a technical report submitted to the faculty of the

DEVELOPMENT AND APPLICATIONS OF AN ANALYTICAL METHODFOR MEASUREMENTS OF HYDROXYACETONE

AND GLYCOLALDEHYDE IN AIR

FANG LIU

A technical report submitted to the Faculty of the University of North CaroUna at ChapelHill in partial fulfillment of the requirements for the degree of Master of Science in PublicHealth in the Department of Environmental Sciences and Engineering.

Chapel HiU

Harvey E. Jeffries, Professor, Advisor

Richard M. Kamens, Professor, Reader

Prof. Donald L. Fox, Professor, Reader

ABSTRACT

An analytical method for measurement of glycolaldehyde andhydroxyacetone has been developed. This method includes derivatization ofhydroxylated carbonyl by 2,4-duutrophenylhydrazine (DNPH) and subsequentanalysis of the derivatives using high performance Uquid chromatography (HPLC).This method has been applied on the smog chamber experiments with satisfactoryresults. Glycolaldehyde and hydroxyacetone can be detected at 0.025ug/ml and0.085ug/ml level, or at lOppb and 29ppb in air with a IL standard air samplevolume using this analytical system.

In this study, a new scheme for reactions of hydroxylated carbonyls with 2,4-DNPH has been proposed based on the interpretation of mass spectra of thesederivatives. Based on this reaction scheme and experimental data, the structures ofDNPH derivatives of dicarbonyls and hydroxylated carbonyls have been foundidentical, but the reaction rates for dicarbonyls and hydroxylated carbonyl withDNPH are varied under different acidic conditions. According to this discovery, aprocedure has been developed for the measurement of both dicarbonyls (glyoxaland methyl-glyoxal) and hydroxylated carbonyls (glycolaldehyde andhydroxyacetone). The problem of interferences ( if both dicarbonyl andhydroxylated carbonyl exist in the same system), therefore, can be solved.

ACKNOWLEDGEMENTS

I would like to thank Prof. Harvey Jeffries for his advice and support in thepreparation of my technical report and for his guidance and financial support duringmy graduate study at UNC-CH.

I am also grateful to Prof. Richard Kamens, Dr. Donald Fox and Dr. KenSexton for their technical advice and useful suggestion during my graduate study andwork on my masters project.

I want express my thanks to my colleagues and friends: Wen Li, ElizabethHayes, Jeff Arnold, Kah-Eng Pua, for their support and suggestion in thepreperation of this technical report.

Lastly, I own my deep thanks to my husband, Zhongqiang, for his patienceand helpfulness during my graduate study.

This work was supported by a Cooperative Agreement from U.S.Environmental Protection Agency #CR818657-01 and from Coordinating ResearchCouncil, Inc. under the contract NO. ME-1.

«7S^-3:t^"B55??J

# TABLE OF CONTENTS

1. Introduction.........................................................................................................................1

2. Experimental Method.......................................................................................................3A. Apparatus and Equipment.....................................................................................3

B. Reagents....................................................................................................................4

C. Chromatograph........................................................................................................4

D. Chromatographic Conditions.................................................................................4E. Purification of DNPH.............................................................................................6

F. Preparation of DNPH stock solution...................................................................8

G. Preparation of carbonyl-DNPH derivatives........................................................8

H. Preparation of standards......................................................................................11I. Peak Identification................................................................................................14

3. Results and Discussion...................................................................................................20

A. The Reaction Mechanism of hydroxylated Carbonyls.....................................20

B. Identification of the products..............................................................................21

C. Optimization of reaction conditions...................................................................27i. DNPH concentration..................................................................................25

ii. Reaction time...............................................................................................25

iii. Acidity............................................................................................................27

iv. Reaction temperature.................................................................................27D. Elimination of interferences................................................................................30

4. Demonstration of the Method on Isoprene, Ethene and Propene Experiments in

UNC Outdoor Smog Chambers.........................................................................................38A. Description of UNC Outdoor Smog Chamber.................................................38B. Apparatus and Equipmental Conditions...........................................................38C. Experimental Operations.....................................................................................40D. Results and Discussion.........................................................................................42

5. Conclusion........................................................................................................................58

6. Future Work.....................................................................................................................59

7. Bibliography.....................................................................................................................61

8. Appendix..........................................................................................................................63

lif^^' ilf *^:^?^3*^'W~

LIST OF TABLES

Table 1. Relative standard deviation test................................................................15

Table 2. The initial conditions of ethene and propene experiment....................41Table 3. The initial conditions of isoprene experiment........................................41Table 4. The analytical data of ethene and propene experiment.......................43Table 5. The analytical data of isoprene experiment............................................44

LIST OF FIGURES

Figure 1. Separation of 15 standard mixture, glycolaldehydeand hydroxyacetone.......................................................................................5

Figure 2. Chromatogram of pure DNPH....................................................................7Figure 3. Chromatogram of glycolaldehyde-DNPH derivative...............................9Figure 4. Chromatogram of hydroxyacetone-DNPH derivative...........................10Figure 5.a Calibration curve of glycolaldehyde.........................................................12Figure 5.b. Calibration curve of hydroxyacetone........................................................13Figure 6. Determination of the detection limit........................................................16Figure 7.a. Peak identification of glycolaldehyde-DNPH derivative......................17Figure 7.b. Peak identification of hydroxyacetone-DNPH derivative.....................19Figure 8.a. Mass spectrum of purified 2,4-DNPH......................................................22Figure 8.b. Published mass spectrum of 2,4-DNPH...................................................22Figure 9.a. Mass spectrum of glycolaldehyde-DNPH derivative.............................23Figure 9.b. PubUshed mass spectrum of glycolaldehyde-DNPH derivative...........23Figure 10. Mass spectrum of hydroxyacetone-DNPH derivative............................24Figure 11. Effect of DNPH concentration..................................................................26Figure 12. Effect of reaction time................................................................................28Figure 13. Effect of acid level.......................................................................................29Figure 14.a. Standard curve of glyoxal at 0.08N HCl acidic condition......................33Figure 14.b. Standard curve of glyoxal at 2N HCl acidic condition...........................34Figure 14.c. Standard curve of glycolaldehyde at 2N HCl acidic condition.............35Figure 14.d. Standard curve of methyl-glyoxal at 0.08N HCl acidic condition........36Figure 14.e. Standard curve of methyl-glyoxal at 2N HCl acidic condition.............37

Figure 15. Schematic of UNC Outdoor Smog Chamber..........................................39Figure 16. Concentration-time profile of glycolaldehyde in ethene experiment..45Figure 17. Concentration-time profile of glycolaldehyde

in isoprene experiment..............................................................................46

Figure 18. Concentration-time profile of hydroxyacetone inisoprene experiment...................................................................................47

Figure 19.a. Chromatogram of a sample collected in isoprene experiment.............49Figure 19.b. Chromatogram of a blank sample prepared in isoprene experiment .50Figure 20. Compound identification............................................................................51Figure 21.a. Chromatogram of a sample collected in ethene experiment................52Figure 21.b. Chromatogram of a blank sample prepared in ethene experiment.....53Figure 22. Isoprene/NOx experiment in UNC outdoor smog chamber................55Figure 23. Ethene/NOx and propene/NOx experiment in

UNC outdoor smog chamber.....................................................................56

1. INTRODUCTION

The study of the emission of volatile organic compounds from anthropogenicand biogenic sources into the atmosphere has continued now for more than 50 years(Isidorov et al., 1985). It has been recognized that these emissions can result in suchapparently diverse effects as photochemical air pollution, acid deposition, long-range transport of chemicals, changes in the stratospheric ozone layer and globalweather modification through very complex chemical and physical transformations(Atkinson, 1990). Scientists have conducted a vast amount of study includinglaboratory, smog chamber, and ambient atmospheric studies to investigate thephotochemical mechanisms of these compounds (e.g., the experimental data ofsmog chamber can be used to test the hypothesized theories). However, there arestill many uncertainties concerning the roles that these complex organic compoundsplay in the atmosphere.

One of the central issues in this area is the concern of "missing carbon"~adiscrepancy between the carbon present in the initial reactants and that measuredin the products (Jeffries, 1990). Several explanations have been developed toaccount for the missing carbon. One is that the measurements of organiccomponents in the atmosphere do not provide information about all the compoundsemitted (Isodorov et al., 1985). Besides the known compounds, there could beothers which currently have not been detected. One hypothesis is that the form ofthe "missing carbon" may be presented as multi-functional compounds such ashydroxylated carbonyls (Jeffries, 1990; Tuazon and Atkison, 1990). Developmentof analytical methods for reliable detection of these compounds in air is needed toidentify these missing carbon compounds and to investigate their reaction

mechanism which is important if we are to understand the chemistry of these

compounds.

For this project, the first two members of the hydroxylated carbonyl

compound series, glycolaldehyde and hydroxyacetone, were chosen as modelcompounds for developing a standardized, regular method for the qualitative andquantitative determination of hydroxylated carbonyls in air samples. The structuralinformation of hydroxylated carbonyl compounds is shown in an appendix.

Several papers have been published in the last ten years concerning the

sampling and analysis of aldehydes, principally formaldehyde, in air. The mostspecific and sensitive analytical method available is based on the reaction of organiccarbonyls (aldehydes and ketones) with 2,4-dinitrophenylhydrazine (DNPH) andsubsequent analysis of the hydrazone derivatives by high performance liquidchromatography (HPLC) (Tejada, 1986; Smith, et al., 1989). There is almost noinformation, however, on the analysis of hydroxylated carbonyls with this or anyother method. Two questions remained: Can this method be applied to analysis of

hydroxylated carbonyls? If so, what are the optimum reaction conditions for

analyzing these compounds?

The objective of this Master's project is to develop an analytical method for

measuring hydroxylated carbonyls in air, focusing particularly on glycoladehyde andhydroxyacetone, and to demonstrate the applicability of this method inisoprene/NOx and ethene/NOx photochemical experiments carried out in anoutdoor smog chamber.

2. EXPERIMENTAL METHOD

The analysis of Ci-Ce carbonyl compounds in ambient air by reaction with2,4-dinitrophenylhydrazone (DNPH) and subsequent analysis of the hydrazonederivatives using HPLC has been reported by Kuwata et al. (Kuwata et al., 1978).This method was based on specific reaction of organic carbonyl compounds withDNPH at the presence of acid to form stable derivatives according to the followingequation:

NO2 NO2

R-_d_R' + HgN-NH-^ ^NOg i^ "\c=N-NH-/ "^-NOg+ Hp

Where R and R' can be any organic radical or hydrogen.Qualitive and quantitative analysis for the carbonyl-DNPH derivatives can beconducted by HPLC. This method is regarded as feasible and convenient for theanalysis of a variety of aldehydes and ketones.

Based on the general principle of his method, I have developed a newmethod based on the new derivatization mechanism that I discovered. This new

method can be used to measure hydroxylated carbonyl under the new optimizedconditions. These will be discussed in the later sections of this paper.

A. APPARATUS AND EQUIPMENT

The following equipment was used in this study:1. Hot plates, beakers, flasks, measuring pipets, volumetric flasks, syringes.2. Mass Spectrometer (available in MS lab, Rosenau Hall, UNC-CH).

B. REAGENTS

A variety of chemical regents are used and are listed below:1. 2,4-Dinitrophenylhydrazme (DNPH) - Aldrich Chemical, 70% soUd in

2. Acetonitrile (ACN) - HPLC grade, best source3. Methanol - HPLC grade, best source4. Water - Distilled and deionized water

5. Hydrochloric acid - Analytical grade, best source

6. Hydroxyacetone - Fluka Chemical Corp., assay -95%

7. Glycolaldehyde - Fluka Chemical Corp., purification > 98%8. Ethanol - Absolute, Aaper Alcohol and Chemical Co.9. Glyxol - Fluka Chemical Corp., 40% in water

10. Methylglyoxal - Fluka Chemical Corp., 40% in water

C. CHROMATOGRAPH

A Varian 5000 HPLC system with a fixed 254nm UV detector, a DuPontZorbax ODS column (4.6mm x 250mm), a 25ul sample loop and a Kipp & ZonenChart recorder are used in this study.

D. CHROMATOGRAPHIC CONDITIONS

The Varian 5000 HPLC system was operated isocratically with a 60% ACN,30% H2O and 10% Methanol mixture and a mobile phase flow of ImL/min.Injection volume of the sample was 25ul. Figure 1 shows the separation of astandard mixture of 15 aldehyde-DNPH and ketone-DNPH deivatives ( thisstandard mixture was acquried from Tejada's lab, EPA). It also includes a second

1. Formaldehyde2. Acetaldehyde3. Acrolein4. Acetone

5- Propionaldehyde6. Crotonaldehyde7. Butyraldehyde8. Benzaldehyde^ Isovaleraldehydelo.Valeraldehyde(l.o-TolualdehydeI? m-Tolualdehydeli. p-Tolualdehyde/i. HexanaldehydeI5-.2, S-Dimethylbenzaldehyde

1.9921.019

1.7630.8190.794

1.0511.072

1.0341.105

I^\^u.\<^\<Uk^ J

^V^''!?U%»<.V<^"^"'**^''

Ti me c wm >S

tn^^^w^

Figure 1. Chromatograms of a standard mix of 15 aldehyde and ketone-DNPHadducts (upper), ancfsynthesized glycolaldehyde-DNPH and hydroxyacetone-DNPHderivatives (lower). The chromatographic conditions are described in text. Range:0.04.

chromatogram of synthesized glycolaldehyde-DNPH and hydroxyacetone-DNPHsamples using this chromatographic condition. This figure shows that thechromatographic condition is adequate to be appUed in this study.

E. PURIFICATION OF 2,4.DNPH REAGENT

Because that the purity of commercially available DNPH is very low (only70% solid in water), purification of DNPH reagent becomes necessary to avoidinterference before it is used as a derivatizing agent. The purification process isdescribed as below.

A supersaturated solution of DNPH was prepared by boiling excess DNPH in200ml of ACN. The supernatant was transferred to a beaker, put under a coverglass and allowed to cool gradually to 40-60° C by putting the beaker on a hot plate.This procedure maximized crystal size and purity (Tejada, 1986). The solvent wasallowed to evaporate slowly at this temperature range until 90% of the solvent hadevaporated (approximately 24 hours). The remaining saturated solution wasdecanted to waste and the crystals were rinsed twice with about three times theirapparent volume with ACN. Then, the crystals were transferred to another cleanbeaker. The recrystllization process was repeated twice. An aliquot of the last rinsewas taken, diluted 10 times with ACN, acidified (with 1ml of concentratedhydrochloric acid per 100ml of DNPH solution), and analyzed by HPLC. Theimpurity level should be comparable to that shown in Figure 2. The crystallizationprocess should be repeated if the impurity level is unsatisfactory (Tejada, 1986).The purified crystals should not be allowed to contact laboratory air because thiscould result in contamination.

.- §r 1 -—j-----r Z3r-a:^ i 4-^----

^ >- I-----1___;___

.______J_______^______i_______(-------------

------------------;-----------------1------------------j----------------\-----------------

,-----

L_________-:—

"—4-------ir

^-----'

1------------.

1-----1------------

-;-----

1-------^-----

! ----j1-------------

1---------H----------\-----------i----------

fcrii^-l-----\-----

1-----------1-----------i-----------1----------

-—1—\—r^=^

^—i—h^!--------! 1 1=^,:----1-----i-----1-----'

-^—^-----'.-----,

1----------ZA---------------------------1-------------,?zE^

1-------------

t——1-------r

1——\----f

t——t-

i—^^4^J;

4=-i—

\-----,i-----------

i-----------

i-----------•T' ——

^-----

1 .....

I---------

^-----

1"Ti'm-e c •mi n )

Figure 2. Chromatogram of pure DNPH reagent. The chromatographic conditionsare described in text. Range: 0,4; injection volume: 25 ul.

F. PREPARATION OF DNPH STOCK SOLUTION

Once the crystals had been satisfactorily cleaned, they were transferred to aglass reagent bottle. Approximately 100ml ACN was added in the bottle, then thebottle was stoppered. The mixture was shaken gently and allowed to standovernight. Clean pipets and rubber bulbs was used when taking aliquots of thesaturated solution. Contact with air was minimized to reduce the contamination.

According to the Tejada's report (1986), the stock solution contains about llmgDNPH per ml at room temperature.

G. PREPARATION OF CARBONYL-DNPH DERIVATIVE

A saturated solution of DNPH in 0.2N HCl (prepared in ACN or ethanol)was titrated with the glycolaldehyde or hydroxyacetone. The mixture was shakenand allowed to stand overnight. The colored precipitate (yellow) was filtered,washed with 2N HCl and water for three times, and allowed to air dry. Thechromatographic purity of the derivative was checked by HPLC analysis of a dilutesolution of the derivative in ACN (Figures 3 and 4). The precipitate should berecrystallized from ACN if the purity is not satisfactory. It should be noted thatpurities of the precipitates can not be determined simply using HPLC analysis,because some impurities may not absorb ultraviolet therefore they can not bedetected by a UV detector.

Although the purity of the carbonyl DNPH derivative can be testedconveniently by checking melting point, the purity of the hydroxyacetone-DNPHderivative and glycolaldehyde-DNPH derivative can not be tested by melting point

Iw^l^l ll ll|^»F%»V«<«fM

2o KO 7 0

Figure 3. Chromatogram of pure glycolaldehy-DNPH derivative. Thechromatographic conditions are described in text. Range: 0.1; inject volume: 25 ul.

. -.—l^

_o.vQ,

___________iQ—--------------- .'r. t

----------H---------]:--

11 '""r

------

~C3___o>

" O--^1

-------Z

----- :------------W^

------

____1____ . '_------ _-.

Tm^ ( mtti)

Figure 4. Chromatogram of pure hydroyacetone-DNPH derivative. Thechromatographic conditions are described in text. Range: 0.1; inject volume: 25 ul.

determination since the theoretical data were unknown. From HPLC and later MS

analysis, however, the derivatives were assumed to be pure compounds.

H. PREPARATION OF STANDARDS

A series of standard solutions of glycolaldehyde-DNPH derivative and

hydroxyacetone-DNPH derivative were prepared by the following procedure: to aknown amount of glycolaldehyde or hydroxyacetone solution, I added 0.2ml ofconcentrated HCl and 2ml saturated DNPH stock solution in a 25ml volumetric

flask; fill with ACN to the mark. The mixture was shaken and allowed to react for

two hours under room temperature before being injected into HPLC.Calibration curves for glycolaldehyde and hydroxyacetone were obtained by

plotting peak heights versus the concentrations of the glycolaldehyde orhydroxyacetone in the standard solutions (see Figures 5.a and 5.b).

Within a certain range, the yield was independent of the glycolaldehyde orhydroxyacetone concentration. This range was therefore determined as the linearrange, i.e the concentration range that can be properly tested by the method. Thelinear ranges of the glycolaldehyde and hydroxyacetone are from 0.02ug/ml to2.5ug/ml and from 0.09ug/inl to 2.6ug/ml, respectively. When the concentrationswere higher than these ranges, the peak heights were lower than expected (seeFigures 5.a and 5.b). This can be attributed to the limited solubility ofglycolaldehyde-DNPH derivative or hydroxyacetone-DNPH derivative in ACN, andthe lower ratio of DNPH concentration to individual derivative concentration (seediscussion of effect of DNPH concentration in the Results and Discussion section).

Calibration Curve of Glycolaldehyde

Range 0.02

Reaction time: 2 hours

Reaction temperature: room temperature

Injection volume: 25 ul

Y = - 0.5378 + 60.28X

R^= 0.9997

concentration (ug/ml)

1 grid = 2 mm

Figure 5.a. Calibration curve of glycolaldehyde (ranee 0.02). Outlier A is excludedfrom statistic computation. Linear range: 0.02ug/im to 2.5ug/ml.

^ 80(Ao

D) 70_c

^ 60E 50OS

Calibration Curve of HydroxyacetoneRange 0.01

Y =-0.1029+ 22.85X

R =0.9997

1 grid = 2 mm

Figure 5.b. Cjilibration curve of hydroyacetone (range 0.01). Outlier B is excludedfrom statistic computation. Linear range: 0.09ug/ml to 2.6ug/ml.

Precision was defined as the variation of the results in a set of replicatedmeasurements. It can be presented as a relative standard deviation (RSD) whichwas defined as:

RSD = (standard deviation) x 100% / mean

The relative standard deviation was determined by testing samples with the sameconcentration of glycolaldehyde or hydroxyacetone. The tests were conducted for

both compounds, and the results are shown in Table 1. The relative standard

deviation of glycolaldehyde derivative at 0.758ug/ml is 2.35%, and hydroxyacetonederivative at 0.428ug/ml is 5.81%.

The detection limit is an important factor in the evaluation of an analytical

system because it can determine whether the method is apphcable for analysis oftrace concentrations of these compounds. In this study, the detection limit was

defined as being a peak of 3 times the noise height. As an example, Figure 6illustrates the detection Hmit of hydroxyacetone. The detection limits of

glycolaldehyde and hydroxyacetone were determined to be 0.025ug/ml and0.085ug/ml using the analytical system described. Assuming the standard air sample

volume is IL ( sample for 20 minutes with flow rate of 0.5ml/min), the detectionlimits of glycolaldehyde and hydroxyacetone will be lOppb (20ppbC) and 29ppb(87ppbC) in air.

I. PEAK IDENTIFICATION

It is necessary to use blank samples to test for absorbing interferences. As

shown in Figure 7.a, a reagent blank (blank 1) was made by adding 2ml saturated

Table 1. Relative Standard Deviation Test

# Sample

Glycolaldehyde Hydroxyacetone

0.758 ug/ml 0.428 ug/ml

0.726 0.486

0.714 0.433

0.693 0.455

Mean 0.711 0.458

Standard Deviation 0.0167 0.0266

Relative Standard Deviation 2.35% 5.81%

o ' Q,C-;

— A*7-^-^tnr.S^'^^ o !

.... -T ^

.._.....\. ——^ '"1.... — " i'

- — — -.-

Figure 6. Determination of the detection limit. Peak height > 3 S/N. Range: 0.005

ai'vr-

r « « »

I A. |2 lo 4-2 0 ,^ ,4- !2 fo 8 6 4 2 0 '^ ic 8 6 4 2 q

Figure 7.a. Chromatograms of glycolaldehyde sample and blanks. Peaks: (1)DNPH; (2) glycolaldehyde (retention time 13 minute).

DNPH solution, 0.2ml concentrated HCl, and ACN into 25ml volumetric flask.

After standing for two hours, 25ul of this blank sample was injected into HPLC.Another blank (blank 2) was made by adding 1 ml glycolaldehyde, 0.2 ml HCl and

ACN into a 25 ml volumetric flask. After standing for two hours, 25ul this mixture

was injected into HPLC. Comparing the chromatogram of glycolaldehyde-DNPHsample with that of blank samples (see Figure 7.a), only one peak corresponding to

the glycolaldehyde-DNPH derivative with the retention time of 13 minutes was

observed. Similarly, Figure 7.b shows the chromatograms of hydroxyacetone-DNPH

sample and a reagent blank (same as blank 1 discussed above). Only one peak

corresponding to the hydroxyacetone-DNPH derivative can be observed. The

retention time of hydroxyacetone is 19 minutes.

In general, the method includes several steps. First, commercially available

DNPH needs to be purified to reduce interferences by recrystalization. Second,

hydroxyacetone and glycolaldehyde are derivatized by DNPH saturated solutionunder controlled acidified condition (0.08N HCl). Finally, after two-hour reaction,

25ul mix is injected into the HPLC column. The retention times of glycolaldehyde

and hydroxyacetone are 13 and 19 minute under chromatographic conditionsdescribed earlier.

4K-j^^-^V'"'

T»»^-<^ OnTvi >

Figure 7.b. Chromatograms of hydroyacetone sample and blank. A: saniple; B:blank. Peaks: (1) DNPH; (2) and (3) contaminates; (4) hydroxyacetone (retentiontime 19 minute).

. -'.,5g??^^^i^

3. RESULTS AND DISCUSSION

A, THE REACTION MECHANISM FOR HYDROXYLATED CARBONYLS

In general, aldehydes and ketones react with DNPH in the presence of acidto form stable derivatives according to the following equation:

NO2 NO2

C-R' + HgN-NH-/ y-N02 -^ "\c=N-NH-^ \_NO2+ Hp

Where R and R' can be any organic radical or hydrogen.

In this study, it was discovered that for the hydroxylated carbonyl compounds,

such as hydroxyacetone and glycolaldehyde, the reaction would go according to thefollowing scheme

HO ONO2

R-HC—C—R' +3H2N—NH-/ V-NO2 H.

R —C=N—NH-^ '^_N02NO2

R —C = N—NH—^ '^—NO2NO2

+ NH3 + 2l-feO + H2Nl-^ '^—NO2Where R and R' can be any organic radical or hydrogen.

While one of the reaction products, bis-(2,4-dinitrophenyl)hydrazone, has beenconfirmed by MS analysis (see discussion below), the reaction mechanism still needsfurther confirmation.

• 20

B. IDENTIFICATION OF THE PRODUCTS

The identifications of the products, i.e. glycolaldehyde-DNPH derivative and

hydroxyacetone-DNPH derivative, were based on the interpretation of the EI

spectra.

Figures 8.a and 8.b show the mass spectrum of purified 2,4-DNPH (as

described earlier) and a pubhshed spectrum of this compound (Wiley/NBS). There

is good agreement between these two spectra. This supports the earlier conclusion:

the purity of DNPH is satisfactory.

In Figure 9.a, the identification of glycolaldehyde-DNPH derivative was

based on the appearance of a molecular ion at m/z 418, fragment ion at m/z 236

[M-(NH-C6H3(N02)2]^» and fragment ions with same numbers of m/z as DNPH(see Figure 9.a). The fragment ions at m/z 401 and m/z 219 appear to arise from

complicated fragmentation and recombination, and therefore do not readily

contribute to the knowledge of the compound structure. Figure 9.b was a published

mass spectrum of the same compound. Both spectra appear to have same fragment

ions at m/z 418, 236, 219, 205,192, 183,164,153, etc. Based on the good agreement

between these two spectra, the assumption that the glycolaldehyde-DNPH

derivative synthesized was a pure compound is supported.

In Figure 10, a weak molecular ion , showing hydroxyacetone-DNPH

derivative (C15H13N8O8), can be observed at m/z 432. The fragment ion at m/z

250 arises from loss of NH-C6H3(N02)2- The fragment ions at m/z 415, 233 showsome analogy to the fragment ions at m/z 401, 219 in mass spectrum of

glycolaldehyde-DNPH derivative. Although there is no published spectrum for

hydroxyacetone-DNPH derivative, the identification of this compound can be based

on the comparison of its spectrum with mass spectrum of glycolaldehyde-DNPH

|rile;V1253 Scaji:51 Mar Def 0.25 Acq: l-oaJl--91 19:41:65-1:0570SEQ EI* Function :Ha<jnet Bp«:63 Bpl:978432 TIC:J1837464trile Text:2,4-DNPH PROBE EI100

ill.'IIlii;|!!l.!i

I'l'Ji'i'

168 180

L9 3E5 11.8 8E5 1

:8 3E5 1L7 BE5

L7 3E5

50 60 70 80 90 l60 110 120 130 140 150 160 1'70 180 190 200 210 220 230 240

8E5 '!

Figure 8.a. Mass spectrum of synthesized DNPH.

I I I I I I I I I I I I I T I I I I I I I I I I'l I I I I 1 I I I I i I I I I I I I I I I I I I I I I I I I I I ' I I I I I I M I 'I I I I I I I I'a 160 17a isa i9a aea aia aaa aba a4a asa asa a?a aea asa saa

" LL'"LL"'i*l!'"'l'!:'''''l'9 IB zSMW:19S

111 iM'I'l'l I I'l'l'l' I' I'I'I'i'i 111 ri'I'i 111111 hi 11111'l-i 111 m6a 7b aa »a lea iia 12a i3a i4e is3a 4a .saHuap&zine, (2,4-dlnltPoph«n||I>-

CeHeN404

Figure 8.b. A published mass spectrum of pure DNPH.

rile:V1635A Scan:104 Met Def 0.25 Acq: 3-SEP-91 00:14:08 +4:29JOSEQ EI+ Function:Magnet BpM:63 Bpl:1059914 TIC:27036650File TextiHYDROXX-ACETONE100_ 63

360* ..l"., -i M

^I.IES

L9.5E5

.9.0ES

.8.5E5

L5 . 8E5

.4.8E5

L4.2E5

L3.7ES

.2 . 6E5

L1.6E5

L1.1E5

400 4S0

O.OEOM/Z

Figure 9.a. Mass spectrum of symhesized glycolaldehyde-DNPH derivative.

418 C14 HlO m 08 im-16-8 SD-1981-0-0 Ethanedial, blsl (2,4-<linitrophenyl)hydraron«);; Glyoxal, bisl(2.4-dinltroph8nvllhvdrazon«l; W/NBS___________________________________________________________58581

,>1, i-,LH k^,f,t*},t}/^,-,4, ikIiio ' 126 HO uo ' 180 260 ' 226 ' u6 ' iio ' iib ' 366 ' 3i6 ' 346 Uo ' iio ' 466 ' 4^6 ' Uo ' U6 ' iio20 40 60 80

Figure 9.b. A published mass spectrum of pure glycolaldehyde-DNPH derivative.

File:V1634 Scan:70 Mer Def 0.25 Acq: 2-SEP-91 23:54:01 +3:0270SEQ EI+ Function:Magnet BpM:63 Bpl;728954 TIC:26577720File Text:hydroxy-acetone DNPH DPROBE100_ 63

«f ^*-i

sio^ fJli, K t^ J^350 400 '4io'

^7.3E5

L6.6E5

L4.7E5

L4.0E5

L3.6E5

3. 3E5

i.2. 9E5

i.2. 6E5

2. 2E5

L1.8E5

11.5E5

.1.1E5

.3.6E4

O.OEOM/Z

Figure 10. Mass spectrum of synthesized hydroyacetone-DNPH derivative.

• 24

derivative. It should be noted that difference between these two spectra at higher

m/z range (14 in unit of m/z) can be attributed by their different molecular weight.These mass spectra support the conclusion that the synthesized compound

(hydroxyacetone-DNPH derivative) was pure.

C. OPTIMIZATION OF REACTION CONDITION

A fast, sensitive, and completive reaction is desired for the derivatization.

Based upon the literature and the experiments conducted, I found that acidity,reaction temperature, reaction time, and DNPH concentration were the factors that

affect the reaction characteristics, e.g. reaction rate and yield. I also studied theinfluence of each factor and combinations of factors on the reaction characteristics,

and I optimized the reaction conditions to achieve the best result in terms of thedesired derivatizations.

DNPH concentration: It is necessary to control the concentration of DNPH

to avoid a huge DNPH peak that would interference the other carbonyl compounds

although higher DNPH concentration could potentially increase reaction rate. The

yields versus the ratio of DNPH concentration to hydroxyacetone concentration

were tested (see Figure 11). My study has shown that the initial DNPHconcentration should be at least 500-fold excess to the concentration of

hydroxyacetone. Similar results were obtained in the glycolaldehyde derivatization.Reaction time: O.Zml concentrated HCl, 1ml hydroxyacetone or

glycolaldehyde and 2ml saturated DNPH solution were added in a 25inl volumetricflask, and the mixture was filled with ACN to the mark. The reaction times were

allowed to vary from 30 minutes to 4 hours prior to injection into the HPLC. The

concentration of glycolaldehyde-DNPH derivative remained constant after 2 hour at

Effect of DNPH Concentration

Range 0.005

(in thousands)Cone, of DNPH/Conc. of hydroxyacetone

1 grid = 2 mm

Figure 11. Effect of the concentration of DNPH on the reaction yield. Reactiontime, 2 hours at room temperature. Chromatographic conditions as described in thetext.

room temperature. The trend of the concentration of hydroxyacetone-DNPH

derivative was increasing with increased reaction time (see Figure 12), This means

that for hydroxyacetone, longer reaction time will increases the reaction yield, andtherefore, increases the sensitivity of the method. By considering the convenience

of its application however, the length of reaction time was decided as two hours.Acidity: According to the reaction mechanism, the solution must be acidic

enough for an appreciable fraction of the carbonyl compound to be protonated, yetnot so acidic that the concentration of the free nitrogen compound is too low(Morrison, 1980). In order to obtain optimum acidity of the reaction solution , the

effect of pH value on the reactivity of DNPH towards hydroxyacetone andglycolaldehyde was examined. The volumes of acid added to the reaction mixturewere varied from 0.1ml to 2ml. Through out this experiment, the concentrations of

DNPH, the glycolaldehyde, and hydroxyacetone were kept constant. The results

are plotted in Figure 13. In this study, acid level was set up at 0.2ml HCl/25 mlreactants.

Reaction temperature: The preliminary experiments showed that increased

reaction temperature could accelerate the speed of the reaction of hydroxyacetoneand glycolaldehyde with DNPH. For example, after the reactant mixture had beenheated at 60«C for two hours, the yield of the hydroxyacetone-DNPH derivative was

found to be 2 times as much as that of under room temperature. Since the reaction

temperature is not easy to be controlled under our laboratory condition, roomtemperature was chosen as reaction temperature in this study.

bJilSgJf f^V^V*^' -

Effect of Reaction TimeRange 0.01

•gCO

(00)Q.

reaction time (hr.)

a glycolaldeliyde + iiydroxyacetone

* 1 grid = 2 mm

Figure 12. Effect of the reaction time on the derivatization of glycolaldehyde andhydroxyacetone with 2ml DNPH and 0.2ml HCl.

Effect of AcidityRange 0.01

00.7 0.9 1.1 1.3

volume of acid (ml)D glycolaldehyde + hydroxyacetone

* 1 grid = 2 mm

Figure 13. Effect of acidity of solution on the reactivity of DNPH towardsglycolaldehyde and hydroj^racetone. Reaction time: 2 hours at room temperature.

^BKS-i-=V3iK^"'Ss; •:

D. ELIMINATION OF INTERFERENCES

There is one potential problem in the method described above. I have

already noted in the development of the derivatizing mechanism that DNPH cansubstitute both hydroxyl (-OH) and carbonyl ( = 0) functional groupssimultaneously. Since the DNPH derivatives of both hydroxyacetone

[CH3C(0)CH2(OH)] and methyl-glyoxal [CH3C(0)CH(0)] are identical, thismethod can not differentiate methyl-glyoxal from hydroxyacetone using HPLC

technique*. The same problem also exists for glycolaldehyde [CH2(0H)CH(0)]and glyoxal [CH(0)CH(0)].

Further studies were conducted to solve this problem. I found that the yields

of derivatization of these compounds varied with different acidity level at a different

rate. Under the 0.08N HCl reaction condition (condition 1), both compounds

reacted with DNPH and had significant yields. Under 2N HCl reaction condition

(condition 2), however, methyl-glyoxal had a higher yield than that of in condition 1,

whereas hydroxyacetone had a much lower yield. This suggests that hydroxyacetone

is unlikely to react with DNPH under this condition.

As described above, when both hydroxyacetone and methyl-glyoxal are

presented in the same sample, they will be derivatized by DNPH and contributed tothe HPLC response in terms of peak area under both conditions. Since the peakwidth is only dependent on mobile phase and column properties, both compoundswill be shown in the HPLC chromatogram with the same retention time and peak

width. Based on this situation, peak height is conveniently adopted in followingcalculation.

* According to Tuazon (1990), hydroxyacetone and methyl-glyoxal could existsimultaneousely in the photochemical products of certain organic compounds suchas isoprene.

Condition 1: Ha + Hb = Hi (1)

Where HA~Peak height corresponding to the concentration ofhydroxyacetone-DNPH derivative under condition 1.

He-Peak height corresponding to the concentration of

methylglyoxal-DNPH derivative under condition 1.

Hi ~ Peak height corresponding to the total concentrations of

both derivatives under condition 1.

Similarly, under condition 2:

Ha'+Hb'=H2 (2)

Since the peak heights corresponding to the concentrations of DNPH

derivatives, they can be expressed as a linear function of the concentrations ofcarbonyl compounds.

HA = a + bXA (3)

HB = c + dXB (4)

HA'=a'+b'XA (5)

HB' = c'+d'XB (6)

Where XA~concentration of hydroxyacetone in the sample.Xfi-concentration of methyl-glyoxal in the sample,

a, b, c, d, a', b', c', d'-parameters of those linear functions.

From above equations, we can find following relations:

(1) = (3) + (4) and

(2) = (5) + (6);

therefore, Hi = a + c + bXA + dXB andH2 = a' + c' + b'XA + d'Xs

Since Hi and H2 can be determined from the experiments (HPLC

chromatograms), and a, b, c, d, a', b', c', d' can be obtained from measurements ofthe standzirds (calibration curves), Xa and Xb can be computed. The same methodcan also be applied to determine the concentrations of glycolaldehyde and glyoxalwhen they are mixed in the same sample.

Based on the above deduction, experiments had been conducted to obtain

the standard curves of these compounds under different acidic conditions. The

results are plotted in the Figure 14.a-e. Under the higher acidic condition, i.e. 2 NHCl, hydroxyacetone is difficult to react with DNPH, therefore the standard curve is

determined as Y = 0. These standard curves can be used in the smog chamber

experiments to quantify these photochemical products.

STANDARD CURVE OF GLYOXAL

Acidity: 0.08N HCl ( Range 0.02)'*\J

35 Reaction temperature: room temperature 1

30Injection volume: 25 ul y 1

15y^ Y =1.628 +75.72 X

y^^ R 2=0.983810

X5- y^0

1 grid = 2 mm

Figure 14.a. Standard curve of glyoxal at 0.08N HCl acidic condition

STANDARD CURVE OF GLYOXAL

Acidity: 2N HCl ( Range 0.02)

35Reaction time: 1 hour y/^ 1Reaction temperature: room temperature y/^ 1

30 Injection volume: 25 ul y^ 1y^° 1

25 y^20 /^15 y^ Y = 0.8601 + 86.51 X

y^ R 2=0.998510

X5 - y0 ^ . . . . . 1

0.2 0.4

1 grid = 2 mm

Figure 14.b. Standard curve of glyoxal at 2N HCl acidic condition

(0(DQ.

'TX' 30

STANDARD CURVE OF GLYCOLALDEHYDE

Acidity: 2N HCl (Range 0.02)

Reaction time: 1 hour

Y = -0.1138 + 20.20X

R 2= 0.9985

* 1 grid = 2 mm

Figure 14.c. Standard curve of glycolaldehyde at 2N HCl acidic condition

:^^^^^Ta W^pj^j^^:^'

STANDARD CURVE OF METHYL-GLYOXAL

Acidity: 0.08N HCl (Range: 0.01)

L Reaction time: 2 hoursr Reaction temperature: room temperature j/^L Injection volume: 25 ul y^

\ y^ Y= 2.9037 + 97.79Xr y^ R^= 0.9949r ^jT

1 r 1 1 1 1 1 1 1 1 1 1 1 1 1..... • 10.2 0.4 0.6 0.8 1 1.2

1.4 1.6 1.8

1 grid = 2 mm

Figure 14.d. Standard curve of methyl-glyoxal at 0.08N HCl acidic condition

180 h170

STANDARD CURVE OF METHYL-GLYOXAL

Acidity: 2N HCl (Range: 0.01)

Reaction time: 1 hour

Y =-1.4740+100.92X

R2= 0.9990

J___I___L

04 06 OS 1 1.2

1.4 1.6 1.8

1 grid = 2 mm

Figure 14.e. Standard curve of methyl-glyoxal at 2N HCl acidic condition

fe^^-^.'SfF" .- "'tH'^^^^^i^—

4. DEMONSTRATION OF THE METHOD IN ISOPRENE/NOx ANDETHENE/NOx CHAMBER EXPERIMENTS

The outdoor smog chamber is a photochemical reactor which can be used to

investigate the photochemical mechanism of the compounds of interest. After Ideveloped the analytical method, I used it in several chamber experiments toidentify and quantify the photochemical products of simple olefin (e.g. ethene,

propene, and isoprene), which play important roles in understanding the chemistryof photochemical smog.

A. UNC OUTDOOR SMOG CHAMBER

The UNC outdoor smog chamber is located at Pittsboro, North Carolina.

The chamber has a Teflon walls and is divided into two identical sides; these are

named Red and Blue (see Figure 15). Both sides have a volume of 150,000L.

B. APPARATUS AND EQUIPMENTS

Following equipments were used in the chamber experiments for sampling

purpose:

1. Air pump, rotameter, mass flow controller

2. Glass bubblers, connecting tubes

3. Teflon tubes

4. Timer, soap bubble flow meter

5. 25ml volumetric flasks, measuring and disposable pipets, glass bar, glass

funnel, etc.

STRUCTURE OF CHAMBERS

bubbler(s)

air pump

bubbler(s)

air pump

Figure 15. Schematics of UNC outdoor smog chamber.

C. EXPERIMENTAL OPERATIONS:

Initial conditions: Two experiments were performed in the dual chamber.

The first chamber experiment was conducted with ethene and propene. The red

side was injected with ethene, NO, and NO2, and the blue side with propene, NO

and NO2. All reactants were injected into chamber and mixed well before sun rise.The second experiment was an attempt to study the isoprene daytime chemistry.

The red side and blue side were injected with isoprene, NO, and NO2 before sunrise. The initial reactant concentrations for both sides are Usted in Table 2. The

analytical method described above was employed to identify the existence of

glycolaldehyde and hydroxyacetone in the isoprene photooxidation products. The

first chamber experiment was conducted with ethene which the red side was injected

with ethene, NO, and NO2, and the blue side with propene, NO and NO2. Allreactants were injected into chamber and mixed well before sun rising. Table 3 haslisted the reactant concentrations in the second chamber run.

Sampling: Because the issue of optimum sampling method is beyond the

scope of this project, the sampling procedures were simplified by taking the samples

directly beneath the chamber and allowing the air to go through standard bubblers.Each bubbler was filled with 10ml ACN. In the first run, two bubblers connected in

series were used for sampling to increase or check on the sampling efficiency. The

flow rate was controlled at 0.52 1/min, and the sampling time was 40 minutes.

During the second experiment, only one bubbler was employed to absorb the air

sample with a flow rate of 0.421/min. Each sample was collected for 40 minutes. In

the second run, two bubblers which connected in series were used for sampling to

Table 2. The initial concentrations of ethene and propene experiments, (October 7,1991)

Ethene

Propene

Red side 0.243 0.108 2.86 ~

Blue side 0.267 0.066 - 2.25

Table 3. The initial concentrations of isoprene experiment, (September 29,1991)

NO ppmC NO2 ppmC Isoprene ppmC

Red side 0.267 0.066 6.00

Blue side 0.267 0.066 2.00

increase or check on the sampling efficiency. The flow rate was controlled at 0.52

1/min, and the sampling time was 40 minutes. After the sample was collected, theabsorbing solvent (10ml ACN) was carefully transferred into a 25ml volumetricflask.

D. RESULTS AND DISCUSSION

The samples were analyzed by the method described before. 0.2ml HCl and2ml saturated DNPH solution were added to a 25ml flask, and the mix was filled

with ACN to the mark. After a 2-hour reaction, the sample was injected into HPLCcolumn. The results are shown in Tables 4 and 5 and plotted in Figures 16, 17 and18.

CALCULATION: The concentration A in parts per million carbon (ppmC,

v/v) of glycolaldehyde or hydroxyacetone is calculated according to the followingequation:

A = (Ci*Vs*RT*Ni)/(t*F*Mi*P)

where Q = concentration in ug/ml of the ith carbonyl DNPH derivativein the sample solution (25ml ACN)

Vs = volume of sample solution in ml (25ml ACN)

R = gas constant in (L-atm)/(deg-mole)

T = Temperature in degree

Ni = number of carbon atoms in a molecular of the ith carbonyl

t = sampling time in minute

F = flow rate in L/min

Mj = molecular weight of the ith carbonyl DNPH derivative

P = total pressure in atmosphere

Table 4. The analytical data for ethene and propene experiments-^

Sampling #in Glycolaldehyde

time senes"^ Peak height Cone, in Cone, in

start 0.005^ ACN ug/ml air ppmC

Ethene 9:50 1 N/d4 ~ ~

2 N/D — ~

11:20 1 7 0.021 0.026

2 N/D ~ ~

13:30 1 10 0.031 0.039

2 N/D •• __

Propene 10:40am 1 N/D -- —

2 N/D -- -

12:35 1 4 0.012 0.015

2 4 0.015 0.015

13:30 1 7 0.021 0.026

2 N/D ~ —

1. During the ethene and propene experiments, no hydroxyacetone was detected.

2. # in series means the position of a bubble in series.

3. 0.005 is the detector attenuation.

2. N/D means that the compound is not detectable by using this method.

Table 5. The analytical data for isoprene experiment

Sampling Glycolaldehyde Hydroxyacetone

time Peak Cone. Cone. Peak Cone. Cone.

start height inACN in air height inACN in air

0.005^ ug/ml ppmC 0.005 ug/ml ppmC

Red 9:11 N/D^ — — N/D ~ —

side 11:57 19 0.08 0.098 21.5 0.45 0.628

14:30 11 0.045 0.054 14 0.30 0.420

Blue 11:10 7 0.028 0.034 12 0.25 0.349

side 12:40 6 0.025 0.030 9 0.18 0.251

15:13 10.5 0.042 0.052 9.5 0.19 0.265

1.0.005 is the detector attenuation.

2. N/D means that the compound is not detectable by using this method.

"toi-+.>

c0)ocoo

Concentration Profile of Glycolaldehyde in Ethene and Propene ExperimentOctober?, 1991

0.04 -

oEQ.Q.S 0.03 -

0.02 -

0.01 -

time (hr.)

a ethene (red side) + propene (blue side)

Figure 16. Concentration-time profile of glycolaldehyde in ethene/NOx andpropene/NOx experiment.

ca>ucou

Concentration Profile of Glycolaldehyde in Isoprene ExperimentSeptember 29, 1991

time (EST hour)a Red side + Blue side

Figure 17. Concentration-time profile of glycolaldehyde in isoprene/NOxexperiment.

c0)ocoo

Concentration Profile of Hydroxyacetone in Isoprene Experiment

September 29,1991

TIME hour

D Red side + Blue side

Figure 18. Concentration-time profile of hydroxyacetone in isoprene/NOxexperiment.

COMPOUND IDENTIFICATION. The carbonyl compounds in the samples

were identified by comparing their retention times with standard samples. Figure

19.a shows the graph of one sample collected in the isoprene experiment (11:57am,

red side). Figure 19.b shows the blank sample prepared for isoprene experiment.

After being analyzed, 50 ul of this sample was spiked by 40 ul hydroxyacetone-

DNPH standard solution ( with a concentration of 10.7 ug/ml) and 10 ul

glycolaldehyde-DNPH standard solution (with a concentration of 18.9 ug/ml).

Then, 25 ul of this mix was injected into HPLC. The peaks corresponding to

glycolaldehyde (retention time at 13 min.) and hydroxyacetone (retention time at 19min.) were increased significantly (see Figure 20). Through this process, it can be

concluded that glycolaldehyde and hydroxyacetone are two of the products of the

isoprene photolysis. Similarly, glycolaldehyde was identified as one of the products

of the ethene/NOx and propene/NOx experiments, but no hydroxyacetone was

observed in both systems. Figure 21.a shows a graph of one sample collected in the

ethene experiment and no hydroxyacetone peak was shown up. Figure 21.b shows

the blank sample prepared for ethene experiment.

THE TREND OF CHANGE OF CONCENTRATION OF THE TWO

COMPOUNDS: Figure 17 and 18 show the changes of the concentrations of

glycolaldehyde and hydroxyacetone in the photochemical reactions of isoprene

during day time. In Figure 18, the patterns of the concentration changes of

hydroxyacetone are quite consistent across different concentrations of isoprene

injected in the chambers. Both reached the highest levels of the day around noon,

• 48

fl^.* jfn2z>

Figure 19.a. Chromatogram of a sample collected in isoprene/NOx experiment.Peaks: (1) and (2) unknowns; (4) glycolaldehyde; (6) hydroxyacetone; others, DNPHand its contaminates. Range: 0.005

I----------------:-_-

ra____,-

fA%^^EE-"ri U 8- 4- 0

Figure 19.b. Chromatogram of a blank sample prepared in isoprene/NOxexperiment. Range: 0.005

-S,o- -X—rp^

DiC;___

Figure 20. Chromatogram of the sample as in Fig. 19.a after spiked with standardsolutions of glycolaldehyde and hydroxyacetone. Peaks 3 (glycolaldehyde) and 5(hydroxyacetone) increase significantly.

"«: =V!5-'=- -'2!^^;-'*'^^'»"«V-^

14- (3 lt> S ^ 4Time ,(rr»in)

Figure 21.a. Chromatogram of a sample collected in ethene/NOx experiment.Peaks: (2) glycolaldehyde; others, DNPH and its contaminates. Note: nohydroxyacetone peak (19 min). Range:0.005

"Tn yr^\ [ l-U'^w |) .

Figure 21.b. Chromatogram of a blank sample prepared in ethene/NOxexperiment. Range: 0,005

and moderate levels at mid afternoon. Those concentrations in the early morning

however were lower than the detection limit(that means the concentration of

glycolaldehyde in the chamber air is lower than 0.012ppmC, and the concentrationof hydroxyacetone in the chamber air is less than 0.033ppmC). The patterns of theconcentration changes of glycolaldehyde during the day are somewhat divergent

(see Figure 17). In the red side of the chamber, where isoprene of higherconcentration was injected, the change of concentration of glycolaldehyde over the

day showed a similar pattern as those changes in propene experiment (Figure 16)

i.e., higher at noon and lower in the morning and afternoon. In the case that

isoprene of lower concentration was presented, however, the change of

concentration over time did not show a clear tendency. Inappropriate distribution

of this sampling times may take account of this ambiguity, because, due to sometechnical problems, the sampling hours tend to cluster around noon. The profiles ofreaction conditions of isoprene/NOx experiment in the outdoor smog chamber are

shown in Figure 22.

Based on the assumption that no interferences existed in the isoprene system,

the experimental results show that the amounts of glycolaldehyde and

hydroxyacetone formed were 4.2% and 17%, respectively, of the isoprene reacted in

the Red side , and 6% and 30%, respectively, of the isoprene reacted in the Blueside of the chamber.

The same method was also applied to detecting the concentrations of

hydroxyacetone and glycolaldehyde during the photochemical reactions of ethene

and propene. The profiles of reaction conditions of ethene/NOx and propene/NOxexperiments in the outdoor smog chamber are shown in Figure 23. Results show

that the formations of glycolaldehyde account for 16.5% of the ethene consumed

SegFile Data

^r^^5^SW^^^95??^!ai^^-:p?^--3*

(/T0)-o

"xOc0D)O

0.7 —

I ' I ' I ' I ' I ' I ' I________^ 6.00ppmC Isoprene._«._____ 2.00ppmCIsoprene

I ' i I ' I M ' I '29-Sep-91

/•"'I. NO,

X I I I I I I I9 10 11 12 13 14 15

HOURS,EDT

-I 0.60.5

-4 0.4

•A 0.1

T3•o

16 17 18 19

d.E(DI-

L' 1 ' 1 ' 1 1 ' 1 ' 1 1 ' 1 ' L' 1 ' 1 ' 1 ' 1 ' 1

UV N. 1TSR \|

1 Dewf

1 1 1 1 1 1 1 . 1 . 1 . 1 . 1 , 1 . K^lF, 1 ,.jX'

9 10 11 12 13 14 15 16 17 18 19

HOURS,EDT

Figure 22. Isoprene/NOx experiment in UNC outdoor smog chamber.

SegRle Data

I ' I ' I ' I ' I ' I 'I--------------------0.243 ppmC Ethene______--- 0.249ppmCPropene

I I • I ' I ' I ' M07-Oct-91 .

5) 0.4 0.4 -6

HOURS,EDT

I I .I ' I ' I ' I ' I ' I I ' I ' I ' I ' I ' I

Temp Temp

DewPDewP

I I I I I

HOURS,EDT

Figure 23. Ethene/NOx and propene/NOx experiments in UNC outdoor smogchamber.

and for 18% of the propene reacted. The concentration of hydroxyacetone was

either beyond the detection limit of this method or not existed at all, since it was notdetected.

5. CONCLUSIONS

Based on the previous results and discussions, several conclusions can be

arrived at: First, unlike reacting vidth simple aldehydes and ketones, DNPH

substitutes both carbonyl (C = 0) and hydroxyl (-OH) functional groups during

derivatizing hydroxylated carbonyl. The speed and yield of this reaction are

critically depended on four factors i.e. temperature, acidity, DNPH concentration,

and reaction time.

Second, result of this study shows that DNPH substitutes both of the two

carbonyls when it reacts with dicarbonyl compounds (O = C-C = O).

Third, since the DNPH-derivatives of hydroxylated carbonyl and dicarbonyl

are identical, presence of dicarbonyl compounds will interferes with the detection

of hydroxylated carbonyl when both of them exist in the same system. In this study,

these interferences are quantitatively identified and therefore, can be eliminated

when calculating the concentration of hydroxylated carbonyls.

In general, an analytical method for measurements of hydroxyacetone and

glycolaldehyde has been developed with satisfactory sensitivity and selectivity. This

method is not necessarily limited to the analysis of the two compounds mentioned

above, it can also be adopted for measuring a series of hydroxylated carbonyl

compounds which are hypothesized to be important intermediately photochemical

products. This method therefore could make potential contribution to the study of

mechanism of air photochemistry.

6. FUTURE WORK

In order to apply this method to the chamber experiment studying the

photochemical reaction mechanism, the following problems needs to be solved.

A. DEVELOPING A SAMPLING TECHNIQUE

Sampling technique plays an important role in the application of this

analytical method. However, there were no well developed sampling method

available in the application of this analytical method. This may explain the

difficulties and problems that this method had encountered during its application.

In order to obtain evenly distributed samples and more information on the

concentration profiles, two air pumps are needed to take the air samples from both

sides of the chamber simulatanously. Sampling recovery also needs to be identified

in developing a sampling technique in the future.

B. IMPROVING THE SENSITIVITY OF THE ANALYTICAL METHOD

Previous experiments showed that longer reaction time and higher

temperature can increase the yields of derivative during derivatization. Sensitivity

of the method can therefore be improved by identifying appropriate reaction time

and temperature. Based on the experience of running chember experiments, Irecommand that the reaction time should be extended to 24 hours and the reaction

temperature could be set up at 60^C when heating box is avaliable. In order toavoid the presence of the contamination peaks, DNPH stock solution needs to be

stored and handled properly because it can easily absorb the carbonyl compounds

existing in lab air.

C. TESTING THE APPLICABILITY OF THIS ANALYTICAL METHOD TO

OTHER HYDROXYLATED CARBONYL COMPOUND

After the discovery of the reaction mechanism of hydroxyacetone and

glycolaldehyde with DNPH, this analytical method is, in theory, applicable to other

hydroxylated carbonyls. Acidity is believed to be a crucial factor to the

derivatization procedure. The applicabiUty of this method to other hydroxylated

carbonyl compounds needs to be tested in the future*.

* During this study, the reactivity of glycolic acid with DNPH was investigated. Theresult shows that DNPH can not derivatize glycolic acid under the acidic conditionsdescribed in the text.

BIBLIOGRAPHY

Atkison, R., "Gas-phase Tropospheric Chemistry of Organic Compounds: AReview", Atmospheric Environment, Vol. 24A, No. 1, pp. 1-41,1990.

Atkison, R., S.M. Aschmann, E.C. Tuazon, J. Arey, and B. Zielinska, "Formation of3-Methylfuran from the Gas-Phase Reaction of OH Radicals with Isoprene and theRate Constant for its Reaction with the OH Radical", International Journal ofChemical Kinetics, Vol. 21, pp. 593-604,1989.

Casazza, J.P. and J.L. Fu, "Measurement of Acetol in Serum", AnalyticalBiochemistry, Vol.148, pp. 344-348,1985.

Edelkraut, F. and U. Brockmann, "Simultaneous Determination of CarboxyUc Acidsand Carbonyl Compounds in Estuaries by HPLC", Chromatographia, Vol. 30, No.7/8, October 1990.

Grosjean, D., Fung, K. and Atkinson, R. "Measurements of Aldehydes in the AirEnvironment", the 73rd Annual Meeting of the APCA.

Isidorov, V,A., I.G. Zenkevich and B.V. loffe, "Volatile Organic compounds in theAtmosphere of Forests", Atmospheric Environment Vol. 19, No. 1, pp. 1-8,1985.

Jeffries, H., Dept. of ESE, UNC-CH, by personal communication.

Kamen, R., Dept. of ESE, UNC-CH, by personal communication.

Kieber, D.J. and K, Mopper, "Trace Determination of a-Keto Acids in NaturalWaters", Analytica Chimica Acta, Vol. 183, pp. 129-140,1986.

Kuntz, R., W. Lx)nneman, G. Namie and L.A. Hull, "Rapid Determination ofAldehydes in Air Analysis", Analytical Letters, Vol. 13(A16), pp. 1409-1415,1980.

Kuwata, K., Uebori, M., Yamasaki, Y. and Kuge, Y. "Determination of AliphaticAldehydes in Air by Liquid Chromatography", Anal. Chem.,Vol.55, pp. 2013-2016,1983.

Kuwata, K, Uebori, M. and Yamasaki, Y., "Determination of Aliphatic andAromatic Aldehydes in Polluted Airs as their 2,4-Dinitrophenylhydrazones by High

Performance liguid Chromatography", J. Chromatogr. Sci., Vol.17, pp. 264-268,1979.

Lamb, B., A. Guenther, D. Gay and H. Westberg, " A National Inventory ofBiogenic Hydrocarbon Emissions", Atmospheric Environment Vol. 21, No. 8, pp.1695-1705,1987.

McLafferty F.W., Stauffer D.B., "The Wiley/NBS Registry of Mass Spectral Data",Wiley-Interscience Pulication, New York, 1988, Vol.1, 5.,

Pigman, W., "The Carbonhydrates", Academic Press Inc. Publishers, 1957.

Smith, D.F., T.E. Kleindienst and E.E. Hudgens, "Improved High-PerformanceLiquid Chromatographic Method for Artifact-free Measurements of Aldehydes inthe Presence of Ozone Using 2,4-Dinitrophenylhydrazine", Journal ofChromatography, Vol.483, pp. 431-436,1989

Tejada S.B., "Evaluation of Silica Gel Cartridge Coated in situ with Acidified 2,4-Dinitrophenylhydrazine for Sampling Aldehydes and Ketones in Air", Intern. J.Environ, and Chem., Vol. 26, pp. 167-185,1986.

Tuazon, E.C. and R. Atkison, "A Product Study of the Gas- Phase Reaction ofMethyl Vinyl Ketone with the OH Radical in the Presence of NOx", InternationalJournal of Chemical Kinetics, Vol. 21, pp. 1141-1152,1989.

Tuazon, E.C. and R. Atkison, "A Product Study of the Gas- Phase Reaction ofMethacrolein with the OH Radical in the Presence of NOx", International Journalof Chemical Kinetics, Vol. 22, pp. 591-602,1990.

Tuazon, E.C. and R. Atkison, "A Product Study of the Gas- Phase Reaction ofIsoprene with the OH Radical in the Presence of NOx", International Journal ofChemical Kinetics, Vol. 22, pp. 1221-1236, 1990.

APPENDIX

Terms used:

I IHydroxylated carbonyl compound: — C — C

glycolaldehyde: CH2CHOH O

hydroxy acetone: CH2CCH3

Dicarbonyl compound: — C — CO O

glyoxal: CHCH

0 01 I

methyl-glyoxal: C H C CH 3

2,4-DinitrophenyIhydrazine (DNPH): H g N— nh —fi \— NO2

a technical report submitted to the faculty of the

Documents

lion fish rov - faculty. leon king rov technical...

thesis submitted to the faculty of the virginia

research report a research report submitted to the faculty

a dissertation submitted to the faculty of the …

major qualifying project submitted to the faculty of

technical university of lisbon faculty of veterinary...

thesis submitted to the faculty of the

by dissertation submitted to the faculty of the for the...

faculty management system - web viewproject report. on....

submitted to the faculty of liberty university in partial

a research report submitted to the faculty in

a dissertation submitted to the faculty of the graduate

submitted to the faculty of the graduate school of...

technical note submitted to gÉotechnique

a thesis submitted to the faculty of the graduate …

submitted to the graduate faculty of texas tech university

a doctoral dissertation submitted to the faculty of

a thesis submitted to letters and humanities faculty in...

attrition a dissertation submitted to the faculty the

restraints dissertation submitted to the faculty of...