Nov 3, 2008 - affected adsorption, specifically for the xylene isomers. In addition ...... was found to form 2-pentene by removal of C2 and C3 hydrogen atoms.
Faculty of Science Department of Applied Chemistry
Adsorbed Organic Species on Inhalable γ-Alumina Particles
Anita D’Angelo
This report is submitted as fulfillment of the requirements for the unit Chemistry Honours Dissertation
The work was performed under the supervision of Dr. Dave Fleming and Dr. Franca Jones
November 2008
Declaration
This dissertation contains no material which has been accepted for the award of any other degree or diploma in any university. To the best of my knowledge and belief this dissertation contains no material previously published by any other person except where due acknowledgement has been made.
Signature:
.....................................................
Date:
November 3rd 2008
Acknowledgements
Before mentioning people directly I would first like to thank the Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC CARE) for awarding me with a scholarship to fund this project. The most important people I would like to thank are my supervisors Dr. Dave Fleming and Dr. Franca Jones. Dave Fleming for allowing me to assist in his adsorption work from which this honours project originates. From his experience and wealth of knowledge I have learnt so much. Franca Jones for her useful advice and tireless efforts to improve and correct my work. Her advice and assistance on both writing and the project was extremely beneficial. I am a better scientific writer because of it. Also, the section leader of the Emergency Response Section of CCWA, Dr. Steve Wilkinson whom allowed me to undertake this big task of studying honours. Finally, everyone in the Emergency Response section for their continued support and much appreciated advice.
i
Abstract Alumina refining is a multi-billion dollar industry in Western Australia (W.A.) where residents living in close proximity to the alumina refineries have raised concerns and health issues regarding refinery emissions. Typically, alumina refinery emission levels are measured separately as either gaseous or particulate matter concentrations which do not include the combination of both such as organics adsorbed onto particulates. The adsorption and desorption of non polar (n-hexane, benzene, toluene, o-, m-, p- xylene) and polar (propionaldehyde, 2-butanone, benzaldehyde, acetophenone, benzyl alcohol) VOCs on γ-alumina particles was invested using thermal desorption gas chromatography mass spectrometry (TD-GCMS). Adsorption of both non polar and polar VOCs was shown to be dependent not only on the strength of intermolecular interactions with surface aluminium cations and oxygen anions but competition with water molecules. Steric factors also affected adsorption, specifically for the xylene isomers. In addition, adsorbed benzyl alcohol catalytically oxidised to benzaldehyde however this is likely to be due to the analytical thermal step. Desorption of non polar VOCs was dependent on both competition with water molecules and the adsorbing VOCs ability to enter γ-alumina surface pores. The polar VOCs were not as affected by their ability to fit into the pore network as they were able to form stronger interactions with the surface. These residual VOCs have been shown theoretically to concentrate on the γ-alumina surface indicating that a new approach to determining exposure standards of air pollutants is critical. The increased concentration of these VOCs have the capability to produce adverse health effects so possible pathways in which volatile organic molecules can enter the human respiratory system are discussed.
ii
List of Abbreviations α-Al2O3
Alpha alumina
γ-Al2O3
Gamma alumina
γ΄-Al2O3
Gamma prime alumina
δ-Al2O3
Delta alumina
η-Al2O3
Eta alumina
θ-Al2O3
Theta alumina
θ”-Al2 O3
Theta double prime alumina
κ-Al2 O3
Kappa alumina
λ-Al2O3
Lamda alumina
χ-Al2O3
Chi alumina
γ-Al(OH)3
Gibbsite polymorph
α-AlO(OH)
Disapore
γ-AlO(OH)
Boehmite
GCMS
Gas chromatography mass spectrometry
LCMS
Liquid chromatography mass spectrometry
PM
Particulate matter
SEM
Scanning electron microscopy
TD
Thermal desorption
TD-GCMS
Thermal desorption gas chromatography mass spectrometry
TWA
Time weighted average
USEPA
United States Environmental Protection Agency
VOC
Volatile organic compounds
XRD
X-ray diffraction
iii
Contents Acknowledgments
i
Abstract
ii
List of Abbreviations
iii
Contents
iv
List of Figures
vi
List of Tables
vii
1.0 Introduction
1
1.1 Bauxite and the Alumina Industry in Western Australia
1
1.2 Alumina Transition States and their Dehydration Pathway
2
1.2.1 Precursor Properties
2
1.3 Adsorption
3
1.4 Emissions
3
1.4.1 The Lung
3
1.4.2 Particulate Matter
4
1.4.3 Particulate Matter in the Lung
5
1.4.4 Particulate Matter and Adsorbed Organics
5
1.5 Perceived Health Effects
6
1.6 Project Summary
6
2.0 Materials and Methods
7
2.1 Thermal Desorption Tube Preparation
7
2.2 Gas Standards
7
2.3 Calibration Standards
7
2.4 Adsorption Studies
8
2.5 Desorption Studies
9
2.6 Quantitation
9
2.6.1 Thermal Desorption Gas Chromatography Mass Spectrometry (TDGCMS) Analysis 2.7 Particulate Characterisation 2.7.1 Scanning Electron Microscopy (SEM) Analysis 3.0 Results 3.1 Statistics
10 10 11 12 12
iv
3.2 Adsorption of Non Polar VOCs
12
3.3 Adsorption of Polar VOCs
13
3.3.1 Benzyl Alcohol
14
3.4 Desorption of Non Polar VOCs
15
3.5 Desorption of Polar VOCs
16
4.0 Discussion 4.1 Adsorption of Non Polar VOCs
18 19
4.1.1 Water Vapour
19
4.1.2 n-Hexane
20
4.1.3 Aromatic Non Polar VOCs
21
4.2 Discrepancies in Non Polar VOC Adsorption 4.2.1 Steric Effects for Adsorption
22 22
4.3 Adsorption of Polar VOCs
23
4.3.1 Propionaldehyde
23
4.3.2 2-Butanone
24
4.3.3 Benzaldehyde and Acetophenone
25
4.3.4 Benzyl Alcohol
26
4.4 Desorption of Non Polar VOCs
27
4.4.1 n-Hexane
28
4.4.2 Non Polar Aromatics
28
4.5 Desorption of Polar VOCs
30
4.6 Inhaled Adsorbed Species
31
5.0 Conclusions
32
Adsorption Behaviour of Alumina
32
Health Implications
32
5.1 Future Work 6.0 References
33 34
Appendix I – Physical Properties of the VOCs used in Adsorption and Desorption Experiments
43
Appendix II – Concentration of p-Xylene Released in one Alveoli per 10 µm Diameter γ-Alumina Particle
44
v
List of Figures Figure 1
Schematic of dust chamber used in adsorption and desorption studies.
9
Figure 2
SEM image of 'typical' γ-alumina particle.
11
Figure 3
Adsorption of non polar VOCs on γ-alumina.
12
Figure 4
Adsorption of polar VOCs on γ-alumina.
13
Figure 5
TD-GCMS chromatogram of benzyl alcohol 1050 ng/µL standard injected onto Tenax TA only.
14
Figure 6
Mass spectrum of benzyl alcohol peak at 13.850 minutes.
14
Figure 7
TD-GCMS chromatogram of benzyl alcohol (13.850 min) and benzaldehyde (14.462 min) in 2 hour adsorption sample.
15
Mass spectrum of benzaldehyde (14.462 minutes) in 2 hour adsorption sample.
15
Figure 9
Residual non polar VOCs present on γ-alumina.
15
Figure 10
Residual polar VOCs present on γ-alumina.
17
Figure 11
Location of aluminium (black) ions and oxygen (white) ions in the γ-alumina structure.
18
Relationship between the boiling point and adsorption after 120 minutes exposure.
22
Probable structure of adsorbed species after benzaldehyde exposure to γ-alumina.
25
Enolate anion formed after acetophenone exposure to γ-alumina.
25
Figure 8
Figure 12
Figure 13
Figure 14
vi
List of Tables
Table 1
Non polar and polar VOC standard concentrations and working VOC standard preparation.
8
Table 2
Adsorption of non polar VOCs on γ-alumina.
13
Table 3
Adsorption of polar VOCs on γ-alumina.
14
Table 4
Residual non polar VOCs present on γ-alumina.
16
Table 5
Residual polar VOCs present on γ-alumina.
17
vii
1.0 Introduction This project describes the pathway in which volatile organic molecules can enter the human respiratory system by means of inhaled γ-alumina particles. The model proposed to describe the process is based on the asthma inhaler in which medication is delivered adsorbed onto particles. Health complaints received from the community in areas surrounding the Alcoa Wagerup refinery cannot be explained by data obtained in monitoring programs as stack emission levels are shown to be within recommended limits (Environment and Public Affairs Committee 2004; Coffey et al 2005). Data present in the 2002 Wagerup Refinery Air Emissions Inventory is evaluated against a time weighted average (TWA) which is defined as the amount a worker can be exposed to daily without adverse effects (Department of Education, Employment and Workplace Relations 1995). These TWAs are for exposure over an eight hour and five day week and cannot be accurately used for evaluation of community exposure. Stack emission levels are also measured separately as either gaseous or particulate matter concentrations which do not include the combination of both such as organics adsorbed onto particulates. Although the data taken from regions surrounding the Alcoa Wagerup plant refinery are below safe guidelines, the presence of organics adsorbed onto particles may provide an alternative explanation for alleged adverse health effects. One of the main alumina products emitted from the Wagerup refinery stacks consists of γalumina particulate matter that is produced from bauxite.
1.1 Bauxite and the Alumina Industry in Western Australia Bauxite is the most common ore of aluminium and is comprised of aluminium hydroxides of gibbsite polymorph (γ-Al(OH)3), boehmite γ-AlO(OH), disapore α-AlO(OH) and various other minerals and organics (Wefers & Misra 1987; Kloprogge et al 2006). The gibbsite polymorph and boehmite are the major constituents of bauxite. Deposits of bauxite (alumina-rich
laterite) in Western Australia are mined at Jarrahdale, Huntly, Willowdale
and Boddington in the Darling Ranges. Seven alumina refineries are currently operated in Australia with four located in the south west of Western Australia. Refineries are located in Kwinana, Pinjarra, Wagerup (Coffey & Ioppolo-Armanios 2004) and Worsley, 13 kilometers North West of Collie. In 2006, Western Australia produced 11.6 million tonnes of alumina accounting for 17 % of global production (Department of Industry and 1
Resources 2008). The gibbsite polymorph is the major aluminium hydroxide constituent of bauxite mined in Western Australia.
1.2 Alumina Transition States and their Dehydration Pathway Gibbsite dehydrates to form aluminium oxide or alumina (Al2O3). Alumina exists in various metastable transition phases including γ, λ (monoclinic), η (cubic), θ (monoclinic), δ (either tetragonal or orthorhombic), κ (orthorhombic) and χ (hexagonal). Other alumina polymorphs such as θ” (monoclinic) and γ΄-alumina, a triple state of γ-alumina have also been identified (Levin, Gemming & Brandon 1998; Paglia et al 2004). These polymorphs are not well defined, for example, γ-alumina has been reported to exist at temperatures between 450 – 700 °C (Paglia et al 2004) and still be present at temperatures of 1150 °C (Bousquet et al 2008). Through dehydration all highly porous oxide phases transform to the thermodynamically stable phase α-alumina (Levin, Gemming & Brandon 1998). The dehydration pathway from boehmite is (Wilson & McConnell 1980; Krokidis et al 2001; Paglia et al 2004): γ-AlO(OH) → γ-Al2O3 → δ-Al2O3 → θ-Al2O3 → α-Al2 O3 Boehmite’s dehydration pathway is extensively reported in literature compared to that of gibbsite. Little work has been done to characterise the dehydration pathway of gibbsite despite it being the primary product in Western Australia Bayer process refineries. Whittington and Ilievski (2004) determined the dehydration pathway to occur through χAl2O3 using gibbsite (91 µm) similar to that produced in Bayer process plants as below: γ-Al(OH)3 → χ-Al2O3 → γ-Al2O3 → θ-Al2O3 → α-Al2O3 1.2.1 Precursor Properties Preparation of γ-alumina through dehydration is influenced by the boehmite or gibbsite precursor properties such as crystallinity, particle size and thermal treatment. These factors effect the structural rearrangement of lattice cations and anions during phase transition which has contributed to uncertainty in the γ-alumina structure. Uncertainty in the γalumina structure is frequently thought to be due to the position of the aluminium ions within the unit cell. Recent work has shown stacking faults occur in the oxygen sub lattice 2
and is thought to result from the retention of some boehmite precursor characteristics (Paglia, Bozin & Billinge 2006). Poor ordering exhibited by γ-alumina has resulted in the characterisation of its structure to be problematic. As a result numerous contradictory assignments of the γ-alumina structure are currently present in the literature. 1.3 Adsorption Adsorption is defined by the International Union of Pure and Applied Chemistry (IUPAC) as “an increase in the concentration of a dissolved substance at the interface of a condensed and a liquid phase due to the operation of surface forces. Adsorption can also occur at the interface of a condensed and a gaseous phase” (McNaught & Wilkinson 1997). Molecules adsorb on a surface via either a physisorption or chemisorption process. Physisorption (physical adsorption) occurs via weak van der Waals forces between the adsorbed molecule and the substrate which is reversible. There is no significant orbital distortion of the adsorbed molecule or substrate. Chemisorption occurs through the formation of chemical bonds or electrostatic interactions that is reversible only at high temperatures. In a gas-particle system covalent bonds are typically formed (Atkins & de Paula 2006). Processes by which adsorption occurs depend on the adsorbing molecule. Molecules which readily adsorb to the γ-alumina surface are present in alumina refinery emissions.
1.4 Emissions Stack emissions from the alumina production process are comprised of particulate matter, alcohols, aromatics, carbonyl and volatile organic compounds. The 2002 Wagerup refinery air emissions inventory monitoring program list these compounds to include toluene, benzene, isomers of xylene, acetaldehyde, acetone, benzaldehyde and particulate matter sizes ranging from less than 2.5 µm and less than 10 µm (Coffey et al 2005). Many of the compounds, including particulate matter, emitted are harmful if inhaled into the lung.
1.4.1 The Lung The lower respiratory system includes the trachea, bronchi and lungs. The respiratory portion is comprised of bronchioles and alveoli. Below is listed the major parts of the human lung (Patton 1996):
3
AIRWAY BRANCHING IN THE HUMAN LUNG Trachea Bronchi Bronchioles Terminal Bronchioles Respiratory Bronchioles Alveoli Ducts Alveolar Sacs
During respiration, different sized particles are removed by a number of mechanisms including deposition as well as the bronchial ciliary escalator. Air passes through the primary bronchi, bronchioles then alveoli. When air is inhaled, gas exchange between the air and blood for respiration occurs through the alveoli (which make up greater than 99 % of the lung surface area) due to their short oxygen to blood diffusion distance (Tortora & Grabowski 2003). If atmospheric pollutants enter the lung they can cause significant adverse health effects (Mazzarella et al 2007).
1.4.2 Particulate Matter The US Environmental Protection Agency first established air quality standards for particulate matter in 1971 as part of the Clean Air Act (United States Environmental Protection Agency 1972). In 1997 separate standards were established for both respirable coarse and fine particulate matter (USEPA 1997). Particulate matter (PM) is particle pollution comprised of many different small particles and liquid droplets. Respirable particulate matter measures less than 10 µm in aerodynamic diameter (PM10) and is produced from fossil fuel combustion in cars, fires and manufacturing and power generation (Bai et al 2007). It includes inhalable particles measuring between 10 – 2.5 µm, fine particles measuring less than 2.5 µm and ultrafine (nano) particles less than 0.1 µm. Particle matter greater than 50 µm in diameter (PM50) is not included in air quality
4
standards for particulate matter as it is generally not considered to cause significant adverse health effects.
1.4.3 Particulate Matter in the Lung Numerous studies have shown an increase in cardiovascular disease, decreased lung function, lung cancer and cardiopulmonary mortality associated with air pollution particulate matter (Arden Pope III et al 2002; Kelly 2003; Bai et al 2007). In the study by Arden Pope III et al (2002), exposure to particulate matter less than 2.5 μm showed an increase of 6 % and 8 % in cardiopulmonary and lung cancer mortality respectively. This was observed for each 10 μg/m3 increase in particulate matter after taking into account lifestyle effects such as weight and cigarette smoking. These greater adverse health effects experienced from smaller sized particles due to their ability to penetrate into deeper regions of the lung and in greater deposition quantities. Deposition of PM10 occurs in the tracheobronchial and alveolar regions of the lower respiratory system whereas irrespirable particle matter (PM50) deposit in the upper airways. Particle surface area and hence diameter has also been shown to have a greater effect on the degree of inflammation, dust translocation and particle clearance from the lung in comparison to particle solubility or toxicity (Cullen et al 2000). The larger surface area of smaller sized particulate matter results in greater interaction with lung alveoli (Bai et al 2007). Particles once in the lung aggregate on the alveoli liquid lining before macrophage engulfment, however insoluble particles are able to cause adverse health effects as they interact with the epithelial cells before removal (Bastacky et al 1995; Dobbs & Johnson 2007). In the alveolar region of the lung 0.1 µm particles can be retained for over 48 hours with 3 % of particles removed within 24 hours and 25 % of the 0.1 µm particles removed within the 48 hours by alveolar macrophages (Moller et al 2007). A greater adverse health effect can occur through the presence of organics which readily adsorb to particle surfaces.
1.4.4 Particulate Matter and Adsorbed Organics Particulate matter in the lung may have increased adverse health effects from the presence of adsorbed organics (Mazzarella et al 2007). Limited work has been carried out in the adsorption of organic species onto mineral dusts, including alumina, where the literature focuses on the adsorption of trace atmospheric compounds such as SO2 (Fellner et al 5
2006), NO2 (Lisachenko et al 2007) and ozone (Thomas et al 1997). As trace atmospheric volatile organic compounds are known to adsorb onto the surface of mineral dusts (Usher, Michel & Grassian 2003), this may well apply to alumina particles. Particles with adsorbed organic compounds have been shown to cause cytotoxic effects to human lung tissue (Mazzarella et al 2007).
1.5 Perceived Health Effects Health complaints received around refineries may be attributed to the detection of refinery odours or enhanced sensitivity of some community members to refinery chemicals (Donoghue & Cullen 2007). An individual’s sensitivity and response has been linked to their perception of the chemical hazard and variables such as age, gender or exposure history (Dalton 2003). Luginaah et al (2002) observed a strong relationship between the reporting of adverse health effects and odour perception. The reporting of adverse health effects was also linked to an individual’s annoyance at refinery emissions believed to cause adverse health effects. The respiratory health complaints experienced by the Wagerup community contradict particulate matter and ambient air quality data obtained from monitoring programs. Although the data taken from regions surrounding the Alcoa Wagerup plant refinery are below safe guidelines, the presence of organics adsorbed onto particles and their photocatalytic products may provide an alternative explanation.
1.6 Project Summary During the processing of alumina, numerous volatile organic compounds (VOCs) are produced which may adsorb onto the alumina surface. Organic compounds adsorbed onto particulate matter have been shown to cause cytotoxic effects to human lung tissue (Mazzarella et al 2007). This project characterises the alumina by scanning electron mircroscopy (SEM) followed by exposing γ-alumina, the major alumina phase, to a range of polar and non-polar VOCs for specified times. Analysis is conducted by thermal desorption gas chromatography mass spectrometry (TD-GCMS). From this work the adsorptive capacity of the alumina particles was determined under ambient temperature and pressure with VOCs known to be produced by the alumina refinery. Although the project does not focus on the cytotoxic effects of organics adsorbed onto alumina, this preliminary work may lead to further studies on potential health effects.
6
2.0 Materials and Methods 2.1 Thermal Desorption Tube Preparation Tenax TA (80/100 mesh, Grace Division Discovery Sciences) cleaned using methanol (≥ 99.7 %, Ajax Finechem) in a Soxhlet apparatus (5 hr) was added (50 mg) to a 0.6 cm O.D. × 8.9 cm length glass thermal desorption (TD) tube (Supelco, part no. 25084). TD tubes were cleaned by thermal desorption and weighed prior to use.
2.2 Gas Standards The VOCs selected are those known to be present in the stack emissions from the Alcoa Wagerup Refinery and the concentrations selected for use are based on stack emission measurements (Coffey et al 2005). NATA certified gas standards of n-hexane (103 ppm ± 2 ppm), benzene (101 ppm ± 2 ppm) and toluene (97.2 ppm ± 1.9 ppm) with ultra high purity (UHP) nitrogen (99.999 %) make up gas were purchased from BOC Scientific (Sydney, Australia). Liquid o-xylene (98 %, Sigma), m-xylene (> 99 %, Sigma), p-xylene (≥ 99 %, Sigma), propionaldehyde (97 %, Aldrich), 2-butanone (99 %, Lancaster), benzaldehyde (≥ 99 %, Merck), acetophenone (≥ 98 %, Merck) and benzyl alcohol ≥( 97 %, Unilab) were used to produce gas standards of compounds not obtainable as gas standards from BOC Scientific. Gas standards were prepared from the liquid compounds by injecting approximately 18 µL of the liquid compound into a 104 L cylinder and pressurising it to approximately 600 psi with UHP nitrogen. Accurate concentrations of oxylene,
m-xylene and
p-xylene standards
prepared
were determined
by gas
chromatography mass spectrometry (GCMS) according to USEPA method TO14A with propionaldehyde, 2-butanone and benzaldehyde determined by high pressure liquid chromatography (HPLC) according to USEPA method TO11A (National Risk Management Research Laboratory 1999). Gas standards were used in both adsorption and desorption experiments.
2.3 Calibration Standards For each non polar VOC used in adsorption and desorption studies five standards (1 to 5) of approximately 5, 50, 130, 260 and 520 ng/µL were prepared in methanol ≥( 99.7 %, Ajax Finechem). Polar VOC standards were prepared in higher concentrations than those used for the non polar VOCs as they adsorbed greater to the γ-alumina. For each polar 7
VOC, standards (1 to 5) of approximately 50, 130, 260, 520 and 1050 ng/µL were prepared in methanol.
Table 1: Non polar and polar VOC standard concentrations and working VOC standard preparation Non Polar VOC Standard Standard Number 3 4
1
2
5
Concentration (ng/µL) 50 130 260
Polar VOC Standard
5
1
520
50
Added to 1mL of Each Non Polar VOC Standard 1 mL toluene-d8 1 mL methanol
1A
Working Non Polar VOC Standard 2A 3A 4A 5A
2
Standard Number 3 4
Concentration (ng/µL) 130 260 520
5
1050
Added to 1mL of Each Polar VOC Standard 1 mL toluene-d8 1 mL methanol
1A
Working Polar VOC Standard 2A 3A 4A
5A
Five working non polar VOC standards (1A to 5A) were prepared from each of the five non polar VOC standards where, 1 mL of non polar VOC standard (1 to 5) was added to 1 mL of deuterated toluene (toluene-d8) (Aldrich, catalog no. 26958-9) and 1 mL methanol (table 1). The toluene-d8 was used as an internal standard and was prepared at a concentration of 25 ng/µL in methanol. Five working polar VOC standards (1A to 5A) were prepared from each of the five polar VOC standards where, 1 mL of polar VOC standard (1 to 5) was added to 1 mL of deuterated toluene and 1 mL methanol. As the concentration of the compounds in the working standard are diluted by three from the addition of toluene-d8 and methanol, when 3 µL of one of the five working standards (1A to 5A) was injected onto a TD tube, the amount on the TD tube was that of the initial standard (standard 1 to 5). An internal standard was prepared with 1 mL of toluene-d8 and 2 mL methanol so when 3 µL was injected onto a TD tube, the amount on the TD tube was 25 ng.
2.4 Adsorption Studies The γ-alumina (0.3 g) was exposed in a dust chamber (figure 1) to an atmosphere of VOC standard for time intervals of 1, 2, 5, 10, 30, 60 and 120 minutes in the presence of water 8
vapour. VOC standard concentrations were prepared at approximately 100 ppm in UHP nitrogen and further diluted to approximately 3 ppm using an Entech Instruments dynamic dilutor, model 4600A. All results were normalised to 3 ppm.
Dynamic Diluter
Dust chamber
Diluent (N2)
VOC
Figure 1: Schematic of dust chamber used in adsorption and desorption studies.
The average concentration of water vapour delivered by the dynamic dilutor was determined by measurement with a photoacoustic multi-gas analyser (Filter SB 0527, Brüel & Kjær type 1302) and was found to be an average of 4000 ppm between 20 – 24 °C. The average laboratory temperature in which all experiments were undertaken was 22 °C.
2.5 Desorption Studies Organics were adsorbed onto γ-alumina (2.4 g) by exposure to a VOC atmosphere for 15 minutes in the absence of water vapour to maximise the amount of VOC adsorbed. A portion (0.3 g) of this γ-alumina was added to the dust chamber and adsorbed VOCs were desorbed with UHP nitrogen in the presence of water vapour (4000 ppm) for time intervals as in adsorption experiments.
2.6 Quantitation After exposure to the VOC standard, γ-alumina (0.5 g) was added to a TD tube between two sections of Tenax TA. Three sample TD tubes were prepared for each time interval. 9
TD tubes were then weighed to determine the amount of γ-alumina added. For each VOCs set of adsorption and desorption experiments, five standard TD tubes were prepared and analysed. To prepare the TD tube standard, a tube was injected with 3 µL of one of the five liquid working standards (1A to 5A). This was repeated with the remaining standards. Each TD tube containing a γ-alumina sample was injected with 3 µL of the prepared internal standard. VOCs present on the γ-alumina from adsorption and desorption studies were quantified using thermal desorption gas chromatography mass spectrometry (TDGCMS).
2.6.1 Thermal Desorption Gas Chromatography Mass Spectrometry (TD-GCMS) Analysis Samples were injected onto the column using a Perkin Elmer ATD 400 thermal desorber. TD tubes were dry purged for 1 minute. Desorption was achieved at a pressure of 35 kPa with a valve temperature of 200 °C. Tubes were primary desorbed at 270 °C at a flow of 50 mL/min for 10 minutes into a secondary trap held for 10 minutes at a temperature of -10 °C. At completion of the primary desorb the secondary trap was heated to 270 °C to transfer the sample onto the GCMS via a transfer line at 220 °C. No inlet split was utilised. Gas chromatographic mass spectrometric (GCMS) analyses were performed using a Varian Saturn 2000 MS/MS interfaced to a Varian CP3800 gas chromatograph. A Zebron Phenonemex 30 m × 0.25 mm i.d. fused silica capillary column coated with a 0.5 µm, 5 % phenyl – 95 % dimethylpolysiloxane (ZB-5MSi) stationary phase (part no. 7HG-G018-17) was used to carry out the separation. The oven was programmed from 30 °C to 250 °C at 10 °C/min with an initial and final hold time of 5 minutes. Helium was used as the carrier gas with a column pressure of 15 psi.
2.7 Particulate Characterisation Gibbsite (C33 alumina hydrate) sourced from the Alcoa World Alumina Australia Wagerup refinery in Western Australia was sieved to 38 – 52 µm size particles. The gibbsite was added into a furnace at room temperature and the temperature left to increase to 900 °C overnight (16 hr) for conversion to γ-alumina. Characterisation of alumina was achieved by x-ray diffraction (XRD) and infrared analysis (D. Fleming [Chemistry Centre of W.A.] pers. comm., 19 April 2007). Images of the γ-alumina particles were obtained by scanning electron microscopy (SEM). 10
2.7.1 Scanning Electron Microscopy (SEM) Analysis Particles were deposited on an aluminium stub using carbon tape and carbon coated before placement in a CamScan CS3200LV SEM coupled to an Oxford INCAx-sight x-ray detector (model 7788). Images were taken of a ‘typical’ γ-alumina particle with the secondary electron detector and beam acceleration voltage of 25 kV using thermionic emission. Figure 2 shows a ‘typical’ γ-alumina particle used in this study for both adsorption and desorption work that was obtained through gibbsite (C33) 1 dehydration. The particle is shown to have an approximate diameter of 50 μm and is an aggregate of smaller flat, tablet and hexagonal-shaped particles that increase the γ-alumina surface area to which organics adsorb.
10 µm Figure 2: SEM image of 'typical' γ-alumina particle.
1
Reference standard used by Alcoa World Alumina.
11
3.0 Results 3.1 Statistics After exposure to a toluene gas standard for 10 minutes, γ-alumina was packed into 15 TD tubes and analysed as in adsorption experiments. The relative standard deviation (% RSD) was calculated as 7 %.
3.2 Adsorption of Non Polar VOCs Adsorption of non polar VOCs n-hexane, benzene, toluene, o-xylene, m-xylene and pxylene on γ-alumina was measured after 1, 2, 5, 10, 30, 60 and 120 minutes the results of which are presented in figure 3 and table 2.
Adsorption (ng/mg)
25 20 n-Hexane
15
Benzene Toluene
10
o-Xylene 5
m -Xylene p-Xylene
0 0
50
100
150
Time (minutes)
Figure 3: Adsorption of non polar VOCs on γ-alumina.
For all non polar VOCs, adsorption profiles exhibit an initial sharp decrease followed by a plateau. The greatest adsorption after 1 minute exposure was observed for p-xylene (20 ng/mg) followed by m-xylene (14 ng/mg), o-xylene (11 ng/mg), toluene (3.6 ng/mg), benzene (2.6 ng/mg) and n-hexane (0.7 ng/mg). After 120 minutes, the same adsorption trend occurred as after 1 minute adsorption time. The greatest adsorption after 120 minutes was observed for p-xylene (14 ng/mg) followed by m-xylene (12 ng/mg), o-xylene (8.4 ng/mg), toluene (2.9 ng/mg), benzene (1.0 ng/mg) and n-hexane (0.4 ng/mg). 12
Table 2: Adsorption of non polar VOCs on γ-alumina Contact Time (min) 1 2 5 10 30 60 120
n-Hexane (ng/mg) 0.7 0.6 0.5 0.4 0.4 0.4 0.4
Benzene (ng/mg) 2.6 2.2 1.8 1.6 1.3 1.2 1.0
Toluene (ng/mg) 3.6 3.3 3.1 3.0 3.0 3.0 2.9
o-Xylene (ng/mg) 11 10 9.3 9.0 8.7 8.5 8.4
m-Xylene (ng/mg) 14 12 12 12 12 12 12
p-Xylene (ng/mg) 20 18 15 14 14 14 14
3.3 Adsorption of Polar VOCs Adsorption of polar VOCs propionaldehyde, 2-butanone, benzaldehyde, acetophenone and benzyl alcohol on γ-alumina was measured after 1, 2, 5, 10, 30, 60 and 120 minutes. The results are presented in figure 4 and table 3. Adsorption data for propionaldehyde was not obtained.
Adsorption (ng/mg)
250 200 150 2-Butanone 100
Benzaldehyde Acetophenone
50 0 0
50
100
150
Time (minutes)
Figure 4: Adsorption of polar VOCs on γ-alumina.
The greatest adsorption after 1 minute exposure was observed for 2-butanone (29 ng/mg) followed by benzaldehyde (20 ng/mg) and acetophenone (0.3 ng/mg). Although initially after 1 minute 2-butanone adsorbed greater than benzaldehyde, between 60 – 120 minutes, the absorbance of benzaldehyde exceeded 2-butanone. After 10 minutes, adsorption of acetophenone was lower than benzaldehyde and 2-butanone however after 30 minutes their adsorption was exceeded. The greatest adsorption after 120 minutes exposure was 13
observed for acetophenone (220 ng/mg) followed by benzaldehyde (51 ng/mg) and 2butanone (47 ng/mg). Table 3: Adsorption of polar VOCs on γ-alumina Contact Time (min) 1 2 5 10 30 60 120
2-Butanone (ng/mg) 29 41 43 43 45 46 47
Benzaldehyde (ng/mg) 20 25 36 40 42 45 51
Acetophenone (ng/mg) 0.3 3.4 8.0 16 52 97 220
3.3.1 Adsorption of Benzyl Alcohol Although adsorption studies for benzyl alcohol were conducted, benzyl alcohol was not identified in the sample TD tube chromatograms although it was added (figure 5 and 6). The standard TD tubes contained Tenax TA only in case of possible VOC alteration if injected onto γ-alumina. Although benzyl alcohol (m/z 108, 91) was not present in the sample TD tubes, a large peak was identified as benzaldehyde (m/z 105, 77) (figure 7 and 8). Benzoic acid (m/z 122) was present at levels slightly higher than on TD tube blanks.
MCounts 45:300
Spectrum 1A BP: 79.0 79.0
1050 ng/µL Standard.SMS TIC 100%
2.0 75% 77.0 1.5
50%
108.0
91.0
1.0 51.0 25%
0.5
78.0 63.0
0.0
92.0
0% 5
10
15
20
25 minutes
Figure 5: TD-GCMS chromatogram of benzyl alcohol 1050 ng/µL standard injected onto Tenax TA only.
50
100
150
200
250
300 m/z
Figure 6: Mass spectrum of benzyl alcohol peak at 13.850 minutes.
14
2 hour adsorption sample.SMS TIC
MCounts 5 45:300
Spectrum 1A BP: 105.0
105.0
100%
4 75% 3
50% 2 77.0 25%
1
51.0 106.0
0
0% 5
10
15
20
25 minutes
50
Figure 7: TD-GCMS chromatogram of benzyl alcohol (13.850 min) and benzaldehyde (14.462 min) in 2 hour adsorption sample.
100
150
200
250
300 m/z
Figure 8: Mass spectrum of benzaldehyde (14.462 minutes) in 2 hour adsorption sample.
3.4 Desorption of Non Polar VOCs The desorption of non polar VOCs n-hexane, benzene, toluene, o-xylene, m-xylene and pxylene from γ-alumina was measured after 1, 2, 5, 10, 30, 60 and 120 minutes and the
Residual Organics (ng/mg)
results are presented in figure 9 and table 4.
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
n-Hexane Benzene Toluene o-Xylene m-Xylene p Xylene
0
50
100
150
Time (minutes) Figure 9: Residual non polar VOCs present on γ-alumina.
15
After 1 minute n-hexane (0.9 ng/mg) had the greatest amount of residual present followed by toluene (0.7 ng/mg), benzene (0.4 ng/mg), p-xylene and m-xylene (0.3 mg/mg) and oxylene (0.1 ng/mg). The greatest amount of residuals present after 120 minutes were for nhexane (0.3 ng/mg) followed by toluene and benzene (0.1 ng/mg), o-xylene (0.02 ng/mg) and p-xylene and m-xylene (0.01 ng/mg). Table 4: Residual non polar VOCs present on γ-alumina Contact Time (min) 1 2 5 10 30 60 120
n-Hexane (ng/mg) 0.9 0.8 0.7 0.6 0.5 0.3 0.3
Benzene (ng/mg) 0.4 0.4 0.3 0.3 0.3 0.1 0.1
Toluene (ng/mg) 0.7 0.4 0.3 0.2 0.2 0.1 0.1
o-Xylene (ng/mg) 0.1 0.02 0.02 0.02 0.02 0.02 0.02
m-Xylene (ng/mg) 0.3 0.03 0.02 0.02 0.02 0.02 0.01
p-Xylene (ng/mg) 0.3 0.05 0.04 0.03 0.03 0.03 0.01
n-Hexane, benzene and tolu ene were slowly removed from the γ-alumina with the maximum amount of VOC desorbed occurring after 60 minutes (0.1, 0.3, 0.1 ng/mg respectively) in which a plateau is observed. o-Xylene, m-xylene and p-xylene exhibit a sharp decrease with little residual present after 10 minutes. Residual p-xylene after 10 minutes was 0.03 ng/mg with residual o-xylene and m-xylene both 0.02 ng/mg.
3.5 Desorption of Polar VOCs The desorption of polar VOCs 2-butanone, benzaldehyde and acetophenone from γalumina was measured after 1, 2, 5, 10, 30, 60 and 120 minutes with the results shown in figure 10 and table 5.
16
Residual Organics (ng/mg)
25 20 15 2-Butanone 10
Benzaldehyde Acetophenone
5 0 0
50
100
150
Time (minutes)
Figure 10: Residual polar VOCs present on γ-alumina.
Initial rapid desorption occurred for both 2-butanone and benzaldehyde with the maximum amount of VOC desorbed occurring after 30 minutes (1.0 ng/mg and 4.0 ng/mg respectively) in which a plateau is observed. Acetophenone was slowly desorbed and a plateau was not observed. Table 5: Residual polar VOCs present on γ-alumina Contact Time (min) 1 2 5 10 30 60 120
2-Butanone (ng/mg) 24 18 3.0 1.0 1.0 1.0 1.0
Benzaldehyde (ng/mg) 4.8 4.3 4.1 4.1 4.0 4.0 4.0
Acetophenone (ng/mg) 18 18 16 14 7.0 3.6 2.1
17
4.0 Discussion The adsorption of VOCs to γ-alumina is significantly dependent on the interaction between the adsorbing VOCs and the surface aluminium cations and oxygen anions. Traditionally γ-alumina has been considered to have a cubic spinel structure (AB2O4) where aluminium cations occupy tetrahedral and octahedral positions with oxygen atoms occupying face centred cubic (fcc) positions. Cations in spinel structures have a cation to anion ratio of 3 to 4 however the Al2O3 stoichiometry is maintained in γ-alumina by vacant cation positions (Wolverton & Hass 2001). A defect spinel structure is created as the γ-alumina possesses a 2 to 3 cation to anion ratio as shown in figure 11.
Figure 11: Location of aluminium (black) ions and oxygen (white) ions in the γ-alumina structure (Paglia et al 2005).
Between two adjacent surface aluminium cations, oxygen bridges form through the dehydration process from gibbsite to γ-alumina (Yang et al 2007). If surface aluminium cations are coordinatively unsaturated, a Lewis acid site is formed and can donate a proton 18
and, in turn, accept an electron pair (Liu & Truitt 1997; Vijay, Mills & Metiu 2002). The γalumina surface can act as a Lewis acid or base. Interactions occur between the adsorbing VOC and nucleophilic oxygen bridges or electrophilic aluminium Lewis acid sites. Possible interactions may occur between the hydrogen atoms of adsorbed water molecules and the adsorbed VOC as physisorbed and chemisorbed water is known to exist predominantly on the γ-alumina surface up to temperatures of 400 °C (Caldararu et al 2001). In this work, interactions between adsorbed atmospheric water vapour and the adsorbing VOC were assumed to not occur. Adsorption to the γ-alumina surface occurs through intermolecular interactions which depend on the adsorbing VOCs molecular weight, polarisability, boiling point and geometric structure. Interactions between the γ-alumina surface and non polar or polar VOC occur through different mechanisms which are also affected by steric factors.
4.1 Adsorption of Non Polar VOCs All non polar VOC adsorption profiles shown there is an initial sharp decrease followed by a plateau where equilibration is reached. Physisorption rapidly occurs as a lower kinetic energy barrier must be over come as there is no significant orbital distortion of the adsorbed molecule or substrate. Physisorption occurs via weak van der Waals forces where the strength of these van der Waal forces increase with increasing molecular weight for compounds in a homologous series. The electron cloud of these larger molecules can be more readily polarised so higher adsorption is observed. The strength of these van der Waal interactions is related in the VOCs boiling point and the presence of water vapour has also been shown to affect the adsorption of non polar VOCs.
4.1.1 Water Vapour The initial decrease in the amount of VOC absorbed with increasing adsorption time is due to the presence of water vapour. Water vapour partially displaces the adsorbed VOC before equilibration between the retained and unretained VOC on the γ-alumina surface is reached. Physisorbed water primarily interacts with the Lewis acid sites through lone electron pairs on the water oxygen. The water molecule is likely to be positioned at a vertical axis to the surface so the water molecules oxygen atoms are positioned so the OAl-O angle is 90° (Ionescu et al 2002).
19
Physisorbed water vapour may also be able to form other dipole–dipole interactions with the γ-alumina surface. Hydrogen bonding can be a secondary interaction between the surface oxygen anions and water hydrogen atoms. Hydrogen bonding is not considered to be the primary adsorption mechanism as the water molecules oxygen atom is more electronegative than its hydrogen atom. A stronger interaction is able to be formed between the more electronegative oxygen atom and the γ-alumina surface causing preferential adsorption to occur between the lone pairs on the water oxygen atom and the aluminium cation. These interactions are stronger than for all of the non polar VOC interactions with the γ-alumina surface. Displacement of VOC molecules by water occurs as VOCs physisorb to the surface through weaker van der Waal forces than the water molecules. Although VOCs are displaced (particularly non polar VOCs), adsorption is increased from interactions of the presence of an aromatic ring and entry into the surface pore network.
4.1.2 n-Hexane Limited surface adsorption of n-hexane is due to it forming the weakest interaction with the γ-alumina surface in comparison to the other non polar VOCs. n-Hexane has little affinity for both nucleophiles and electrophiles, such as the γ-alumina surface oxygen bridges or Lewis acid sites. The small amount of adsorption that occurs may be due to a chemical reaction followed by a physisorption mechanism. n-Hexane may adsorb to the γ-alumina surface by the initial removal of two hydrogen atoms to form hexene. Weak dipole induced van der Waal forces occur between the double bond of a newly formed hexene molecule and the γ-alumina surface. At an aluminium and oxygen terminated γ-alumina surface, hexene’s C and H atoms can interact with surface oxygen atoms. If the surface is only oxygen terminated, after initial formation of hexene and removal of the n-hexene’s allylic position hydrogen, the carbon atom can interact with the γ-alumina surface oxygen atoms. This mechanism was found to occur by Cai, Chihaia and Sohlberg (2007) where ethane and pentane adsorbed to a γ-alumina surface through formation of their corresponding alkenes. Pentane was found to form 2-pentene by removal of C2 and C3 hydrogen atoms. Adsorption of these alkenes was found to not occur via Lewis acidity of surface aluminium sites as the aluminium surface atoms may not interact with the adsorbing alkene (Cai and Sohlberg 2006). Likewise, they observed that methane adsorption to γ-alumina occurred through 20
interactions between the methane carbon and surface oxygen atoms. Although interactions are usually proposed to involve the surface aluminium cations only and not surface oxygen atoms, this mechanism may provide a possible explanation for the greater residual nhexane present in the desorption experiments. n-Hexane adsorbed less in comparison to the aromatic non polar VOCs as adsorption of aromatic non polar VOCs is due to van der Waal forces which increase with increasing molecular weight.
4.1.3 Aromatic Non Polar VOCs All aromatic non polar VOCs (benzene, toluene, o-xylene, m-xylene and p-xylene) adsorbed in a greater quantity to γ-alumina than n-hexane. Similarly to n-hexane, the aromatic non polar VOCs interactions with the γ-alumina surface occur from van der Waal forces. Adsorption increases with increasing molecular weight and polarisability showing the general trend in adsorption being xylene isomers (106.17 g/mol-1) > toluene (92.14 g/mol-1) > benzene (78.11 g/mol-1). Specific interactions may also form through the aromatic ring delocalised π electrons. For the benzene and toluene molecules these interactions form between the Lewis acid sites and the nucleophilic π electrons above and below the plane of the aromatic rin g. These interactions occur through the positioning of the aromatic cloud either edge on or face on to the γ-alumina surface. Larger aromatic molecules typically adsorb edge on to the surface as the aromatic π-π bonds are stronger than the π-surface interactions (Haghseresht et al 2002; Hurt et al 2002). Benzene and toluene may adsorb edge on to the surface as this allows for the aromatic ring to interact with Lewis acid aluminium sites (Ahmad et al 1998). Zaitan et al (2008) however has suggested interactions between the γ-alumina surface and o-xylene may occur between the γ-alumina surface and o-xylene molecule or o-xylene methyl group and γ-alumina Lewis acid sites. Suggestions that interactions with the surface occur with o- xylenes methyl group and not the aromatic group may be a result of the two adjacent methyl group electron clouds strengthening the interaction more than interactions with the aromatic ring. This may be also possible for m-xylene. In addition for p-xylene, steric effects from the methyl groups will hinder aromatic ring interactions with the γ-alumina Lewis acid sites.
21
4.2 Discrepancies in Non Polar VOC Adsorption Although n-hexane (86.18 g/mol-1) has a greater molecular weight than benzene (78.11 g/mol-1) it adsorbs less to γ-alumina than benzene. In addition to van der Waal interactions, molecules may adsorb onto a surface through dipole-dipole interactions from the presence of a polar functional group. For the non polar VOCs including n-hexane, adsorptive affinity of VOCs for γ-alumina increases with increasing boiling point (appendix I). The adsorption curves of the non polar compounds show adsorption increases as xylene isomers (138.3 – 144.5 °C) > toluene (110.6 °C) > benzene (80.0 °C) > n-hexane (68.7 °C) as shown in figure 12. Boiling point can be used as an indicator of adsorbing ability for all VOCs studied. The relationship between boiling point and adsorption ability, however, is more apparent for the polar VOCs.
250 n-Hexane
Adsorption (ng/mg)
200
Benzene 150
Toluene o-Xylene
100
m -Xylene p-Xylene
50
2-Butanone Benzaldehyde
0 0
50
100
150
200
250
Acetophenone
Boiling Point (°C)
Figure 12: Relationship between the boiling point and adsorption after 120 minutes exposure.
4.2.1 Steric Effects for Adsorption As boiling point increases for the xylene isomers, toluene, benzene and n-hexane, so too does the VOCs ability to adsorb. This trend however does not occur for the xylene isomers individually. Boiling point increases according to the following trend: o-xylene (144.5 °C) > m-xylene (139.1 °C) > p-xylene (138.3 °C). The amount adsorbed increases in the reverse order where at 120 minutes p-xylene > m-xylene > o-xylene. The data suggests adsorption is affected by steric factors from the position of the two methyl groups on the 22
aromatic ring. Unlike p-xylene, methyl substitutes on o-xylene are sterically hindered from their close proximity to each other. Steric constraints inhibit adsorption as the separation distance between the adsorbing molecule and the γ-alumina surface is increased, in turn the strength of the interaction is reduced. Steric effects were found to have an effect on adsorption in a study by Díaz et al (2004). Their work showed cyclohexane and cycloheptane inhibited adsorption on Zeolite 5A compared to adsorption of n-hexane and heptane respectively. In addition symmetrical molecules, such as p-xylene have a greater affinity for the γ-alumina surface than a less symmetrical molecule such as o-xylene or mxylene. This results from a lower loss of rotational and translational freedom during adsorption (Reitmeier et al 2008). Adsorption may also be due to the VOCs ability to enter surface pores.
4.3 Adsorption of Polar VOCs Adsorption curves of the polar compounds vary depending on the properties of the adsorbing VOC. Both 2-butanone and benzaldehyde exhibit a rapid increase in initial adsorption followed by a slower continuous increase. This initial increase occurs from these VOCs chemisorbing to the surface. Chemisorption slowly occurs as a higher kinetic energy barrier must be over come as new chemical bonds or electrostatic interactions are formed. Unlike the non polar compounds, adsorption does not increase with increasing molecular weight as in addition to van der Waal interactions, molecules may adsorb onto a surface through stronger dipole-dipole interactions of the carbonyl group. Dipole-dipole interactions form between the γ-alumina surface and either the carbonyl group carbon or oxygen atom. This leads to chemisorbed species rather than physisorbed. For all polar VOCs the strength of interactions are reflected in the VOCs boiling point. Adsorption of the polar VOCs increases with increasing boiling point where boiling point increases as benzyl alcohol (205.35 °C) > acetophenone (202 °C) > benzaldehyde (179.0 °C) > 2-butanone (79.59 °C) > propionaldehyde (48.8 °C). Also, as polar compounds form stronger interactions with the surface, multilayer adsorption can occur.
4.3.1 Propionaldehyde Although adsorption studies for propionaldehyde were conducted, propionaldehyde was not identified in sample chromatograms. This was reflected by its molecular weight (58.08 23
g/mol-1) and boiling point (48.8 °C). Aldehydes may adsorb and react readily on acidic metal oxide surfaces by undergoing conversion via Aldol condensation where the saturated propionaldehyde molecule may self couple to form higher molecular weight unsaturated compounds. Adsorption of propionaldehyde on α-alumina was shown to form 2-methyl 2pentenal (Li et al 2001). Finocchio et al 1997 has also shown that propionaldehyde adsorption on the metal oxides Co3O4 and MgCr2 O4 forms unidentified acrylate species at room temperature. Neither 2-methyl 2-pentenal nor acrylate compounds, however, were tentatively identified in the propionaldehyde adsorption experiment chromatograms. In addition to reactivity, propionaldehyde’s volatility requires a low volume of gas to elute it from the Tenax TA, i.e. breakthrough volume. In a study by Rothweiler, Wäger and Schlatter (1991) the recovery of propionaldehyde from a TD tube containing 155 milligrams of Tenax TA after 1 litre of air had passed through was 85.2 % compared to 93.8 % for toluene. When 5 litres was passed through the recovery was 37.8 % and 99.7 % for propionaldehyde and toluene respectively.
4.3.2 2-Butanone The adsorption profile of 2-butanone shows there is no initial decrease in adsorption due to removal of the VOC by water vapour. Both water vapour and 2-butanone compete for Lewis acid sites. Also, if both water and polar VOCs form chemisorbed species they are directly competing for the alumina surface, meaning the concentration of both species will increase over time with the favoured species dominating at longer times. Adsorption continues to increase until equilibration is reached and maximum adsorption occurs. When maximum adsorption occurs there are no active Lewis acid sites for interactions to form. Adsorption after 120 minutes is lower than the other polar VOCs but higher than the non polar VOCs studied. This is reflected in 2-butanones boiling point where, except for propionaldehyde, is lower than the polar compounds VOCs. Despite having a boiling point only slightly higher than n-hexane, its ability to adsorb greater than the non polar VOCs, is due to the carbonyl group, see section 4.3. Adsorption for 2-butanone is less than the adsorption for benzaldehyde as the ketone functional group is less electronegative than the aldehyde functional group. It also does not have strengthened interactions from the presence of an aromatic ring.
24
4.3.3 Benzaldehyde and Acetophenone The adsorption of benzaldehyde is similar to all polar VOC adsorption behaviour where there is no initial decrease in adsorption due to the removal of the VOC by water vapour. Similarly to 2-butanone, both water vapour and benzaldehyde compete for Lewis acid sites. Adsorption continued to increase beyond 120 minutes suggesting that equilibration was not attained, hence maximum adsorption did not occur. Adsorption begins to slow when approaching 120 minutes and is proposed to be due to benzaldehyde’s larger molecular size and bulkier aromatic group. The bulkier benzaldehyde molecules may require a greater amount of time to orientate around other adsorbed molecules to find active adsorption sites and thereby equilibrate. With increasing adsorption time (120 min), molecules rearrange to a more ordered and geometrically stable structure (Mukti, Jentys & Lercher 2007). Although equilibrium is attained for benzaldehyde, in comparison to 2butanone, adsorption will be increased because of its aromatic ring and more electronegative aldehyde functional group. The adsorption profile of acetophenone did not resemble that of the other non polar or other polar VOCs studied. As adsorption increased linearly with increasing adsorption time we suspect that acetophenone may adsorbs in multilayers. Acetophenone also does not appear to compete with water vapour for active Lewis acid sites unlike 2-butanone and benzaldehyde.
O O Al
-
H O Al
O Al
Figure 13: Probable structure of adsorbed species after benzaldehyde exposure to γalumina.
CH2
O Al
H O
Al
Al
Figure 14: Enolate anion formed after acetophenone exposure to γ-alumina.
Benzaldehyde and acetophenone chemisorption may occur between the Lewis acid sites and oxygen bridges or through abstraction of the aldehydic hydrogen as suggested by 25
Lichtenberger, Hargrove-Leak and Amiridis (2006). Removal of the most acidic hydrogen allows the carbonyl carbon cation to interact with γ-alumina oxygen bridges (figure 13). Similarly their work showed acetophenone formed an enolate anion after removal of a methyl group hydrogen and subsequent interaction with magnesium cations in MgO. Chemisorption of acetophenone may occur through this mechanism where the oxygen interacts with the γ-alumina Lewis acid sites (figure 14). Adsorption may also be increased by electron withdrawing effects from aldehyde and ketone functional groups. The moderately deactivating carbonyl substitutes are slightly electron donating. γ-Alumina surface Lewis acid sites are in turn an electron pair acceptor. Benzaldehyde’s adsorption may be more assisted than the 2-butanone as it has a more electronegative carbonyl substitute. Electron donation or withdrawing effects are, however, considered to have a minor effect on the adsorption affinity of the VOC in comparison to intermolecular forces. This is evident by benzaldehyde’s decreased adsorption compared to acetophenone.
4.3.4 Benzyl Alcohol The presence of benzaldehyde in the benzyl alcohol adsorption sample showed a typical acid-catalysed reaction of oxidising an alcohol had occurred. Oxidation of benzyl alcohol may be either a result of analysis via TD-GCMS or at room temperature through photocatalysis from the presence of ultra violet (UV) light. These atmospheric reactions occur through a hydroxyl radical (•OH) (Kleindienst et al 2004; Lim & Ziemann 2005), NOx or O3 (Seinfeld & Pankow 2003) but are unlikely. For example 0.26 % of 1-pentanol was photo-oxidised over Al2O3 under irradiation at 323 K in O2 compared to 11.7 % on TiO2 (Ohuchi et al 2007). Without radical formation, benzaldehyde has been shown to be a product of benzyl alcohol oxidation over an alumina catalyst where the amount of benzaldehyde increased with increasing reaction temperature (Jayamani & Pillai 1983). Investigations into the oxidation of adsorbed species are proposed as future work in section 5.1. Oxidation of alcohols with a catalyst yields its corresponding aldehyde which may further oxidise to form the corresponding carboxylic acid. γ-Alumina may act as this catalyst for benzyl alcohol oxidation to benzaldehyde because of its large surface area and surface reactive sites. The properties of the γ-alumina used in this work are determined by the
26
precursor’s (gibbsite’s) growth and morphology which is determined by production conditions (Belaroui, Pons & Vivier 2002). The benzyl alcohol may undergo disproportionation or dehydrogenation reactions on the γalumina surface. Disproportionation produces benzaldehyde and toluene, however, toluene was not identified in sample chromatograms at levels greater than blank TD tubes containing Tenax TA only. For oxidation of benzyl alcohol to occur via dehydrogenation, the γ-alumina surface requires Lewis acid–strong Brønsted base site pairs in order to form and stabile alkoxide intermediates. These site pairs are characteristic of the γ-alumina surface where a strong Brønsted base is provided by the oxygen bridges (Di Cosimo et al 2000). Chemisorption occurs between benzyl alcohols hydrogen atom and surface oxygen anions where chemisorption is dependent on the number of vacant sites in close proximity to the adsorbing alcohol as vacant sites increase the available space for the adsorbing molecule. If the γ-alumina surface oxygen is close to an aluminium vacancy, dehydrogenation at a Lewis acid site is possible with one alcohol OH hydrogen interacting with the surface oxygen. The alcohol OH however must interact as one entity with the surface if no vacancies are present for dehydrogenation to occur (Cai & Sohlberg 2003).
4.4 Desorption of Non Polar VOCs Desorption of VOCs from the γ-alumina surface is significantly dependent on the adsorbing VOCs ability to enter the γ-alumina surface pores. VOCs with a kinetic diameter greater than or equal to the surface pore diameter will not be able to enter surface pores (Almazán-Almazán et al 2007). These VOCs adsorb to the γ-alumina surface only. Parallel pores separate the γ-alumina crystallites into 2 – 3 nm lamellae and parallel to these is a system of pores that are 1 nm irregular shaped slits with pore diameters less than 2 nm (Wefers & Misra 1987). Pore diameter, however, is determined by the precursor’s properties and the γ-alumina preparation route. Entry into these pores is determined by the adsorbing molecules size and geometry; where entry of larger sized and more branched VOCs is hindered. Desorption of VOCs is also affected by the rate of diffusion into (and out of) pores and repulsion effects between two adjacent adsorbing VOCs. Molecules not able to sufficiently enter the pore network initially adsorb to the γ-alumina surface. Here water vapour is able to readily displace the VOC. After equilibration is reached and the amount of VOC desorbed is maximised, residuals on the γ-alumina increased in the order n-hexane > toluene and benzene > o-xylene > p-xylene and m-xylene. 27
4.4.1 n-Hexane A loss in the amount of n-hexane adsorbed occurs primarily from the competition with water vapour. n-Hexane cannot compete with water vapour for adsorption sites and is readily displaced (Ruiz, Bilbao & Murillo 1998). It was removed slowly from the γalumina surface. Although n-hexane adsorbed the least of all non polar VOCs it had the greatest residuals in comparison to the other non polar VOCs. The greater residual n-hexane is proposed to be due to the molecule entering the surface pores. The linear geometry and smallest kinetic diameter of n-hexane (4.30 nm) allows for accommodation in the slit shaped lamellae pores and accessibility to adsorb in the pore structure with limited repulsion. Short range repulsion between two n-hexane molecules adsorbing adjacent to each other in a pore will be reduced as n-hexane is not branched and possesses C-H atoms with similar electronegativities. Also diffusion into the pore network will be the fastest for n-hexane due to its neutral charge distribution and linear geometry. Both of these factors will also permit easier packing of n-hexane molecules in a pore so the number of molecules in an individual pore can increase. Water vapour may not be able to remove n-hexane adsorbed in the pores as readily as it displaces surface adsorbed nhexane. Although surface adsorbed n-hexane is displaced by water vapour, n-hexane molecules adsorbed onto the pore surface may obstruct a water molecule from entering the pore channel. If water molecules are not able to interact with the γ-alumina pore surface, nhexane cannot be desorbed. Desorption is effected by molecular geometry for n-hexane much less than for the non polar aromatics.
4.4.2 Non Polar Aromatic VOCs The overall residual VOC’s present after 120 minutes desorption were also lower than nhexane. Only molecules adsorbed to the surface may be removed by water vapour leaving molecules adsorbed in the pores. In comparison to n-hexane, the majority of non polar aromatics can interact with the surface. The number of molecules entering the pore network is also reduced and desorption is higher. Aromatic non polar molecules have a rigid structure which is also not as accommodated around other aromatic non polar molecules in a pore. The close proximity of one molecule next to another in a pore may have a short range repulsion effect from the aromatic ring π electrons which inhibits adsorption to a pore surface. Repulsion between the pore walls and
28
the electron cloud of toluene and xylenes methyl substitutes can also occur (Mukti, Jentys & Lercher 2007). The size and geometry of toluene and benzene’s aromatic ring can inhibit the number of molecules able to enter a pore. As residual toluene is still present on the γ-alumina after 120 minutes, we hypothesise that it is able to enter surface pores. The adsorption of toluene in surface pores is less than n-hexane and has been attributed to its larger molecular size (Van Bavel et al 2005). Toluene may be able to enter pores and adsorb but also adsorb to the γ-alumina surface from the presence of the aromatic ring. Benzene’s lower residual loss than toluene is expected as it does not have the added size, geometry and repulsion effects from the added methyl substitute. Entry into surface pores will not be reduced as much as the xylene isomers, in particularly o- and m-xylene. The amounts of residual xylene (o-, m-, p-) on the γ-alumina were lower than for all other VOCs (non polar and polar) with almost all xylene completely desorbed. Desorption profiles show all isomers were rapidly desorbed in the presence of water vapour whereas toluene, benzene and n-hexane were gradually desorbed. The larger sized xylene isomers may not able to enter surface pores so the molecule only adsorbs to the γ-alumina surface. Although interactions with the γ-alumina surface are strongest for the xylene isomers, their in ab ility to enter su rface p ores allows water to interact with the γ-alumina surface and desorption of xylene to occur. If the xylenes molecules are not able to enter surface pores, their equilibration desorption value should be the same as their equilibration adsorbed value. This is not the case and may be attributed to the shorter initial adsorption time of 15 minutes in the desorption experiments. The xylene molecules may not be able to penetrate into the pore network in this period as deeply as in the longer timed adsorption experiments. If however, xylene molecules are able enter surface pores, their large molecular size and electron cloud will slow diffusion into the pore. Their size will also limit the number of other xylene and water molecules able to enter the pore. Adsorption onto the γ-alumina surface can increase and, in turn, water vapours ability to be able to desorb xylene. Repulsion effects inside of the pore will also be the greatest between two xylene molecules as they have an electron cloud larger than the other non polar VOCs. This will also limit the amount of xylene able to enter a surface pore. Comparing individual xylene isomers after two minutes desorption, p-xylene had the greatest residuals which can be attributed to its planar geometry. The presence of two 29
methyl substitutes may reduce the number of molecules able to enter the pore network which also has greater repulsion effects. Diffusion of the p-xylene molecules into the pore network will also be slower from its large molecular size. Diffusion of o- and m-xylene molecules may be slower than p-xylene from their larger kinetic diameter (both 6.80 nm) due to the position of the two methyl substitutes. The smaller kinetic diameter of p-xylene (5.85 nm) can allow the molecule to enter surface pores whereas in comparison to o- and m-xylene these isomers may not be able to enter the pore network at all.
4.5 Desorption of Polar VOCs All polar VOCs are retained longer than the non polar VOCs. Although water vapour may remove the polar VOCs, it cannot remove the volume of polar VOCs comparable to the non polar VOCs. Higher residuals result from their ability to form stronger interactions. Desorption profiles of 2-butanone and benzaldehyde are similar and show fast rapid removal of the VOC to reach equilibrium and maximum desorption quickly. Acetophenone is gradually removed due to it forming the strongest interaction. It is not displaced readily. Interestingly, a greater amount of acetophenone is desorbed than benzaldehyde. Initially, acetophenone was expected to have the greatest residuals as it adsorbs more than the other polar VOCs, however, these unexpected results may be due to the desorption method. In the adsorption work the amount of acetophenone adsorbed only exceeds 2-butanone and benzaldehyde after 30 minutes. Adsorption occurs slowly for acetophenone, resulting in less molecules adsorbed on the surface. This increases the water molecules ability to desorb them. Polar VOCs are assumed to have a greater kinetic diameter and are either too large to enter the pores or diffuse too slowly into the pore network. The initial higher amount of desorbed 2-butanone compared to benzaldehyde and acetophenone may be due to it entering the pore network as it does not have an aromatic group. If 2-butanone does enter surface pores it will have stronger repulsion between two adjacent adsorbing molecules than the largest non polar VOCs, the xylene isomers. After 120 minutes however 2butanone is desorbed more than benzaldehyde and acetophenone as it forms the weakest interaction with the γ-alumina surface. The retention of polar VOCs after desorption can, thus, be explained using a similar rationale to the non polar VOC adsorption. As non polar VOCs form weaker interactions with the surface, they adsorb less and also desorb more readily than polar VOCs. 30
4.6 Inhaled Adsorbed Species VOCs were found to adsorb significantly on γ-alumina. Although 50 µm diameter particles were used in adsorption and desorption experiments, a greater amount of residual organics (per weight of alumina) will be present on respirable particles which are less than 10 µm in diameter. Calculations have shown that γ-alumina particles concentrate VOCs to levels greater than the original concentration of the VOC delivered to the γ-alumina (appendix II). Using pxylene as an example, the amount of p-xylene, if all desorbed, in lung alveoli was determined as 8.5 ppm (D. Fleming [Chemistry Centre of W.A.] pers. comm., 23 October 2008). This calculation has shown theoretically that VOCs concentrate on the γ-alumina surface as the original concentration of the VOC delivered to the γ-alumina in the adsorption experiments was 3 ppm. Based on this calculation, a review of the current exposure standards is needed to include the combined effects of particles with adsorbed VOCs. It is proposed that these concentrated VOCs may enter the lung by two different mechanisms. As non polar VOCs are more readily displaced by water vapour, the high moisture content of the lung may also cause organics adsorbed to the γ-alumina surface to desorb and enter the body through gas exchange. Polar compounds may interact directly with the lung cell wall if not desorbed by water vapour. It is known that organics adsorbed onto particulate matter cause cytotoxic effects to human lung tissue (Mazzarella et al 2007). The work by Mazzarella et al (2007) showed lung cell death occurred within a few hours after simulation with diesel PM1.0 at concentrations of 0.2 and 0.4 mg/mL. Diesel particulate matter is composed of carbon particles that adsorb other species whose effects may be similar to γ-alumina particulate exposure with adsorbed VOCs.
31
5.0 Conclusions Adsorption Behaviour of Alumina It has been shown that organic compounds present in alumina refinery emissions adsorb on to alumina particles. The relative amounts of the VOCs adsorbed were found to be a function of molecular weight, polarisability, van der Waal interactions, steric factors and polarity. This was reflected in the VOCs boiling point. For the non polar VOCs, adsorption increases with increasing boiling point however this trend did not apply for the xylene isomers which were affected by steric factors. Adsorption was also found to be effected by water vapour present which displaced adsorbed non polar VOCs. Similarly to the non polar VOCs, adsorption of the polar VOCs increased with increasing boiling point. Greater quantities of polar VOCs than non polar VOCs adsorbed as they were able to form a stronger interaction with the γ-alumina surface. Desorption was proposed to be a function of the water vapour present and the VOCs ability to fit into the γ-alumina surface pores. The larger and more highly branched VOCs cannot enter the pore network considerably as diffusion into these pores is slowed and repulsion effects between two adjacent adsorbing VOCs occur. This was observed for the non polar VOCs as residual organics increased with decreasing boiling point. For the polar VOCs, greater residuals were present and concluded to be due to formation of stronger interactions than the non polar VOCs.
Health Implications Adsorbed non polar and polar VOCs on alumina particles (< 10 µm in diameter) may readily enter the alveola portion of the lung where the amount of adsorbed species will increase with decreasing particle size (per equivalent mass). Respirable γ-alumina particles of 10 µm diameter, for example, may concentrate VOCs by factor of approximately 3 for p-xylene and far greater for more polar compounds. A model of the fate of inhaled respirable alumina particles with adsorbed VOCs is proposed. Non polar VOCs being more readily displaced by water vapour will desorb in the lung and enter the body through gas exchange. Conversely, polar VOCs are less affected by water vapour and can interact directly with the lung cell wall causing cytotoxic effects. These results show that a review of the current exposure standards is critical. 32
5.1 Future Work 1. Further characterisation of the γ-alumina produced by gibbsite calcination at 900 °C to determine the average pore diameter. This will allow more accurate conclusions to be made for desorption studies as to a VOCs ability to enter surface pores.
2. Using a mixed gas standard in adsorption experiments to observe preferential adsorption. 3. Exposure of adsorbed species on γ-alumina to ultra violet (UV) light and subsequent analysis by liquid chromatography mass spectrometry (LCMS). Many VOCs will oxidise on the catalytic surface of γ-alumina to form other VOCs and these may be more harmful to humans. 4. Cytotoxicty studies involving exposure of lung cells to γ-alumina with adsorbed organic species to determine the health effect on humans.
33
6.0 References Ahmad, I, Anderson, JA, Rochester, CH & Dines, TJ 1998, ‘Infrared study of adsorption of acetophenones on silica-alumina’, Journal of Molecular Catalysis A: Chemical, vol. 135, pp. 63-73.
Almazán-Almazán, MC, Pérez-Mendoza, M. Domingo-García, M, Fernández-Morales, I, del Rey-Bueno, F, García-Rodríguez, A, López-Garzón FJ 2007, ‘The role of the porosity and oxygen groups on the adsorption of n-alkanes, benzene, trichloroethylene and 1,2dichloroethane on active carbons at zero surface coverage’, Carbon, vol. 45, pp. 17771785.
Arden Pope III, C, Burnett, RT, Thun, MJ, Calle, EE, Krewski, D, Ito, K & Thurston GD 2002, ‘Lung cancer, cardiopulmonary mortality and long-term exposure to fine particulate air pollution’, The Journal of the American Medical Association, vol. 287, issue 9, pp. 1132-1141. Atkins, P & de Paula, J 2006, Atkins’ Physical Chemistry, 8th edn, Oxford University Press, Oxford, UK.
Bai, N, Khazaei, M, van Eeden, SF, & Laher, I 2007, ‘The pharmacology of particulate matter air pollution-induced cardiovascular dysfunction’, Pharmacology and Therapeutics, vol. 113, pp. 16-29.
Bastacky, J, Lee, CY, Goerke, J, Koushafar, H, Yager, D, Kenaga, L, Speed, TP, Chen, Y & Clements, JA 1995, ‘Alveolar lining layer is thin and continuous: Low-temperature scanning electron microscopy of rat lung’, Journal of Applied Physiology, vol. 79, pp. 1615-1628.
Belaroui, K, Pons, MN & Vivier, H 2002, ‘Morphological characterisation of gibbsite and alumina’, Powder Technology, vol. 127, pp. 246-256.
34
Bousquet, C, Elissalde, C, Aymonier, C, Maglione, M, Cansell, F & Heintz, JM 2008, ‘Tuning Al2O3 crystallinity under supercritical fluid conditions: Effect on sintering’, Journal of the European Ceramic Society, vol. 28, pp. 223-228. Cai, S & Sohlberg, K 2003, ‘Adsorption of alcohols on γ-alumina (110C)’, Journal of Molecular Catalysis A: Chemical, vol. 193, pp. 157-164. Cai, S & Sohlberg, K 2006, ‘Adsorption of 1-hexene on γ-alumina (110C)’, Journal of Molecular Catalysis A: Chemical, vol. 248, pp. 76-83.
Cai, S, Chihaia, V & Sohlberg, K 2007, ‘Interactions of methane, ethane and pentane with the (110C) surface of γ-alumina’, Journal of Molecular Catalysis A: Chemical, vol. 275, pp. 63-71.
Caldararu, M, Postole, G, Hornoiu, C, Bratan, V, Dragan, M & Ionescu, NI 2001, ‘Electrical conductivity of γ-Al2O3 at atmospheric pressure under dehydrating/hydrating conditions, Applied Surface Sciences, vol. 181, pp. 255-264.
Coffey, PS & Ioppolo-Armanios, M 2004, ‘Identification of the odour and chemical composition of alumina refinery air emissions’, Water Science and Technology, vol. 50, issue 4, pp. 39-47.
Coffey, P, Ioppolo-Armanios, M, Cox, S, Jones, M, Logiudice, A & Gwynne, K 2005, Wagerup Refinery Air Emissions Inventory – Final Report 2002, Alcoa World Alumina, Perth.
CSIRO 2007, BET Surface Area by TriStar 3000, Analysis Report, Division of Minerals, Waterford, Perth.
Cullen, RT, Tran, CL, Buchanan, D, Davis, JMG, Searl, A, Jones, AD & Donaldson, K 2000, ‘Inhalation of poorly soluble particles. I. Differences in inflammatory response and clearance during exposure’, Inhalation Toxicology, vol. 12, issue 12, pp. 1089-1111.
35
Dalton, P 2003, ‘Upper airway irritation, odor perception and health risk due to airborne chemicals’, Toxicology Letters, vol. 140-141, pp. 239-248.
Dean, JA 1992, Lange’s Handbook of Chemistry, McGraw-Hill Inc., USA.
Department of Education, Employment and Workplace Relations 1995, Adopted National Exposure Standards for Atmospheric Contaminants in the Occupational Environment [NOHSC:1003(1995)], Office of the Australian Safety and Compensation Council, Commonwealth Government, Canberra.
Department of Industry and Resources 2008, Opportunities in Western Australia – Aluminium, Investment Attraction Branch, Government of Western Australia, Perth.
Di Cosimo, JI, Apesteguía, CR, Ginés, MJL & Iglesia, E 2000, ‘Structural requirements and reaction pathways in condensation reactions of alcohols on MgyAlOx catalysts’, Journal of Catalysis, vol. 190, pp. 261-275.
Díaz, E, Ordóñez, S, Vega, A & Coca, J 2004, ‘Adsorption characterisation of different volatile organic compounds over alumina, zeolites and activated carbon using inverse gas chromatography’, Journal of Chromatography A, vol. 1049, pp. 139-146.
Dobbs, LG & Johnson, MD 2007, ‘Alveolar epithelial transport in the adult lung’, Respiratory Physiology and Neurobiology, vol. 159, pp. 283-300.
Donoghue, AM & Cullen, MR 2007, ‘Air emissions from Wagerup alumina refinery and community symptoms: An environmental case study’, Journal of Occupational and Environmental Medicine, vol. 49, issue 9, pp. 1027-1039.
Environment and Public Affairs Committee 2004 Report of the Standing Committee on Environment and Public Affairs in Relation to the Alcoa Refinery at Wagerup Inquiry, eleventh report (C Sharp, chairman), Parliament of Western Australia, Perth.
36
Fellner, P, Jurisova, J, Khandl, V, Sykorova, A & Thonstad, J 2006, ‘Adsorption of SO2 on alumina’, Chemical Papers, vol. 60, issue 4, pp. 311-314.
Finocchio, E, Willey, RJ, Busca G & Lorenzelli, V 1997, ‘FTIR studies on the selective oxidation and combustion of light hydrocarbons at metal oxide surfaces. Part 3. Comparison of the oxidation of C3 organic compounds over Co3O4, MgCr2O4 and CuO’, Journal of the Chemical Society, Faraday Transactions, vol. 93, issue 1, pp. 175-180.
Haghseresht, F, Finnerty, JJ, Nouri, S & Lu,GQ 2002, ‘Adsorption of aromatic compounds onto activated carbons: Effects of the orientation of the adsorbates’, Langmuir, vol. 18, pp. 6193-6200.
Hurt, R, Krammer, G, Crawford, G, Jian K & Rulison, C 2002, ‘Polyaromatic assembly mechanisms and structure selection in carbon materials’, Chemistry of Materials, vol. 14, pp. 4558-4565. Ionescu, A, Allouche, A, Aycard, JP & Rajzmann, M 2002, ‘Study of γ-alumina surface reactivity: Adsorption of water and hydrogen sulfide on octahedral aluminum sites’, Journal of Physical Chemistry B, vol. 106, pp. 9359-9366. Jayamani, M & Pillai, CN 1983, ‘Hydride transfer reactions. VIII. Reactions of benzyl alcohol over alumina: Dehydration and disproportionation’, Journal of Catalysis, vol. 82, pp. 485-488.
Kelly, FJ 2003, ‘Oxidative stress: Its role in air pollution and adverse health effects’, Journal of Occupational and Environmental Medicine, vol. 60, pp. 612-616.
Kleindienst, TE, Conver, TS, McIver, CD & Edney, EO 2004, ‘Determination of secondary organic aerosol products from the photooxidation of toluene and their implications in ambient PM2.5’, Journal of Atmospheric Chemistry, vol. 47, pp. 79-100.
37
Kloprogge, JT, Duong, LV, Wood, BJ & Frost, RL 2006, ‘XPS study of the major minerals in bauxite: Gibbsite, bayerite and (pseudo-)boehmite’, Journal of Colloid and Interface Science, vol. 296, pp. 572-576.
Krokidis, X, Raybaud, P, Gobichon, AE, Rebours, B, Euzen, P & Toulhoat, H 2001, ‘Theoretical study of the dehydration process of boehmite to γ-alumina’, Journal of Physical Chemistry B, vol. 105, pp. 5121-5130.
Levin, I, Gemming, T & Brandon, DG 1998, ‘Some metastable polymorphs and transient stages of transformation in alumina’, Physica Status Solidi (a), vol. 166, pp. 197-218.
Li, P, Perreau, KA, Covington, E, Song, CH, Carmichael, GR & Grassian, VH 2001, ‘Heterogeneous reactions of volatile organic compounds on oxide particles of the most abundant
crustal
elements:
Surface
reactions
of
acetaldehyde,
acetone,
and
propionaldehyde on SiO2, Al2O3, Fe2O3, TiO2, and CaO’, Journal of Geophysical Research, vol. 106, issue D6, pp. 5517-5529.
Lichtenberger, J, Hargrove-Leak, SC & Amiridis, MD 2006, ‘In situ FTIR study of the adsorption and reaction of 2'-hydroxyacetophenone and benzaldehyde on MgO’, Journal of Catalysis, vol. 238, pp. 165–176. Lide, DR (ed.) 2002, CRC Handbook of Chemistry and Physics, 83rd edn, CRC Press, Boca Raton.
Lim, YB & Ziemann, PJ 2005, ‘Products and mechanism of secondary organic aerosol formation from reactions of n-alkanes with OH radicals in the presence of NOx’, Environmental Science and Technology, vol. 39, pp. 9229-9236.
Lisachenko, AA, Klimovskii, AO, Mikhailov, RV, Shelimov, BN & Che, M 2007, ‘Lowpressure chemical and photochemical reactions of oxides of nitrogen on alumina taken as a model substance for mineral dust in relation to air pollution’, Catalysis Today, vol. 119, pp. 247-251.
38
Liu, X & Truitt, RE 1997, ‘DRFT-IR studies of the surface of γ-alumina’, Journal of the American Chemical Society, vol. 119, pp. 9856-9860.
Luginaah, IN, Taylor, SM, Elliott, SJ & Eyles, JD 2002, ‘Community reappraisal of the perceived health effects of a petroleum refinery’, Social Science and Medicine, vol. 55, pp. 47-61.
Mazzarella, G, Ferraraccio, F, Prati, MV, Annunziata, S, Bianco, A, Mezzogiorno, A, Liguori, G, Angelillo, IF & Cazzola, M 2007, ‘Effects of diesel exhaust particles on human lung epithelial cells: An in vitro study’, Respiratory Medicine, vol. 101, pp. 1155-1162.
McNaught, AD & Wilkinson, A 1997, IUPAC Compendium of Chemical Terminology, Royal Society of Chemistry, Cambridge, UK.
Moller, W, Felten, K, Sommerer, K, Scheuch, G, Meyer, G, Meyer, P, Haussinger, K & Kreyling, WG 2007, ‘Deposition, retention, and translocation of ultrafine particles from the central airways and lung periphery’, American Journal of Respiratory and Critical Care Medicine, vol. 177, issue 4, pp 426-432.
Mukti, RR, Jentys, A & Lercher, JA 2007, ‘Orientation of alkyl-substituted aromatic molecules during sorption in the pores of H/ZSM-5 zeolites’, Journal of Physical Chemistry C, vol. 111, pp. 3973-3980.
National Risk Management Research Laboratory 1999, Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient Air, United States Environmental Protection Agency, Cincinnati.
Ochs, M, Nyengaard, JR, Jung, A, Knudsen, L, Voigt, M, Wahlers, T, Richter, J & Gundersen, HJG 2004, ‘The number of alveoli in the human lung’, American Journal of Respiratory and Critical Care Medicine, vol. 169, issue 1, pp. 120-124.
39
Ohuchi, T, Miyatake, T, Hitomi Y & Tanaka, T 2007, ‘Liquid phase photooxidation of alcohol over niobium oxide without solvents’, Catalysis Today, vol. 120, issue 2, pp. 233239. Paglia, G, Bozin, ES & Billinge, SLJ 2006, ‘Fine-scale nanostructure in γ-Al2O3’, Chemistry of Materials, vol. 18, pp. 3242-3248.
Paglia, G, Buckley, CE, Rohl, AL, Hart, RD, Winter, K, Studer, AJ, Hunter, BA & Hanna, JV 2004, ‘Boehmite derived γ-alumina system. 1. Structural evolution with temperature, with the identification and structural determination of a new transition phase, γ΄-alumina’ Chemistry of Materials, vol. 16, pp. 220-236. Paglia, G, Rohl, AL, Buckley, CE & Gale, JD 2005, ‘Determination of the structure of γalumina from interatomic potential and first-principles calculations: The requirement of significant numbers of nonspinel positions to achieve an accurate structural model’, Physical Review B, vol. 71, issue 22, pp. 224115, 1-16.
Patton, JS 1996, ‘Mechanisms of macromolecule absorption by the lungs’, Advanced Drug Delivery Reviews, vol. 19, pp. 3-36.
Reitmeier, SJ, Mukti, RR, Jentys, A & Lercher, JA 2008, ‘Surface transport processes and sticking probability of aromatic molecules in HZSM-5’, Journal of Physical Chemistry C, vol. 112, pp. 2538-2544.
Rothweiler, H, Wäger PA & Schlatter, C 1991, ‘Comparison of Tenax TA and Carbotrap for sampling and analysis of volatile organic compounds in air’, Atmospheric Environment Part B. Urban Atmosphere, vol. 25, issue 2, pp. 231-235.
Ruiz, J, Bilbao, R & Murillo, MB 1998, ‘Adsorption of different VOC onto soil minerals from gas phase: Influence of mineral, type of VOC, and air humidity’, Environmental Science and Technology, vol. 32, pp. 1079-1084.
40
Seinfeld, JH & Pankow, JF 2003, ‘Organic atmospheric particulate material’, Annual Review of Physical Chemistry, vol. 54, pp. 121-140.
Thomas, K, Hoggan, PE, Mariey, L, Lamotte, J & Lavalley, JC 1997, ‘Experimental and theoretical study of ozone adsorption on alumina’, Catalysis Letters, vol. 46, pp. 77-82.
Tortora, GJ & Grabowski, SR 2003, Principles of Anatomy and Physiology, John Wiley & Sons, Hoboken, N.J.
United States Environmental Protection Agency 1972, Environmental Protection Agency Progress Report, United States Environmental Protection Agency, Washington D.C.
United States Environmental Protection Agency 1997, Final Revisions to the Ozone and Particulate Air Quality Standards, United States Environmental Protection Agency, Washington D.C.
Usher, CR, Michel, AE & Grassian, VH 2003, ‘Reactions on mineral dust’, Chemical Reviews, vol. 103, pp. 4883-4939.
Van Bavel, E, Meynen, V, Cool, P, Lebeau, K & Vansant, EF 2005, ‘Adsorption of hydrocarbons on mesoporous SBA-15 and PHTS materials’, Langmuir, vol. 21, pp. 24472453. Vijay, A, Mills, G & Metiu, H 2002, ‘Structure of the (001) surface of γ alumina’, Journal of Chemical Physics, vol. 117, issue 9, pp. 4509-4516.
Wefers, K & Misra, C 1987, Oxides and Hydroxides of Aluminium, Alcoa Laboratories, Pittsburg.
Whittington, B & Ilievski, D 2004, ‘Determination of the gibbsite dehydration reaction pathway at conditions relevant to Bayer refineries’, Chemical Engineering Journal, vol. 98, pp. 89-97.
41
Wilson, SJ & McConnell, JDC 1980, ‘A kinetic study of the system γ-AlOOH/Al2O3’, Journal of Solid State Chemistry, vol. 34, pp. 315-322.
Wolverton, C & Hass, KC 2001, ‘Phase stability and structure of spinel-based transition aluminas’, Physical Review B, vol. 63, issue 2, pp. 024102, 1-16.
Yang, X, Sun, Z, Wang, D & Forsling, W 2007, ‘Surface acid–base properties and hydration/dehydration mechanisms of aluminium (hydr)oxides’, Journal of Colloid and Interface Science, vol. 308, pp. 395-404.
Zaitan, H, Bianchi, D, Achak, O & Chafik, T 2008, ‘A comparative study of the adsorption and desorption of o-xylene onto bentonite clay and alumina’, Journal of Hazardous Materials, vol. 153, pp. 852-859.
42
Appendix I Physical Properties of the VOCs used in Adsorption and Desorption Experiments 2 Molecular Weight (g/mol-1)
Boiling Point (°C)
Dipole Moment (D)
Polarisability (× 10-24 cm3)
n-Hexane
86.18
68.7
0
11.9
Benzene
78.11
80.0
0
10.74
Toluene
92.14
110.6
0.375
12.3
o-Xylene
106.17
144.5
0.640
14.1
m-Xylene
106.17
139.1
0.33
14.2
p-Xylene
106.17
138.3
0
14.2
Propionaldehyde
58.08
48.8
2.72
6.5
2-Butanone
72.11
79.59
2.78
8.13
Benzaldehyde
106.12
179.0
3.0
-
Acetophenone
120.16
202
3.02
15
Benzyl Alcohol
108.15
205.35
7.71
-
VOC
2
Molecular Structure
(Dean JA 1992; Lide 2002)
43
Appendix II Concentration of p-Xylene Released in one Alveoli per 10 µm Diameter γ-Alumina Particle
Parameters Amount of p-xylene adsorbed after 120 minutes:
14 ng/mg
p-Xylene molecular weight:
106.17 g/mol-1
Area of one p-xylene molecule which was calculated from bond lengths and does not consider repulsions between two adjacent adsorbing molecules (D. Fleming [Chemistry Centre of W.A.] pers. comm., 23 October 2008):
0.359 nm2
γ-Alumina surface area (CSIRO 2007):
38 m2/g = 0.038 m2/mg
γ-Alumina density:
3.65 g/cm3
Mean size of a single alveolus (Ochs et al 2004):
4.2 × 106 µm3 = 0.0042 µL
Calculation
44
45
