Simultaneous determination of sucrose and reducing sugars using indirect flow-injection biamperometry

239

Anulytrca Chuntca Acta, 271(1993) 239-246 Elsevler Science Publishers B V , Amsterdam

Simultaneous determination of sucrose and reducing sugars using indirect flow-injection biamperometry Jacek Mlchalowslu and Anatol KOJ~ Instrtute of Chemrstry, Warsaw Umversrty Branch, Bzalystok (Poland)

Marek TroJanowlcz and Bogdan Szostek Department of Chemutry, Unrversrty of Warsaw, Pasteura I, 02-093 Warsaw (Poland)

Ehas A G Zagatto Centre for Nuclear Energy and Agnculture, Unrversu’yof Sio Paula, Rraclcaba (Bra&) (Recewed 1st Aprd 1992, reused manuscript received 17th July 1992)

The method developed for the determmatlon of sugars 1s based on reduction of hexacyanoferrate(II1) by simple carbohydrates m a strongly alkaline medium at elevated temperature Changes m hexacyanoferrate(II1) concentration are momtored amperometrlcally Hrlthplatinum wire electrodes polarized at 200 mV Sucrose determmatlon requires on-line hydrolysis m hydrochlonc acid The detection sensltivlty for dtierent simple sugars and dlsacchandes and the influence of several mterferents were examined Sucrose and glucose were determined with a sampling rate of 40 h-’ m natural samples from a sugar production process spiked with glucose Keywords Amperometry, Plow mJectlon, Glucose, Sucrose, Sugars, Reducmg sugars

Numerous flow-mJectlon methods have already been reported for the determmatlon of sugars m foods, blologlcal materials and physlologlcal fluids [l-16] Although m routme analysis, especially for the simultaneous determination of several species, liquid chromatographlc methods predommate, m many instances those methods can be replaced by fast flow-mJectlon procedures for the determmatlon of one or two components In reported flow-mJectlon procedures for the determmatlon of sugars, electrochermcal detecCorrespondence to M TroJanowuz, Department of Chemistry, Umversity of Warsaw, Pasteura 1, 02-093 Warsaw (Poland) 0003-2670/93/$06

tlon methods are most frequently used, although owing to the relatively low electrochemical actlvity of sugars, they are mostly determined mdlrectly In the potentlometrlc determmatlon of simple sugars platinum [ll and metallic copper [2] electrodes have been used Indlrect amperometnc detection of snnple carbohydrates at platmum electrodes m alkaline solutions has been performed by apphcatlon of a triple--pulse potential waveform [3], although indirect amperometnc detection has also been reported [l] There has been rapid progress recently m enzymatic flow-mJectlon methods for the determrnation of sunple sugars and dlsaccharldes with amperometrlc detection For this purpose either

00 0 1993 - Elsevler Science Pubhshers B V All nghts reserved

240

oxldases [4,5] or dehydrogenases [6,7] of sunple sugars are utilized In systems wth regeneration and recychng of coenzyme and enzyme, a repetltlve enzymatic determmatlon can be carried out [8] In multi-enzyme systems the determmatmn of sucrose m the presence of glucose [9] and the simultaneous determmatlon of sucrose and glucose [lo] have been performed Spectrophotometnc [ll-141 and chemllummescence [15,16] detection have also been employed m the flow-mJectlon determmatron of sugars The aim of this study was to develop a flow-mJectlon method for the determmatlon of sugars usmg blamperometrlc detection wth two polarized electrodes using a hexacyanoferrate(III)hexacyanoferrate(I1) mdlcatmg system In prevl01.1sapphcatlons of flow-mJectlon blamperometry mostly an I,-1 - mdlcatmg system has been employed 117-191 The reduction of hexacyanoferrate(II1) by carbohydrate aldehyde groups has already been apphed to flow-mJectlon determmatlons urlth potentlometrlc [1,20,21], amperometrlc with one polarized workmg electrode [ll and spectrophotometric [14] detection In this study, a umvarlate optlmlzatlon of the manifold and operating conditions was performed for blamperometrlc detection of sugars with hexacyanoferrate(III)-hexacyanoferrate(I1) mdlcatmg system

EXPERIMENTAL

Apparatus

The measuring system consisted of a multichannel pump (MS Reglo, Ismatec, Zurich), a laboratory-made rotary mJectlon valve and a flow cell connected to an a c -d c polarograph (PLP 225C, Zalmed, Warsaw), which was m turn connected to a potentlometnc strrp-chart recorder (TZ 4620, Laboratomi Piistrqe, Prague) The flow-through cell used was the same as described previously [17], the platmum wire electrodes were 0 6 mm m diameter and 13 mm long A constant polarmng voltage of 200 mV was applied from the polarograph The system used is shown m Fig 1 The mamfold was made of polypropylene tubmg (0 7 mm

J MIchafowskr et aL /And Chun Acta 271(1993) 239-246

NaOH

35

SM

-

ml/mm

Fig 1 Schematic diagram of the optmuzed flowqection system used for the simultaneous determmation of sucrose and reducmg sugars with bmmperometrx detection D - flowthrough amperometnc detector, DB = debubbler, V,, V, = mlectlon valves

1d ) with laboratory-made Perspex connectors The heating cods, made of PTFE tubmg (0 7 mm 1d , wall thickness 0 2 mm), were nnmersed m boiling water under reflux, for this purpose a heatmg mantle with a l-l round-bottomed flask and a 50-cm reflux condenser were used The standard Techmcon fitting BO from an AAII AutoAnalyzer was used as a debubbler The mamfold 1s equipped with two 1nJectlon valves, V, and V, The sample solution mJected wth valve V, reacts with hexacyanoferrate(II1) and after coolmg m a 100~11 delay loop and debubbhng m DB it is transported to the detector When mJected with valve V,, first the sample compounds are hydrolysed m a 250~cm toll and then react wth hexacyanoferrate(III), and the resulting hexacyanoferrate(I1) 1s detected Polarunetnc reference measurements were carried out using a Polamet-S polarnneter (Carl Zeiss, Jena) Reagents D-C+ )-Glucose was obtained from Polfa (Krak6w) and ~-(-)-fructose from Ublchem (Easthght) All other reagents used were of analytical-reagent grade from POCh (Gliwxe) Stock standard solutions of normal and invert sucrose, D-C - I-fructose and D-C + )-glucose were prepared as described prewously [13] and stored m a refrigerator All solutions were prepared m trlply distilled water Workmg standard solutions m the 50-300 mg 1-l mvert sucrose range were

J Mchabwsk~ et al /Anal

241

Chm Acta 271 (1993) 239-246

of the flow-through cell, the area of the mdlcatmg electrodes and the magnitude of the polarlzmg voltage apphed The choice of the geometry of the flow cell and the area of the mdlcator electrodes was based on previous work [17-191, whereas other factors were optlmlzed during this study The effect of temperature on hexacyanoferrate0111 reduction was studied usmg the system shown m Fig 1, but without the HCI lme and mJectlon valve V, used for sucrose hydrolysis Samples of 150 ~1 of 100 mg l- ’ glucose solutions were injected mto the stream of water pumped at a flow-rate of 5 5 ml mm-’ and merged with streams of 1% hexacyanoferrate(II1) (4 3 ml mm-l) and 2 M NaOH (3 2 ml mm-‘) A slgmflcant increase m slgnal magnitude with increase m temperature was observed over the whole range of temperature exammed (Fig 2) In the flow system without the debubbler an increase m temperature also had a strong adverse effect on the precision of measurements The use of the debubbler improved the preclslon 5-6-fold, whereas there was only a slight increase m the total daperslon in the system (about 3%) At 100°C the relative standard devlatlon (R S D > of measurements with multiple (n = 5) mJectlons of 100 mg I-’ glucose solution was 0 5-O 6% This temperature was found to be the optimum, consistent with previous observations [14], and was used m subsequent studies The determmatlon of sucrose with blamperometric detection requires its prehmmary hydrolysis to simple carbohydrates Accordmg to prevlous fmdmgs [141, this was done m hydrochloric

freshly prepared by dllutton of the stock standard solution with water White-beet Juices and syrups from various stages of sugar production were provided by the IAPY sugar factory

try

Procedure for determnatlon of sucrose and glucose m natural samples Determmatlons were made using the flow system shown m Fig 1, where two 150-~1 portions of sample were inJected with valves V, and V, with a tune delay of 30 s Cahbratlon graphs were obtamed for both parts of the mamfold with and without a hydrolysis step based on mJectlons of 100, 150, ZOO,250 and 300 pg ml - ’ invert sugar solutions Fresh Juice and thm syrup after flltratlon were diluted lOOO-fold and thick syrup 4000-fold with dlstllled water and then InJected mto the system Results obtained m the part of the manifold mvolvmg hydrolysis were subtracted from those obtained without hydrolysis

RESULTS AND DISCUSSION

Optrnuzatwn of the flow-vqectlon system The sensltlvlty of detection for different simple sugars usmg the proposed system with indirect blamperometrlc detectlon depends substantially on the reactlon tune of hexacyanoferrate(II1) reduction, temperature and the concentrations of reagents used Other essential contrlbutlons to the detectlon sensltmty come from the condltlons of blamperometrlc detection such as the geome-

TABLE 1 Optmuzatlon of experrmental concbons Length of reactton cod (cm)

for on-hne hydrolysis of sucrose m the flow-mJecuon system a

Peak height for different HCl concentrations (PAI OSM Sucrose

Invert

Sucrose

Invert

51 53 52

55 53 52

sucrose

200 250 300 a Injected

52 53 54

Sampling rate (h - ‘1

10M

56 55 54

sucrose

100 mg 1-l sucrose or Invert sucrose wtth valve V, (Fig 1)

45 40 33

242

J Mchabwskr et al /Anal Chrm Acta 271 (1993) 239-246

1600 1

OOO,..~.,..,.,.,,.......,.....,~~~ 45 00

65 00

65 00

Temperature,

105 00

‘C

Fig 2 Effect of temperature on flow-mJection response for the mjection of 150 ~1 of 100 mg I’-’ glucose solution m the system with biamperometnc detection See text for details

aad solution Table 1 shows the influence of the length of the reaction cod and the HCl concentration on the efflcleficy of hydrolysis m the flow system shown m Fig 1 In this study the results obtained under different condltlons for sucrose and invert sucrose solutions of the same concen-

tration inJected wth valve V, were compared A 250-cm long reaction cod and a 1 M concentration of hydrochloric acid for hydrolysis at 100°C were found to be the optunum conditions Although under these conditions a slightly smaller signal was obtained than wrth a 3OO-cm reaction cod, the sampling rate was higher With higher HCl concentrations a decrease m the signal magnitude was observed, and the nsk of suger caramehzatlon must also be taken mto account The same measuring set-up was also employed for the optlmlzatlon of the concentration of hexacyanoferrate(II1) and sodium hydroxide solutions for the detection of simple sugars The system with one mJectlon valve and sphttmg the sample prior to the reaction cods was replaced with the system with two independent mJectlon valves because of unfavourable reproduclbhty of results The optlmlzatlon was done usmg 150-4 lqectlons of 100 mg 1-l glucose and fructose standard solutions injected v&h both valves V, and V, The results obtained are given m Table 2 The first peak corresponds to the mjectlon with valve V, and the second peak, which is smaller owing

TABLE 2 Effect of concentration of reagents on flow-mJection response obtamed for mJections of 100 mg I-’ glucose and fructose m the system with two lwection valves shown m Fig 1 NaOH

concentration (MI 3

5

IWe( concentration (%I 05 1 2 3 4 5 05 1 2 3 4 5 05 1 2 3 4 5

Peak height, mm Fructose

Glucose 1st peak

2nd peak

1st peak

2nd peak

112 110 112 115 111 115 120 125 123 13 5 128 135 13 9 140 145 140 142 142

39 46 49 52 48 47 47 50 51 55 58 64 42 42 43 42 43 45

13 0 125 129 13 0 125 125 13 1 13 5 13 3 14 0 13 0 13 5 12 3 124 126 124 124 12 6

65 57 59 62 59 56 54 54 54 60 60 65 50 51 50 52 52 55

J hhchaiowskl et al /And

243

Chm Acta 271 (1993) 239-246 TABLE 3

Relative sensltmty of the flow-mJectton response obtained m the optnmzed measurmg system urlth blamperometrx detection for 100 mg I-’ solutions of different carbohydrates Injected species

50

0 50

100

150

200

invert sucrose,

250

300

350

Relative sensltlvlty Blamperometnc detection

Spectrophotometrlc detection [141

Glucose

loo0

1000

Fructose Xylose Rlbose Galactose Sucrose Maltose Lactose

1007 1166 0890 0903 0097 0 421 0 655

1004 1052 0 892 0 879 1035 a 0398a 0675’

ppm

Fig 3 Cahbratlon graphs obtamed for Invert sucrose at polarlzmg voltages of (1) 50, (2) 100, (3) 150 and (4) 200 mV Injected 150 ~1 of Invert sucrose solution with valve Vl

to the larger dlsperslon of the injected sample solution, corresponds to the mJectlon with valve V, The optnnum conditions were considered to be those for which signals correspondmg to the injections of glucose or fructose solution were identical with inJection with either valve V, or V, This was observed with 5% hexacyanoferrate(II1) and 4 M NaOH at flow-rates of 2 5 and 3 5 ml mm-l, respectively For the optnmzed manifold and dlmenslons of the mdlcatmg electrodes mentioned under Experimental, the effect of the polarizing voltage m the range 50-200 mV on the signal magnitude observed for invert sucrose concentrations ranging from 25 to 300 mg 1-l was examined (Frg 3) The change m the polarlzmg voltage affects both the signal magnitude and the extent of the linear response range The optnnum polarizing voltage was taken as 200 mV, at which a linear response was observed m the range 100-300 mg 1-l invert sucrose injected with valve V, and for 100-250 mg 1-l invert sucrose injected with valve V, With such a system the R S D obtained for multiple (n = 5) injections of sample containing 10 5% (w/w> sucrose and 17% (w/w) glucose was 0 63% for inJection with valve V, and 0 77% for mJectlon with valve V, The sensltlvlty of detection for different sugars was also examined under the same condltlons

a After hydrolysis

(Table 3) All results reported m Table 3 from this study were obtamed for mjectlon into the manifold without a hydrolysis step Those results which were obtained under similar condltlons were virtually identical with those observed prevlously for flow-mqectlon analysis with spectrophotometrlc detection [14] Snnple carbohydrates other than glucose and fructose exhibit some differences m signal magnitude up to 16% sensitlvlty for glucose, but their content m the natural samples examined 1s negligible Snmlarly to the previous studies, disaccharides showed also a response, the smallest signal was observed for sucrose

60

0’

100

150

200

250

Invert sucrose, ppm Fig 4 Cahbratlon graphs obtained for 150 ~1 of invert sucrose solutions mected with valve (A) VI and (B) Vz m the flow-mJectlon system Hrlthblamperometnc detection at a polaming voltage of 200 mV

244

J Mchabwskl el al /Am!

TABLE 4

equlhbruun of the mdlcatmg redox couple Therefore, the effect of the presence of several transItIon and alkaline earth metal ions was mvestlgated (Table 4) In spite of the different type of detectlon applied m this study m comparison with that m previous papers, the tolerance level of metal cations examined was slmllar to that one observed with spectrophotometrlc detection [ 141 Because of the reductive effect also found for sucrose (Table 31, the Influence of sucrose on glucose determmatlon was exammed Solutions of 100 mg 1-l glucose were spiked with different amounts of sucrose and injected with valve V, A positive error of about 2% was observed with a 170 mg 1-l sucrose level in the injected glucose solutions This level, however, 1s much larger than observed m the real samples reported on below

Tolerance levels of the presence of metal IOIIScausing an error not exceedmg 2% m flowqecfion determmatlon of 100 mg 1-l glucose usmg mdlrect blamperometrlc detectlon Metal

Tolerance level (mg 1-l)

Co, NI, Mt.1011 Fe(H) Fe(III), Cu, Ca, Mg

12 15 15

Chm Acra 271 (1993) 239-246

Interferences Interference with the proposed mduect blamperometrlc detectlon with the hexacyanoferrate(III)-hexacyanoferrate(I1) mdlcatmg system can be produced by the presence of species that bmd any of the reagents used or Influence the

TABLE 5 Results of biamperometnc flow-mJectlon determmatlon of sucrose and glucose m natural samples from sugar productlon spiked with known amounts of glucose Sample

Fresh white-beet JUlCe

=

Thm syrup a

Thick syrup b

Glucose (%, w/w)

Sucrose (%, w/w) Found by polanmetIy

Found by FIA

Error (a)

r2

Added

Found by FIA

Error (o/o)

12

14 1 13 9 145 125 14 1 139 145 125 13 5 138 13 6 112 135 13 8 13 6 112 57 7 610 642 520 57 7 610 642 520

140 140 146 13 0 143 13 7 14 8 120 13 9 13 9 13 1 111 13 4 13 4 13 1 109 57 0 58 3 63 9 52 9 539 62 1 648 519

-07 +07 +07 +40 +14 -14 +21 -40 +30 +07 -37 -09 -07 -29 -37 -27 -12 -44 -05 +17 -66 +18 +09 -02

0 958

100 100 100 100 15 0 15 0 15 0 15 0 100 100 10 0 100 130 13 0 13 0 13 0 400 40 0 40 0 40 0 60 0 60 0 60 0 600

105 98 99 102 145 15 1 15 6 15 0 107 97 99 99 135 13 1 128 129 420 410 39 1 400 62 1 598 592 61 1

+50 -20 -10 +20 -33 +07 +40 0 +70 -30 -10 -10 +38 +08 -15 -08 +50 +25 -23 0 -03 -03 -13 +18

0999

0 993

0 943

’ Dduted 1 1000 prior to mjectlon b Dduted 1 4000 pnor to mjection

0999

0 999

J Mtchalowskr

et al /Anal

245

Cham Acta 271 (1993) 239-246

correlation coefficients showed satisfactory agreement In order to obtain a baselme separation of the two peaks resulting from 1nJectlons with valves V, and V,, mectlon with valve V, was made 30 s after inJection wth valve V,, and then after a further 60 s the next solution was injected with valve V, Such a procedure leads to a sampling frequency 40 h-’

6 In,” -

Fig 5 Flow-InJectIon response recorded for multiple mjectlons of five thick syrup samples with added known amounts of glucose I= 12 4% sucrose, 18% (w/w) glucose,2 = 172% sucrose, 9 8% glucose, 3 = 15 4% sucrose, 8 7% glucose, 4 = 14 6% sucrose, 3 7% glucose, 5 = 118% sucrose, 3 2% glucose

Determmatlon sample3 The method

of sucrose and glucose m natural

was applied to the analysis of fresh white-beet Juice samples and syrups from different stages of sugar productlon Because of the very low natural content of reducing sugars m these samples (m contrast to snmlar samples from sugar-cane processmg 11411,they were spiked with known amounts of glucose Cahbratlon of the measurmg system was done using the same 150-111volumes of the invert sucrose solutions injected with both valves The analytical signal obtamed for mJectlon wrth valve V, corresponds to the content of reducmg sugars m the mjected standard or sample solution (first peak) The signal correspondmg to mjectlon wrth valve V, 1s related to the total content of reducmg sugars present m the raw sample and formed from the hydrolysis of sucrose Figure 4 shows the cahbratlon graphs obtained and Fig 5 presents an example of the flow-mJectlon response recorded for multiple mJectlons of several real samples The results obtained were correlated Hrlth sucrose determmatlon by polarnnetry and v&h the known amounts of added glucose (Table 5) The

c0nc1uswns The flow-mJectlon blamperometrlc method developed here 1s an alternative for the rapid, slmultaneous determmatlon of sugars to the spectrophotometrlc procedure reported previously [14] The chemical conditions are similar The replacement of optical detection wth snnple amperometrlc detection with two polarized metallic electrodes should be advantageous m routme apphcatlons m the sugar industry when control of sucrose together with reducing sugars is needed From an instrumental point of view, the signal magnitude 1s m a convenient range of nucroamps, and the detector cell with two platinum wire electrodes does not require frequent mamtenance

REFERENCES 1 K. Brunt, Analyst, 107 (1982) 1261 2 P W Alexander, P R Haddad and M Trojanomcz, Anal L&t, 18 (1985) 1953 3 S Hughes and DC Johnson, Anal Chun Acta, 132 (1981) 11 4 H Lundback and B Olsson, Anal Lett , 18 (1985) 871 5 W Matuszewslu, M Trojanowlcz and L Ilcheva, Electroanalysis, 2 (1990) 147 6 G Marko-Varga, R Appleqvlst and L Gorton Anal Chnn Acta, 179 (1986) 371 7 K. Matsumoto, 0 Hamada, H Ukeda and Y OsaJlma, Anal Chem , 58 (1986) 2732 8 P Roehrmg, CM Wolff and J P Schwmg, Anal Chun Acta, 153 (1983) 181 9 B Olsson, B Stalbom and G Johansson, Anal Chum Acta, 179 (1986) 203 10 M Masoom and A Townshend, Anal Chum Acta, 171 (1985) 185 11 J Toe] and N Baba, Bunselu Kagaku, 35 (1985) 411 12 M J Medma, J Bartroh, J Alonso, M Blanco and J Fuentes, Anal Lett , B17 (1984) 385 13 EA G Zagatto, IL Mattos and A 0 Jacmtho, Anal Chum Acta, 204 (1988) 259

246 14 IL Mattos, EA G Zagatto and A0 Jacmtho, Anal Chum Acta, 214 (1988) 247 15 CA Koerner and TA Nleman, Anal Chem , 58 (1986) 116 16 M Maeda and A TSUJI,Anal Sa ,2 (1986) 183 17 M Trqanowlcz, A Hulamckl, W Matuszewslu, M Palys, A Fukslewlcz, T Hulamcka-Mlchalak, S Raszewskl, J Szyller and W Augustymak, Anal Chum Acta, 188 (1986) 165

J Mchabwskr et al /Anal Chun Acta 271 (1993) 239-246 18 W Matuszewslu, A Hulamclu and M Trojanowicz, Anal Chum Acta, 194 (1987) 269 19 M TroJanowxz, W Matuszewskl, B Szostek and J Mlchalowskl, Anal Chum Acta, 261 (1992) 391 20 H Ohura, T Imato, Y Asamo, S Yamasalu and N Ishlbashl, Bunselu Kagaku, 35 (1986) 807 21 H Ohura, T Imato, S Yamasakl and N Ishlbashl, Anal SCI , 3 (1987) 453