the United States - National Museum - Smithsonian Institution

Proceedings ^of

the United States

National Museum ^{^^^^g^c^}

SMITHSONIAN INSTITUTION

^•

WASHINGTON, D.C.

THE FREEZE-DRY PRESERVATION OF BIOLOGICAL SPECIMENS

By ROLLAND

O.

HOWER

^{^}

Officeof Exhibits

Introduction

VisitorstodaytotheUnitedStatesNational

Museum

seespecimens of small animals that accurately represent the form and color they possessedinlife. Thisimprovedrepresentationisthedirect result of the

new method

ofspecimenpreservation that hasbeen brought about

by

the process of freeze-drying.

By

this process, the specimen retains most of its original characteristicswithout further need for preservation.

This

new

technique is based upon sublimation of frozen fluids.

One

of the earliest papersdetailing the

phenomenon

of sublimation was delivered to the Royal Society of

London

in 1813

by

William

Hyde

^{Wollaston, a} physicist. Wollaston's

commentary

(1813) on water passmg from the frozen to the gaseous state, apparently by- passing theliquid phase, discussed

what

was already a well-known phenomenon. Itwasnotuntilthe1890's,however, that theremoval

1Asfarasisknown, Mr.Howeristhefirsttoapplythe freeze-dry process to

museumwork. Developingmoresophisticatedapparatusandestablishingtime schedulesforthe freezingofspecimens,heisnowinvestigating freeze-dryproblems ofcolor retention,fat stabilization,andthe freezing of larger specimens.

1

2 PROCEEDINGS OF

THE NATIONAL MUSEUM

^{vol. u9}

of fluid from tissue

by

sublimation at low temperatures had been accomplished

by

Dr. Richard

Altman

ofLeipzig (Glick, 1949, p. 4).

Onlyinrecent years,with the advent of

modern

refrigeration and

vacuum

techniques and equipment thatis widely available,has the process

become

practical for commercial or professional purposes.

The

process offreeze-drying,

now

greatly refined, is used currently for food processing and for the preservation of pharmaceuticals,

human

bone, tissue,andblood plasma.

The

freeze-drying techniques utilized at theSmithsonian Institu- tion for the preservation of biological specimens began in the late 1950's (Hower, 1962). There is

now

a well-established freeze-dry program to preserve

many

types of natiu-al history specimens for exhibit. Ciu-rent exhibits include freeze-dried rodents, reptiles, crustaceans, insects, andfishes.

Reflectingoneventsofpast years thatled to

work

onthefreeze-dry process,Iwishtoexpress

my

gratitude toDr.H.T.

Meryman

ofthe Biophysics Divisionof the

Navy

Medical Research Institute,

who

pioneered

many

of the techniques and has

made

himself always available for consultation.

SpecialthanksalsoareextendedtoJamesC.Nyce,

who

contributed invaluable aid in the preparation of this paper, to Mrs. Constance Minkin,forher writingcontributions, and toJohn Hackmaster,

who

preparedthe illustrations.

Principlesof theFreeze-dryConcept

Freeze-drying consists bascially of dehydrating tissue while it is frozen. Whereas some tissue dried from a nonfrozen state becomes shrunken andconsequently^ distorted intheprocess, tissuedriedfrom a frozen state virtually retainsitsoriginalappearance;thus,theprocess

is not only valuable in preserving the quality of food but also in maintaining the appearance of museiun specimens.

An

additional advantage is that the dehydrated tissueisnot subjecttodecay.

Compared

to other preservation methods, the shrinkage and dis- tortionoffreeze-dried

museum

specimensisminimal,butthereis

some

distortion in certain types of animals, particularly some fishes. In mostcases,however, thedistortionisnegligibleor correctable. Fm-ry orfeatheryspecimensandthosewith exoskeletons appear completely natural

when

freeze-dried,provided theyhave beenproperly prepared.

Such preparation includes the positioning of the specimen before dehydration, the actual dehydration, installation of artificial eyes, and the painting of

mucous membranes

or exoskeletons (mucous

membranes

becomewhiteafterdehydrationandcrustaceanslose their coloration)

.

NO.3549 FREEZE-DRY PRESERVATION

— HOWER

3 Sublimation.

—

^Ifbiological tissueisfirstfrozen to giveitmechanical rigidityandwateristhenremoved bysublimation,mosttissuecan be dehydrated without apparent physical change.

Considerthisprocessintermsofplain frozenwater: ice.

Each

ice crystal is a geometrically sound structure,

made

up of myriads of water molecules thatareretainedintheirpositionsbythe gravitational field ofthesurroundingmolecules. Withintherestricting confines of this lattice, each molecule

moves

randomly, andthereisapossibility thatone of themotionsof a surfacemoleculewillbe greatenoughto propel it out of the confines of the lattice.

The

greater the crystal mass, thegreatertheprobabilityof suchescapes.

As

temperatureis increasedandmolecularmotion becomes accelerated, the probability ofescapebecomesgreater; thus,theaverage

number

ofwater molecules that will escape from a given mass at agiven temperature can be

statisticallypredicted. Ice within abiological specimen behaves in nearlyanidenticalmanner.

Sublimation beginsattheoutersurface ofaspecimenandcontinues at the boundary between the frozen and the dried tissue. This boundary recedes toward the center of the specimen as di'yingpro- ceeds.

As

water molecules continuetoescapefromicecrystalsonthe boundary, they

move

aboutathighvelocity, collidingconstantlywith othermoleculesandwiththe structm'eofthe dried tissuesurrounding them. (Astheyarebuffetedfromcollisiontocollision,theyare virtu- allyindependentofexternalforces.) Thereisaheavyconcentration of water-vapor molecules at the sublimation boundary, due to the gi'eat

number

ofmoleculesescaping;consequently,there aremorecolli- sions,whichricochetmolecules alongtheline ofleastresistancetoward theoutershellofthespecimen.

The

forceofthemolecules'collisions following their escape from the ice crystals on the sublimation boundarysupports their drivethrough the dried tissue of the speci-

men

and into the atmosphere beyond its outer shell. For ice to sublimeefficiently, the vaporpressure within the specimen chamber

must

be lower than the vapor pressure of ice within the specimen

itself.

Vapor

pressure.

—

^{If the} temperature of a

vacuum

chambercon- tainingiceis

mamtained

at

—10° C,

evacuated with a

vacuum pump,

and then valvedoffso thatno external

ah

canenter, moleciiles will begin to escape from the ice crystals within the chamber (fig. 1).

Some

ofthesemoleculeswillricochetabout,collidingwith one another and with the sides of thechamber, while others willrelocate them- selvesuponothericecrystals.

As

the concentrationof the vapor formed

by

these freemolecules reaches aspecificpoint(whichisdependent uponthetemperaturesof theiratmosphere), therate atwhich molecules returnto theicewill

4

PROCEEDINGS OF

THE

NATIONAL

MUSEUM

^vol.119

become

equal tothe rate atwhich theydepart, andthe watervapor

isthensaid to beinastate ofequilibrium (fig. 1).

The

vaporpres- sm'e at which this equilibrium occurs is referred to as equilibrium vapor pressure.

With

the temperature of the chamber maintained at

—

^10°

C,

^the pressure indicatedonits

vacuum

gaugewillbe 1.950

mm.

Hg. Thisis the equilibriumvapor pressure (consequently the vaporpressure) ofwaterat

—10°

C.

Water-vapormolecules

may

be

removed

fromthe areaimmediately smTounding the ice in a

vacuum

chamber, thereby upsetting the equilibrium and permitting

more

molecules to escape and allowing fewerto retm-n. Thisprincipleapplies to biologicalspecimensaswell as toice.

The

most effective

method

ofcontinuouslyremovingwater-vapor moleculesfromaspecimenchamberistocreate alowervaporpressure elsewhere. This can best be done

by

estabhshing a colder surface (condenser) nearby.

Water

vapor will diffuse to a colder surface andrecondensetoform

new

icecrystals.

Figure ^1.

—

Water-vapor equilib- rium(^=refrigeratedchamber;

5=

vacuum gauge;

C=

cut-off valve;Z)=vacuum pump).

A

refrigerated condenser serving as an efficient water

pump

is

employedforthispurpose.

As

previouslystated,

when

the tempera- ture of a specimen chamber is

—10° C,

the vapor pressure of ice within itis 1.950

mm.

Hg. If the temperature of the nearby con- densing surface is

—40° C,

^the vapor pressure is 96 micron Hg., creating a vapor-pressiu-e differential of 1.854

mm.

Hg.

Water

molecules will collect on the cold condenser surface.

With

^this

FREEZE-DRY PRESERVATION

— HOWER

much

vapor-pressiire difference, the only limitation to the effective- ness ofacondenseristhesurfacearea (fig.2).

Pressure

reduction.

— When

vaportransferoccursatatmospheric pressure,watermoleculesareimpeded

by

collisionswithairmolecules.

At

atmosphericpressure, the

mean

freepath (theaverage distance a vapormolecule can travelbeforecollidingwith another molecule) is

approximately .005 microns (5X10~^ mm.).

The mean

freepath of awater-vapor moleculewithrelationtopressureisasfollows

:

pressure meanfreepath

0.0034mm.

0.034mm.

0.34mm.

3.4mm.

34.0mm.

The

transfer ofwatermolecules from anicecrystal to acondenser can be accelerated

by

increasingthe lengthof their

mean

freepaths;

this can be done

by

reducing the pressure in the chamber with a 10

6

PROCEEDINGS OF

THE NATIONAL MUSEUM

^{vol. 119}

Instrumentationof theFreeze-drySystem

The

basic equipmentnecessaryto establish a freeze-dryapparatus consists ofaspecimen chamber, acondenserchamber, anda

vacuum pump. The

Smithsonian system also incorporatesrefrigeration con- trolcomponents thatarevery desirable formaintaining the depend- abihtyofthe system.

Figure ^3.

—

Continuous copper tubing(%^in.)connectedatalternateendsbyreturnbends withintliespecimen chamber.

Specimen

chamber.

— ^The

^basic requirements for the specimen chamberare that ₍₁₎ it

must

be

vacuum

tight; (2) it

must

berefrig- erated; and ₍₃₎ it

must

have a large accessopening.

The

ultimate

vacuum

within the chamber can be as low as 10 microns (1X10"^

mm.

Hg.) of pressure, producing a differential between internal and externalpressure of approximately 15 pounds

NO. ³⁵⁴⁹ FREEZE-DRY PRESERVATION

— ^HOWER

7 per squareinch. Sturdy construction is required if achamber^isto withstand such^force.

At

the Smithsonian, thestructm-alrequirementsfor the chambers were

met by

the selection of a 60-gallon paint-spray pressure tank for the specimen chamber and a 30-gallon tank for the condenser.

These tanks required modification but proved bothsatisfactory and economical.

The

Smithsonian specimen chamber was

mounted

horizontally on a base shaped to the contour of its walls, with allowance for the thickness ofaninsulatedouterplasticshell.

The

doorontheSmith- sonian chamber is two feet in diameter and withstands a total ex- ternal pressure of 6780 pounds.

The

chamber ^itself, two feet in diameter and 36 inches long, withstands a total external pressure ofalittle

more

than 20 tons.

Figure^4.

—

Detachablerefrigerationassembly:A,outside view; B, inside view.

The

vaporline, a piece of steel

tubmg

with an inside diameterof 3^2inches, was weldedinto a3%-inchhole

midway down

the side of thechamber.

The

inside ofthechamber waslinedwith^s-inchcontinuouscopper refrigeration tubing, filling one end and running the length of the walls, whereitwasjoinedatalternateends with2-inchretiu-nbends.

The

tubing was soldered to the steel wall of the chamber, anda

filletof metaUic plastic

compound

was added to improve heat con- ductionbetweenwallandtubing (fig.3).

8

PROCEEDINGS OF

THE

NATIONAL

MUSEUM

^vol.119

Due

to the difference in the coefficients of expansion of the two materialsatoperating temperatures,however, fracturing ofthesteel- to-copper attachments developed.

To

overcome this difficulty, therefrigeration fine was soldered to the outer siu-face of a copper sleeve, negating theneed forattachment between the copper tubing andthesteelwall (fig.4). This innovation not only solved thefrac- tm-ing problem but also established the refrigeration line as a de- tachable unit within the system, allowing easy removal from the chamberformaintenanceandrepair.

Since the refrigeration line entered and left the chamber through two small openingsin the rear ofthe tank, the openingsaround the tubing were sealed with soft solder.

When

installation of the re- frigeration line was completed, the chamber was tested thoroughly for

vacuum

leaks.

Refrigeration of the specimen chamber.

— ^Withm

^anevacuated

system, heat is transferred principally

by

radiation. Radiation travels inaIme-of-sight path between a specimen and surroimding surfaces. There is no heat produced within the specimen chamber;

therefore, the only appreciable load on the refrigeration system is

ambientheat thatleaks inthrough theinsulation. This load canbe calculatedin

BTU's

perhourfromtheinformation given_{(p. 17).}

Any

refrigeration compressor (q.v.) serving thespecimen chamber shoiddbecapableofproducingcontrolledtemperaturesatthechamber walls of

—15°

_to

—30°

_C.(afterheat leak hasbeentakenintoaccount).

When

selectingarefrigeration compressor, it is desirable todouble the calcidated heat load (primarily heat leak) to allow a sufficient marginforerror (table 1).

Table1.

—

Refrigerationcompressorcapacity inBTU's perhour (basedona90°

F. ambient temperature; r-12 (freon-12)=Dichlorodifluoromethane; F-22 (freon-22)=Chlorotrifluoromethane)

Horsepower

FREEZE-DRY PRESERVATION

— ^HOWER

9

At

theSmithsonian,alengthofcoppertubing,3^inchesindiameter (laterreplaced with steel), wasbrazedinto an opening, 3%inchesin diameter, nearthetopofthe30-gallon tank.

The

oppositeendofthe tubingwas connected to the specimen-chamber vapor line.

At

the baseofthetank was anopeningintowhich al}^-inch steel pipe

was

welded, connecting ittothe

vacuum pump.

Insidethe condensingchamber,a largecomplexofcoUs withincoils

was fashionedfrom 200 feet of copper refrigeration line % inchesin diameter.

The

expansion valvewas

mounted

insidethechamber and the refrigeration linewas broughtintoandoutofthechamberthrough openingsin the side ofthechamber, which weresealedwith softsol- derasin the specimen chamber (see fig. 4).

Refrigeration of

the

condenser.

— The

force underlying the motionofwater vapor fromtheice,through thespecimenchamber^to the refrigerated condenser, is the vapor-pressure difference. This force is produced

by

the difference in temperature between the ice crystalswithinthespecimenandthe refrigerated condenser. Table2 Table 2.

—

Relationship between temperature ^(°C) andvapor pressure (mm. Hg.)

°c

10 PROCEEDINGS OF

THE

NATIONAL

MUSEUM

It is clear, therefore, that any effort to produce extremely low condenser temperature is not economical since the cost per

BTU

ofrefrigeration increases rapidly as temperature is reduced, but the vapor-pressure differential increases only moderately.

A

tempera- ture of

—40°

C.issatisfactoryandiswithin therangeofconventional freon-12 orfreon-22 refrigeration units.

Another factor to be considered in designing a condenser is that the rate of waterlossfrom the specimen wUl bevery low^and, thus, theheat loadon thecondenser'srefrigeration unitwillbeslight.

The

heat leak through the insulation of the condenser chamber shouldbecalculated exactly, as for thespecimen chamber.

TemperatureControl: Temperature

may

be controlled

by

asimple mercurycontrol switch with a2- or 3-degreelow-temperaturediffer- ential.

The

control employedin theSmithsonian system is a sealed mercury switch, triggered

by

a vapor-pressure-actuated Bourbon tube. Exterior adjustments with visible settings over a calibrated dialareemployed.

Figure^5.

—

SchematicofSmithsonianfreeze-dryapparatus (^=specimen chamber;

B=

condensingchamber;

C=

recompression valve;Z)= vacuum pump; E, /"=refrigeration compressors;

G=

vacuumgauge).

Vacuum

pump.

— The vacuum pump,

veryimportantto the freeze- dry system, should be selected with care. There are

many

suitable

pumps

available.

When

a refrigerated condenser is employed as a vapor trap, the only functionof the

vacuum pump

istoreduceairpressuremechani- cally within the chamber and evacuate the noncondensable gases releasedfromthe specimens during thedryingprocess.

The

capacity of a

pump

is described in terms ofdisplacement.

Displacement, expressedin literspersecondorcubicfeetper minute,

is ameasure ofapump's capability toproduce a

vacuum

in agiven time, as well as tokeep the system evacuated as gases are released

FREEZE-DRY PRESERVATION

HOWER

₁₁

from volatile materials within the system.

Pump

displacement or

pumping

speed

must

be calculated according to the volume of the system andpressuredesired.

The

time required to

pump

thesystem

down

totheultimatepressure(about10to 100microns Hg.) should be established.

The pump-down

chart (fig. 6)

may

be used to calculate

Figure ^6.

—

Pump-downfactor chart.

displacement requirements.

By way

of illustration, the following two examples are cited:

(1)

To

determinetherequired

pumping

displacementtoevacuate a system with a totalvolume of 15 cubic feet to a pressure of 100 micronsin 5minutes, reference to the

pump-down

chart shows that at 100 microns the factor is 10.9. Multiply 10.9 by the volume (15 cu. ft.) to obtain the total

number

of cubicfeet to be

pumped

(163.5).

To

determine therequned

pump

displacement, divide this total

by

the

number

ofminutes(5) allowedfor theevacuation.

10.9X15:

163.5

163.5

32.7cu.ft./mm. or 55,563.6liters/sec.

12 PROCEEDINGS OF

THE

NATIONAL

MUSEUM

^vol.119

(2)

To

determine

how

long itwill take a

pump

with a displace-

ment

of27 cubicfeetperminutetoevacuate asystemwith avolume of 160 cubicfeet to a pressure of200 microns (at200 microns the factoris9.50), calculate as follows:

9.50X160=152

^ ⁼

^56.29min.

Under

certain conditions the refrigerated condenser can be eliminated and water vapor can be removed directly through the

vacuum pump,

provided the

pump

iscapable ofremovingthevapor at a rate equal to its release from the specimen.

To

determine whether the condenser can be omitted, the approximatequantity of water vaporbeingreleasedeach minute

by

the specimens within the chamber

must

becomputed.

If, forexample, a specimen

chamber

contains three squirrels, two

flickers, atoad, and agartersnake (at various drying stages and all ofaverage weight), the averagedaily release ofwater vaporwall be approximately 12grams.

If 150 micronsof pressiu-eweremaintained within thechamber, it

would benecessaryto

pump

200 cubic feet of Avater vapor foreach

gTam

^ofwaterremoved from the system, or2400 cubicfeet perday (1.66 cu.ft./min.). Ifthe

vacuum pump

cannot handlethis volume ofvapor, the pressure within the specimenchamber willslowly rise;

and, asitrises,thevolumeoccupied

by

a unitweightofwater vapor decreases. At a pressure of 300 microns, two grams of water will

occupy the same volume as would one

gram

at half that pressure.

Itisobvious, therefore, that overloading achamberorproviding an inadqeuate

vacuum pump

willcauseanincrease in operating pressure in the specimen chamber.

To

^{avoid this, a} refrigerated condenser should beused.

Since thedifferentialbetweenthewater-vapor pressure within the chamber andthevaporpressureoftheicewithin thespecimen provides the drivingforce for the

movement

ofwatervaporthrough the system, an increase

m

^{pressure wdthin the} chamber causes a decrease in efficiency.

An

increase in chamberpressure from 150 to 300 microns with a specimen temperature of

—10°

_C. would result in a vapor pressure differential offrom 1.80 to 1.65

mm.

Hg., or a reduction ofa little

more

than 8 percent.

While thisexample demonstrates thatincreasingwater-vapor load

isnotas deleterious asexpected,itnevertheless establishesa relation- ship between water-vapor load and

vacumn-pump

capacity and

PROC. U.S.NAT. MUS..VOL. 119 HOWER

—

PLATE

Rctiii^ciailngapparatusfor thespecimen andcondenser chambers (Smithsonian photo).

PROC. U.S.NAT. MUS.,VOL. 119

HOWER—

PLATE2

PROC. U.S.NAT. MUS..VOL. 119

HOWER—

PLATE3

PROC. U.S.NAT. MUS.. VOL.119 HOWER

—

PLATE4

FREEZE-DRY PRESERVATION

— ^HOWER

13 indicates that a

vacuum-pump

limitation can limit the

number

of specimens that can be producedina reasonable span oftime.

A

2-stage gas-ballast

vacuum pump

^is

recommended

for all freeze- dry^use.

Gas

ballast onthe

pump

enablesit toremovecondensable vapors suchaswatervaporfroma

vacuum

chamberwith a

minimum

of internalcondensation.

In the gasballast,avalve provides a

means

forcontrollingtheentry ofairintotheexhauststage of thepump's operation.

The

ballast of airacts asa transporting

medium

carryingdilutedvapors through the exhaust partto theatmosphere.

Such

pumps

are generallyratedwith ultimatepressurerangesfrom

1 to 10 microns; however,

pumping

speedusuallybegins todecrease atabout 100 microns,which isconsidered asatisfactorypressure for freeze-drying.

Oilshouldbe changedfrequentlyinany

pump

usedinafreeze-dry system.

Even

thoughthe

pump

isgas-ballasted,acertain

amount

of watervapormil contaminatetheoil and

damage

the

pump.

Table3.

—

^Vacuumtable

Terminology

14 PROCEEDINGS OF

THE

NATIONAL

MUSEUM

^vol. ng Table4.

—

Metric conversiontable

1in.

1ft.

1in.2 1ft.2 1in.3 lft.3

10mm.

010 cm.

100mm.2 100cm.2 010ml.

010cl.

010dl.

0101.

002.5400 cm.

NO.3549 FREEZE-DRY PRESERVATION

— HOWER

15 thewholesectionof thetube inunittime (seconds, asapplied here), the pressure differentialat each endofthe tube, the radius, and the lengthof the tube.

F= Volume

(cm.Vsec.)

_p= Pressure (dynes/cm.^) ^r*

r

=

Radius (in cm.)

~

8z,

Z=

Length (incm.)

The

formulastates thatthevolumeofvaportransferred througha tubeinaspecificperiodoftimeisproportionaltothefourthpowerof the tuberadius. This

means

thata smallincrease in tubediameter willproduceaconsiderableincrease invaporconductance.

To

utilize the formula, it is necessary to

Imow

the coefficient of viscosity,whichforairisapproximately1.7X10~^poises atapressure of 1 dyne/cm.2= 7.5X10-"

mm.

Hg. 1.

For example, to calculate the conductance of vapor through a tube thatis 193 cm. long, 7.62cm. in diameter, and^^dth a pressure differential between the specimen and the condensing surface of 2000 microns, the following values are:

F=cm.^/sec.

^=2666.7

dynes/cm.2 (2000)lipressuredifferential)

r=3.81cm.

i=193

cm.

z;=1.7X10"*poises

y^

^{(3.14)(2666.7)}(3.81)"

^

1737861.723 (8)(193)(.017) 26.248

F=

66209 cm.Vsec.

Testing for leaks.

— The

overall system

must

be tested for

vacuum

leaks. Leaksina

vacuum

system canbe detected easUy-with a Peroni pressure gauge used in conjunction with acetone.

The

vaporpressureof acetoneissohighthat,

when

jointsor surfaces are brushed with it, any leak will cause a

marked

increase of pressure within the system.

The

gauge head should be installed at a point near the

vacuum pump

andthenwatchedforpressure increases after the systemhasbeenevacuatedandtested.

Anothertestisto injectfreonunderslightpressureinto thesystem and locate the leaks with a refrigeration leak detector.

When

the system isinitially

pumped

down, the pressure in the system

may

be higherthan anticipateddue to gases being given off

by

zinc coating orother materialsofhighvaporpressure that

may

beinsidethesys- tem.

The

pressure can be effectively reduced

by

allowing air to reenterthe system and by

pumping

it

down

repeatedly until these vaporshave beenevacuated.

16 PROCEEDINGS OF

THE NATIONAL MUSEUM

^{vol. 119}

Repairs should be

made

whilethesystemisevacuated becausethe

vacuum

helps a sealant penetrate a leak. Unless the pressure is

unusually high

when

the

pump

isoperating, it

may

beassumed that a leak is small and can be sealed with glyptol varnish or lacquer applied

by

brush.

The

instant that either material reaches the

vacuum

through a pinholeleak, it dries and plugs the leak. Ifthe pressure remainshigh, the leak islarge and should be plugged with soft solder.

Refrigeration compressors.

— The

major consideration in refrig- erationis that constant temperature be maintained.

The

hermeticallysealed refrigeration unit is composed of a motor andcompressorshaft of 1-piece construction.

The

motor(cooledby the flow of refrigerant gas) and compressor assembly _are within a gas-tight

housmg

thatis weldedshut. This

method

of construction eliminates the need for certain parts (pulley, belt, compressor fly- wheel, and compressor seal) found in an open unit and, of course, avoids the servicingandreplacing ofthoseparts.

One

objection tothistypeofunitisthat,underfreeze-dryoperating conditions, thereis

some

danger ofthe motor's burning out because ofthevery small

amount

ofrefrigerantbeing cu"culated.

The

Smithsonian's compressors, however, have not overheated despite continuous operation for several years in a

room

with an average temperatureof about 90°F. If overheating shouldoccur,it

could be remedied by installing a water-cooled condenser in the dischargeside ofthe compressor.

A

second t3rpe, the air-cooled compressor, usually operates witha belt andpulley; the motor is in the open, where itis cooled

by

air circulatingaroundit.

Operatingcharacteristics of both types of refrigeration compressors are,otherwise, essentially thesame.

The

selection of a fractional-horsepower refrigeration compressor should be based upon calculation of heat loadon thermal msulation

(seep. 17and alsotable1).

Expansion valves.—

For

maxmium

vaporizationofthe refrigerant, it is important to select thermostatic expansion valves of correct capacity. Itisequallyimportantthat thevalves beinstalled at the properlocations,smce bothfactorscaninfluence the success or failure of theentiresystem.

The

thermostatic expansion valve selected for the condensing chamber should be a type designed to control temperatures below

-40°

C.

Although thethermostatic expansion valvesin eitherchamber

may

be

mounted

in any position, they should be installed as near the

NO. 3549 FREEZE-DRY PRESERVATION

—

SHOWER 17 evaporator inletas possible.

The

valvesin the Smithsonian system are

mounted

insidethechambers,where

maximum

efficiencyisgained.

Bulb Location: For satisfactory expansion-valve control, good thermal contactbetween the bulb and the suction line is essential.

The

bidb, which controls the expansion valve, should be fastened securely

mth

two metal straps to a clean section of the suctionline inside thechamber.

The

bulb should be located nearthemidpointof thelinearoundthe coU. Itshould not be nearthebottomoftheline because a refrigerant-and-oilmixture isusually present there, which woiddresidt inmcorrectcontrolof theexpansionvalve.

FilterandDrier:These should beinstalled in the liquidlineahead

of thethermostatic expansionvalve.

SightGlass:Further protectioniseasilyandinexpensivelyprovided with a sight glass throughwhich the refrigerant level can be deter- mined

by

thepresenceor absenceofbubblesinthe liquidline.

Bub-

blesindicate thatthe refrigerant levelislow.

Thermal

insulation.

—

^{Insulation material is required} to sub- stantiallyreduce heatleak, whichloads the refrigeration unit. This material

may

beglasswool, wdtha

K

factorof 0.29, rock wool, wdth a

K

factor of 0.26,compressedcork,with a

K

factor of 0.30, orother similarmaterial.

The

Smithsonian freeze-dryapparatusisinsulated with a foam-in-place plastic (polyurithane). Foam-in-place plastics rangeindensityfrom2 to 10poundsper cubicfoot; they offer

K

fac- tors offrom 0.02 to 0.24. These plastics canbeused\\dthout great difficulty andcan be

made

toconformto anycontour. Suchproper- ties

make

plasticfoam anidealmaterial touseas insulation.

Calculation for the determination of thermal impedence andheat loadisasfollows:

K=

insulation factor (BTU/hr./

A = Ambient

temperature (dif-

sq.ft.) ference between the inside

S=

surface area of insulation and outside of chambers in

(sq.ft.) ''F)

IT=

thickness of insulation (in.)

<M2A _{= BTU/perhr.}

Temperature

reading

and

recording.

—

Temperatures through- out the drying process should be carefully watched and accurately recorded.

The

record is indispensable for establishing drying times andfordeterminingideal temperaturesforvarious typesofspecimen material.

An

automatic temperature-recording device is valuable for indicating the temperature variations as the refrigeration com- pressors are cycling, andit

may

also caU attention toproblems that would otherwise go unnoticed, such as a faulty temperature-control switch.

18 PROCEEDINGS OF

THE NATIONAL MUSEUM

The

system at the Smithsonian incorporates a telethermometer that utilizes a thermistor probe as a sensmg element and provides temperature readings directly from a temperature scale.

The

tele- thermometer's range is from

—40°

_to 150° C.

Both

surface and

air probes are used and the probe leads are connected through a multiple-selector switch to permit convenient temperature reading

atseveral places.

Figure ^7.

—

^{Simple freeze-dry apparatus (.:/=specimen chamber;}

5=

commercial deep- freeze;

C=

recompressionvalve;Z)= vacuum pump; £==wood-spacerinsert).

Recorder. — Any

versatile general-purpose D.C. voltage and cur- rent recorder

may

be used to record temperatures continuously.

One

with a low-to-high-level range (200microvolts to 500 volts and 200 millimicroamperes to 100 microamperes) is suitable for freeze- dry purposes. Itis important that the recorderbe compatiblewith the temperature devicein scalerange andvoltage output.

Constructionof aFreeze-drySystem

The

simplest apparatus suitable for freeze-drying consists of a commercial deep-freeze containing a small chamber connected to a gas-ballasted

vacuum pump

^(fig. 7).

NO. 3549 FREEZE-DRY PRESERVATION

— ^HOWER

19

The main

requirementforthedeep-freeze chestis thatitmaintain a constanttemperatureof

—

^15°to

—20°

C.

The

onlyalterationneeded

isremovalofitshingesandlatchandtheinsertion ofa

wood

spacer betweenthe chestanditslidtopermit passageofavaporlinebetween them.

The

vaporline

must

pass through the

wood

spacer.

The

hinges and latchshould be relocated for the raised top; the thickness of the

wood

spacer is determined

by

the diameter of the vaporlinethatwillpass throughit.

A

smallspecimenchamber

may

be

made

froma5-gaUon paint-spray pressure tank roughly18 inchesin diameter and 21 incheshigh.

A

hole

must

becutintheside ofthechamberfora

vacuum

line,which can bebrazedinto place.

The

majorconsiderationinselectinga gas-ballasted

vacuum pump

is the volume occupied

by

the water vapor at thetemperature and pressure being used. Assuming that the chamber temperture is

—20° C,

thevaporpressm'eis0.8

mm.

Hg.

At

avaporpressureof0.8

mm.,

one

gram

of water occupies a volume of 1200 litters. If the pressure gradientisapproximately0.6

mm., we

findone

gram

ofwater willoccupyapproximately 1400liters.

A

gas-ballasted

vacuum pump

with a25-liter-per-minutecapacity willtake approximately onehour to

pump

one

gram

ofwater.

A pump

withacapacityof79litersper minute wUl removeapproximatelyone

gram

ofwaterin18minutes.

If thevapor lineis 150cm. (5 ft.)long, the diameteriscalculated tobeapproximately2.5cm.or1 inch. Tubing_{IK inches}in diameter willallowa sufficientoperatingmargin

when

usingthe largercapacity

pump.

The

chamber

must

notbeloadedbeyond

pumping

capacity. In a systemofthis size,it issuggested that asinglespecimen beplacedin thechamber andasecondoneadded twoor threedayslater,followed

by

a thu'dspecimen onthecompletionofthefirst, andso on, in this pattern. Observations of drying times during experimental drying cycles will determine the capacity of each individual system.

The

dryingcyclecan beinterrupted,providedspecimensarenot permitted tothaw.

The

apparatus describedaboveislimited

by

twobasic characteris- tics of its construction: the vapor-handling capacity and the size of the specimen chamber (fig. 8).

As

stated, the mostefficient device forremovingwatervaporfromthespecimenchamberisa refrigerated condenser. If an external refrigerated condenserof sufficient size is used, the onlylimitation to the size of the specimen chamber^is the sizeofthedeep-freezeunit that holdsthespecimen chamber.

A

refrigerated condenser can be constructed within a paint-spray pressuretanksimilar tothe oneusedforthespecimenchamber.

The

condensingsurface

may

beacoilof^s-inchcopperrefrigerationtubing

20

PROCEEDINGS OF

THE NATIONAL MUSEUM

withanexpansion valve

mounted

atthe topor inletside ofthecoil.

The

coilshouldhaveas

many

turns asspace permits.

A

coU-within- a-coilarrangementismostdesirable (fig. 9).

Figure^8.

—

^{Simple freeze-dry apparatus utilizing a} condenser surface (.^= specimen chamber;

5=

commercial deepfreeze;

C^

condenserchamber;Z)= vacuumpump).

A

vacuum-tight drain should be installed at the bottom of the condensing chamber to conveniently remove the condensation from thechamber during defrosting.

The

condensingchamber shouldbe encased in an insulated box with openings in appropriate positions forvaporlinesandrefrigeration tubing.

An

importantrequirementin anyfreeze-drysystemisa valve (see fig. 5) thatwill permit atmosphericair to enter the systemin order

Figure ^9.

—

Condensingcoil.

that thechamber canbe opened. This valve shouldbesolocated as toinsurethat the condensingchamberissituatedbetween thevalve and the specimen chamber, thereby assuringcondensation of atmos- pheric moisture on the condensingcoilrather than on the specimen.

The

valve

must

be

vacuum

tight

when

closedand neednot belarge;

a ){-inch intake is sufficient to shut

down

the system. Itis recom-

FREEZE-DRY PRESERVATION

— ^HOWER

21 Table ^6.

—

Freezing mixtures (PS

=

primary substance and its proportion;

SS

=

secondary substance anditsproportion; TS

=

tertiarysubstanceandits proportion;

A =

temperature (° C throughout) ofsubstance before mixture;

B

=

temperatureofmixture; C

=

total reduced temperature;

D =

temperature whenallsnowismelted(whensnowisused);

E =

heatabsorbed(calorieswhen

A

is grams); ^*

=

lowesttemperatureobtained. Table modifiedfrom "Smith- sonianPhysical Tables," 1959,9threv.ed.,byW.E.Forsythe.

PS

22

PROCEEDINGS OF

THE NATIONAL MUSEUM

^{vol. 119}

mended

thatthis valvebe openedwhile thepiunp isstillrunningto prevent atmospheric ah'fromforcingthe

pump

to run backward or forcing^oilfromthe

pump

into thecondensingchamber.

Figure 8 illustrates assembly of the modified apparatus using a refrigeratedcondenser.

PreparationofSpecimens

The

freeze-dry process will not improve a poor specimen. If a specimen is emaciated, slightly deteriorated, or otherwise inferior

when

itenters thechamber, itwillbeinno bettercondition

when

it leaves. Sagged tissue, however, can often bereshaped ^^dth subcu- taneousinjections ofwater, andsagged

body

cavitiescanberestored with cotton beforefreezing.

The

artand skillof the taxidermistwillhelp decide thesuccess or failure of the freeze-dry technique.

Many

of the same tools and methodsofconventionaltaxidermycanbe employed advantageously in the freeze-dry process. Wires, supports, and other tricks of the trade are useful. Also rapidfreezingwithliquidnitrogen orfreezing mixtureswillhold aspecimeninposition.

Initialfreezing of

the

specimen.

—

^{Of all}

^compounds

^{in animal}

tissue, water is the most abundant. For the average, it constitutes 70 percent of the animal's total weight.

Water

is found in cellular andvascularspaces, and, insmall quanties, inprotein andcarbohy- drates. Ofthe totalliquidcontent,20 percent is usuallyinextracel- lularfluid;approximately 25 percentof thisextracellularfluidisplasma, and the remainderis interstitial fluid, mostly water.

About

76per- centofmuscletissueiswater.

The

Ro^\^ltreedata(seeHarrow, 1951) relating to the biochemistry of

man

is a good general guide to the distribution of tissuewaterin

most mammals.

Water

in tissueisfoundalmost alwaysincombination^^^thnaturaUy occurring salts;for this reason, freezing shouldbe brought about in the shortest possible time. Slow freezing invariably leads to the formationof eutecticsthrough the concentrationofsalts.

As

freezing waterseparatesitselffroma solution

by

theprocessofion diffusion, the salts

become more

highly concentrated than theywereinthe original solution.

Eutecticshavelowerfreezingpomts andlowervaporpressuresthan water; thus, theirformation during the process slows do^\^l drying.

Due

to unfrozen saturated fluid in the tissue and resulting surface tension, shrinkage occurs. Rapid freezing, which reduces the formation of eutectics, can be accomplished

by

using the freezing mixtureslistedin table 6, liquidnitrogen, or afreezing chestwith a

NO.3549 FREEZE-DRY PRESERVATION

—

SHOWER

23

temperatureof lessthan

—25°

_{C. Rapid}freezing also createssmaller icecrystals, causinglesstissue distortionduringthe process.

Techniques

for faster drying.

— The

greatest

amount

ofwater

is removedearlyin the dryingcycle

when

a specimen'sdriedshellis thinnestandoffers theleastimpedencetothe escape ofwater vapor.

The

rate of weightloss dueto the escape ofvapor approacheszero with the completionofdrying. Ifaspecimenchamberisloaded with a great

number

of fresh specimens, the water-vapor atmosphere

will be very great during the first few days, possibly exceedmg the capacity of the system. If, however, specimens are introduced at spaced intervals, the same

number

of specimens can probably be processed without exceeding the system's vapor-removing capacity, and a constant level of water-vapor removal

may

be approached.

By

drillingholes inthebackorbottom portions ofspecimens,their drying surfaces areincreased and the area oftheir epidermal layers isdecreased, allowingwater vaporto escapemorerapidly. Skinning one side of a specimen serves a similar function in that itremoves epidermal tissue that would otherwise act as avapor barrier.

Evisceration of a specimen also reduces itswater content consid- erably;however, theremoval ofwater-laden organsrequiresincision,

removal oftheviscera, fillingthe

body

cavity

mth

cotton or similar material, all of which requires time and involves

some

of the less desirable features of taxidermy.

A

record should be maintained of each specimen's weight since observmg theweight reduction (by removingthe specimenfrom the chamber and weighingit) is thebest

method

fordeternunmg

when

a specimeniscompletelydry.

A

specimenshouldbeleftinthechamber for one or two days after the loss of weight ceases tobe apparent.

Thisis especiallytrue forrodents such as miceor rats with a scaly epidermis on their ^tails.

The

tails dry slowly and it is therefore advisable to perforate their undersides and keep the specimens in the chamber beyond their apparent drying times.

Conclusion

As

technology advances, it becomes apparent that freeze-dry will be an ever-growing field. Plans are already underway for increasing the Smithsonian facilities. Theoretically, there is no limitation to the ultimatesize or numbers of specimens that atone time could be processedin thismanner. Itishoped that thispaper willhelp to encourage thegrowth and exchange offurther ideas on the freeze-drying process.

24

PROCEEDINGS OF

THE

NATIONAL

MUSEUM

^vol.119

References Bkaddick, H. J. J.

1954. The physics ofexperimentalmethod.

New

York: John Wileyand Sons.

Daniels, F.; Mathews, J. H.; Williams,J. W.,etal

1949. Experimentalphysicalchemistry,4thed.

New

York: McGraw-Hill BookCo.

Glick, David.

1949. Techniquesofhisto-andcytochemistry.

Harrow, B.

1951. Textbook of biochemistry. Philadelphia and London: W. B.

Saunders Co.

HOWER, R. O.

1962. Freeze-drying biological specimens. Smithsonian Inst. Inf. Leaf!., no. 324.

1964. Freeze-dryingbiologicalspecimens. Mus. NewsTechn. SuppL, vol.

1,no.1.

Meryman, H. T.

1960. The preparation of biological museum specimens by freeze-drying.

Curator,vol.3,no. 1.

1961. Biological museum specimens. Naval Medical Research Institute, Bethesda, Md.

1961. Thepreparationof bilogical museumspecimensbyfreeze-drying,2:

Instrumentation. Curator,vol.4,no.2.

Strong, J.

1953. Proceduresinexperimentalphysics.

New

York: Prentice-Hall.

Wollaston, William Hyde

1813. On a method of freezingat a distance. Roy. Soc. Phil. Trans., pp. 71-74.

U.S.GOVERNMENTPRINTING