More From This User
Related Docuements
A As
From Yudi Permana
A As
From Ali Haider
VARIANINSTRUMENTS
AT WORK
ATOMIC ABSORPTION
The Measurement of Silicon, Tin, and Titaniumin Jet-Engine Oil
Number AA-59
April, 1986.
Michael Knowles,
Varian Techtron,
Mulgrave, Victoria, Australia.
Introduction
The analysis of used engine oils by spectrometrictechniques for the early detection of machine componentfailure has been an important area of investigation sincefirst applied by railroad companies in the 1940s (1).The long term monitoring of wear metal levels inlubricating oils has been undertaken by the armedservices, major
road transport companies, tractormanufacturers, airlines and railways.A wear metal profile is often prepared for each engine.Typically an early wear-in period of high metalsconcentration is observed, followed by a plateau of lowmetals concentration (2). Impending component failurewill be indicated by a rapid increase in wear metalconcentration, or the sudden appearance of a metal. Themetal type and concentration may also indicate to theanalyst and engineers which part of an engine is failing.A change in the concentration of sodium for example,may indicate an antifreeze leakage, the increasedpresence of lead may indicate the wear of bearings in adiesel engine (3).Many different analytical techniques have been applied towear metal investigations. The most widely usedtechnique for wear-metal analysis in the last decade hasbeen flame atomic absorption spectroscopy (FAAS) (4).In more recent times, graphite furnace AAS and atomicemission spectroscopy (AES) methods have beenincreasingly employed.Despite this interest and continued development, caseshave been documented where a Spectrometric OilAnalysis Program (SOAP) has failed to predict aircraft oil-wetted component failure (3). This has been attributed tolimitations in the particle size detection capabilities ofmajor SOAP techniques such as rotating disk electrodeAES and flame AAS.Typically wear-metal particles less than 1 micron indimension can be readily determined by spectrometrictechniques, but particles greater than 5 microns(indicative of more severe wear) may go undetected. Thisproblem is usually caused by the inability of the sampletransport system to move such particles or to the failure ofan atomization process to break up the particles so thattheir components can be atomized. In addressing thisproblem C.S. Saba and co-workers have developedParticle Size Independent Methods (PSIM) for wear-metals by the reaction of the used oil sample with amixture of hydrofluoric, hydrochloric and nitric acids
(1 and 5-9).
In a recent publication, however, Saba et al. (1) noted in acomparative study of spectrometric analysis techniques,that only a graphite furnace atomization AAS method wascapable of analyzing iron particles from the submicron to20-30 micron (unfiltered) size, as was the PSIMdeveloped by these workers. Flame AAS, inductivelycoupled plasma AES, rotating disk AES and DC argonplasma spectrometry all showed serious particle sizedetection limitations.In the study described here, the capabilities of the VarianGTA-96 graphite furnace AAS system were examined forthe determination of silicon, tin and titanium in jet-engineoils. These elements were selected because:
(a)
They are reported relatively infrequently in the
literature (4).
(b)
They exhibit relatively poor sensitivity whendetermined by conventional flame AAS techniques
(10).
(c)
They are of vital interest in jet-engine oil analysisprograms.Table 1 (2, 3, 13) indicates the possible sources of theseelements in engine oils and lists the expected flame AASworking ranges for these elements.The use of graphite furnace atomization withautosampling in this work, allows greater sensitivity withsmaller sample consumption.
Instrument Parameters
Silicon (flame)
Spectrometer
Instrument modeCalibration modeMeasurement modeFlame
Acetylene
Nitrous oxide
Wavelength
Slit widthLampLamp currentSilicon (furnace)
SpectrometerAutosampler
Instrument modeCalibration modeMeasurement modeFurnaceGas
WavelengthSlit width
LampLamp current
Tin (furnace)SpectrometerAutosamplerinstrument modeCalibration mode
Measurement modeFurnace
GasWavelength
Slit widthLampLamp currentTitanium (furnace)
SpectrometerAutosampler
Instrument modeCalibration modeMeasurement mode
Furnace
GasWavelength
Slit width
LampLamp currentVarian SpectrAA-40AbsorbanceConcentrationIntegration (3 seconds)Nitrous oxide-acetylene6.51 litres/minute
11.8 litres/minute
251.6 nm0.2 nmVarian; silicon HC
10mA
Varian SpectrAA-40VarianAbsorbanceStandard AdditionsPeak heightVarian GTA-96Normal: Nitrogen251.6 nm0.2 nmVarian; silicon HC
10mA
Varian SpectrAA-40VarianAbsorbanceStandard AdditionsPeak heightVarian GTA-96Alt: argon235.5 nm0.5 nmVarian; tin HC
7mA
Varian SpectrAA-40VarianAbsorbance
Standard Additions
Peak height
Varian GTA-96
Alt: argon364.3 nm0.5 nmVarian; titanium HC20 mA
Standard Preparation
Multielement organometallic standards in Mobil Jet-Oil IIwere supplied by an Australian organization operating aSOAP for aircraft maintenance. The concentrationsranged from 5 to 200
µ
g/mL, and contained the elements:Ag, Fe, Mg, Cu, Al, Si, Cr, Sn, Ni and Ti. Theseconcentrated standards were less than five months old.For the determination of silicon by flame AAS, standardswere diluted four-fold (1 part standard plus 3 partssolvent) with MIBK (methyl isobutyl ketone, BDH Analargrade) to achieve the required concentrations. Standarddilutions were identical to ensure matrix matching.For the determination of silicon, tin and titanium byGTA-AAS, the standard additions method of calibrationwas used to ensure matrix uniformity. Standards werediluted as necessary with MIBK.
Sample Preparation
A sample of used oil from a jet-aircraft engine, was dilutedfive fold in MIBK (1 part sample plus 4 parts solvent) forthe determination of tin and titanium.
Sample Handling Notes
The handling of both thick oils and low viscosity volatileorganic solvents for wear-metals in oils analysis requires
different sampling protocols.
Some points to consider are:
(i)(ii)
(iii)(iv)
(v)
(vi)
Oil samples and standards will settle with time.Agitating the sample oil in an ultrasonic bath forseveral minutes prior to or during sampling providesa convenient method of overcoming this problem.Two techniques are used for the reproduciblemeasurement of sample size — weight and volume.Sample weighings offer convenience in terms of
reproducibility. Sample volume measurements
require a strict protocol regarding pipette filling anddraining procedures. A typical drain time could be aslong as one or two minutes for viscous fluids. Caremust be taken to ensure no fluid is carried on theoutside of the pipette.Dilutions in this work were carried out by volume
measurement.
Samples and standards should be stored at lowtemperatures to guard against evaporation of
volatile solvents. Fresh dilutions of concentrated
standards should be made before the start of theanalytical program of each day. Low concentrationstandards (less than 1
µ
g/mL) may deteriorate overa period of days.Contamination and toxicity also present specificproblems in oil analysis. Volatile solvents should behandled in a fume cupboard.The autosampler should be protected with its dustcover and in many solvent handling situations theoperator would be advised to wear gloves.Safety practices for the handling of organic solvents,particularly for flame AAS, must be strictly followed.A good summary of these practices can be found inReference (11).Reproducible sampling techniques are a major steptowards achieving reproducible results. Whensampling from the aircraft it is recommended (2) thatthe sample be taken when the engine is at normaloperating temperature. A sampling tube is inserted
2
into the dipstick port to a specified depth and thesample withdrawn by syringe (5-20 mL).
Results and Method Development
To achieve optimum results care must be taken with thegraphite furnace sampler alignment when using organicsolvents. The low viscosity of the organic solventsrequires the positioning of the dispensing tip much closerto the furnace wall than would be the case for aqueoussamples.This ensures that the solvent does not creep back up theoutside of the sampler tip, which would cause errors in thevolume dispensed. A further problem of creep can occur ifthe dispensed sample expands as the tube is heated anda small amount rises out through the tube hole, aftercreeping along the tube walls. This increases tube wearand adversely affects precision. The problem can beovercome by careful sample alignment. The tip should bekept low and central in the tube and small samplevolumes should be used (10-20 microlitres). The graphitetube was of the partitioned type which is especiallysuitable for organic solvents.The autosampler rinse solution was made up ofapproximately 0.001% Triton X detergent and 0.002%nitric acid. Use of this rinse solution ensures thoroughcleaning of the sampler tip to give proper dispensing
without contamination.Silicon In Jet-Engine Oil By Flame MS
For the determination of silicon by flame AAS, the uptakerate for the diluted samples and standards was set at 2.7mL/minute. Samples and standards were all prepared in
MIBK.
Selection of the uptake rate was a compromise betweenminimum flame flare-up, absorbance noise, avoidingflame lift-off in the absence of solvent, minimizing sampleconsumption and achieving maximum sensitivity.Background correction was selected, but it is not strictlynecessary as the background from the flame was virtuallyeliminated by the optimization of conditions.Carbon build-up on the burner slot may occur especiallywhen using fuel-rich nitrous-oxide acetylene flames. Thiswas not found to be a problem with the silicondetermination, but several precautions were taken tominimize the possibility of build-up. The burner wascleaned thoroughly before analysis in a detergentconcentrate in an ultra-sonic bath. The burner was thenrinsed with tap-water and distilled water. The burner wasdried with a blast of air. After installing the burner andigniting the flame, the burner was allowed to warm upcompletely on a lean flame prior to setting the gas flowsrequired for the analysis. When developing a suitablemethod, the preferred conditions for minimum carbonbuild-up are: high total gas flows and the leanest flamepossible. In the subsequent silicon determination novisible carbon build-up occurred in the time required toproduce a calibration graph and sample results.Results obtained are shown in Table 2. Acceptable levelsof precision for silicon determination could be predictedfrom this data for levels above 10
µ
g/mb. The calibrationgraph in Figure 1 shows a high degree of linearity even atlow levels. Error bars are shown in this graph whichreflect the precision of the replicate measurements.From these results a characteristic concentration can becalculated for silicon analysis in jet-engine oil. Thecharacteristic concentration (CC) is that concentrationwhich gives an absorbance of 0.0044. The value obtainedfrom this analysis (1.5
µ
g/mL) is the same as theexpected literature value (12) for aqueous samples.
Silicon ln Jet-engine Oil By Furnace AAS
The determination of low levels of silicon in jet-engine oilis greatly facilitated by the use of graphite furnace AAS.Although individual measurement times are longer thanthose for flame AAS, the following advantages enhance
its applicability:
(i)
After initial program development and system set-up, virtually unattended operation is possible.
(ii)
Improved sensitivity.
(iii)
Reduced risk from the use of flammable solvents.
(iv)
Greatly reduced sample consumption.A deuterium background corrector was used. The furnaceprogram developed is presented in Table 3. The programconsists of a two-stage dry phase, ashing at 1000°C andatomization at 2700°C. The sheath gas used wasnitrogen, it would be expected that argon would givegreater sensitivity. The other instrumental conditionswere the same as used for the flame determination.The automatic sampler parameters are presented inTable 4. The programmable sample dispenser used withthe GTA-96 automatically mixed the standard additionsfor deposition in the furnace. The total volume dispensedwas 20 microlitres. The standard solution was 25 ng/mLof silicon.Figure 2 is a signal graphics trace for a typical standardadditions signal. It is evident from this trace thatbackground levels are negligble, while the atomic signalis sharp and symmetrical. The temperature profile is alsoshown on these signals graphics.Table 5 shows the results obtained from this analysis.The precision of the measurement (%RSD) is good. Thecalibration graph is presented in Figure 3. Thecharacteristic concentration calculated from this data is7 pg. This compares favorably with the literature value of25 pg (13). Based on these figures, analyses could becarried out in the low parts-per-billion range.
Tin in Jet-engine Oil By Furnace AAS
The development of a method for tin by graphite furnaceAAS indicated that a short dry time would be possible inthe furnace program. This permitted the development of atemperature program which is complete within oneminute. The possibility of background interference wasthoroughly investigated, and it was found that backgroundcorrection was not necessary.
3
of 00
Application Note_The Measurement of Silicon Tin and Titanium in Jet-Engine Oil
2,548 Reads
Info and Rating
| Category: | Uncategorized. |
| Rating: | |
| Upload Date: | 01/16/2009 |
| Copyright: | Attribution Non-commercial |
| Tags: |
Leave a Comment