coulter counter p00197-p00211.pdf

7/24/2019 coulter counter p00197-p00211.pdf

http://slidepdf.com/reader/full/coulter-counter-p00197-p00211pdf 1/15

j. Soc.Cosmetic hemists 7 197-211 1966) •) 1966 ocietyfCosmeticl•emistsfGreat ritain

Particlesizeanalysis sing

Counters

Coulter

W. M. WOOD and R. W. LINES*

Presentedt theSymposium n PhysicalMethods, rganisedy theSociety

of Cosmetic hemists f GreatBritain, in Bristol on I6th November 965.

1tyno10sis--A elatively new instrument for the size analysis of most forms of particulate

material is described. The instrument sensesparticles suspended n an electrolyte by their

momentary displacement of electrolyte, causing an increase of resistance o applied current,

as each passes hrough a small hole, or orifice, in an insulator. Passage s essentially singly,

although correctionscan be applied for any coincidenceoss. Counts at up to $,000 particles

per second re possible, nd the resultingsizedistribution s built up in some wenty minutes,

including calculation time, although this can be reduced on a routine basis.

A technique for the size analysis of wide range powders, i.e. those wider than the range

of resolution of any particular orifice, is discussed.

The range of new models available working on this same basic principle is reviewed.

These reduce calculation time considerably,as well as ncorporating many other refinements.

A series of particle size distribution on a wide range of cosmetic materials illustrate the

usefulness and versatility of the instrument.

The realization that the size distribution of powders, and other

particulate material, is of critical importance o the final product of that

material has been achieved in every industry using powders. Particle

size affectssuch things as the definition obtained from phosphors sed to

coat television tubes, the workability of a metal alloy, the grittiness of

food products, the colour of pigments, the solubility and efficiencyof

pharmaceuticalpreparations, and the responseof control systems n

modern high speed aircraft.

No less mportant are the particles n the cosmeticsndustry. Simple

* Coulter Electronics Ltd., Dunstable, Beds.

197



198

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

powder ormulationssuchas face powder, alc and rouge depend or their

efficiencyon the finenessof the powder. Being too fine, however, means

a possiblehealth hazard from inhalation.

Colour n lotion is a function of particle size and a colour can be made

to appear to have two different shades by a change in particle size.

Liquids such as perfumes, colognesand toilet waters must be filtered

efficientlynot only for their good appearance,but also for health reasons

since poorly filtered products can contain substancesharmful to the

skin..

The earliest form of size analysis equipment was the sieve, followed

by the microscope,he sedimentation echniques, nd many other methods.

All of these methodshave their good points but in the main it is fair to

say that they are not suited to the demandsof a modern ndustry,which

requiresan automatic method which s independentof operatorerror, and

which measures ach particle individually and accurately.

Figure 1 Diagram of Coulter Counter principle

Coulter (1) in 1956 described n instrument or the automaticcounting

and sizingof blood cells. It has sincebeen shown hat the sameprinciple

can be applied to a wide range of other particulate materials,with equal

speedand accuracy. The sizerangecoveredby the instrument s approx-

imately 0.5•-400• and the only imitation s that the sample o be analysed

must be suspendedn an electricallyconductive iquid. A stirring system

can be employed o avoid settlingeffectsduring the size analysis.

A diagramof the instrument s shown n Fig. 1. The sample o be



PARTICLE SIZE ANALYSIS USING COULTER COUNTERS 199

analysed s suspendedn a suitableelectrolyte n a beakerand placedon

the beaker platform. A glassorificetube, having an accuratelymade

aperture n the lower end projects nto the beaker, he insideof the tube

being filled with the same electrolyte as in the beaker. On either side

of the orifice s an immersed lectrode. By applyinga controlled acuum

to the orificeand the mercury manometersituated behind it, liquid and

the suspendedarticlesare drawn through he aperture. The passage f

the particle through he orificecauses momentary ncreasen the resis-

tance to the current which is simultaneously assing hrough the orifice.

This increase s detectedas a voltage pulse, proportional o the volume

of the particle. This pulse s then amplified,scaled,passed hrough an

adjustablehreshold nd counted,f it exceedshat thresholdevel, when

desired.

Closing he top tap on the control piece cuts off the vacuumand the

returningmercurycolumncontinues ample low through he orifice. Set

in the manometer are a series of electrodes which enable one to obtain

a particlecount n varyingsample olumes 0.05,0.5 and 2 ml). A single

count aboveany givensizecan be made n some15 sec,and by a repetition

of this process sizedistributioncan be made over as many pointsas may

be necessary n a very short time.

Since he Coulter Countermeasureshe volumeof the particlesdirectly,

conversionof particles to volume per cent or for particles of uniform

densityweight per cent, presents ew calculationproblems. An example

of a typical data sheet s shown n Fig. 2.

With the Coulter Counter one of the essential eatures s the quality

of the dispersion, hich is, of course,common o all other sizingmethods

using a suspension.The Counterwill count and size anything that is

presentedo the orifice,so t is vital that one decideswhetherone equires

the size of the particles n the powderor liquid as they are in the original

sample, or the size of the discreetparticles, and accordinglyselect the

appropriatemethod of dispersion.

Many methodsare available for this purpose,but the one which is

currently finding fayour is the use of ultrasonics. Small laboratory baths

are now commerciallyavailable into which are placed the beaker, the

sample ogetherwith the electrolyte lussomeotherdispersant,f needed.

The time interval needed o obtain a reproducibledispersions usually

between 15 secand 2 min, according o the easeor difficulty of dispersion.

An exampleof the reproducibilityof the methodof dispersion, ampling

and the Counter is shown in Table I:



2OO


ii

Figure œ Typical data sheet



PARTICLE SIZE ANALYSIS USING COULTER COUNTERS

201

Table I

Reproduced results on a sample of powder

Cumulative weight % above stated size

Particle diameter

p Run 1 Run 2 Run 3

35 1.1 1.4 1.7

25 9.8 10.4 10.7

20 23.8 26.9 29.5

15 48.0 50.1 52.0

10 78.3 80.1 80.5

7 91.2 93.1 93.4

5 96.8 98.0 98.1

3 99.4 99.6 99.7

2 99.75 99.83 99.93

This was a medium-pricedpowder and the electrolyte used was 5%

trisodiumphosphate, nd the time in the ultrasonic ath was20 sec. From

Table it will be seen hat the reproducibilitys within some3 or 4% on

a weight basis at the most sensitivepart of the distribution.

Once he sample s dispersedt is essential hat no flocculationof the

particlesoccurs,as the Counterwill accuratelycount and size any particle

presented o the orifice. In order that this can be prevented t is usual

to add some dispersant,preferably a nonionicsuch as Nonidet P.42 to

the dispersed ample,whilst one can alsomechanically tir the suspension

during the analysis. This fact enables one to study the stability of

emulsions, olubility rates and the effect of dispersants,locculants,etc.,

with an ease which has previouslybeen unobtainable.

The choice of the electrolyte is also of considerable mportance, the

selectionbeing dependentupon the nature of the material under investi-

gation. One does not therefore select a solution which would either

dissolve or flocculate the material under test.

Wherever possiblea sodiumpolyphosphate olution,such as Calgon,

is used as this is itself a dispersantat a concentrationof 2-4% by

suchascan be usedon the Counter tself. This type of electrolyte s ideal

for the majority of insoluble owders uchas alc and rouge. For emulsions

of the o/w type sodiumchlorideat 1% by weight can be used. If the

material s water soluble hen t is possibleo usea non-aqueouslectrolyte.

A material of this type is calcium carbonate, he electrolyteused being

5% lithium chloride n methanol.

Fig. 3 shows he resttitsobtainedon a seriesof typical toothpastes,

the electrolyte n questionbeing the one for calcium carbonate. The

slight solubilityof the materialunder est is counteracted y prior satura-

tion in the electrolyte.



202


CUMULttTIVE Wœ1½I4T PERCENT A•OVE $Tt•TœP .•IZE

Figure 3 Particle size distributions of typical toothpastes and powders (Logarithmic--

probability plot)

The response f the CoulterCounter s relatively independentof particle

shape, n that the true volume is always measuredover the size range

covered by each orifice.

It has been shown hat the change n apertureresistance ausedby

a particle passing hrough the orifice is

1

Po.V 1

AR- A Po•--A

1--•-

Where: Po = electrolyteresistivity

A = aperturearea normal to axis

V,P,a = particle volume, effective esistivity and area normal

to aperture axis

X = particle dimension atio =

1/d lengthlongperturexis

diameter of equivalent sphere




Thus for any given electricalconditionand aperture size, responses

essentially inear with particle volume, giving insignificanterror (1%, on

a volume basis)provided that the maximum size measuredby any one

orifice tube is below 40-50% of its physicaldiameter.

Particle resistivity has also been proved to have no significanteffect

on instrument esponses. All powders,once hey are in suspensionehave

as non-conductors, his being attributed to either an oxide film on the

particle surface or the Helmholtz Double Electrical layer, surrounding

each particle. It can also be seen from the above equation that particle

density cannot affect response. However, data reduction from volume to

weight per cent cannot be made in the caseof a mixed powder of varying

densities unless the ratio of densities is known and can be attributed at

each particle diameter. Temperature changemainly affects the response

by the change n electrolyte esistivity. Under normal workingconditions

this is negligiblebut a simple correctioncan be made if necessary.

Particle concentrationused in the Coulter Counter s that which gives

a maximum number count below the limit of coincidence effects for the

orifice tube in use. In terms of sample amount used this is usually of

the order of 20-30 mg in 150 ml of electrolyte.

As the method involvesthe countingof particles suspendedn a liquid

it is obvious that for any desired accuracy the particulate counts must

be significantly higher than that of the base electrolyte. Ideally this

solutionshouldbe completelyparticle free but this is a practical mpossi-

bility.

In practice, filtration is carried out with celluloseacetate membranes

of the Millipore ype, and theseare satisfactory or aqueous olutions, ut

for non-aqueousmedia a glass ibre paper must be used. This, of course,

must be supported on a glass sinter to prevent fibres from the paper

getting into the electrolyte. Calibrationof the CoulterCounter s usually

madedirectlyagainstparticlesof a knownsizesuchas polystyreneatices,

spores nd pollens,all of which have a fairly low standard deviationabout

their mean.

As particle volume is measureddirectly by the instrument settings

(thresholddial (t ) and aperturecurrent switch (I)) then the calculation

ofparticleiameter, sequaloK x 3•/instrumentettings,being

the calibration constant or that aperture tube and electrolyte system.

Since the instrument producesnumber versussize it can be used for

the on-stream ontrolof solutions uchas eau-de-colognes.A typical set

of results are shown in Table II.



204


Table II

Particle count obtained on various toilet waters

No. of particles per 100 ml of solution

Particle diameter

/• Eau de cologne Lavender water Perfume

15 1,500 0 20,000

10 9,800 80 113,500

8 11,300 1,000 233,300

6 11,500 1,800 450,000

4 88,000 19,400 1,230,000

2 270,000 355,000 9,000,000

1 60,000,000 5,000,000 --

The smallerthe size the more particleswere present. In the caseof

the perfumewe wereunable o count he vast numberspresentat 1•.

•00 ' ' '

90'

70'

+ 'PUFFER'AL

7

CaHULATIVE •fl•flT PEREEN• ABOVK5TATKP SIZE

TALCUM POWDERS

The particle size of all talc is usually ess han 300 mesh (53•). In

particularmostbaby powders re usually ess han 325 mesh 44•). The

optimum range is 10-60•. Fig. 4 shows he resultsobtainedwith the

Coulter Counter on various grades of talcum powder.




It will be seenthat they range in size from 65• down to 1.5• with a

wide range of mean sizes.

A 30.0[•

B 24.6[•

C 15.8[•

D 12.2[•

E 8.7[•

The baby powder is sample D and the range covered s 36-2• with

a mean size of 12.2•. This would appear to be somewherewithin the

specification.The electrolyte nd dispersion assimilar o thosepreviously

mentioned.

With some materials the size range lies beyond that covered by one

orifice tube, the range of one tube being approximately 1 or 2-40% of

its stated diameter. Clearly when the particle countsare low this upper

limit can be extended o 50% or so since he statisticalerror (q- •/•_)

outweighs he lack of accuracy of responseof the sizing circuit. It is

possible hat somematerials have more than 5-10% by weight outside

this range on the most suitableaperture. Below this level extrapolation

techniques an be used with some degreeof accuracy. Extrapolation

techniques,however, usually assume hat the sample n question ollows

a log-normaldistributionwith no secondpeaks or other variations rom

normal. This assumptions usuallyquite correct,but there will be times

when an analysishas to be made of the entire size range. Several tech-

niques or this type of analysis xist wherebyone combineswo or more

orifice tubes to cover the range. Basically the techniquesare analogous

to a microscopenalysis a largefield is scanned nder ow magnification

followedby a smaller ield under high power, and the distributionbuilt up.

The standard method is as follows: A suspension f the material under

test is made up as usual and analysedon a tube suitable for the largest

particlespresent. When the lower imits have beenreached, he remaining

sample s removed and a smaller tube fitted to the Counter. The sus-

pension s sieved or allowed to settle in order to remove the oversize

particles, which would tend to block the smaller tube. Sieving is best

accomplished ith electro-formedmicro-mesh ieves,and theseare most

efficientwhen usedto wet-sievedilute suspensionsf the type used on

the Coulter Counter. The analysis s then continued,overlapping he last

few points of the larger tube down to the end of the particle systemor

until the lower limit of the smaller tube is reached.

It can be argued hat whatever method of removing he large particles



206 JOURNALOF THE SOCIETYOF COSMETIC HEMISTS

from the system s employed omesmalleroneswill alsobe removed. In

practice, owever,he error s insignificantsuallybeing ess han 2% in

termsof number,and decreasing, hich s negligible n a numberor weight

basis. The calculation hen proceeds s normal.

The time taken or a typicalsingle ube analysisncludinghe calcu-

lation is 20 min. Two tubes require perhapsanother 5-10 min. Times

can normally be halved on a routine basis.

Figs.5 and6 show he sizeanalysis f variousothercosmetic aterials

obtained with the Coulter Counter.

IOO

qo

90.

70.

6o.

fro-

7'

6

cunuL•rtvœ [vœtc•lr p•l•cœ/vr ,•ovœ 5rArc• stzœ

Figure 8 Particle size distributions of various cosmetic materials

The Coulter Counter being akeady partly automatic is ideally suited

to further automation. New Models B and C now lend themselves to

this system.

The Model B has two counting circuits so that a frequency number

distributioncan be obtaineddirectly. Amongthe many other advantages

is that the Model B responses independentof electrolyte resistance o

that one calibration factor will hold for changes n electrolyte resistivity

causedby concentration hangeor temperature ncrease. It also lends

itself to easierand faster data reduction. Using a Model J plotter coupled




•oo

•o

7

CUPIUL•TI•E •/EIC#T ?œ•[EAIT fiOVœT•TEP SIZE

Fiõure • P•rt•cle s•ze distributions oœ¾•r•ous cosmetic

to the B Model a frequencyor cumulativenumber histogramcan be auto-

matically obtained in 110 sec over 25 points with no operator attention.

Each plot is only over a relatively narrow range,about 3-1 on a diameter

basis,a wide rangedistribution an be built up if necessaryy the pro-

duction of several such plots.

A further development s to use the informationobtainedby the above

system,and by coupling he numbersand particle volume,automatically

and instantly compile he cumulative volume or weight distribution. This

is achieved with the Model M.

The Coulter CounterModel C is the most elaboratemodelyet com-

mercially available. By utilizing advancedcircuitry it is possible o

obtain complete a six, nine or twelve point size distribution in as little

as 15 sec. Recording f the data can be done n many ways, e.g. simple

digital print-out and direct connection o a computor. The model has

many other advantagesover the original Model A.

Correlationof the CoulterCounter o other methodsof particle size

analysis as beenwell studied. This presents o problemsf one appre-

ciates he limitationsof the various echniques. For instance ravitational



208 JOURNALOF THE SOCIETYOF COSMETIC HEMISTS

sedimentation annot generallybe relied upon below some4-5[• when such

effects as Browntan movement of the particles by convectioncurrents

(even in thermostatically controlled baths) and the actual method of

following he analysisall help to reduce he limit one can reach.

We hope that this paper has made clear how this instrument can be

used for the size analysis of the majority of cosmeticmaterials from

liquids, creamsand emulsionso pastesand powders.

(Received: st September 965)

REFERENCE

(1) Coulter,W. H. Paper presentedbefore he National ElectronicsConference, hicago, ll.,

U.S.A. (3rd October 1956).

DISCUSSION

MR. A. MoSs: When one has hydrophobic particles suspended n an electrolyte the

quality of the dispersion s not good becauseof flocculation of the particles. The

addition of a nonionic wetting agent such as polysorbate 80 decreases he surface

tension and facilitates the dispersion. Do you think that the use of such a wetting

agent modifies he di-electric constant of the electrically conductive iquid ?

MR. 1•. W. LINES: The addition of a wetting agent in no way affects the response

of the Coulter Counter to particles. Since most materials, whatever their nature,

need to be analysed n a fully dispersedstate it is necessary o add wetting agent to

get the ultimate particle size. The only effect that suchaddition might have is that a

very high concentration might increase the aperture resistance and therefore the

calibration constant would change for Model 'A'. This can be overcome by simply

finding the calibration constant with your system, including wetting agent.

A problem might arise with a wetting agent of the polysorbate ype. Some of

these surfactants are not completely water-soluble, and are in a very, very fine

colloidal state; in that event a very, very high background of particles would be

obtained so that the sensitivity of the size distribution is lost.

MR. A. Mo/is: May one usea solutionof ionic dispersantdirectly as an electrically

conductive iquid in which to suspend he particles to be analysed?

MR. R. W. LIN•S: One may use such surfactants directly, but in practice their

conductivity in an aqueousmedium would be so low in terms of the conductivity

needed for the Coulter Counter Model 'A' that the size distribution might suffer at the

fine end. It might well be that if a low concentrationof sucha surfactantwereused,

one could get to perhaps 5[z, possibly 2-31z,but one certainly could not get to the

very small sizes that probably interest you.

MR. A. Mogs: Since particle concentration s a critical point when the Counter s

used, how is it possibleto dilute liquid o/w emulsionswithout giving rise to the

coalescenceof dispersed droplets?

MR. R. W. LINES: One has to dilute carefully, and a slightly different technique

must be employed,dependingon the system. Probably the most satisfactoryway

is to make a two stage dilution, one stage nto the de-ionizedwater, and the second




stage into the salt solution, to which the wetting agent has been added previously.

This technique has been described by Marshall and Taylor (2). Their emulsion

systemwas 50 • liquid paraffin in water, with 2 •o w/v methyl cellulose 0 as emulgent.

They diluted 1 ml emulsion taken from a mixed suspensionwith filtered de-ionized

water to 100 mi. They then took 0.4 ml of this and diluted it to 100 ml with filtered

0.9• NaC1. They commented that the results were strongly comparable to those

from the microscope,within the limits obtainable from both techniques. To quote

the dilution techniques used induced no detectable change in the particle size

distribution. A similar conclusion as recently been publishedby Rowe (13).

MR. A. Mogs: If w/o emulsionshave to be tested, what kind of liquid must one use

to dilute them ?

MR. R. W. LINES: According to our knowledgew/o emulsionshave not yet been

analysed satisfactorily with a Coulter Counter. As one must mix the sample into a

liquid which conductselectricity, this liquid must be a solvent to take the oil phase.

At the same time, its dielectric constant must be reasonably high to get some salt

into it for use with the Counter, and a high dielectric constant means that it is also

going to take up water. There are two possibilities hat could be tried. One must

try to coat the water droplets with something hat will not go into the solvent electro-

lyte. Alternatively, one might freeze the systemand measure he size distribution

of the ice crystals which will be the same particle size to the Coulter Counter.

MR. A. Mogs: What is the best way to use the threshold circuit when an unknown

sample has to be tested ?

MR. R. W. LINES: Fig. I indicates that the size distribution on the Coulter

Counter appears on a screenas a seriesof vertical pulses, the height of each pulse

being the size of the particle going through the orifice at that instant. Apart from

one or two rather minor functions it is the purpose of this screen o indicate the

approximate size distribution. In this way one can adjust the sizing controls, the

threshold dial and the aperture current switch, in a way that enablesone to build up a

size distribution over the required range.

MR. T. A. BROCK: n Table II, figures are quoted for particle size measurements

on colognes nd a perfume, and presumably the products were mixed with an electro-

lyte before measurement. If so which one, and was the resultant mixture clear or

cloudy due to the perfume oils being thrown out of solution?

MR. W. M. Woo•): The electrolyte used was 1 •o NaC1 and we were very careful to

ensure hat the solution was clear; the presenceof a cloudy solution would indicate

that somethingpeculiar had happenedwith the electrolyte system, e.g. precipitation.

To the best of our knowledge we were counting only the particles present in the

cologne.

MR. R. W. LINES: One of the chief advantagesof the Coulter Counter is that if

one is in any doubt as to whether the distribution that is being counted is the

(2) Marshall, K. and Taylor, J. CoulterCounterusers'meeting,Nottingham, 30/9/1965.

(3) Rowe,E. L., J. Pharrn.Sci. õ4 260 (1965).



210


proper one, or whether t is changing,one can always go back over pointspreviously

counted. One can seeby a change n countsat varioussize evels,whetheranything

such as aggregation, flocculation, etc., is occurring.

MR. T. A. BROCK:n Fig. 5 a distribution s shown or particlespresent n a lip-

stick. What do you consider hat you were measuring, as presumably the system

presented o the Counter was a suspension f the pigmentspresent n an artificial

emulsionof the base in the electrolyte?

MR. W. M. WOOD:As the electrolytesystemwas lithium chloride n methanolwe

have assumed hat this would probably dissolve he lipstick base,and that we were

actually counting the pigment dispersion.

DR. M. I•. W. BROWN: S it not true that a particle passes etween he electrodes,

displacesan equal volume of the electrolyte solution, which causesa change n the

electricalpropertiesof the system, n this case he resistance, nd this is recorded

MR. W. M. WOOD: That is so.

DR. M. R. W. BROWN:Every deductionwhich you make from your measurements

thereforedepends pon he premise hat a particle s displacing n equalvolumeof an

electrolyte, and that an inert particle is displacingan equal volume of an electrolyte

solution. Do you have any experience f the electrolytesolutionbeing displaced

by somethingwhich s not inert, suchas a cell coatedwith a chemicalwhich s intended

to kill it, and give it a charge?

MR. W. M. WOOD:The response f the CoulterCounter s in no way affectedby the

particlesgoing through the orifice. It doesnot matter whether it is a biologicalcell

or a ceramic particle. The chargeson the particle itself, or the resistanceof the

particle, in no way affects he response f the instrument becauseone is measuringa

volume displacement. The U.S. National Bureau of Standards claim that once

particlesgo into the suspensionhey form a very thin electronshieldor oxide film

aroundhem,which sperhaps nly1-22khick,and his endersheparticle elatively

inert within a chosenelectrolytesystem. Biologicalmaterial in no way affects he

count, and it is possible o size blood cells,bacteria, the larger viruses,etc., without

getting a wrong answerdue to this chargeon the particle itself.

MR. G. PROUT: In our laboratories, the Coulter Counter has been used to count

tissue culture cells and bacteria.

The bacteria usedwere a Leuconostocpecies pproximately0.8p by 0.Sp,often in

chainsup to 4 or 8 cells ong. As no satisfactorymethod could be found to separate

these cells into single units the suspensions ere counted without any attempt to

achieveseparation xceptstirring, using he stirrer attached o the CoulterCounter.

The electrolyteusedwas 0.9• sodiumchloridedissolvedn nt•trient broth and this

medium was used to culture the organism. Treating the bacterial population with

1.0 o phenolor tannic acid solution n order o affect surface otential, did not affect

the reproducibility of the count.

The bacteria were also counted using serial dilution and plating out to give a

viable count, and by hemocytometer fter dilution and subsequent taining by

Gram'smethod,givinga total count of viable and dead cells.

Invariably the viable countwas slightly ower (about 10 o) than the hemocyto-




meter count. The hemocytometer count varied between 97•o and 101.8•o of the

Coulter Count.

Sample Results

Tube 50[•

Aperture resistance 21.1 K•

Electrolyte 0.9 •o NaC1 n nutrient broth

Coincidence factor 3.125

Manometer volume 0.5 ml

Gain index 3

Calibration factor 2.42

Count per ml 60.4 x 106organismsper ml

Hemocytometer count

(i) 61 X 106 organismsper ml

(it) 59.7 X 106 ......

(iii) 61.5 X 106 ......

Viable count 56.4 X 106 ...... 93.4

•o of Coulter Count

101.0

Many references are available, mainly from the U.S.A., of using the Coulter

Counter for biological particles, and in several of these correlation has been shown

between Coulter results and other counting techniques.

coulter counter p00197-p00211.pdf

Documents