed by di lut ing an am poule to 10 mL with dis tilled wa ter.This stock so lu was used to pre pare sev eral di lu tions from 10% to 20% (w/v).The re -Ta ble 3. Cal i bra tion equa tions with the op ti mal flow-injection assem bly for dif fer ent a-amino ac ids.

Amino Acid
Equa tion (c in µg mL sults ob tained were quite con sis tent with the cer ti fied values, namely: 102.5 mg/tab let for Okal (cer ti fied con tent 100 mg/tab let, rel a tive er ror 2.5%) and 507.9 mg/am poule (cer ti fied con tent 500 mg/am poule, rel a tive er ror 1.6%).The re sults ob tained were com pared with those ob tained with a ref er ence method 16 , in which the de ter mi na tion of glycine was car ried out by AAS us ing the same solid-phase re ac tor with out a der i va tion re ac tion.The re sults ob tained were: Okal (99.85 mg/tab let) and Actilevol (497.1 mg/ampoul).

Con clu sions
The pro posed FIA con fig u ra tion uses a solid phase reac tor that pro vides a di lute, re pro duc ible stream of con stant con cen tra tion that acts as a cat a lyst in a sub se quent re action.The prod uct of such a re ac tion is used to in di rectly obtain the con cen tra tion of the spe cies re leased by the metal ion act ing as the cat a lyst.
The pro posed method al lows mon i tor ing pro ce dures for or ganic com pounds based on atomic ab sorp tion measure ments (of the re leased metal ion) to be con verted into mo lec u lar ab sorp tion pro ce dures, which are much more com mon place in phar ma ceu ti cal anal y sis.In ad di tion, the use of a cat a lytic pro ce dure is a pos si ble way to im prove the de tec tion lim its and sen si tiv ity of the method.
Ta ble 1.The ef fects of pharmaceuticals on the re lease of Cu(II) from the re ac tor by dis place ment re ac tion.