Automated Particle Size Analysis.
KAWLEWSKl, RON, STEVE AAKRE, and JIM SCHUELLER, Southern MinnesotaBeet Sugar Cooperative, 83550 County Road 21 , Renville, MN 56284. Methods forautomated particle size analysis.ABSTRACTSouthern Minnesota Beet Sugar Coop utilizes a particle size analyzer, Rotex's Gradex2000, to automate sieve analysis for shipped sugars. We have added an Ankersmidparticle size analyzer that is used on our production sugar and a number of othermaterials. The unit has two modes of operation, laser Time-of-Transition and imageanalysis using a video camera. This paper will discuss the implementation of these twoinstruments and correlation of sieve data to 'the standard method hand sieving.Particle analysis is an important part of evaluating process efficiencies and productquality. Increased throughputs and a customer awareness of the impact of granulation onthe handling characteristics of sugar has necessitated investigating alternatives to theclassic sieve analysis. Automation of the sieving process was investigated and concludedwith the purchase of a Rotex 2000 in 1996. Alternative methods based on microscopyand video imaging techniques, and laser light instruments were recently reviewed forcost, ease of use, and data comparison to sieve data with an Ankersmid CIS-IOOpurchased in December 2005.Sugar is typically measured for size and distribution using a stack of woven wire testsieves. These can be used for hand sieving or placed in a mechanical shaker whichincorporates a motion that moves the sample across the screens with a secondary motion.TIlls secondary motion is usually provided by a large piece of steel on top of the stack.This tapping action will help Hft the sample off the screens. The weight retained on eachsieve is obtained and is typically used to calculate the Mean Aperture (MA) andCoefficient of Variation (CV). Various methods bave been developed to detennine thesevalues. The Butler method is what we use which calculates the weighted average valuedirectly from a nine sieve set without graphing of values.Sieve analysis is the standard method referenced by ICUMSA GS2-37 and is probablythe most widely used technique. The wide spread use of sieving for particle distributionis due to its low cost, reliability, and simplicity'. Sieving covers a wide range of particlesizes from 100 millimeters to 20 micrometers 2 and is well suited for free flowing drypowders. The reproducibility of this method can be influenced by factors such as the sizeof the sample charge selected, which can blind the screens. Additional factors includePercent moisture, particle shape, or agglomerates. This last factor can allow easierfracturing during subsequent bulk handling, such as pneumatic unloading, yielding a finerproduct.Sugar is essentially a non-friable material under the conditions specified for sieving andwill give consistent results across several test fUllS. Several samples were run in triplicateto test the durability of sugar under the standard gyratory motion with tapping for tenminutes.70
Triplicate Test 140.035.0.-- -- --- ------.-.--.------- --------- -- -. - ----.- ----. - - - I!30.0- --- - - -- ---- - - . -/--' - - - - - -- -- - - -- - 120.0 - -,.- - -- . -c/ -- I-.----- - :\:\ -- - - - - - --- -- -125.0j 15.0--;;e7/10.0 -- - -- - -5.0\-- -- ---.-- ---- -- --- --- -- ----.,0.0 -.Sieve1-- Test 1 Run 1 - - Test 1 Run 2Test 1 Run31One of the principle issues with sieving by hand is retention of the sample duringpreparation and weighing. A low relative humidity can impact the static charge presentand cause two issues. The first is the sugar particles will tend to cling to the sieves andgive results biased to the coarser sieves. Second a build up of static electricity can impartthe sugar particles with a tendency to repulse from each other, shooting particles out ofthe funnel and weighing cup. The use of metallic spoons and cups versus plastics greatlyassist in diminishing this effect.Recent years have seen increased demands for particle size analysis data. Customerrequirements for certificates of analysis include ever more stringent sugar particle sizerequirements. The use of a mechanical shaker with manual sieve cleaning and weighingof the fractions had long been utilized. As more screen tests were required, a concernwith repetitive stress injuries was becoming an issue. To resolve these issues, some wayto automate sieve analysis was considered.The Gradex 2000 used at Southern Minnesota Sugar is fitted with nine standard halfheight sieves.USS20USS 70USS 30USS 80USS 35USS 100USS 40USS 50Pass US 10071USS60
The sieves are fitted into aluminum ringswhich are in tum bolted to a belt. This beltprovides for attachment of the sieves fordumping and brushing and yet has enoughflexibility to allow for rotary shaking andtapping. The picture to the right shows thesieves mounted on the belt. The sieves areclamped into a frame for the shaking phaseof the analysis. This method of sieveattachment allows the Gradex 2000 tosupply the same rotary shaking/tappingaction as the more traditional RoTapshaker.After shaking, each sieve is dumped thenmechanically brushed. The contents ofeach sieve are then weighed and the data issent to a Pc. Using the weight data, thefractional and cumulative percentagesretained on each sieve are calculated asshown below.72
The fractional! % retained and the cumulative % retained results are then printed andimported into the daily lab sheets or the sugar shipment program for generation of thecertificate of analysis. Below is an example of the input sheet into which the percentcumulative data is imported with the calculation of the MA and CV for a strike sample.SusiaranptiData et - .1- .- --- -.J8adc1- s.lfm1WllE9t.12PI/OO9.n!I.l!K"'"45f.:!BWI 7!U)AA;.24HBIII'lIu fj 11.N1J111TeDy'IlII1Il1: T1Ie.I8y. Fenry 11. DJ1, '),0,. HItlII H:.lthon 11 dill I;UU.,.UkC.UJiO77.074.231A'J. 5.107 t Si. 9M"" In;mM !iWo'I1 PrICUwc"' - 'f.! 51.1:. f'fi,y,c:W.\. . m1714.00714.0071G.80000001. 2.28I10.7517.4811.! 26.3039.(t127.59,67.!!i81.5088.7691" ' .49S!.IGMl'ROOaMt'il- 69.0077." '81.88187.02 1!!t.9625-S!i lI,// mOIHlJi!!I' IIOCSOOMPROOP.\5SSOftfJIIllO.S'J(W, )IoI!\ru:wl1'P.1'»C'!!li0411I97. .1712.91]51o.(t1 09152 O140321434tom1922.4SI,ll.t,ftP!'1OD ""' lIIIIt 20.707220001.7588.749114 .29l ,:".,75. 41.01.019'!l 0H4 ':., ·saell1'll '1!ftlO 1lIIt't,.,·Jmj MG81ll!1111 2?!)9.XlAA?"n,68197683714.301101.00023nfDI2.7f1 :v:wt · ·CUti t'f'111 SOOM FItOO iH:! I L'UOO8192::aFJJ:; .l».l 19F14 &1.lAA,., I'fi'lilm :w.:i\0700111.01.000261.10568191.01.0Q.l1OlIfA TClDl.a 14.1L 681775.6o.lDI}; ·L rnI';'lIfi H 681674.2ImAi'Al, Illlltlia.1.01.000Il2.01500231.205SlftaTIILO. - 4.2200010013.8394.7997.41:1170703982.5689.029119 .292.5'.100151340(1144li84.76!!t: 171 ' - - ·11Below is an example of Gradex results for a shipment Lot which has been imported intothe su ar shi ment database for use in the eneration of the certificated of analysi s.Package Quality Analysis - New[iW:KAGi L BlAJ( [l! Pro!lrl:\1 Sblrt 0Me:1Yl/21/07Lot 1'Urber:su .2HSIf' ltBf A.--o:J It,."--0- .1-:u.,--iW140.6. .f,'11\ .1\1l1.!l:tJ,11ti3'r bn 'S. tHO'ilB)JlUJI.JA!11.5l7)- 2.5" .,.0.l1li9f"":''' ,."".D.0,GJ.lnt'V»ThIS I73lib lat ana' 15 number L.
Benefits of sieve analysis using the Gradex 2000:The use of the Gradex 2000 allows us to automate this procedure while still using amethod based on classic sieve particle size analysis. The software allows for the use ofapproximate sample weights for analysis with automatic calculation of % retained. Thisallows for simpler sample preparation especially when using riffled samples. The Gradex2000 has consistent dumping and brushing with improved reproducibility when comparedwith manual weighing of samples. Using the Automatic Sample Feeder allows theanalyst to set up and run up to seven samples unattended. This automation has allowedfor the increased frequency of analysis of various sugar samples such as production sugarfrom each white pan several times a day using the results from a nine sieve screen test todetermine the MA and the CV. Such an increase in frequency and number of sieves isalso a safety issue when considering the potential for ailments arising from the repetitivemotion involved in manually brushing out a set of nine sieves by an analyst. Comparisonto data from the Rotap has been very consistent with a small increase in the amountcaptured on the US 30 and US 40 screens, giving a slightly higher MA. One interestingnote is the same samples were analyzed on a vibratory shaker. These results gave ahigher MA than both the Rotap and the Gradex with larger fractions retained on thelarger sieves.A verages 0 flO 014233.1Shaker0.014932.6Ankersmid0.014336.2Some challenges using the Gradex 2000:The automation of the Gradex requires many moving parts which must be monitored,kept clean, and in adjustment for good operation and dependable results. In tens years ofoperation, learning and modification by Rotex Inc., most of the, initial problemsexperienced have been eliminated. Weekly cleaning of the sieves and the rings whichhold the sieves in place seems to be important for trouble free operation. Monitoring theperformance of the various air cylinders used in the analyzer, for leaks, is important fortrouble free operation. Frequency of cleaning of the analyzer can also vary depending onthe amount of "fines" in the sugar samples. More dust requires more frequent cleaning.Cleaning frequency and even sample results can be affected by static electricity chargedsamples. Low moisture, production sugar fresh from the cooler/dryer or handlingsamples in certain plastic containers can develop a static electrical charge strong enoughto prevent the sample from freely falling through the screen stack resulting in higher %retained on coarser sieves. This has been a problem a few times in the years of Gradex2000 use.Particle analysisWhen sugar shipments and production reached a level that required separate analysts, aninvestigation of alternative methods for granulation was initiated. Recognizing the needfor particle analysis in a variety of wet and dry samples over an extended size rangenecessitated an instrument with flexible application setup. Particle Size distribution74
(PSD) using microscopic and video evaluation along with laser diffraction and laser Timeof Transition were evaluated for their range of applications, ease of use, andreproducibility including comparisons to sieve data.Particle size, shape, and distribution are important to many factory operations. Sugar endPan Operations call for larger crystals and a narrow distribution. A wider distribution ofcrystals in the wall of sugar in the centrifugal becomes less porous and centrifugal washbecomes less efficient and effective. Longer cycle times and greater amounts of washwater required are the result. Conglomerates, with the inclusion of mother liquor, arealso difficult to wash and will contribute to higher color and ash.In storage and handling particle size, shape, and distribution affect the ability of thematerial to flow. Conglomerates will break apart easily exposing surfaces which mayrelease trapped moisture. Conglomerates breaking up, along with micro-particles, cancause dust hazards, including explosive potential.Customers want a product that will flow freely out of storage without lumps. Particlesize, shape, and distribution, along with other factors such as moisture migration, have alarge effect on flowability. All else being equal, larger particles will flow with less forceapplied than smaller particles, as win particles approaching spherical shape as opposed torods or other shapes. In the case of the centrifugal washing, a narrow distribution ofparticle size was desirable, but from a flowability standpoint a mixed distribution willflow more easily. The smaller particles get in between the larger ones and allow them topass by each other with less friction.Particle size and shape can also have an effect on our customer's process. One exampleis a powder mix used to make a flavored drink. If fine particles increase too much, alower density of the loose powder may result. If packaged by weight, the product may nolonger fit in the packaging. Large particles will I)ot dissolve as fast in water and finepowders may float on top or form clumps, both of which are difficult to dissolve.Ankersmid CIS-IOOLaser diffraction, laser Time-of-Transition, and video image analysis were thetechnologies demonstrated. Laser diffraction uses a system with a laser, lenses, andmultiple detectors or detector zones. The theories used are the Fraunhofer and Mietheories. When the laser beam passes thru the particle zone the beam is diffracted byparticles and the detectors produce a pattern which is interpreted by the Fraunhofer orMie theories. This type of analysis assumes spherical particles. Because of thisassumption it is difficult to correlate with sieve fraction data for non-spherical shapedparticles like sugar.The ability to perform particle analysis on dry or liquid suspensions from submicron tomillimeters was very attractive. A number of different manufacturers were asked todemonstrate their particle size analyzers. Based on features, performance and price wepurchased the Ankersmid CIS-I 00.Laser Time-of-Transition is one of the analysis forms used by the Ankersmid unit wepurchased. It is much more straightforward. A laser beam is passed thru a wedge prism75
rotating at a fixed rate (200 Hz) creating movement of the laser beam in a circular path.Samples are place in the laser path and a photodiode detector is placed at the end oppositethe laser. Gravity fed or liquid suspended particles are detected by interruption of thelaserlphotodiode signal. Since the laser is moving at fixed rate,the time that the laser isobscured is used to determine the size of the particle. The shape of the signal Transitionis used to reject particles encountered out-of-focus or off-eenter. If the sample particlesare transparent or translucent the analysis can be setup to handle the signal differently aswell. The use of Time of Transition makes the analysis independent of the particles'shape or its real or imaginary refractive index that is generaUy required for LaserDiffraction. We have used the laser ToT for analysis of liquid samples includingcarbonation, centrifugal wash water, and wastewater samples. This method suffers fromthe same shortcoming of assuming spherical particles.lA.C Dl'r-- - -- - - -- - - - - -- -.- IA,.Diagrams from Ankersmid B. V.Production sugar is analyzed using the video channel. Images are acquired in real-timeand can be analyzed for many size and shape parameters. One parameter that may be ofinterest is the shape factor. This is indicative of how smooth or rough the sampleparticles are and may indicate the presence of agglomerations. Another shape parameteris the Aspect ratio. This parameter indicates the ratio of a particle's width to its length. Ahigher aspect ratio indicates a spherical particle, while a lower ratio indicates a rod shape.This Transition in particle shape may be indicative of the presence of crystal habitmodifiers like raffinose or dextrans. The variation of the sample's aspect ratio can yieldinteresting difference in the size distribution results of sieve analysis versus laser or videoimaging techniques. Particle analysis with sieves uses a two dimensional aperture todetermine the size of a three dimensional particle. Particles wiU pass through a sievebased on its 2nd minimum diameter. The probability of a rod shaped particle passingthrough a screen depends on its length in addition to the motion and force applied to thescreen set. The gyratory motion combined with a tapping force of standard methods winhelp fluidize the bed of sugar particles and allow them to transition to the verticalposition needed for passage. Factors such as the size of the charge load change the depthof the particles on a screen and can affect the efficiency of fluidization.76
Laser and video image analysis also take a two dimensional look at the sugar particle. Asthe sugar moves across the detector it leaves a shadow that is captured for analysis. Thisdifference will be seen as the orientation of the particles' axis win vary and the particlesize willi be seen as an averaged result. By using the video mode, the CIS-IOO can be setto analyze based on the Minimum Ferrets or diameter. The use of the minimum ferretparameter allows us to come close to the second smallest diameter found in sieveanalysis.Sampling methodsThe initial testing of the Ankersmid was done with duplicate samples using a chute typesample riffier. The variation of the initial results was disappointing. White repeatedanalysis on the same sample was very reproducible, the comparison between hand sieveand the video analysis had a high level of variation One step was to improve oursampling and splitting methods. Various splitting techniques were tried and the samplestested to measure the level of variation introduced by this step of the process. Thefollowing table lists the expected level of variation from several standard methods ofsample splitting.Reliability of selected sampling methods using a 60:40 sand mixtureSampling techniqueStandard deviation6.81Cone and quarteringScoop sampling5.14Table sampling2.091.01Chute slitting0.146Spinning rifflingRandom variation0.075.Table from Table 1.5 page 38, T.Allen PartIcle SIze Measurement, 5th EdItIon Volume 1Chapman and Hall 1997 ISBN 0 412 729504The solution to variation in sample splitting was to run the same sample on both the videoinstrument and the Gradex. When analysis for production sugar was run on theAnkersmid unit we would coHect a composite of four runs. This composite was thenanalyzed on the Gradex for comparison with the average of the four from the Ankersmidanalyzer.The evaluation of MA and CV data derived from sieve and video analysis has beensatisfactory and yielded good results even with the differences in handling of the crystal'sphysical differences.One addition is the ability to provide particle distributions based on vo lumes, which isused to compare to sieve analysis, but also on numbers. This distribution provides asignificantly different outcome. The impact of the large number of fme particles in agiven volume or weight of a sample is clearly evident. A particle size distribution by77
numbers will give a double peaked graph showing the large number of particles smallerthan 150 microns.One measurement of concern was the results on individual screens. Granulation controlfor optimizing operations is important but is normally based on MA and CV data. Inaddition the coarse (US 30) and fine screen (Pass 100) data is also important to meetingmany customer specifications. In this instance matching data between the two methodswas difficult to correlate, especially for the sub 150 micron particles of the pass 100screen. Samples of the fractions caught on individual screens were analyzed on the CIS 100. These samples often showed residual particles that did not pass through the sieves,and was collected on coarser screens but which were found using video analysis.Overall the data provided matched well when averaged, but determining the results forindividual ranges was effected by how the instruments handled the rod shape of the sugarcrystal and the finer particles. In order to provide specific screen data, the video andsieve analysis were used to develop a correction factor to reliably predict actual sieveresults.ClS-100 I Gradex FractionalComparing a 4-Strlke CIS average with composite Gradex0.0175 .-- - - - - - --- - - --- - - - - -·f 50- -.0.0170 -- -- - - - - - - -- -- - ---
Coefficient of Variation (CV). Various methods bave been developed to detennine these values. The Butler method is what we use which calculates the weighted average value directly from a nine sieve set without graphing of values. Sieve analysis is the standard method referenced by ICUMSA GS2-37 and is probably the most widely used technique.
Particle Properties Particle size By far the most important physical property of particulate samples is particle size. Particle size measurement is routinely carried out across a wide range of industries and is often a critical parameter in the manufacture of many products. Particle size has a direct influence on material properties such as:
property tests. The best particle model had a particle coefficient of restitution of 0.6; particle static friction of 0.45 for soybean-soybean contact (0.30 for soybean-steel interaction); particle rolling friction of 0.05; normal particle size distribution with standard deviation factor of 0.4; and particle shear modulus of 1.04 MPa. Keywords.
Particle counter channels are set based on a particle size standard. Realistically, particle standards seldom size exactly within a particular size channel. Using 0.3 µm particles as an example, most particles are . Cleanroom air often has laminar flow with velocities ranging from 45 to 90 ft/min. Some probes are sized to
Sieve Analysis Particle-Size Distribution Curve End of Part 1 Particle-Size Distribution Curve The results of mechanical analysis (sieve and hydrometer analyses) are generally presented by semi-logarithmic plots known as particle-size distribution curves. The particle dia
q C to heat exchanger for test and evaluation Solex/VPE/Sandia particle/sCO2 shell - and - plate heat exchanger Heat duty 100 kW T particle,in 775 q C T particle,out 570 q C T sCO2,in 550 q C T sCO2,out 700 q C I 6 0.5 kg/s High - Temperature Particle Receiver Particle receiver testing at the National Solar Thermal Test Facility
zation six samples taken at random were subject to particle size analyses using a Horiba LA-950 Light Scattering Particle Analyzer. For each sample, the particle size mean and range, and the percentage of particles below a given size was calculated. On average, the particle size of CRM AO
Helps determine degree of reduction based on material being processed Screen TyPe Helps regulate particle output within a specified size range roTor SPeed Works with screen to regulate particle output within the size range The BeneFiTS oF conTrolled ParTicle reducTion Particle size affects any number of characteristics in the manufacturing process.
1/4” 4garta size 3/8” confdy size.250” conirsb10 confdr110 size.110” confdb187 size.187” conifdb110 size.110” conbmr size male conbmb size male conifdy size.250 conifdb205 size.205” conifdb187 size.187” conbfr size female conbfb size female conifdb size.250” confdr size.250” confdb si
