0003-2670_92_ - Continuous flow method for the determination of by dfsdf224s


									Analytza Chlmrca Acta, 258 (1992) 253-258                                                                                   253
Elsevler Science Pubhshers B V , Amsterdam

            Continuous flow method for the determination
                       of aromatic aldehydes
                         Mayda Lopez-Nleves I, Peter D Wentzell * and S R Crouch *
                       Department of Chemrstry, Mtchrgan State Unrverslty,East Lansmg, MI 48824 (USA)
                                                 (Recemed 21st August 1991)


          The reactIon of o-d~amsxhne and aromatic aldehydes was adapted to an air-segmented contmuous-flow analysis
      system Stanmc chlonde was used as the catalyst Figures of ment of the automated method (e    g,sensltmlty, hmlts of
      detection, and accuracy) are reported for 13 aromatIc aldehydes The results of selectwlty and Interference studies
      are also reported
      Keywords Flow system, Aldehydes, Aromatx aldehydes

    Many colonmetrlc and fluorlmetrlc methods                     the development of the analytlcal method Re-
have been developed for the determmatlon of                       sults of these studies are reported elsewhere [6,7]
aromatlc aldehydes [1,2] Unfortunately, mterfer-
ences are common with these methods, especially
from ketones, and most are difficult to automate                  EXPERIMENTAL
because they require such condrtlons as concen-
trated acid mixtures, high temperatures or long                       Apparatus
reactlon tunes                                                       A schematic diagram of the modular ASCF
    One reactlon that has not been extensively                    system employed m this work 1s shown m Fig 1
used for the determmatlon of aromatic aldehydes                   The signal from the detector 1s directed to a
1s the Schlff base formation reaction with o-dl-                  “bubble gate” clrcmt which allows measurements
amsldme [3] In this paper, the analytical reaction                of absorbance only when the cell 1s filled wth
between aromatlc aldehydes and o-dlamndme 1s                      hqmd This electronically removes the effect of
adapted to an ax-segmented contmuous-flow                         au- bubbles without physical debubbhng [8] The
(ASCF) system [4,5] In the method described,                      detector and the bubble gate were constructed
stanmc chloride IS used as a catalyst and anhy-                   locally 19,101 A Model IP-12 vanable-speed, 12-
drous ethanol 1s the solvent The kmetlcs and                      channel perlstaltlc pump (Brmkmann Instru-
mechamsm of this reaction were studied prior to                   ments, Westbury, NY) was modlfled to reduce
                                                                  pump pulsations by replacing the standard roller
                                                                  assembly (8 rollers) wth one containing 16 rollers
                                                                  Wavelength selection (380 nm> was accomplished
’ Present address Umverslty of Puerto F&o, Aguaddla Re-           wth a 1 27 cm diameter, narrow bandwidth (= 8
  glonal College, P 0 Box 160, Ramey Station, Aguadllla, PR
  00604, USA                                                      nm), three-cavity, interference filter (No 15-
’ Present address Dept of Chemntry, Dalhoune Umversity,           10060, Dltrx Optics, Hudson, MA) A flowcell
  Hahfax, NS, B3H 453, Canada                                     with sapphire windows, a 10 cm pathlength and

0003-2670/92/$05 00 0 1992 - Elsevler Science Pubhshers B V Ail rights reserved
254                                                                                            M LOPEZ-NIEVESETAL

an internal diameter of 0 05 cm was obtained               tube m the particular solution (1 e , “peckmg”)
commercially (PN 178-13724-02, Techmcon In-                An segmentation was carried out by the dual
struments, Tarrytown, NY)                                  pump tube method described by Hablg et al [ll]
   All manifold cods were constructed from 0 10               The filter photometer was operated m a sm-
cm 1 d X 0 17 cm o d PTFE tubing (No N-06417-              gle-beam mode because of relatively low light
41, Cole Parmer, Chicago, IL) The reaction cod             intensities at 380 nm Baseline drift was about
was wrapped around a glass rod (125 cm dlame-              l-2% per h after a warm-up period of 1 h In this
ter) which was mounted m a water-tight Plext-              system, a slgnal of 5 V was obtained for a 100%
glass Jacket so that it could be thermostated by           transmittance reading
means of cn-culatmg water from a water bath
The sample and reagents were added via Plexr-
                                                              Samples and reagents
glass tee connections Fisherbrand solvent flexr-
                                                              Commercial abfolute ethanol was dried by
ble pump tubes (Fisher, Sprmgfleld, NJ) were
                                                           passing It over 3 A molecular sieves (Davidson,
used mstead of the standard polyethylene pump
                                                           Baltunore, MD) After decanting, it was filtered
tubes because the latter were not mert to the
                                                           through a 0 40-pm pore filter (M&pore, Bed-
ethanol solvent
                                                           ford, MA)
   A QUICK BASIC program (Microsoft, Bellevue,
                                                              The o-dlamsldme was recrystallized four to
WA) was developed to acquire, process and store
                                                           five times from ethanol, activated coconut char-
data with an IBM data acqulsltion board
                                                           coal (50-200 mesh) was used as a decolonzmg
(Mendelson Electronics, Dayton, OH) and an
                                                           agent (Fisher Scientific, Sprmgfield, NJ) White
IBM-compatible personal computer (Bentley T,
                                                           to off-white crystals were obtained with a net
Round Rock, TX)
                                                           yield of approximately 38% The o-dlamedme
                                                           crystals were vacuum dried and stored m a deac-
                                                           cator m the dark These crystals are stable for
  Procedure                                                more than SIXmonths if all solvent IS removed
    All solutions were filtered with Mdlex 0 5-pm          Solutions of o-dlanuldme were prepared fresh
filters (No SLSR 025 NS, M&pore, Bedford,                  dally
MA) prior to being asplrated mto the mamfold                  AI1 aldehydes and ketones were of commercial
Three mtersample air segments were mtroduced               orlgm Solid samples were recrystalhzed from
at the start of each sample and wash interval by           ethanol and then vacuum dried Melting points
repetltlve wlthdrawal and msertlon of the sample           were determmed m order to check Identity and

Fig 1 Modular au-segmented contmuous-flow system used m the determmatlon   of aromatic aldehydes   Stream flow path IS
denoted by -     and the electrical slgnal by - - -
DETERMINATION    OF AROMATIC   ALDEHYDES                                                            2.55

purity   Lqmd samples were dlstllled Workmg          the highest value that avoided preclpltatlon at
standards were prepared by dllutlon of stock so-     room temperature (0 100 g/20 ml ethanol =
lutlons usmg anhydrous ethanol as solvent            0 0204 M) The upper lnmt of the stanmc chlo-
    Furfural was determined m Creme de Menthe        ride concentration (100 x 10m2 M) was estab-
using the o-dlamsrdme method after steam dlstl-      lished by the decomposltlon observed at higher
latlon of the sample and an ether extractlon A       concentrations [6], whde the lower limit (3 00 x
25 OO-mlsample of Creme de Menthe was steam          10V3 M), was the mmunum concentration of
distilled and about 200 ml of the dlstlllate col-    stanmc chlonde needed to observe an apprecla-
lected All the collected dlstlllate was extracted    ble reaction The range of reactlon coil tempera-
with rune 20-ml portions of ethyl ether The ex-      tures for the optlmlzatlon was 20-50°C This
tractions were collected m a 200 O-ml flask and      avoids problems of mstabdlty m the system due to
diluted with additional ether Alter drymg            excesslve solvent vapor pressure at higher tem-
overnight with magnesium sulfate, the ethereal       peratures
solution was filtered by gravity usmg a coarse-          Prehmmary optlmlzatlon studies showed that
porosity fretted disc funnel Exactly 175 0 ml of     with ethanol as solvent a regular bubble pattern
the ethereal solution were concentrated m a ro-      was only attamable at pump speed settmgs corre-
tary evaporator The residue left was quantlta-       sponding to nommal flow-rates of 0 25-O 63 ml
twely transferred to a 50 O-ml volumetric flask      mm-’ (settings 28 to 70 on the Brmkman pump)
usmg anhydrous ethanol Sur samples were pre-         At these high pump speed settings, dlsperslon
pared by standard addition of 0 00, 0 300, 0 600,    was a problem when using a 2 50 m long manifold
100, 130 or 160 ml of a 114 x 10m3 M furfural        tube With a 2 00 m long manifold tube, less
solution to 8 00 ml of the above ethanohc solu-      dlsperslon was observed, but the response of the
tion m 10 O-ml volumetric flasks Each sample         system was about 10% lower To reduce dlsper-
was filtered through 0 5-pm Mdlex filters prror to   slon and noise further, the o-dlaneldme was
the furfural determination                           rmxed m a 3 2 ratio (v/v) mth the stanmc chlo-
                                                     ride solution to be tested This mtiure was then
                                                     dispensed by means of a single pump tube (0 23
RESULTS                                              ml mm-’ nominal flow-rate) For a 2 00 m long
                                                     reaction coil, the best experunental conditions
    Optwmation                                       found from the modified simplex method were a
    Snyder’s dlsperslon model [121was used m the     pump speed setting of 50, a temperature of 45”C,
design of the ASCF mamfold Overall dispersion        and a stanmc chlonde concentration of 7 00 X
was mmumsed by optmuzmg the followmg key             10m3 M Under these condltlons the flow-rate
variables mamfold tube internal diameter cd,),       excludmg air was 0 86 ml mm-‘, the residence
air segmentation frequency (n), and flow-rate (P)    tune of the sample from the point of asplratlon to
At 4O”C,the best dlmensrons for the reaction co11    detection was about 135 s, and the au-segmenta-
were calculated from the model as 0 10 cm I d ,      tion frequency was 2 2 s-l
and 2 5 m length for a reaction time of about 5          Another parameter that was studied, was that
mm                                                   of sampling tune At a feasible wash tune of 15 s,
    Next, an experunental approach to optlmlza-      sampling rates of 34, 48, 60, 80 and 120 samples
tlon was carried out using the modified simplex      per h can be achieved wrth high reproducibility of
method [13-H] The parameters chosen for opti-        the response slgnal at sampling tunes of 90, 60,
mlzatlon were the photometric response (I e , ab-    45, 30 and 15 s, respectively At a sampling tune
sorbance), flow system dispersion and system         of 15 s the response is only reduced to 93% of the
noise The three variables momtored were tem-         maxlmum obtained at sampling tune of 60-90 s
perature, flow-rate, and stanmc chlonde concen-          In the analytical method, a samplmg time of 90
tration                                              s was used to prepare cahbratlon curves m order
   The o-dlanlsidme concentration was fixed at       to assess the true sensltlvlty of the method. Dur-
256                                                                                                            M LOPEZ-NIEVES    ET AL

mg the selectlvlty and mterference studies, a sam-                 to dlstmgmsh an an bubble from the sample, and
pling tnne of 60 s was used to increase through-                   erratic trlggermg results The upper hnut of the
put The wash time was 30 s m both cases                            hnear dynanuc range for most aldehydes was
                                                                   taken as either the highest concentration that
   Method                                                          exhibited lmearlty or the highest concentration
   Once the ASCF system was optlmlzed, cahbra-                     tested that showed an absorbance of less than
tlon curves for 13 aromatic aldehydes were con-                    10
structed to assess the usefulness of the o-dlamsl-                    The precision and accuracy of the method for
dine method The figures of merit of o-dlamn-                       the determmatlon of each aldehyde m an Ideal
dme as a reagent for aromatic aldehyde detectlon                   sample (a sample of known concentration of the
are presented m Table 1 For each aldehyde,                         particular aldehyde) were obtamed by makmg
seven standards were prepared and four determl-                    eight absorbance measurements of the samples
nations were made for each standard                                which were chosen to have absorbances of about
   The cahbratlon sensltlvlty of the o-dlamn-                      0 50 Column seven m Table 1 reports the relative
dine-aldehyde reaction was calculated for each                     standard devlatlons of the eight absorbance mea-
aldehyde as the slope of a weighted linear regres-                 surements for each aldehyde, and column eight
sion of the absorbance at 380 nm versus molar                      reports the errors m concentrations found from
concentration [16] The reaction does not go to                     the cahbratlon curves relative to the known con-
completion Therefore, the sensltmlty reported is                   centrations
not equal to the molar absorptlvlty of the partlcu-
lar adduct formed The hmlt of detection for each                     Selectmty studres
aldehyde was calculated as the concentration of                      The selectlvlty of the method was tested by
the aldehyde that produced an absorbance equal                     reacting several ketones and ahphatlc aldehydes
to twrce the standard devlatlon of the baselme                     under the same condltlons used for the aromatic
    Due to constramts m the design of the bubble                   aldehyde determmatlons In general, ahphatlc and
gate, absorbances higher than 10 cannot be eas-                    aromatic ketones showed the least response to-
ily measured m this system At higher ab-                           wards o-dlanaldme The average sensltlvlty for
sorbances it becomes difficult for the electronics                 ketones was 7 1 1 mol-’ cm-‘, about 400 tnnes

TABLE    1
Figures of ment of the analytlcal method

Aldehyde                       Sensltwty   Error of    Intercept   DetectIon    Lmear                           RSD       Accuracy
                               (1 mol-’    the slope    (AU )      hmlt (M)     range (M)                       (%)       (So)
                               cm-‘)       (1 mol-l
p-Amsaldehyde                  1362        14            0009      63x10-’      63~10-~-6OxlO-~                 075       207
Benzaldehyde                   2024         9            0 002     78X10K6      7 8 x 10-6-4 3 x        10-4    0 71      0 14
5-Chlorosahcylaldehyde         5354        45            0011      12x10-”      12 x 10-6-l 2 x         10-4    0 45      088
Cmnamaldehyde                  4490        22            0010      17x10-6      17 x 10-6-2 0 x         10-4    0 41      240
2,4-Dahlorobenzaldehyde        3085         7            0 005     4 1 x 10-e   4 1 x 10-h-3 0 x        10-d    0 97      0 83
2,6-Dlchlorobenzaldehyde       1630        15          -0001       11 x 10KS    11 x 10-5-5 5 x         10-4    056       057
2,SDlmethoxybenzaldehyde       3740        15          -0009       70x lo+      7 0 x 10-6-2 3 x        10-4    072       56
   benzaldehyde                1359        13          -0013       20x 10-5     20   x   lo-‘-6    5x   1O-4    0   79    009
Furfural                       2820        10            0 002     57x10-6      57   x   10-6-3    2x   10-4    0   52    039
p-Hydroxybenzaldehyde           976         5          -0001       19 x 10-s    19   x   10-5-8    0x   1O-4    1   13    112
m-Nltrobenzaldehyde            2602         9          -0005       87x10+       87   x   lo-+-3    5x   1O-4    0   68    051
Sahcylaldehyde                 5916        22            0006      19XlOV       19   x   10-6-14    x   10-4    0   57    183
p-Tolualdehyde                 1870        10            0003      77x10-6      77   x   10-e-4    7x   10-4    0   94    144

less than for aromatic aldehydes Ahphatlc alde-        the solution density rather than to preapltate
hydes exhlbrt an average sensltlvlty of 15 6 I mol-’   formatlon
cm-‘, about 184 times less than aroma@ aldehy-            An aldehyde sample contaming 0 30% of BrlJ
des The response of cY$-unsaturated carbonyl           35 was also tested BrlJ 35 is a surfactant used in
compounds with the o-dlamsldme reagent 1s ap-          aqueous ASCF systems to Improve the repro-
preciable Consequently, a#-unsaturated          car-   duclblhty of the bubble pattern No nnprovement
bony1 compounds are possible interferences m           m the bubble pattern was observed m this system
any Intended apphcatlon of this method                 urlth ethanol as the solvent In turn, a 3 2%
                                                       decrease m the signal was observed

  Interference studEs                                    Apphcatlon to the determmatwn of furfurai m
   To mvestlgate further the usefulness of the         Creme ak Menthe
o-dlamsldme reagent for the determmatlon of              The o-dlamsldme ASCF method was evalu-
aromatic aldehydes, the effect of concormtants on      ated on the determination of furfural m Creme
the slgnal mtenslty was mvestlgated The alde-          de Menthe and compared to direct UV spec-
hyde chosen for the interference studies was fur-      trophotometry For both methods, the hquor was
fural m a 1 86 x 10e4 M concentration                  steam dlstllled to separate the furfural from other
   Water was found to be a major mterferant,           components present m the sample, and the stan-
causmg as much as a 78% decrease m the slgnal          dard addltlon technique was used to compensate
at concentrations of water as low as 8% The            for possible mterferences
presence of large amounts of water m strong acid           In the direct spectrophotometnc method the
medium 1s probably the reason that this reaction       absorbances were measured at a wavelength of
has not been successfully used m the past 117,181      277 nm on a Hltachl 200-Perkm-Elmer spec-
   Acetlc acid and benzolc acid at concentrations      trophotometer     (bandpass = 100 nm) In the
of 608 and 589 mg 1-l respectwely, were used to        ASCF determmatlon, the reactlon co11was ther-
test the mterference of carboxyhc acids m the          mostated at 45”C, samphng and wash times were
method Acetic acid caused a 3 0% mcrease m             60 and 30 s, respectively The stanmc chloride
the signal, while benzolc aad gave a 18% m-            and o-dlamadme reagents were prepared as de-
crease                                                 scribed previously and their concentrations were
   Tests have shown that the Interferences noted       set to the optimum values as determined from the
here from water and from carboxyllc acids can be       optimization expenments
overcome by using the standard addltron method             Prehmmary comparisons were somewhat dls-
m which aldehyde standards are added to the            couragmg m that the ASCF values for furfural
sample m small volumes so that the matrix re-          (obtamed after ether extractlon of the steam ds-
mains approxnnately constant in composition            tllled samples) were consistently lower than the
   A possible lmplementatlon of the o-dlamsl-          results obtained by UV spectrophotometry (ob-
dme method as an hquld chromatography post-            tamed after only steam dlstrllatlon) To dlscover
column detectlon reaction prompted us to test          the reasons for this discrepancy and to fmd which
the mterference of hexane m the method Fwe             procedure was glvmg accurate results, UV spec-
ethanohc samples of the aldehyde were prepared         trophotometry was done on three sets of samples
m which the hexane content was 0, 10, 20 and           one where standard addmons were made prior to
40% No net effect of hexane on the signal was          steam dlstllatlon, a second where standard addl-
observed, but, as the concentration of hexane was      tlons were made after steam dlstlllatlon, and a
increased, the baseline was observed to drift up       third m which standard addltlons were made af-
somewhat The lack of precipltatlon m the waste         ter an ether extraction of the steam dlstdled
contamer and the flat shape of the signals regard-     sample
less of hexane concentration are mdlcatlve that            The results of the determmatlons are summa-
the observed drift was probably due to changes m       razed m Table 2 The nearly ldentlcal furfural
258                                                                                             M LOPEZ NIEVES ET AL

TABLE 2                                                      derrvatlzatlon wth o-dlamadme Other reported
Determmatlon of furfural m Creme de Menthe                   colonmetrlc and fluorlmetnc methods [1,2] for
                                           mg furfural I-’
                                                             carbonyl groups are m general, also more sensl-
Standard addition            Detection
                             mode                            tlve than the o-dlamadme method The most
                             W             60*05
                                                             nnportant advantage, however, of the proposed
Before steam dlstdlatlon
After steam dlstdlatlon      UV            59*03             o-dlamsldme method IS the selectlvlty for aro-
After ether extractlon of                                    matlc aldehydes Such selectlvlty IS very lmpor-
  a steam dlstdled sample    W             20*05             tant m the analysis of foods and waste waters,
After ether extraction of                                    where trace amounts of ketones and/or ahphatlc
  a steam dIstIlled sample   o-Dlamsldme   20f 0 1
                                                             aldehydes could mterfere with direct spectropho-
                                                             tometrlc determmatlons or with colorlmetrlc or
                                                             fluorunetrtc methods
values obtained by UV spectrophotometry when
the addmons were made prior to and after steam
                                                                The authors gratefully acknowledge the fman-
dlstlllatlon mdlcate that the steam dlstlllatlon
                                                             clal support of the National Science Foundation
procedure achieves nearly 100% recovery of the
                                                             through NSF Grant No CHE 8705069
W absorbing species The ether extraction effi-
ciency was tested m separate recovery expen-
ments on Creme de Menthe samples spiked with
known amounts of furfural UV spectrophoto-
metric results indicated a nearly 100% ether ex-              1 E Sawlckt and CR Saw&, Aldehydes Photometnc
traction efflclency Also, the UV absorption spec-               Analysis, Vols l-5, Academic Press, London, 1975-78
trum of an ether extracted Creme de Menthe                    2 S Nakamura, M Toda, H Sate and Y Ohkura, Anal
sample more closely resembles the spectrum of a                 Chum Acta, 134 (1982) 39
                                                              3 J A Attaway, R W Wolford, M H Dougherty and G J
pure furfural sample than does a sample after
                                                                Edwards, J Agr~c Food Chem , 15 (1967) 688
only steam dlstlllatlon Hence, we conclude that               4 W B Furman (Ed ), Contmuous Flow Analysis Theory
the higher values obtamed by UV spectropho-                     and Practice, Marcel Dekker, New York, 1976
tometry prior to ether extraction are most hkely              5 L R Snyder, Anal Chun Acta, 114 (1980) 3
due to concomitants m the sample that also ab-                6 M Lopez-Nleves, P D Wentzell and S R Crouch, Anal
sorb at 277 nm The ether extraction effectively                 Chem ,62 (1990) 304
                                                              7 M Lopez-Nleves, Ph D Thesis, Mlchlgan State Umver-
separates the furfural from these mterfermg                     slty, East Lansmg, MI, 1989
species As shown m Table 2, the o-dlamadme                    8 WE Neeley, S C Wardlaw, T Yates, W G Hollmgsworth
results are ldentlcal to the UV spectrophotomet-                and M ET Swmnen, Clm Chem ,22 (1976) 227
ric results on ether extracted samples                        9 C J Patton, Ph D Thesis, Michigan State Umverslty, East
                                                                Lansmg, MI, 1982
                                                             10 C J Patton, M Rabb and S R Crouch, Anal Chem , 54
   Conclzmons                                                   (1982) 1113
   The o-dlamadme method has several advan-                  11 R L Hablg, B W Schlem, L Walters and R E Thlem,
tages over many of the common colorlmetrlc or                   Chn Chem , 15 (1969) 1045
fluorlmetrlc methods [1,21 for determmmg aro-                12 L R Snyder, J Chromatogr , 125 (1976) 287
matic aldehydes The method 1s automated, uses                13 W Spendley, G R Hext and F R Hunsworth, Technomet-
                                                                ncs, 4 (1962) 441
simple mstrumentatlon, and IS much faster than               14 S L Morgan and S N Demmg, Anal Chem , 46 (1974)
most of the colorlmetrlc or fluorlmetrlc methods                1170
reported m the literature      The o-dlanlsldme              15 KW C Burton and G NIckless, Chemometncs and Intel-
method produces results consistent with W                       hgent Laboratory Systems, 1 (1987) 135
spectrophotometrlc determmatlons On the other                16 J D Ingle and S R Crouch, Spectrochemlcal Analysis,
                                                                Prentice Hall, Englewood Cbffs, 1988
hand, direct spectrophotometrlc determmatlon of              17 R B Wearn, W M Murray, Jr, M P R-y              and N
aromatic aldehydes m the ultraviolet region 1s                  Chandler, Anal Chem , 20 (1948) 922
about 5 to 10 times more sensltlve method than               18 V Cerdi and C Mongay, Analusls, 4(2) (1976) 94

To top