A PULSED NMR RELAXATION AND DIFFUSION

STUDY

OF WATER IN TREATED AND UNTREATED WATERLOGGED WOOD

By David A Bannister

Thesis for

submitted

to the University

of Nottingham

the Degree of Doctor January 1990

of Philosophy

Proton N.M.R. Relaxation and Self Diffusion in WaterloggedWoods. Contents. Abstract. Acknowledgemenis. Chapter 1. Chapter 2. 2.1 2.2 2.3 2.4 2.5 Chapter 3. NMR and the Study of Water in Waterloggedgood. Waterloggedgood and its Conservation. Introducion. The Structure of Wood. The Decompositionof Wood. Methodsof Conserving WaterloggedWood. Stabilisation of WaterloggedWood using Polyethylene Glycol. 6 7 10 12 14 Uv) (vi) I

The Theory of NMR Relaxation and Diffusion Measurements in Heterogeneous Systems.

3.1 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 3.3

Introduction. FundamentalConcepts. The Block Equations. The Rotating Frameof Reference. Spin Lattice (TI), and Spin-Spin (T2) Relaxation. Pulsed Nuclear Magnetic Resonance. The Interpretation Molecular Events. of Relaxation times (TI and T2) in, Termsof

21 22 24 26 28 29 35

3.3.1

Mechanisms of Relaxation and the Bloemberg, Foundand Purcell Theory.

35

3.4 3.4.1 3.4.2 3.4.3 3.4.4 3.5

Heterogeneous Systems. The Discrete Multiphase Model. A Distribution of Correlation Times.

40 41 45 49 50 52

Cross-Relaxation Effects. Anisotropic Motion. Diffusion.

(1)

5.4.2 5.4.3 5.5 5.5.1 5.5.2 5.6 5.7 5.7.1 5.7.2

Controller. TheTemperature TheBruker Minispeck. TheDiffusion Equipment. TheField GradientCoils. ThePulsedField GradientUnit. Data Analysis. Preparation. Sample PolyethyleneGlycol Solutions. Xylem.

112 113 113 113 116 117 120 120 120

Wood. Studyof WaterIn Waterlogged Chapter6. A PulsedNMR 6.1 6.2 6.3 6.3.1 6.3.2 6.3.3 Introduction. FreezingCurves. NMR Relaxationof Waterin Waterlogged Woods. Single Component Measurements of TI and T2. Multicoaponent Woods. Relaxationof Waterin Waterlogged NMR RelaxationTimesof Waterin Waterlogged Woods as a Function of MoistureContentandTemperature. 6.4 NMR Studies on Waterlogged Impregnated Wood with Polyethylene 172 Solutions. -Glycol 6.4.1 6.4.2 ThePolyethylene-glycol/Water System. Impregnated Studyof Waterlogged Wood A PulsedNMR with Polyethylene-glycolSolution. 6.5 A Studyof Self-diffusion for Waterin Untreatedand Treated Woods UsingNMR Field Gradient Techniques. Waterlogged 6.5.1 6.5.2 6.6 Wood. Self-diffusion In UntreatedWaterlogged Wood. Self-diffusion in TreatedWaterlogged Conclusion. 193 204 207 193 173 181 124 125 133 133 142 145

QiD

ABSTRACT

Freezing curve. NMR elax3tion data. 5nd Steady fieid

jr3dient and Puised fieid gradient

experiments were conoucted on samplesoi water-logged woocs excavated from the TUdor warship. impreqn; P-oiyethylene-Glycol solutions; samples tted the flar- Rose. and on 3; witl iaiLar polymer used as a buiking agent to prevent decay. 16 'he

Freezing found in At ic3s,. two distinquis, i-, wood. molecules are aole populations of water curves indicate the ?resence of approximately 0.38 g/g of hydration water. close to that 00servea in iresh timoers. Reiaxation measurements on pre-treatea samplesprovides evidence hydration 0.14 below of q/g contents of at water present Above this

of a very tightlY tound fraction.

-Aater. with a secondpopul3tion of hydration water being present upto MIS 9/9. *,., alue a thitd. 'free' "., ozuiation is observed.

Therelaxation deC3YS in longitudinal and transverse oitection havebeenanalysedin terms 01 3 sum of exponentials. Theseindicate the presenceof two populationsof water
inicn do not correspond to FaFulations observed in freezin; curve analysis.

in behaviour Exchange dominate the dependency the temperature mechanisms relaxation ot

pre-treited sampis. Aich is similar to that observed in other fibrous materials such as

to the PhYsicalcharacdo not appearto correspond meat. however.the difierent components teriitics or the wood.and the spin populations cannot be associatedwith a distribution be-

tweenioentifiable -compartments within the. system.

In FEG impregnated the contribution to the signal from the polymeris not resolvsamples
able on the equipment used. Samplestreated with FEGsfor which the degree of polymerisation is ireater than 1540showa dependencyof relaxation characteristics on the water content Of

the sample. At low vater contents PEGs low of molecularweight impart a mobility to the

Qv)

1 %

ABSTRACI

Freezing curve. NMR eiax3tion

data. and Steady field

jr3dient

and Pulsed field

gradient

trom the Tudor uarship, excavated experiments were conoucted cn samplesof vater-logged wooos
the Mary Rose. and on 3laiiar Polyethylene-Glycol itft impregn; 3amoies tted . soiutions: the

buikin? as a used agent to prevent dec3y. poiymer

At lc-3st two distinguis, 61aolepopulations of water molecules are found in wood. Freezinj curves inucate the presence of approximately 0.38 g/g of hvdration water. close to that ODReiaxation me3surements on "pre-treated samplesprovides evidence hydration 0.14 below 0i g/g of present at water contents Above this

rea in iresh timoers. ser,, oi a very tilhtlY

bound fraction.

'mater. uith 3 secondPopulation of hydration water beinj present upto 0.1.8 9/9. value a third,. 'free' PozulBtion is observed.

Therelaxation deC3YS in longitudinal and transverse citection havebeenanaiysedin terms ot a sumof exponentials. Theseindic3te the presenceoi two populationsof water
inicn do not Correspandto oplations observed in freezing curve anaiysis.

in behaviour Ex. dependency the the temperature dominate -han? e mechanisms relaxation ot
Pfe-tt,? ited samcles. which is similar to that observed in other fibrous materials such as

to the physical characdo not appearto correspond meat. however.the difterent components teristics Or the wood.and the spin populations cannot be associatedwith a distribution tetween icentifiable compartments within the system.

In FEG impregnated the contribution to the signal from the polymeris not resoivsamples treated with FEGs able on the equipment for which the degreeof polymerisatiOn used. Samples
is greater than 140 showa dependency of relaxation characteristics on the water content Of

the sample. At low water contents FEGs low of molecular weight impart a mobility to the

Qv)

"bound" water moleculeswhich is not seen in untreated samples.

3eIr dirfision .

coetticients ror water maieculesin woodare anisotropic, and are reauCed water. Ihis reau. -tion is 'brought aboutbecause ; ater

itom those observed. in distilled

mol-cuies are coth held in a hyd-ationla, ver, ana boundeaby the cellular structure Oi the wood.

In impregnatedsampiesthe diffusiOn rates are lowered by a factOr Oi 10- r hOlJzq, this is e not reiiected in the relaxation benaviour. 'the anisotropy is reduced,and proton excnan? mecranisms ate biocked.

(y)

5 E! IT ,%CKN0'jLEDGE?

I ouid' like to express my appreciation to the 'Ioliowi'g;

Frof. 'W. Derbyshire - for his supervision and the provision of laborato ry facilities. Dr. J. Harvey - ior initiating the project and indentitving sampie species.

Fror. -F. Cloqh - for his patience. ''' . Ir. S. Booth - for his incredible hard work in keeping the equicme, rt operational. Tle113ryRose Trust - 'for providing samples. C3Snand heiprui injor-i3tiOn. --The S.E.R.C. - for rinancing the case award. Bunen Phonon T ana 'or. F. Zelaya - for friendsnip. en,,, usi3sm and assistance. the R. Underwood'skitzhen table - for its support! Mum and Dad.

I NO

CHAPTER VATER-LOGGED VOOD. 1 NMR OFVATERIN '-", AND THE STUDY

-I

There have been numerous reports on the study of water in heterogeneoussystems using a variety of techniques 11-11. Despite this, an understanding of the nature of water within these systems, and Its Interaction with these systemms, has not been well established. system which has received little attentici in the literature lsthat`of One

water-! oggedwood.

Curre: rit procedures lsej In the p, -i3tion-'of -eservatiCn and conse. are, at best., pragmat!c, ard are certairly'rot, developed over the past thirty

wate: -logged artefaCts

'Meth, been have that ' well underst:,cd. cids

beIn manycases pr: ve, u-,sat! s1actory,,! arSelY years or so with the materials which they are

cause It is not knownhcw the preservatives'lnteract employedto preserve.

!n order t- opt.,mllse the processes used at present, It Is obviously

that important Uat the penetration prcpertles be determined it !s for thles, considered and,
someknowledge'of the'dynamics of both the systam and the praservatIve Is necessary.

Thedynamic led f: omthose 11, properties of water In biolcgical 'systemsare modif ulk .f

degrees, to depending the watercontent, t. varying upon water of the system and13 or envir n e 'M Lt that to explainthe propertiesof,,iater In thesesystems it hasnot yet been established !t to Invoke is necessary 3 biological effect and/orthat explanations 'cased uponthe systems do noCprove beingdispersed andheterogeneous sufficient. ' In general,the applicationof InvesUgative technijues In termsof to suchsystems Is-non-trivial, sometimes the standard
tecbrilques,but moreoften In terms cf interpretation of observation.

Some experkental, techniquescapableof detecting different types of physiof'the common in figure 1.1. 'The first class encompasses the thermodynamic techcal processes are shown, niques suchas sorption Isothermsand calorimetric a, nd volumetric analysis', useful in'the study of bulk phercmena suchas phasechanges, energyrelationships and-the determination of

to do asus thermodynamic fractions of unfreezablewater. However, not enable measurements kinetics, the Information of, do they on to offer sign water molecules specific sites, not molecules of binding processes. cr mechanisms

Lai Tint fs)

OASipinj Ot solias, Chtwal. reaction Diffasm in solid P419ykers chmi: &I reaction-' Rotation in Sollas
DiffiLsion - PoiUmos in 5DILAtkv

*vat] bEfraotum HIM scattering

"T,

Acoustic
loui

MI 1

ic -+- -4

Palmation Smf Lt
-

Iq. Pa Spec Covj

-L I

riuupt U fin scalesof "Chcular Trocesms anaapplication"wits d amerimntal tedaims.

nelltrOnMel

tr ing T sra

ic I

_r

Dif rnslan m lecutar TvIation Small,moleculQs

Ionavibration

the molecularstvuctural parameters The second and pertLrbagroupof techniquesexamines tions thereof, together iith the dynamic properties of wate- and zolecular substrates. The
dielectric includes group elastic scattering, relaxation, Infa-red reflectance, photo-acoustic effects, neutron

nuclear magnetic resonanceand electron spin reso. nance. These techniques and translational diffusion discan and of water molecules 2otion. Of this

to be monitor rotational used can

In the bulk Is between water and water which unde:joing restricted criminate

in the Resonance Magnetic (NMR) is the study Auclear useful and possibly most powerful group,

the their It Important has over advantages, and woods preservation, as several of water-logged other methods.

high NFIR selectivityoffers ,

to study,,, As a result, the NMR spectroscopistcan choose

The that they moment. proton, with a magnetic possess atomic, nuclei of-any elementprovided , to its high natural abundance a nuclear spinnumber 1: 1/2 Is the most-commonly exploited,, due. In Include Others 2.673 101 CKg)., and,large magnetogyric (gammacommon use x ratio ,, deuteriumoxygen-17, and carbon-13.,, -,",:, ,., I"II -

The magnetic field experienced by field the is (B) applied a nucle; s

(Bo),

pertufted by

I local magneticeffects. smal

! ated tc It,

honce field Some the 7 rethese al and induced exte",, of are in ; ts emer. , Eystem, Including all other

4hile others are due to aal, stlc

nun!o! ?osses3inj magnetic mozents. The local fields associated with these perturbations have both steady and fluctuating the of Cons width I 'he and compcnents. ",csition ejuent y, examination, under nuclei

spectral 'lines are extremely sensitive to the environments

Ic I. e. to the structure and dynamicsof the molsc,,, r. les contain-ng the nuclei I'l est ,q U

Vithin

Efferent

systems there 3re 3 variety on, diffusicn,

of inter-moleciflar

Interactlons,

such as

molecular reorientat! properties

and chemical exchange, which JIVe rise

to different of

of the NMRsignals.

T, Nse interacti^. ns may be : ez-.Ived -; slnj types and ! r.fcr-. at! or. itzlit

the versatility

NFR. Cre can deduce ! Ifferent

a seysts: y changing elthe. - the

type :f NMRmeasun-ment, or the system itself.

Is that it is essentially a -,onNMR one ,f t1he most Importantadvantages using .n invasive and non-destnctive probing technique. Samples ZaYbe sublectel to a series of NMR

damage. without structural or chemical experiments

In the use of NMR. Firstly, the parameters Thereare two major Jisadvan'tages of Interrather than on decayrates, are phencmencn est, suchas chemicalshifts and magnetisat! InterpretaIn the NVR other words, res-ilts extensive of measurezents require system-based. In to the properties of the systemexpressed tion 'Anorder to relate these derived 'parameters familiar -inits. Further complicationsoccur because of the fifferential dependences of the

different NMR techniq1jes uponthe different interactions.

ghilst this has beenrelatively In those of greater com-

direct In the case of simple systems,It has provedvery difficult PIexI ty.

Secondly,whilst the selectivity of MR Is high, the sersitivitY Is low. Since the

technique Involvestransitions between energylevels for populations 'Alch are distributed

those levels accordingto Boltzmann amongst statistics, and the available field strengths are low, i. e. the magneticenergy2g.B is much less than the thermal energyKT, the signal II

large strengths are Inherently poor. NMR samples and are not experiments relatively require suitable for trace analysis.

The low'sensitivity

does not present difficulties

in the study of wite r -saturated timbers

in a complex signals can readily be obtained. However, andNMR systemsuchas,that of waterloggedwood;one Is operating at the extremelimits, J not beyond limits, the of, apIf plicability theories. In attemptingto improveour understanding, of'current NMR of the Sys-

testing the applicability of current tes, weare simultaneously, theories. 111MR

It should be stressed that in this work It is primarily the properties of the water in

the systemthat are of Importance to us, the emphasis being on the manner In Alch Its Interactions with the cellular material of the wocd,and the prese: vatives jsed to stabilise the behaviourfrom that of bulk water. It Is consideredthat the woods modify its dynamic detailed compositionof wood on a molecularlevel is relatively unimportantfrom this point Yhat Is of interest to us however, is the macroscopic of view. structure of the ucodand the by which water preservatives maypenetrateand,reside in the wood.Onewouldexpect method the compartmentation of the water to haveprofoundeffects on Its NMR properties. A description of the macroscopic structure is therefore given in chapter 2.1 1

Vooddiffers from other cellular, systems*inthat the water contents normally encountered by weight, or below. Consequently, are low; ' usually, in the order of 38% water In woodIS normally found In close association with the woodcell wall, or in vapour form. 'in contrastf "lary Rose", are highly saturated, having samples of woodfrom the Tudorbattleship, the ", water contents which are similar in magnitude to those of other biological cellular systems. One behaviourof the water moleculesin water-loggedtimbers, therewouldexpect the dynamic tore, to be markedlydifferent from that of water In fresh wood,and that this should be

reflected In Its NMR-propertleg-Onedoesnot necessarily expect to observethe behaviour Is not In equilibrium with because which is shownIn other cellular systems the water In wood 'Its environment.

In this work measurements of TI and T2 and 'he self diffusion

coefficients

of the water-

logged woodhave been measuredas a furction oil water content, degree of degradation -o'and temperature of the samples lin order to characterlise its nature and contrast1ts behaviour

In with that observed for vater in fresh wocder 'Limbers. The NMR samples of water properties treated with, different Irades-of 'the stabilislng agent Polyethylene-glycol have also been ob-

seryed,, and the modification of the dynamiic properties of-dater in the treated samples 15 discussed. I11, ' . -..

Refererces

1.

Mathur-de-Vre R, "The studies of vater in Biological systems. " Frog. Blophys. Molec. BiOl. Vol 35, pp 103-134., (11979) ---:,, , " , ,, "'I ,I

2.

Belton P-S-, NMR - Ratcliffe R,G. ", in Biological Ussues." Frog. andCompartmentation Spectroscopy. Vol 17, pp 241-279.(1985) In NMR -,

3.

Lillford F.J., Clark A.H., JonesO. food V. 'Distribution of water in heterogeneous ' Vater In Polymers,Chapter10. Pub. American Chez.Soc. (1980) and modelsystems.

4.

R.,, Kuntz-1. D. "Propertles of water in Biological systems. Cooke " Ann.'Review Bloph. and Bloch. Vol 3,, p 95. (119714)

S.

FranksF. "Vater -A comprehensive treatise. '

Pub. Plenum Press,,New York.'(1972)

CHAPTER 2 VATER-LOGGGED ITS CONSERVATION AND UOOD

2.1 Introduction

In an archaeological coniext water-logged woodmaybe defined as that which has been completely filled with water instead of'air and has been chemically broken downby the action of

'causing considerable weakeningof, the the mood. structure of micro-organiims,

to absorbwater Is amazingly Theability of different speciesof wood variable; alder, beechand maple, for'example, take up water very rapidly and can become saturated within a few hours,,while in other species suchas cak the processcan take weeks,cr evenyears. ''In
the case of porous woodsthe chemical breakdownoccurs almost simultaneously throughout"the In less hydroscopic species the breakdownis likely ' sample?whereas to occur from the out-

layer cell one at a time. side,

'Despite the quite fresh appearance of water-loggedwood,normaldrying of mostsamples lead destruction of the samples to the complete Athin a matter of usually procedures out hours. Theconservationof vater-lolged wooden artefacts is aimedatstablllslnj the size

the'samplest"and"at durability, without destroying'their aesproviding some and'shape'of thitic value.

The following chapter begins with a brief description of the structural and chemicalcomafid,the waysthat'these are degraded as a result of water-loggini." position's of wood short historlhl A

to' conserve, havebeenmade accountIs presenteddescribing attempts which a

saturated wooden artefacts, followed by an appraisal of the processof conservationby impregnationwith polyethylene glycols, which is the mostwidely and successfully used method to date.

2.2 TheStructure of Vood. -, -

-There Is a great'diversity I-II, of wood'

In the appearanceand physical properties of different

types

and yet the ul tra-structure

ct the dif f erent species is basical ly the same. - Each cel I

Al I 'woods are madeup of cel Nlar tissues composed of three main constituents. maybe thought of as a lattice luloses. of crystalline

cellulose encrusted by lignins and hemicel-

Cellulose is a straight,

fibrous and extremely long polysaccharide madeup of several In wood, cellulose forms a ropeto each other to makeup strands threaded together form The macrof!brils give form

thousand, beta-glucosemolecules that are rq-1 C-4 linked. like structure. - About forty-polymer chains align parallel knownas elementary fibrils. microfibrili,

Several thousand of these fibrils

which themselves intertwine to form macrofibrils.

and strength to the cell wall by binding together in alignment. electron microscopes, and appear as thin striations

They are visible using

In the cell wall,

The elementary fibrils different have which

are surrounded by shor'ter' chain molecules called hemicelluloses

'with d*e-g'rees monomeric units' of polymerisation of around 100. The hygroscopic form the most system, and

hemicelluloses are adapted to the cellulose crystalline fraction of the cell wall.

Lignin Is depositedonto hemicellulose and Is bonded to it. It doesnot bondwith the cellulose itself. and gives rigidity It is an a3orphous building blocks substancederived from phenyl-propane to the wood.

- -1 1.

''I

-.

1'4.

7
I

Thecell walls are constructedof, four layers;, the primaryuall, and the outer, middle and inner layers, of the secondary wall. Theyare distinguishedby the orientation of their inner the deposited Is on figure 2-1). The thick secondary side of macrofibrils (See, wall the primary wall and provides by far the majority of the cell wall substance.

known lignins Between'the as with walls of adjacent cells is an amorphous region packed to eachother.vii pores-in-the call, walls the.middle lamella. ThecellAumina are connected In fresh wood, qits act as valves and allow for the passage of, water frcm one called pits. but, dead to wood,mostpits are closed., another, cell -In

S3 S2
si

Mmarm wall
Figure21 Organisation in filres. of cell walls wooa
the tanins, fats of Thewalls contain resins, and waxes,which occupya proportion between fibrils, open capillaries existing and cavities fibrils. microfibrils and elementary

For Is each the 40% Thecapillary system extensive and makes cell wall volome. up about of 100-2-00 It Is is there that of cell wall, estimated squaremetersof surface cubic centimetre below: ", to the dimensions is in A guide given of area. pores wood rough

Fissures within elementaryfibrils Capillaries between fibrils elementary

Inz IODM

8
-i

fibrils Capillaries between Pares In pit membranes Diameterof-a water molecule

up to 80nm up to Isonm 0.2nm

There Is a great diversity another. The different

In size, shapeand juxtaposition of cells from one species to

species of woodmaybe broadly divided Into two groups. These are

ang!osperms, knownas hardwoods, and gymnosperis,known as softwoods. Both are constructed

trat Ing systems by two Inter -pene cf ceils;

Zitudinally,

one oriented radially and the other lon3paces caUed resin canals.

which are Interspersed by intercellular

The'Iwogr^. ups can be distinguished by their different cellular ccmpcsiticns. The large
pores,' evident only in angiosper: s, are called'vessel3. These are madeup of smaller tubular

that : an vary In length from a few cen. units that are-joined-end tO end to fcrz passageways tIzetres to several meters. They are used In the 46 raI.sportation of water. lost are oval in species.

shape, and their spatial arrangementsare fixed, makinj them useful when Identifying

I, he tissue between vessels Is madeup predominantly of tjo types of cells. These are thin

to both anwalled trachleds, -and thick walled fibres. Thesetypes of cells are common In gymnosperms, and gymnosperms. glosperms as much as 90% of xylemsmade up of vertically lengths-of trachleds have that which are elongated are and arranged cells stacked uniformly, 3 and 7 millimetres. Fibres resemble trachleds but have thicker walls, fewer pills, between

Fibres lumens. termthe bulk of the solid matterfoundin angiosperms. andsmaller

Thereare a number of other types of cells found in woodsuchas the parenchyma, and the resin-producingepithelial cells. gitudinal direction. Themajority'.cf cells in wood are oriented In the lon-

The radial elements are called rays. Theseare composed of passageways

whichradiate outwardslike the spokesof a wheel. Theyare usually 10 to 15 cells wide, and are often associated with horizontally oriented resin canals. (Seefigure 2.2)

Homogeneous rivs composea entireig of paaialIg orientatea cells.

Heterigeneous rags untaining UPI-ight cells.

Fivre

2.2 lag Patterns in wood

2.3

The DecOmpositionof 'Jcod

There are three methodsof attack : ausing the deccmpositlon o! -jood'";

a chemical attack

causedby -,xygen ard water, a vegetable attack by fungi, mic: o-crianisms and bacteria, and an animal attack caused by insects.

In the first

process, the long cellulose molecules are broken downAt h the result that The cel-

the mechanical strength of the cell walls is weakened brittle. and the woodbecomes

Julose undergoes a slow, natj: al, decomposition In the presence of oxygen, heat and light. oxygen is the main deterioration agent, and tte process is accelerated by the presence of

U.V. Ilght and high temperatures. Lignin decomposes in the samemanner, but at a muchslower rate.

Vater causesthe hydrolysis of certain compounds In xylemand can also be the basis for factors, suchas fungal growth. A number other decomposing of fungi are able to live from' wood,using it as a nutrient. This results in the physical decayandweakening of the wood, whlc gives"rise to both orfzontal and vertical cracks.

to

Attack from micro-organisms mayoccur aerobically and araerobically.

Anaerobic processes

Occ'jrunderwater, or in the mud micro-orlanisms at the bottomof salt or ! -9sh water.,, The, Into ethyl alcohol, acetic acid and, finally, attack the cellulose, breaking It down carbon

dioxide and water. Methane fermentationprocesses of anaerobicmicromayoc=r by means can lose up to 65-85% organic attack, and the wood of Its solid cortent. Lignin is rot 'If it stays fermentation,but xylemcan be convertedto humus directly by methane decomposed in water long enough. Enzymic attack fis least effective In acidic =rditions.

Animal decomposition arises becausemanyInsects, Or'their

larvae, dilest'xylem.. ""

TheProcessof deterioration of wood in wet araercbic c: -dIticns is a combination of the above. " Firstly, the readily soluble materials In the vocd ! Issolve into the surrounding

followed by readily hydrolysedcompcunc's medium, suchas pe:tIns and pentosans., Mcrobi3logical degradationof the morestable cellulose and hemicellulosewill then follow. Finally, lignin remains, thoughthis too may *--y in an anaerobic be deccmposed micro-organisms enviro-nment.Both fungi and bacteria causethe breakdown :f xylem, andboth use extra-',, ', to hydrolyse the cellulose, hemicalluloseand lIgnin. ' Degradation cellular enzymes spreads throughthe capillary systemof the wood can penetrate. as far as the ext: i-cellular enzymes

degradation hemicellulose degree be by thus the of can expressed cellulose, -The remaining
and 119nin"'. Scanninjelectron microscopy'"' reveals that degradationoccurs Initially

In the secondary wall, starting in the lumen,and progressingInto the cell wall. Structures in whichdegradationis extensive maycontain only 30% of the original cell wall substance, of the, carbohydrate with 90-95% material having disappeared.; Eventhe middle lazella*maybe from t.he fibrils corrodeJand dissolved. As carbohydrates are removed cavities are produced which fill of the cell wall,, --

with water. In these structures, water acts as a bulking

11

the agent, which keeps the woodIn shape long by remaining preventing kept is it -jet, as as hydroxyl cellulose molecules from combiningwith bonding of adjacent eachother via hydrogen groups.

The result of the decay of cell material Is a considerable loss of original the wood Is extremely decomp", He uah: content can be as muchas 9095. start to collapse.

strength. Den In at

It at

IS r'elative humidity, such woodAl!

Id jat? r-logled -icod hen drying, :., as the

' disintegraticn, strong shrinkage, and S'l. very warping cracking, complete and even hows strength of the capillary evaporating water. bY the contracting capillary system is exceeded

forces Of the

In manytimbers, the decay is nor-iniftrt, tifled.

Idenbe deiradaticn three an -. of and regions aa

The decay Is mcst advancedat the s=face ct these samples, which become extremely Below thi's lies 3 thin ! 1brous layer In which thle break71, ': luner core cf these

soft, and greyish-brown In colour.

downof the cell walls Is less advancedthan the cuter layers. sampleshas the apppar:nue of fresh Ulbe,, st is

Collapse ompletely saturated with wate:. .

due to In the layer Is the greater cell wall, the voids much greater cells on outside c! a the by removal of celluflose. caused lamella and someprimary wall. : ells In this rellon ccwntalnlittle a more than the -middle

In the two less degraded layers the secondary wall is still

frequently 1osened from its contracted and position. although Present,

Water-logged Wood 2.4 Methods of Conserving

Vhenrestoring water-loggedartefacts, it is the aim of the conservationist to producea finished product that Is stable in the envircnient In ihich it Is to be kept, while maintaint of the original aesthetic qualities of the artafacts as possible."' Ing as many

12

!n the : age of

woodartafacts,

to Iftat of drying out ced the proble3 !s red, -,

Once dried, Ahe the samples whilst preserving their overall shape, artefacts appearance. and must be able to withstand changes in the surrounding temperature and relative protection against fungal and organic attack Is also desirable. humidity. - Some

Re first

the alum was successfully employedmethodof -.onservaticn, cf. water-logled wood

In The boiling jorgensen"' 11859. In mby potassil; Introduced samples were soaked method, the' In the 1 had the been the by and samples alum,, water sulphate unt'j. replaced aluminium the The left to in then the of alum solidified cool. were pores and capUlaries samples bulking thereby agent prevenUngsubsequent and as a acting collapse. wood,

be had several disadvantages. It : equired that the surface of the samples Themethod cleanedand treated Ath varnish, or some such protective suter layer. It wasnot reversible. It was found that the water vas rezovedfrom the sa., ples faster than the alumcould

Furthermore, the the structure, leadingto'Internal Stresses Impregnate a, 14 Some collapse. .. by the the the depended on method abscrbed of: amount alum of was and success unpredictable, itself. and-theconditionof the wood wood, , I -- .?"' . 1.. I

forces Is the dries in to deogree It of The shrinkage, capillary of proportional as wood , liquld, and shrinkageIs less pronounced drying Is fast. the evaporating, when It is possible

to reducethese forces by replacing the water with volatile, low surface tension liquids tertiary Ethyl forces the far and ether on : ell exert wall. smaller which, uponevaporation, butanol were successfully,used in early trials In a number of caseswherethe degradationwas

not severe.

' An alternative wayof reducingthese forces Is by freeze-drying the samples., The technique dependsupon the sublimation of ice within the Yoodbeing suchthat-the drying-water phaseIs avolded. -III,, 1 .-II., 11 1 "I. -I I- II

13

Freeze techniques. these Wrequired both "'In general, some of 'consolidation will with the by vhen set up drying from iater often gives rise to Innumerable stresses 'caused cracks, bulking If be' agents, the ' These with freezes. samples are pre-treated can avoided water method. using the alcohol exchange while a final treatmentmaybe required when

"Consolidation of the'saturated'xylemcells has beencarried'out using a variety, of 'bulkhigh involved the I impregnation Early in by at hard ing agents. embedding wood a attempts wax the in the Substitution the for where temperatures. of water wood alcohols wasnecessary be in that'are insoluble totally In this 'Objacts can way wate. r. usedwire nated waxes iimprel. fixed )or, lr to &'periods of time, but the surfaces usually rereiain, a state ove"r expected quire extensive cleaning, which is normally difficult the surin, but cf simplest on anyth, -s

do not readily penetrate the wood. faces. The'processcan be slcw if the waxes

"he to Impregnate To speedup the rate of Impregnation,several attemptshavebeenmade glycol, -and suchas vinyl acetate, methylmethacrylate, and zoncethylene with monomers wood then to polymerise'theseinside the cell walls using catalysts. Thetoxicity of the monomers difficilt ire be It however, " drawbacks fire to serious can method, and risk using'this and in is The technique some result surface. non-: can eversible, and obtain a satisfactory the In wood uponpolymerisation. shrinkage 'to

Voodusing PolyethyleneGlycol. 2.5" Stabillsation-of Vatei-'Iogged

to date is by Impregnation Themost effective zethodof conservingwater-loggedwood with Polyethyleneglycols (PEGs)"". Theseare water soluble polymers of ethylene oxide. PEGs

are produced in different gradesthat are designateda number representing the average molecularweight of the grade. In general, PEGs 200 between and of veights vith molecular

14

600 are clear, viscous liquids between 1000 those weights with of at roomtemperature,while and 20,000 are white, waxy, solidsM,. solubility, At higher molecular weights, there Is a decrease In

vapour pressure and hygroscopicity. The changes in the polymer properties with whendeciding which grade to use In the preservation of

molecular weight are significant water-logged timbers.

Low hygroscopicity of the finished the but. desirable is at product grades can

sametime, the conservationist must consider the ease with which the different

penetrate the water-loggedartafacts, as well as the effectiveness of the different gradesIn stabillsing the timbers.

, The treatment of water-logged artefacts using Polyethylene glycols was first

Introduced

In the early 1950's by Centerwall and Moren"'. They adopted a technique wherebythe artefacts were ImmersedIn PEGsolutions at 65 degrees centigrade. The concentration of the

by increments of 15%. solution was increased daily from 0%to 100%

PEGS In early experiments, between 15,000 'Jith molecular treated using weights samples
hollow cheeksafter being allowed to dry for several days. Samples and20,000developed little treated with molecularweights betweei1,500and 3,000 underwent or no dimensional

became but it wasnoticed that these samples greasyand increasedin weight whenthe changes, humidity Of the air rose. -

havingmolecular PEGs weightsin the intermediate range,I. e. fro: 4,000to 10,000,

dimensional stability good provide to the water-loggedwoods,which showno tendencyto take

4,000 were favoured, up water at relative humidities of less than 83%. Treatments using PEG as the lower molecularweights could penetrate the woodfaster than the larger molecules.

15

Today the conservationist

Is still

dependenton the PEG-vater system. lt'Is

relatively

it"". health to those dces with cheapand risk working not provide a 4,000 is still molecules will

In most cases, PEG

the be Soo PEG !, as smaller for chosen; may preserved wood, well used or, freezefollowed by 400, PEG Ath into better the wood. -Treatment penetrate

drying, has also proved useful.

is to The important factor in addressirl'the questionof how?EGs'prevent'sh. -Inkageof different Molecularweights occupythe cell wall determinethe degreeto whichPEGs Is the bound how, how of replaced, much water or and effectively prr'-occupiescell spaces, lumina.

impregnate PEGs either the cell wall, the lumina, cr both. Thebulking effect then in te the drying. PEGs the collapse weights of on prevents of molecular cellular strilcture It is 600 200 during to treatments, bound assumed soaking and the replace water range 3ome of PEGs that they remain in the cell wall upondrying. These offer substantial dimsional the the the to below fibre of wood point saturation evenat concentrationswell stability wood.

in the molecularweight range1000to 3000are also expectedto infIltrate PEGs

the

bulk'the to the higher cell of expected vood-but, -at are also concentrations, ultrastructure to 3000 than resist luminaas well. PEGs expected with molecularweights s: are eater aa by bulking forces ol'the'cell shrinkage allow themto penetrate the cell wall. too Is large to 1umIna their sphere solution only, -as

16

The grade of PEGthat is mcst suitable for the preservation cf water-logged timbers dependsheavily on the state of the wooditself. A detailed study of the preservation of

water-logged wascarried out by Hoffman"O', in whichhe attemptedto rewood,using PEGs, late the degreeof wood degradationandPEG PEG taken by molecularsize to the amount up -of
the system, the stabilltY achieved, ard the resulting hygroscopicity.

Hoffmannused PSGS with different

molevilar weights from 200 to 4000 oa voods with water .

580%. 120% between The fluorescence tests develcped by young and 'Jainwright'"' and content the to Is extent the assess PEGs. fluorescence of penetration used of AYI to were Ir een ellow

fre: inducedin woodby shortuavelight Of wavelength 450-490n. The fluorescence m. origirates

lignin. PEGstained with cobalt thio-cyanate suppresses this natural fluorescence, which can

be observed inder a microscope.

The anti-shrink

efficiency

(ASE) of a treatment was def! ned as:

(BO BD/BOx 100%

the is 30 shrinkageobservedin untreated wood,and 31 Is the shrinkageobservedIn where treated samples. A goodindication"U'the rel-atlonship between Initial water content and the

degreeof shrinkageto be expectedfor the different treatments is given in figure 2,3.

that some Theresults showed PEG penetrated the wood,whichevergradewasused, but that the penetration, using low molecularweights, was superior to that when high molecular using a weights. The various degreesto which the fluorescencewassuppressed wasindicative of the easewith which PEGs penetrateddifferent regions of the cell. Thelost accessible regions

are the corners of the cells and the middle lamella. Next is the primary cell wall, and the 'Bostelusive region for the PEG moleculesIs the secondary wall.

17

U-A

C"Oss

Sectional Shrinkage Wl" IdI1.4. tp@atnent

40
Ix..

321

/, -, , 11 .

-I

_ I

PEG 49g

699 PEG PLC1590

130
rigure M

Stabilitg acheivea treating water-loggea waacls PECSWferent with

2ig 363 40 Max mcisture cantantof imod

plase

3090 440

qoffmann concluded with the following:

ID

PEG 1,5010 wasCOnsistentlypoor at effectin., the ji. -e-S, of waterstability 40nal loggedwoodto: all water contents.

Thehygroscopicity wasgreatest whenlow molecular 4eight ts were usedbut, even in these cases, relative humidities were'1'0%-80% before there wasevidenceof leaching.

In relatively fresh timbers, wheredegradationis not severe, stabilisation is brought about mosteffectively by low molecular. i. e. small molecules welght PEGs, which enter Into the'large parts of the capillary systemof the cell walls, middle lamella andeel I, keep the drying, Here they replace the 1water, wood on and,

-corners in a swollen state. Voods having water contents.of up.to 160% can be treated In this way.

is

In heavily: degraded'woods having moisture,contents upwardsof, 400%, - especially where the secondary'wall'has been destroyed, there Is evidently not enoughwoodsubstance ! eft' to be kept'swollen by s2all'FEG moleculesand, thereby, stabillse pus.- In this case, stabillsation higher'molecular weights. the whole -car-

is more effective If the wcod Is bulked using

In these Concentrations, 70% cases. are required of around

(V)

highly two both His 1% there 3etween these' extremes a region of -joodcontaining fferent d! grade degraded and well preservedcells, eachcomponent a of whichrequires apbe used, or consecutive of PEG.In this case, 111tute5of different jradeszay mayprovide the solution. plications of PEG

Referenl-as

I. 2. 3.

Stamm U.

Woodand Cellular Ecience. Press, New " Pub. Ronald York. (1964)

MorleyP.R. "PowTreesGrow. ' Pub. Edward Arnold Ltd. (1963) PanshinA.J. and deZeeuw Pub. Chapters Tech, ' 4,5,11. C. "Textbook 'Jood "nonlogy. of McGraw-Hill,New Yerk. (1980)

BorginK. "Themechanism d,., to 'he environmental e wood of of breakdown structure of factors." J. Inst. Vood Scl. Vol. 5, pp 26-30. (1971)
Jesperson large K. "The decomposition Symposium Int. of on conservation of wood., 1977) objects of water-loggedwood. Amsterdam.(Scept Borlin K. "Theuse of the electron scanningmicrOSCOPe f or the study of weathered " J. of Microscopy.Vol 92, pp,47-55. (1970) wood. 7. Findley G.IJ.D. and Levy J. F. "Scanning the to of as study an aid electron microscopy 57-63. woodanatomy I(1969) ' Inst. decay. 4, Yood Vol Scl. pp and 8. OddyV.A. IED 45-49. (19715) Maritime Monographs 16, pp Vol and reports.

19

9.

Christensen B.B. "The conservation'of water-logged wood in the National Museum of, Denmark. Tech. Copenhagen. (1970) w Studies in Museum

10.

Murray H. ' 'Conservation of artifacts

from the Mary Rose.",

Jagels R. "A deteriorwCondition and

tion evaluation procedure for waterlogged wood. ", dimensional stabillsation "Stabilisation

Barbour R.J.

", of highly deteriorated samples.

Foffzan P.

of water-logged woodwith PEG.- Molecular si: e v degree of

degradation. " Proceedings of the ICOnwater-logged woodworking group conference. ' Ottowa. 11. (1981)

Union Carbide Material data safety'sheet No. F-48039A-GB.

20

in Heterogeneous ISystems. CHAPTER RelaxationandDiffusion Measurements, 3 The Theoryof NMR

3.1

Introduction.

The Nuclear Magnetic Resonance phenomena occurs whennuclei possessing magnetic moments interact with an applied magnetic field. is founded is well estabThe theory on which NMR

lished and has beenadeli; ately describedby a number -.f authors'"".

io fract 4 i n A CC -le nS erab

its ideas this be understandirg. theory the classical and explained using simplifies can of of is given and the concepts In, this chapter a brief accountof the resonance phenomena "his introduced times by are adoptingwherepossible, szpin-lattice and spin-spin relaxation classical approach.,

The Interpretation

ecular events of NMR relax3tion times in terms of real, physical, mo!

behaviour :3 for ttere the types that is and complicated, account cf, can all no unique model a that have been observed. An underst3nding of how molecular motions lead to an observed time Aas,first relaxation predict the characteristic prcpcsed by Bloemtergent Poundand Purcell"'. Their theory could

behaviour of relaxation times observed In bulk water, but could systems and polymer

for those cbserved in less uniform samples, such as cellular account, not

solutions.
1,1. III

For water In these morecomplicated of effects"', systemsthere are a number

most

bulk to increase in water, relative relaxation, an relaxaticn rates notably multi-component behaviour dynamic dapendencles, from the stem modified of and a variety of-temperature whlch These from the have the the water molecules,and in some substrate. of cases magnetisation Bloambergen forward by Ideas led the to theories which extend put et necessarily a number of In water is the al. Some of these are describedin this chapter. Althoughproton resonance is kept as general 35 mainconcernof this work, the description of the relaxation phenomena Possible.

21

Self-diffusion

ccefficients

for water in heterogeneoussystemsare conveniently measured

using NMR techniques'&-". steady-field

Twomethodsare used during the course Of this work, namely the technique, and descriptions

gradiert techrique and the pulsed fleld-gradient

necessary for the Interpretation

of the results from both methodsare presented in this chapcoeffi-

ter. The effects of anisotropy and compartzertation on the observed self-diffusion cients are also considered.

3.2.1

Furdamental Concepts.

A nucleus is a compound syst?m which mayhave a non-z-ar: angular moment=hl, and a mag. The two juantities netic moment are related by the equation

x hl gamma

Theconsiant,of proportionality g'amma'is characteristic of the nuclear species. Nuclei other than those with evenatomic andmass a nuclear magnetic possess numbers moment whenin their groundstate.

theory predicts that for a nucleuswith spin quantimnumber Quantum 1, the component of resolved In any specific direction (e. g. the z dire-ction),mayhaveany one angular momentum of 21+1observablevalues given by m(l), wherem(l) can adopt.Integral values from m(l) to m(l) : fl. Vhenplaced in a magneticfield of flux Bo the corresponding permissible energiesp3o are given by

Em = gamma x hm(D.Bo

3.2

idherethe energy levels are separatedby an energy

22

9E = Es(l)

EMU 1) Bo(m(l)-m(l t D) 3.3

= gamma ihx

gamma xhx Bo

The probability

to the transitions is between levels proportional spontaneous of energy

cube of the frequency of the energy associated with such an event. The frequencies encoun'ds'avallable by limited the NOR tered In fiel and are experiments are size of the magnetic n typically a factor of 101 downon those of optical spectroscopy, and thus the probability are negligible. Transitions maybe Induced by the presence of the rescnance condition of

spontaneoustransitions

electromagnetic radiation of angular frequency 'Jo, satisfying

hVo=U : 'iana xhx So vo = gamma x So 3.4

For a typical laboratory field the frequencyis in the r. f. range.

A first order perturbation t: eatment yields a probability of t: ansition proportional to the strength of the perturbing r. f. field together with the selection rule of tz = 1.

of nuclei under the Influence of a static homogeneous an ensemble magneticfield BOat T. the 21+1energy levels are occupiedat equilibrium accordingto the Boltzmann temperature factor

h. Bo)/KT exp (-m(l). gamma.

Therelative populations Nm(Dof nuclei occupyingthe by m(Dth energy level is then given a

23

h. BO)IKT , Nm(l)/Nm(l D. 2 exp (gamma.

3.5

ThereIs there! ore.a surplus of nuclei in the lower energylevels a consequently a net magfrom the a field. the 3f excitation As the direction, in applied, probability of, netisation Mo levels lower the are to the that energy Is population and of state, given state proportional 3.1) figure (See be there a net absorptionof energyon radiation. will morehighly populated

PCpalation

x 1 4/2

to them amongst aistrautei accorain, Boltzmann's theors.

statistical

2 PossJ619 enerS9levels with spins

(a)

figure 3.1 Theaistrilution of Parinetic spinsU. - a nucleus with 1 1/2.

3.2.2

TheBloch Equations

is the vectorial sumof th inTh In NMR e observablequantity measured spectroscopy M. The time dependence dividual nuclear magneticmoments vector in a of the magnetisation For Bo Is described field by famous 31: the a magnetic classically ch"' equations. magnetic by a single vector equation, field in the z direction these can be summarised

dM(r,Q/dT

(M^Bo) (?, 4,: xi + Myj)/T2 - (, gamma. -Mo)k/Tl

3.6

24

T he M^Badescribes the effect of the torque experienced by the magnetic mo3entin ; term gazzma. the field. This torque causes the precession of the magnetisatiOn vector about the z axis at

the Larmor frequency Vo, 'which is identical to the angular frequency of the electromagnetic radiation required to match the : esonancecondition and indice %transitionsbetweenenergy levels.

The system has classical a classical analogue in a gyroscope. !Jlnena gyroscope rotating about Its axis initially vertical Is Ideflected ! rcm the vertical through an angle theta, the The r3te

gyrosccpe, while continuint to rotate abcut Its axis, precessas about the vertical. cf precession is dependson the massdistribution the gravitational a field.

within the gyroscope and on the strength of

In an analogous the Individual nuclear magneticmozents manner andtheir resultant, the disturbed from equilibrium alcng the applied magneticfield will net magneticmoment when _M_Oj precessabout the applied magneticfield Bo at a frequencyproportional to that field, and dependent uponthe 'mass' distribution within the nucleus, and therefore will be Onaracteristic of it. In the event, the rate of precessionis given by jamma. has'the h.Bo where gamma numericalvalue as in equation3.1. Hence is frequently termedthe gyro-magnetic same gamma ratio.

it is this phenomena that allows the nuclear magneticresonance to be observed.It Is unto consider the nuclei Individually; they can be by the net magnetic necessary represented Mo. If this is disturbed from its equilibrium along the direction of the applied magmoment netic field Ro, Mowill precessabout Ro. If a coil of wire is placed with Its axis in the x-Y plane perpendicular to the t, axis, In the direction of the applied field direction, it will havea voltage inducedin It as the flux passing through it changes as Moprecesses. This Induced voltage signal which Is proportional to the magnitude of Moand the rate of precession,can be amplified and detected.

25

The second and third terms on the right of equation 2.6 were Introduced by Bloch to account for the effects of spin-spin and 3pin-lattice relaxation respectively. They indicate the com-

that the componentcf magnetisation in the x-y plane decays in time to zero, uhilst

ponent In the direction of the magneti: field relaxes to its equilibrium value Mo, as Is determined by the Boltzmann'distribution of nucle! amongstthe available energy levels.

3.2.3

TheRotating Frame of Reference.

Theapplication of an alternating magneticfield Bly perpendic-flar463 the static field 3o Bo, will induce transitions between the and at the resonant frequencyI;o = gamma. energy a leyels and c3usea net absorption of energyfrom the radiation field. If the amount of energy by the spin systems is sufficient to equallse the populations, then the systemis absorbed said to be saturated,

In addition to disturbing the equilibrium Boltzmann dist: ibution of spins amongst the by the field BI also causesthe individual available energy levels, the peitur'bati-on caused spins to precessabout the direction of the external field 2-oin phasewith eachother to a net magnetisationIn the x-y plane. produce

the effects of the field Lt rotating at the Larmorfrequencyin the x-y To'understand 0 plane it Is convenientto consider the behaviourof the magnetisatIon vector M In a rotating frameof reference rather than the laboratory frame. Observed from a reference framerotating at a frequency'Vr'aboutthe z axis, a disturbed magnetIsation wouldappearto precessat a frequencyVo i Vt. Equation3.6 Is still valid provided that the magneticfield Bo is re-

placedby an effective field B,. suchthat LL,

B.tt : Bo- Vr/gamma

3.7

04 26

Thus for a frame of reference W, f,, frequency 'do, the effective ), rotating at the 'armor L zp

field B.,, is zzero,and the mapetisation vector is stationary, existing as MoIn the z' direction at equilibrium.

The application of an oscillating

magnetic field of strength 2.31 and frequency WoIn the

x-y plane of the laboratory frame of reference is equivalent to applying two counter-rotating ! ields each of strength 31 and rotating about the _: axis with angular frequency +Qoand -1jo.

Oneof these components rotates In the samedirection is the nuclear spins and the other r-,tat? s In the opposite direction. The latter can be Ignored as Its Interactions with the

to zero over each complete cycle. Viewed in the rotating frame of : e! a7encet cancel nuclei however, the componentrotating wlt%, the nuclei is statinn-a-: y, and Its lnteractln with 10 I's

Lo zero. aurngad not

No

no init

up

Ma after 90#pulse
xo;; o D1 I-J. No"
, Mo

Altel. IN4 NIS@

3.2 vierts onma riem it. at gooanamo paisea maimetic

The laws of mechanics are independent of the frameof reference. ThusIf B-I Is applied along a specified direction (e. g. the x' axis) then the magnetisationprecesses about the xf Bl. In time t, the magnetisation axis with an angular frequencyof gamma. vector precesses about 31 through an angle

81A gamma.

3.8

27

A puI se BI of suf fI cient' durat Ion to Up the zagneti sat Ion vector Into, the xI -y" pI ane from the equilibrium along the z' axis Is termed a 90 degree pulse. A 180 degree pulse rotates the figure 3.2) 31. (See 180 degrees the through direction about of magnetisation vector

12. '4

Spin-Lattice (TI)l

and Spin-Spin (M-lelaxation.,

The either completely or partially an of return ,, the removal of a perturbing field LI

following to equilibrium saturated system

occurs via two simultaneous processes.

Spin-Lattice Relaxation.

'Jhenattention is focused on a given spin system, all other nuclei and electrons In the sampleare collectively referred to as the lattice. ! rcm 'he latA transfer of energy to or %,

tice from the spin system, resulting between ! ower transitions of nuclei upper and s1rom energystates, can be Inducedby the presence ef f luctuating fields in the lattice, If there is a component at the appropriate frequencyto Ind-icetransitions. This enablesthe return to distribution of the nuclel about their energylevels at a rate which is characa Boltzmann terised by the spin-lattice relaxation rate IM. -.11,1 1"I" 1-11 ". I

Relaxation. Spin-Spin

Thespins return to equilibrium with eachother via'spin-spin interactions bet1jeen, in-' dividual, nuclei. 'This Involves an exchange of energybetween-the nuclei of the same spin System so-thata transition,. from an upper to a lower state Is, matched by'a transition from a lowerAo an upper state. The simultaneous transitions cause'a loss of, phasecoherence in the spin systemwlthouta changeIn the net energyof the system. Further loss of, phase, ), coherence arises from local field anomalies.Eachnucleus in the spin systemwill Interact

28

with non-zero magnetic fields producedby the lattice

and other spins of the samesystem. As I'M, experienced by dif-

a result of these Interactions there 13 a spread in the local fields, ferent, nuclei, which gives rise'to a distribution

of precessional frequencies.

The'effect of spin flipping

and local field anomalies on a syste:a of nuclei precessing in

2. phase following the application of an r. f. field LI, Is a loss of phase at a rate gamma. the Hencethe voltage induced In the deteciing coil will decrease, representing a decrease .n of the net magnetisation in the x-y plane. There will also be an additicnal component loss of

signal in the x-y plane as Individual nuclear magnetic 3omentsreturn to there equilibrium distribution ,, with the lattice. The spin-spin relaxation rate 1/(T&*), is defined as the rate

of Icss of magnetisation in the x-y plane, whIchlis therefore given by

I/T24 , Samma. 0 + I/Tl

3.9 ,

3.2.5 ' 'Pulsed Nuclear Magnetiz. Resonance.,

In pulsed NMR Is applied to a spin systemIn a series of short duration, measurements'B-l high powered " Usually these are in combinations pulses at the Larmorfrequency. of 90 degree and 180degreepulses along the x1 or j! axes of the rotating frame. The response of a 'iystes to a sequence TI and T2. for measuring nuclear s'pin of-such pulses provides methods Phase to monitor the behaviourof the component or diode detectors are employed of magnetisation' in the'xl-y' plane. Diodedetection records the amplitudeof the net voltage inducedin the detecting . coil and'is in turn proportional to the net magnetisationIn the V-y''plane, : - using the r. f. 'source from which 81 is derived as a reference, whereas phasedetection, detects the component of magnetisation along some particular directionln the x'-y' plane.

Z9

Q)

The Free induction,Decay.

A. e. 'I/&, and -112. For a nucleus of spin 41/21 there are onlytwo possible mi components, Thesehave different two states-is magretic energies, *Po and -gEo. The ratios of the populations of these

given by

II

Nu/Nl z exp(-2.4o/KT)

3.10 -

In practice PBo<< KTand the ret magnetic Is given by moment

(Nu-ND- N12x [(I+; iBa/KT)-(I-pBo/KT-... )] N/2 x 2go. g/KT (i4A-'Bo)iKT

Theamplitude of the signal immediately after a 90 degreepulse Is therefore proportional to N, the number of resonantnuclei containedwithin the sample YoNme.In time the signal decaysto zero as the nuclei return to equilibrium. The, form of the decay,! by s determined T1,,T2*, and experimentalconsiderationssuchas the homogeneity, of the applied magnetic field. This is the-free induction decay(FID).

TheMeasurement Tiu, 11. of the Spin-LatticeRelaxation

Spin-lattice relaxation times In this work weremeasured pulse sequence. using a 180-1-90 The application of a ISOdegreer. f. pulse In the x1y' plane to a spin systemIn equilibrium inverts the magnetisation, ft, so that it lies In the -z' (-z) direction. Following this-pulse spin-lattice relaxation begins andnuclel. relax back to the +z' direction. After a time T, the z' component has a magnitude of the magnetisation given by

30

?, z (T) = Mo xII,

Zexp (-,r/Tl) 1 t.

3.12

To zake the observationa 90 degreepulse is applied whichenablesan irstantaneous measumentof M;.o be cadeby rotating 11.Into the x1y' plane. Thusthe Initial height of

the decaycurve following the 90 degreer. f. pulse is proportional to Mzat tIMe T.

After a time of not less than S-x 1% the spins have effectively

Ibrium to equil returned

the be different, the From of the may repeated using sequence values of T. and. measurement initial signal amplitude, and ising, expresslon, 112 a ial,.,e of TI maybe determined.

(H)

The Measurement Spin-SpinRelaxationTime, T2. of 1161na.

An initial

90 degreer. f pulse applied along the 1! axis to a spin systemin equilibrium

rotates the magnetisationMoI.nto the ) direction. Thephasi coherence is lost as a consequence of the combination in the and the inhomogeneities of SFID-lattice relaxation processes static field at the site of the nuclear spins, some of which are due to other surrounding nuclear magneticmoments to the ImperfectionsIn the applied magneticfield. The oband some servedFID doesnot therefore give a true measure of the spin-spin relaxation time T2, which shouldbe independent of the applied field. 1

The loss of magnetisaticnin the x-y plane has beengiven by 1/T&^* : gamma x EB+ IM 3.13

Incorporating the No sourcesof the spread In the local fields at the sites of the individual nuclei, i. e. changes in the local field_producedby other neighbouringnuclei and field Inhmogenelties In the applied field, equation3.13 maybe expanded thus; II

I,.,

-I,

I/M

- gamma x Minhom. +'EBs)+I/Tl-

3.14

31'

T "Bs Is determined by the sample. The jhe term t'Binhom. Is dependentupon the equipmentwhilst e gBs + I/Ti. sPin-SPin relaxation rate I/T2 is defined as gamma. To determine the spin-spin

relaxation rate correctly somemechanism must be devised to eliminate the effects of applied 2eld InhomogeneRles, I. e. 931nhom. This can be 3chieved '. "/ the Hahi Spin-Echo experiment"".

The sample is exposed to 3 90 degree :. f, pulse causing the net magnetisation to rotate from the +z' direction to the X1 direction. Nuclei will the about subsequently precess ;.s higher than the 'Jill mean a Viewed In the rotating frame of

axis. Nuclei In the sample'wherethe applied static field precess ! aster than those in weaker regions of the field.

reference the'spins dephaseaho-jt the :e axis. An indIvid, Ual,-it:cleus will precess through an angle 0 relative to the mean.

9=

7amma x I.Hinhoz. a

'23) xT + I-

Ci

the the the ncc agnetisation spin-lattice along 11 and rescivinj and effects orporating axes

MX'(T)

MOx Isin(g3mmax Minhom Mo x sin(e) x exp (-, r/Tl)

SEW x T)l x exp (-T/Tl)

3.16 and Myl(i) = Mox cos(O) x exp (-T/TAI, ). 3.17

but W Is 'If a 180 degreepulse is applied about the xv axis, Mx1remainsunchanged Is SB! deorpi has the the 1.50 reversed; becoming r uncharnd, p1se nhom as reversed -11,0. 2 -but 11
,I

due to 14 the is Us component all of of the fi. also resonant nucle,,, reversed. .s .. gamma x Minhom- 3s)

Thenew precession rate is;

32

After a further time

Mo x tcos(-O Minhom -t Us) x T)i x eip

Mo x Iccs(gamma x2x

t'Es) x Ti x exp (--! -riT1)

3.13 have been Uinhom The nuclei, eliminated, and suczing over all resonant effects c. becomes:
MY'(2T) 2 Mo x exp (-ZT/T2*)

Theva!ue of T2* could be derived by repeating the pulse sq,, erce at. different values of Carr Instead, spin-spin relaxation tizes in 'Nzz wo: k weremeaswed lta and of --ethcd usinj Purcel I Gill"21 (CPMG). and inccrporating Meibcom the and modificatlon, ri. r4

Initial An ,,

90 degreepulse 'Is applied to the spin sYste-'in equilibrium to rotate the direction. Theeffects of static field inhomogenelties an TZ*

MaInto the f magnetisat'lon

if after a is f. This the degree 180 can be removed r. Pulse applied along j! axis. ra time rotates the spins about the V axis so that-the faster spins are nowlaggIng-In phase,while the slow spins are leading In phase. It requires a'further time before the*quickly"precess-

In& spins recover their phaseloss and the slowly precessingspins lose their phaselead. At the 4! axis and this All ore', the spins will refocus along',, a time 2T theref, of the magnetisation. echoIn the magnitude be seenas aLn

This processcan After the rephasinghas occurredthe spins onceagain begin dephase. -10 be reversedby the same 180 r. f. pulse alongthe V axis. A train of IN degreepulses can be applied at Intervals of 2T, and echosignals observedIn a cne-shot experiment.This Is the Carr-Purcell sequence.

33

This'particular sequence Is susceptible to a cumulativeerror.' 'If'the 180degreepulse 0, or Is Inhcmcgenecus, then an error of 3 degrees hasbeen3is-set by an amount will reduce Is 20 he the by and ' After error net the'amplitudeof the first echo cos(O). pulse a second ,. Is Is After Costno). the reduction in cumulative cos(23). n pulses amplitude reduction signal degree 180 the the Carr-Purcell the of phase Meiboom-Glll of the modification sequence , Jsing t%e'f Induced by the the that about rotation pulse occurs ref oc-Issing, pulse is shifted so thro 90 n3l After Itial g degree the in uhae the instead of precesses pulse x1 axis. a nucleus , be 19 f Followinj would 160 jerreis the to'the 'axis. a : ntatlon of phaseang of +jrelative the 90, then less be than to 180 degrees by the angle of rotation were an angle and if been has be SO canAfter time the deph'asing the plane. above a would x1y' -r magnetisation Is the 90 the but is'at signal magnetisaticn an angle above'the x1y' plane as celled out, by a factor cOS(93).At a further' time r, the phaseangle for tine'nucleusIs again +1 reduced 180-90'about Rotation Vy' of the an angle relative to the X! axis, but Z"O plane. of above back into the the f axis usi,. 180degree*. f. pulse will take the magnetisatIon Ig the same x yl plane and the effect of the mis-setting will havebeenreversed. (SeeFigure VD.

va 990pube rotates NoAm V, e', in Do variation causesSpindiituslen lm" tatst reger-Ses lacl lag at t=1 phaie

IL

91
KA

91 Xkl
i t: 2T

91
0 Effug row xw -Was o=IPS fcl%

Tpat, '"-Islesa uez, foratin CM ligure 3.3a nw the a spin of eobi) palse 5eqwrcl! ttsing mId m, In -e 4lileotion. apylLod an enserlleof eins . EXI. ernal ritU 31)
990Pulse 1891 Pulse

t=at=7t =27 time bs 980. Pulse figure 3.3h CPKG is 1800 shifteJ sequence. phase pulse

34

The sPin-spin relaxation ti3e may then be calculated vla the relationship MM z "o x exp (-t/T2) , 3. Z 10

(n "IT, t2'. the time Is height the spine echo observed at , 4here 11(t) of

3.3

Ev9ntsMolecular Relaxation times (TI T! 11, in terms The Interpretatioon of of and 4

A. 1cemberogeen, Pound Mechnisms of Relaxaticn and thes and Purcell"'

Theory.

Following the applicaticn of a sequenceof pulses, a spin system will , librium Boltzmann distribution.

return to an equi-

Interact is throug, The principle mechanism -he of : elaxation

the field tion between nuclear magnetic moments having correct of magnetic a component a and frequency and polarisation to induce transitions between '11re energy levels of the spin sys-

tem. In the case of dipolar spin-spin interactions these f! elds stem frcm the Interacticyn of Its by nuclear magnetic moment reighbou: inl, nuclear magnetic mo-7 the fields generated with ments.

In the presenceof a large static magneticfield, the energyof a nucleus. in the field due to Its, nelghboursIs represented as the sumof energiesof that nuclear magneticmoment due to its magneticInteraction with each. The field at the nucleus I due to a neighbouring by nuclei j Is liven a pi. bj (Lpj lrlj, ) x (3COSIM-1) h. 11 x gamma. h. li gamma.
-----------------------rij, 1 OCOS20H),

3.20

rij

is the separation of the two nuclei, and 0 Is the angle between the magneticmoment of

the Ith nuclei and the jth nuclei.

35

For nuclei of spin 1/2, mi can be +/- 1/2, leading to two quantumstates which maybe ftur total to leads this of combinations of 31. For a two callel ii and neighbouring, nuclei

spin states; lliai, ajoi, Iab Ifl-

Transitions can occur betweenall levels. The transit! on from alaJ to ii$j

or $1-ij, and

that $Uj Olaj to in the nucleus transitions fros alfl or representchanges state of a single for the in two the h9o. state energy of nuclei of change a requires

The transition energy No.

aimi to ifl

involves the simultarecus t: ansitic.., of bcth nucle;. and an the involve to from states !: 3! Izj to -xij --2i$j and om xj

The transitions

transition simultaneous the lattice is required.

of both nuclel, and no net transfer in energy betweenthe system 3Td

The local Internal field has sever3l sources Of arijin.

The magnetic mc. -en'.s of relaxing

be Influenced by the fields produced by other magnetic nuclei -may particles suchas other nuclei of similar or different spin systems,or unpairedelectrons. The linteracUonsbetween the moments maybe direct dipolar couplings or Indirect scalar couplings via electrons. Each 'fiucleusin a sample will experiencea field due to its neilhbours which mayadd or subtract from the applied field. Thedipole - dipole Interactions predominate in the caseof proton spectroscopy. resonance

The local field Is determined by statistical width of the distribution is the order ef

i consider3tions. At normaldensities Ithe

IaL! will to movo : -. ve practice'-nuclei mT. each -In

by a given nuclei will changs.Onecan define an autoother and the field experianced 'of fuction the forim, correlation

B(T) X NOT) G(T) = ------------BIM M

36

whereBM is the local field experienced by a given nucleus at time t, G(T) describes the coherenceof randommolecular processes with time. The spectral density at a frequency W,

G(T); transform Fourier frequency is the 9 of representing the component motion at a of

J(w)

1-0) G(,r)exp(-iVt)

dt

3. 21 "-

it is usual to assumethat, G(T) is expandedin character so that

GW = G(O) x exp(-,T/Tc),

leading to;

2TC J(W) = G(O) x --------1+ w2TC2 3.24

fiVre 3.4 the J(W) fcr form the TCis a function of the -temperature and shows of, sample, of different values of Tc. The component below locally frequency the field oscillating any at of the correlation frequencywe quiCkly falls to, zero. bovewe t,1, e spectral density function rc-1 Is constant. A.

&"Si"

Sectral "
im

Low t4mpooatape

aensity for MO. lov fuction R4)skova fivure3.4 ]It spectral lemperaturvs. ani intentaiate

37

Bloembergen. for Purcell Pound the a systemof similar : considered elaxaticn and magnetic *. I= 1/2, undergoing c. ctrcpic moticn. They spins with an angular spin quantum randcm number that the only contribution to th fluctuating lattin assumed dipole field wasfrom the dipole-

the spins, which-culd te between by a single correlation time interactions ddescrfled

Using time-deFendent perturbation"'I

-I theO"Y &,. ey s'nowed that for such a system, . om-

1ccal field f: 14c 2"w'o the at and equencies of could contribl. -te to the spin-! &ttice pon. ents that such : elaxatlon

---

Z-

7-,. t

Jwo)

AVII-40)

12.1 AS v

TI
T

--------3

---------

jammaz- whereuz = 9/160. rij-4,

between distance the the ith and th interacting i is and :,.,

spins. zz is termedVanVleck's rigid lattice second -2cment..

In addition to the ccmpo,,, ent of the sFectral density at 'Jo and No, there is a : ont: ibution to the spin-spin relaxation time from the low frequencycomponent of J(O).

TZ

; 2.13 J(O) + SJ(wo)+ 2J(i'6'wo)l L331

rII.

3.27

11 ,,

- ------------

3(l WQZTZ) 3(l + 4woZT2)1

- -----------

3.28

The functions of equations3-26 and'3,8 Z35 are describedby the log-normalplot in figure

..

It has beenassumed that the correlation time has an Arrhenius temperature dependence, i. e.

7=TX

eXp

(E/RT)

3.29

33

, jheie

R is

the

niversal ,; tor

ps

constant.

T is the absolute oy the nuclei

temverature. le3dinj

and E is an apparent local

actiation field.

energy

Ine motions

describej

to the Huctuating

Relaxation times T1,12 as) (Secon, I

TI at low frequencti Tl ana T2

frequencV high 11 at

12 Curye

Figure 3.5

Correlation time (7) (seconas) Diem. Behayiourof T1 anJ 72 with Correlation time (1) - BBP 18-19 Wolz1 12-5 I

Thesecurves havebeenusedto predict the behaviourof T1 andTZ over a rangeof temfield peratures in liquids where3 single correlation time maybe used to describe the IOC31 fluctuations. It can be seenthat TI passesthrougha minimum at the point wherethe rate of molecuiarmotionsare closest to the resonantfrequency. i. e. when wcT 1. At lower values of T. TI is equal to T2. This is a consequence of the extreme motional narrowingin this the spin-spin region. As r increases, the J(O) term of equation3.27 increasingly dominates relaxation. and causesT2 to decrease until the rigid lattice condition sets in.

Thecomponents to rapid motionsdo not contribute to the dephasing of J(w) corresponding is dephasing T2. The they time because the occurs are averaged out. scale over process which If dephasing were the sale factor determiningrelaxation. then because

T2

AH

Bi(t). dt gamma.

is.,th, that. e dephasing angle of nucleus i, at a. time t, T2 wouldbe defined such.

39
N

T2

Cos[gamma. 31(t). dt] >=

I/e

3.31

in dephasing the be effective andonly components of motion ;p to a frequencyof. II/T21 would process. IPIIZ is proportional to the integral oil J(w) ! rc2 0 to IM. "

frequency Is As the..temper3ture,is decreased a cond!'11onIs reached w1here-the correlation less than the static rigid lattice line width 1/TZ, which Is prcpartional to 19, the sq-jare root of the -. In the correlation C' I , result ! requncy dces ". 10t

sec: nd moT. ent. Further

4, any change In the. M -rtegral,

ad r, t', e sp, In-spin relaxation rate remains ncharged-

3.4 i.,

Heterogenevis Zystems

11theory of Thebehaviourof TI andT2 for bulk water Is adeqately JescriteJ by the BPI ,, section Hydraticn wate: in heterogeneo,,, s systems,however,exists In a variety CI

Its A single 'zorrelainteractions environment. from with states of molecularmotion arising systlOn time Is no longer sufficient to characterlse the kinetic prcpertles of such complex T2 T1 the for to tems, and a number obbeen and have acz^wunt theories proposed of extended servations. IIII-, I. 11 111.

Thecharacteristics are that the relaxation rates are considerablyenhanced over bulk, T2 values at all temperaTl values exceed water.values at all frequenciesand temperatures,, tures including, those abovethat at which the TI value is a 2inimu, 2, T1 exhibits a frequency dependence that is Inconsistent with simple BPPtheory. Thereis also a detectable unfrozen water componentInmany belowOIC, which cannotbe explained simply sa2plesat temperatures by assuming that the water moleculesexhibit a restricted mobillty, because the relaxation rates are considerableenhanced over bulk water values.

40

f-,

4.1

T11-1 discrete Mfultlphase model.

in that water in heterogeneous I It hasbeenproposed a number of 'discrete exists systems ?phases', "' . Twosuchphases are ncrmally envisaged;one containing 'free' water molecules with properties resembling'those3f water In the bulk, 3M the other containing -dater degreeAth the substrate zaterial. The molecular3otions of iolec,iles which interact to some it this For be restricted or modified 1-1 this interacting water may reason a r, mbero* ways. -, has beentermed'boundwater'.

Exchangeof nuclei between the two phases proceeds-at rates which determine the NMIR relaxation characteristics of the system. In general an NIR : easurementcannot resolve a experience all possible are

population of -nuclei into separate phases 14' the nuclei ef! ictively

The these two "Etions the during cc.. phases -der states relaxation. period of zagnetlic . ! /TI-js'given' 'Id to be in ite : relaxation said, : ap, exchange, and an observed average of the relaxation rates cf'the individual' phases.

by the weighted

Pa

Pb 4

+ TI Ta Tb

3.3Z,

wherePa and Pb are the populations, andTa andTb are the : elaxation times of the boundand free phases.

Watermoleculesin the bulk typically teorientate at a : ate Vc v 1011Hz. I/T1 is proportional to Vc and has a value of about I sec.,. In bulk ice the vater moleculemobility is

restricted and We%104Hz. The spln-sin relaxation rates are therefo'r'eIn the order of 10, sec-', giving TZ values of a fev :aicroseconds.

41 -

When the adjacent to a surface cf slow moving macromolecules water moleculesexhibit an Hz, 101 intermediate in the by Vc of order corcharacterised rate values reorientation if Hence intrinsic to times a systemexists responding spin-spin relaxation of milliseconds. then 4here1% of the water is boundand exchanleIs rapid betueenthe boundand free phases by the bound the relaxation will be dominated phase.

In the limit of slow exchangethe lifetimes of the spins in each environment are long to the respective relaxation Umes, and each environment effectively compared relaxes inde-

3Pare If The two be the separately conditions phases will observed pendently. spectrol: eter is ! The a sumof No decays such that ecaY propriate. observed magnetisation

MR) = M(O)x

pa.exp(-t/Ta) + Pb.exp(-t! 'Ib) 1,3.

where for longitudinal

M(t) ! transverse (O)-Mz(t), ?, : elaxaticn and or relaxation MM :

My'('6j. In favourable conditions the two components of relaxation may be resolved and the true value of Pa, Pb, Ta and Tb found.

3.32 forvard, Is TheaI 'phases to equation thr'ee discrete and straight xtenslon of or more tecomes n Tobs i--l ?I TI

fCr the case of rapid exchange, atlon eq,., and In the limit of slow exchange 3.34 becomes ,1,

NO

nti K(O)xE Pl. exp(-t/Ti).

42

For 'Intermediate rates of exchangeAere the lifetimes of the spirs within each envircnment are comparable with the respective intrinsic relaxation times. the decay of magnetisa-

tion mayest!'11be described by equation, 3.26 FrovUed that the tne values zf Pi and Ti 3-e 'Ill. f: mes by PI, replaced apparent 3ctional populations and apparent -elaxaticn-t!

I1-1.1 The case cf Intermedlate exchangehas been discussed by Zimmer-, an and Brittin. ; mmims W: ne obsened relaxWon bemme s5per
case

The ex-

Qn

Ta')>_W, and ?a

PS A AM

Fb
%ios --------(7b 4-Tb) +

Fa
33 -a

the Cund t,,. he lifetime the 71hase. -I spins w, rere Tb represents -of

In the two-phase ciso numbe: Of special

R: Pa', Fbn Ta' and7b' law' t: -n etermired for a Te , esullu are complex the tve pa:am. eltars, the

A !: the phase. of each resonant equency I! fet,.. Tes Of the Spins 'In eachphase,*a and lb, and Is It these aswhers s, sn, Pramet9: description nuclear exchange 1:aphlcal of the effects mof below. is mmechanis:, presented that the act of transfer is not itself a, relaxation sumed

the apprent Iractional, population of In figure 3.6, the effects of nuclsar exchange con
41

RTb transfer : rate in te: msof a reducednuclear the slow relaing comporent Pal Is shown RTb, reachesa?Pb. Tb/Tb, for a systemin whichFra= Pb, Ta : ! 00.Tb, nd lJoa: 'Job. Uhen proximately3, Pa' tends to unity, and Since Pa+Pb' 1, the apparentfractional population

of the fast relaxing phase'Pb'tends to zero. In other wordsthe ftst relaxing state appears to be emptyingand in general if RTb)2 it Is not possible to detect two separaterelaxation
times experimentally"".

43

populations P'a, P'L

Fa Statewith longrelaution time. la--1221h

0.5
State with short relaxation time M4 19+1
CM

i Ig -3

,- fate of exchange Figure 3.6 Apparent fractional Populations as a function of exchangerate.

Fizure 3.7 shows the effects of the reduced nuclear transfer rate RTb. on the ratio of the apparent spin-spin relaxation times to the true relaxation times. Similar curves are found for spin-lattice times. The most striking feature is that both apparent relaxation

times decrease with increasing exchangerate. the effects on the long relaxation time being more pronouncedat lower exchangerates than for the fast relaxing component. It is possible for Pal. Pbl and Tb' to be close to their true values while Tal is several times smaller than Ta.

Figure 3.7 Theeffect of nuclear exchinge on apparent relaxation times two state s9stem, of a with TV
Th T Ma = A, Pa : Ph, anJ Ta = IBM.

T T

Tal Nucleiexchange hettween states a anab at a rate Ca,anaviceYersa at a rate Ch.

(ITb

44

'a 'function The behaviour oi Ta' and TV over a, range be II 'the of temperatures vi of separate temperature dependences of the correlations times, for phases a and b, modiiied by the eftects at nuclear exchangedescribed above. if nuclear exchangedominates then the relaxation times will decrease with increasing temperature in contradiction to the behaviour predicted by EFPtheory.

Since T1 is usually greater than'TZ for nuclei in absorbed systems It'is

possibleAhat'

it In this T2. case the exchangeis rapid in comparison with T1, and yet slow compared with is possible for two-phase behaviour to be observed for spin-spin relaxation but not for spin-lattice taneously. relaxation. I )) both Pb. T2Zb/ib Pb. Tlb/, simulI (( rb satisfy and since rb may,

In general, numerical' solutions'can be found'tor systems in which discrete phases ex, Identified is Bound water often the transfer changemagnetisation via of n0cleifreezing water. Typically relaxation times of non-freezatie water ,I ample. The TI'minimum for-non-freezable water is typically with non-

SS as used in our ex-

between -IO'C and -SO'C and is

bound ' between the TI the frequency dependent. Room is and"exchange minima, temperature above T1. 'dependence is frequency of free T2, >> 'and why'there a and phases cannot explain why TI

3.4.2

A Distribution of'Coirelation Times.

I In the discrete multiphasemodelit is assumed that a quantity of highly immobilised water determinesthe relaxation behaviourof the bulk throughexchange processes.A variation on BPP'theoryin'contrast with-this is one-in which all the water moleculesAnheterogeneous systeasare consideredto be incontinuouslong-rangeinteraction Atli the substrate.'This 11 -I-I. 1 ... I- ,

45

leadsto a continuum by a distribution of corat motional properties whichmay be-described relation times"""'. that; This distribution may be expressed as a density function J(T) such

0(,r). dr

3.37

rhe relaxation times TI and TZ are then obtained by averaging the EPPexpressions over the distrintion to gi-je;

zTI and

UZ.

T.

O(T) 1'r
Zrz W1,

31+
00

+4

-'r I+

0(,r)

4W, ZlrZ

dT

r. O(T). dT +5r. Td' 0 'f ,,:, 0,,

0ir)dr + W.ZT

r. O(, t)dr_ 3j 1+ 4W, Zlr -: 0' -

I", 4. '3'9

The density function which has been most frequently applied Is that of a log-normai distribution. distribution Here the binding energies of the water molecules are represented by a Gaussian function. (See figure 3.8). The Gaussiandistribution is specified in terms of a

width parameter 6, and a median correlation time rf

i. e.

0(,r)dT

exp (-OIZZ)

3.40

where

In

3.41

Other types of distribution

functions have been proposed to describe the variation

in

binding energies of the molecules in the system under investigation. that are analogous to r* and 0.

All introduce variables -

46

Variations of r*, with temperature are governedby thermal activation have been largely considered;

laws. Two such laws

(a) the Eyring form given by r* , (b) the Arthenius form given by T,

r*.,. exo(E/RT) KT 1. r#.. exp(E/RT)

In times, distril6utlon Figure3.8 The of correlation GaussioO

width Theeffects of increasing heterogeneityin the systemis reflected In an Increasing, inj decrease ect level increase a without notably 0. An in the hydration will parameter ?eif

0 on TII and T2-as a function Figure 3.9 showsthe effect of changing the width parameter to of reciprocal temperature,T, wherethe Eyring form of thermal activation law is. assumed describe the behaviourof ir*. from the. slope of T2 As the heterogeneity increasesthe activation enthalpy as determined. curve decreases.This Is matched by an increase in the ratio TIM at and abovethe temperature of the T1 minimum. Theminimum less pronounced itself becomes as the curves become shallower with increasing 0. _
I'

47

71,72

P2

II in TI an&12 caused by an inuease in the ralve of listrilation of Figure 3.1 Changes ceprelallentimes,T. FD F2) FL.

Ll

lembpel-auLvIl

--

I-,

I,

'.

in In somecircumstances a mayhave a temperature dependence,and this-CauseS the minimum

the T1 curve to be shifted from whereWT*2 0.615a. In addition TI may vary asymmetricallY aocut the, minimum,and the temperature dependence on T2 will to the be increased as compared

case when0 is a constant. In modelswherea temperature dependent j) has been required to

matchexperimentaldata with theory, an inverse relationship has beenused.

Resing"51.defined a cut-off value of 'the correlation time To, abovewhich rigid lattice in which there is a continuous behaviouris observed.On lowering the temperature of a sYstem distribution of correlation times there is a gradual transition to the rigid lattice state, with those moleculeshaving the longest correlation times reaching the rigid lattice cut-off point first. The result is a decrease In the observed phase population of the heterogeneous

and has beenconsequently describedas a phasetransition. To incorporate this effect into the caicuations,for TI and T2 the rangeof integration in equations3.38 and 3.39 are to the cut off value r,. With this modification the modelhas beensuccessful in prechanged dicting the behaviourof TI andT22. in a number of systems.

In systemswherea distribution of correlation times exist, values for VanVleck's second moment. the distribution width parameter 0, and the activation energyof relaxation mechanisms can be theoretically determined.

48

3.4.3 - Cross-Relaxation Effects.

.I-I,,

In addition

to the, transier

of magnetisation

via the exchange of nuclei can result

bet-ween phases.

It has been acknowledged that a transfer tions between nuclei face results

of magnetisation

from magnetic interacacross the interbetween the phases. of magnetisation rate to act as a process of mag-, _,

at the interfacel2l-23).

Coupling by spin interactions magnetisation transfer

in cross-relaxation

effects, including transfer

The mechanisms of magnetisation within

at the interface,

and diffusion relaxation

the bulk phases enable the population

with the greater The driving

magnetic sink for the rest of the population. netisation transfer is the difference in 'spin

force for this

temperatures'

intrindiffe! en't. created by the rates of, the two Popula' are selectively

sic relaxation

rates of the two phases. If the intrinsic then there is no detectable magnetisations. effect

relaxition

tions arc similar, excited to different

unless the populations

Vhenthe populations are eachkept isothermalby rapid diffusion of spin mainetisation, which in liquids takes place via molecularor nuclear diffusion and exchange, and in solids wherethere is sufficiently strong spin-spin Interaction, via spin diffusion, the contribution, from the cross-relaxation effects maybe IncorporatedAnto, the relaxation equationsfor the two populations I and S to give; ', III: 1,

d( I-lo) -z dt

-(R, - RT

I-lo)

+RL

(S-50)

3.42

-2L F]

d(S-So) dt

-[R, + Rj x (S-so) + Rf(1-1o)

3.43

49

ghere lo and So are the 2agnetisation of the two phases at equilibrium,

R, and Re are the

relaxation rates appropriate to isolated I and S spins, R, is the, rate of transfer of-magnetisation between I and 5-spin systems, and F is the ratio of. the spin populations. At equilibrium lo = So/F

The solution of these coupled equations is a two-componentspin-lattice and 3 components.

decay for the I

by data has been incorporating the cross-relaxation effect, analysed cases most where -In the exchangeof magnetisation and a resulting modification to the spin-lattice. relaxation It

time have generally been considered-to be the only consequences of the cross relaxation.

has been demonstrated experimentally, however, that direct spin-spin relaxation is also a

consequence.

3.4.4

Anisotropic Motion.

have In been discussed The section NMR couplings which govern relaxation mechanisms 3.2-1. Theyare vector quantities. In liquids, the constant tumbling motionsof the. spins the local, tield excauses an averagevalue,for their couplings.to be observed.For example depends periencedby=e nucleus as a result of the dipole moment uponthe angle of a second 0 between two the inter-nuclear vector and the external field. have of values one may -It depending on the relative orientation of the two spins, given by

9B -- t/- Ko.(3cosl(O)-I)

3.44

so

whereK, is the dipolar coupling constant. An isolated water molecule would therefore yield a doublet resonance spectrum. In liquid water the local field effects are dynamically averaged to zero, and a single resonance is observed. The randomorientations of the molecules produce no preferred orientations iplitting of the inter-proton vector and so there is no angular dependence or

of the spectrm.

When water molecules interact with molecular system'ssomemotional anisotropy maybe imposed on the water. If there is Macroscopicanisotropy' of the substrate such'as occurs in fibrous and laminar materials, the anisotropy will be can extena throughout the sampleand

observed as a dependence of the magnetic resonance properties on the orientation of the specimenin the magnetic field"I"I".

In the case of dipolar couplings the local fields given by equation3.44 mayhavea non by these dephasing zero time averagevalue. Since the rate of T2 relaxation is determined fields I/TZ2 dependence on the angle 0. exhibits the same

Anisotropic effects are morenoticeable In TZ than TI, since it Is the low-frequencY cOm'the Ie. local If there exists macroscopic fi Ids which havean angular dependence. pon'ents'af If type be anisotropy but no long-rangeorder in the sample obtained. powder spectra will IeImany'domains the scale of'the' anisotropy is such that theI water moleculesexperienc in the time scale of the NMR be there isotropic no obwill experiment will occur, and averaging an servable angular dependence properties. of the magneticresonance

sl

3-5 ''

Ditfusion.

Diftusion may be described as the general transDattation of matter fro& one part of a system to another, such that molecules or ions mix, through normal thermal agitation. there exists two fundamental types of diffusion. sion causedby initial diffusion is 'static' the first being 'dynamic' or 'driven' In wood diffu-

concentration gradients of the diffusing or 'self' diffusion,

species. The second type of

wherebymolecules diffuse with respect to one

another without any changes in the local concentrations, and is a consequence of randommotion.

In this work it is the latter which has been investigated.

The self-diffusion

coefficient

D is defined by Fick's law; in a single of multi-component system where there exists uniform macroscopicconcentrations, the one-dimensional form of Fick's, 21, second law states that;

LP (X,V dt

t) Dx d2P(X, dxz

3.45

whereP(x,t) is the probability that the position of one particular moleculeof the species underconsideration is at a position x at a time t.

In an isotropic medium this equationmaybe extendedto a three dimensional form;

dP (L, t) dt

D del. ZF(r., t)

3.46

In an anisotropic medium the value of D-is often orientationly dependent

52

and the equation 3.46 needs modification. Choosinga rectangular system of co-ordinates (xj, x:, xj) for example, with Di, D2, and D3 as the principle diffusion coefficients, 3.46 becomes; equation

LP(rt) it

DILI P+ dxI

D2LIP + D3dzP
dx22 dxl

3.47-

Solutions to the Diffusion Ejuation.

for found be a Solutions to equation 3.46 describing diffusion in an isotropic medium may variety of initial and boundary conditions"". is Themethod of separation of variables used

the probability density function P(r, t) is expressed as the product of a spatial and whereby three in two, the isotropic function. and temporal For one, medium solutions a an unbounded dimensions are given below;

(a) Onedimensionalcase,-, jx,: x, t)---

(4nDt)'IexpE-(x-x4)2/4Dtl -A. 48

whereP(x,: x, t) is the probability that a moleculeinitially In time t.,

at x. will diffuse to position x

(b) Twodimensionalcase, where = (x, y)

P(rl.:. rt)

= (4nDt)*Iexp[-(r-_rL)1/4Dtl 3.49

(c) Threedimensionalcase, F(rL: rt) . !', - t' .I(X'Y'Z) wherer

z (4xDt)*3'2exp1-(r-ra)I/4DtI 3.50 11 11 I ' 1

53

The one dimensional solution describes diffusion

in a one dimensional medium.and also

diffusion in any specified direction within an isotropic plane (i. e. two dimensions), or the-shape describes of-the probabilisotropic Figure 3.10 dimensional three medium. within a ity density iunction for any specified direction at three successive values of time t.

displacement r2 of molecules diffusing The.mean-square calculated irom the density functions ;7z +0 rz. P(P,: Lt)

in an Isotropic media is

dr

3.51

Xx:

t) X(@),

L,

".

IIXIIIx,

"

''-

X(O)

rioure3ja Theshare or the probability aensity runction P( x: x(S), t t3, tit ta, three ti'"s to', successijo top a moleculeinitially at X(Q)

For the one dimensionalcase this yields xz-.

In a plane;

71 71

4Dt,

rz

X, + Y,

3.52

and in three dimensions.

61)t,-

(71 X1

+-Y2 + zz').

Solutions to Fick's secondlaw for anisotrople mediadescribedby equation 3.47 maybe obtained in a similar manner to that employed In the isotropic case. For unrestricted diffuSion in a direction parallel to one of the principle axes xi, x2, or xj, the probability density function takes the form;

54

p(l., : X, t) ,

(4,tD, 0- 1expl-(x, -x. PAR], i

!: l, 2,3

3.54

which is equivalent to equation 3.48. For the more general three-dimensional case;

rX-

(4nD, D3D3t' )-I exp I L

1-1102

(XI-XI'O');

Ott

-, 4D2t

4DIt

3.55
diffusion along for found then 3.51, T11-je displacements are using equation and meansquare one at the main axes this yields xI, Dit, where i

5.3

Self-Diffusion Measurementslb-il. The theory of Spin-Echo

the the when echo of field attenuate amplitude In the Hahn gradients spin-echoexperiment diffusion. translational to the resonantnuclei are susceptible Consequently pulsed NMR Twomethods

lends itself to the measurement of translational self-diffusion coefficients. during the course of this work. wereemployed

Q)

TheSteadyField Gradient (SFG)technique.

the 90* and 180* r. f., pulses in the spin-echoexperiment,the spins dephase ' '.' Between as a by the individual nuclei. result of differences in the local magneticfields experienced Thesedifferences are caused by Inhomogenelties in the applied magneticfield, and by fields by resonantand non-resonant produced nuclei in the vicinity. Inverting the population of

resonantspins by application of the 180' r. f. pulse causesthe dephasing angle of spins due to variations in the field experienced by the nuclei to be reversedtime 2T these effects are

55

cancelled and the height of the spin echo is theoretically

therefore a function of the spin-

spin relaxation rate due to interactions with other resonant nuclei, the local field due to the other resonant nuclei also having been'rever'sedby the 180' pulse.

If howeverthe molecules diffuse during the time 2, r, then the average fields experienced betweenthe'dephasing and rephasffig priods will no longer cancel, resIting in an-attenuation of the spin echo height given by;

my,(2,r)
Where -j(2-T) is the phase shift

2 MO CCSI'1(2T)l

3.556
their to due spins dlffu-

in radians experienced by individual

sion in time ZT, and My"(2,r) = Mo In the'absence of diffusion. term sil Inifie s an ensemble average. of Cos re-written as; over all the nuclei

The line above the COSIK2T) in the'sample. Ns may be

0 My'(2,r) di (1) Mo cos(f) p = 3.57

" I I is the wherep(j) probability density function describedby the phaseshifts due to diffu-

Sion.

IeId I unbound in a. linear field gradient, spins undergoing diffusion experiencephaseshifts for which the probability density function is Gaussian.i. e.

P(P

exp

'12

3.58

where

is. the mean averagesquare, phaseshift.

Substi'tuting.this into equation 3.57,,

My,, (2, T)

r Moexp

56

rhe 3ppiication to lifterences Jitrusion

'at

3 known field

gradient G 31lows,. the reduction experienced by the nuclei at the gradient.

in height

to be related to the self

in the magnetic fields

and therefore

-cefticient ,.

D in the direction

The phase shift direction is;

for a spin at time 2T in a continuous field

gradient applied

in the z

r
gamma
L

Zr
z(t) dt P ZM dt

3.60

-.nere nt)

is the displacement of the spins in the : direction

in time t, and the negative For un-

the due is to phase shift two tet-ieen the integrals iign diftusion, oounded
vient suostitution

following the 180' pulse.

its 3.60, subse12%. from Evaluation and (, Gaussian. is of equation z(t)

into equation I. E9 leads to", " rA

Ln

(G) AtG: O)

Z iamma2GZD-TI

3.61

-1

11:

I,

11

'where AtG) is the amplitude of the echo in field

gradient

G, the the gradient. of magnitude pulses, 'r, or ariation of either the spacingbetween via this relationship. allows the diftusion coefficient to be determined

rhere are two main limitations in using this technique. It requires the application ot short duration, high power901 and 160' r. f. pulses to ensurethat eachnucleus Is rotated an because Furthermoret equal amountin the presence field of the wide gradient. of a constant rangeof precessional frequencies inducedby the field gradient, the spin echo is itself of short duration, and to detect it a wide receiver bandwidthIs required, causinga reduction in the signal to noise ratio.

57

fA) Steadq

Field

Grbadient Experiment, 18010pulse

900 pulse

(B) Fulsea Fiebl GraAient Experiment,

Two pulsed fiela graclients applie(l.

tj

tl+ 6Tt1

ZT

Figure 3.11 The SFGancl PFC spin-echo Aiffusion experiments.

7hp

s,i FaiJ]:

adierit

FFG

iclue. -In

i--u! ties

ncoljn--red

3 3F, 2xper ment can 3e 3rgel'i the in ontinuous

'See figure

overcome

wo '. ul -:eS

,Dr , ie i i i 3j ia nt 3Ti Is ej is iovosed to 3

pliEc et-4een the 10' 3nd 1301 r. t. -: oulses.

f ie J-, r adient.
3 '1:. .. it

The ; rSt OUISe 13 3p3 phase shift ! or

induces

fte : csonant nucleus irii: n is determined by its position In ne field.

The second pulse of

f ied beiore is 180' ifst ne j iowino,, ol Oe ine puise, 3ppl Q. uai : ntEnsity and mration to ,?
rormaion ot the spin cno. and at a time HLTA after the first
oriase -rjjjts produced by the
ilrSt Pulse He

field

4t3dient pUlSe. 71he

reversedtor ail Ion-oijtusing nuclei.

3 xperiment, n the case 3i the PF,

equ3tiOn

11.60

O@C"ClIeS

rtIt, gamma i Ltt ztt) dt

+DELTA+ z(t) dt +bELTA

3.62

whIM upon solvirie jor j2 :,, leacs to, ', I

58

FI

Ln

A(G) A(G--O)

gammaIG21)SI(DELTA-913) 3.63

Variation,of either G, S, or DELTA can therefore lead to the determinationof D. Using the this technique, 'r. large,field inhomogeneities. t. pulses -may be aPPlieaand the ! pin echoobservedin the absence Ot

Restricted and Anisotropic Diffusion.

Theexpressionsquotedpreviously for the spin-echoattenuation in SFG experiandPFG mentsare only applicable if the diffusion In the direction of the field gradient is physithe translational motion of moleculesis In many biological systems cally unbounded. diffusion. to barriers restricted by the presence of

A suitable criterion for unrestricted diffusion is that the r, m.s. distances travelled by the moleculesduring the time over whichdiffusion Is observedshould be substantially less than the separation of the barriers confining them. If this cndltlon is not'met'then the the amplitudeof the spin-echowill depend dimensionsupon ofthe confinement.

In PFG experimentsthe time during which the effects of diffusion contribute to the attenuation of the spin-echoIs conveniently describedby the Interval )> 9, then the condition for unbounded diffusion maybe expressed as (Delta + 9). If Delta

Delta'(( al/D

ldherea is the'se'parat'ion of - the barriers

59

In the case of restricted the true self-diffusion

diffusion

the application of equation 3.63. no longer yields diffusion coeffl-

coefficient,

but maybe used to define an effective

cient D'. The dependence of Don the'experimentally adjustable parameter Delta often enables intrinsic diffusion coefficients to be measured and the'nature of barriers to be

investig3ted"2031. 0 Tanner'311 has considered the ef-iects on the spin-echo attenuation of molecules diffusing betweentwo infinite-, "paraliel irmeable' barriers, and'imp separated by a distance a; 1

A(G) A(G--O)

exp(-gammazSZG-Dsinz(9)).21

cosxcI +

2. Z expE-n2n2D!all. ll-(-I)Ilcos(, tcl)I [I-(nicl)212 nz-1

where cl

gamma. aG-, cos(Q) x

between field the the 0 is angle and

3.64 gradient and the normal to

the barrier.

This expressionis valid providedthat Delta >) S. but gives an approximate representation evenwhenthis condition doesnot hold. Theequation reducesto the usual form for the PFG 0z xJ2, Le. the diffusion is parailel to the barriers. experimentwhen

D', Figure 3.12 shows diffusion the form of the measured as the incoefficient apparent terval of observationDelta, is increased. and when by the spin-echoattenuation Is governed equation3.50, using 0 for the case when 0. In contrast, the behaviourof D' is also shown

the barriers are permeable to some degree. Thediagramsillustrates that for larle values of delta, DI is reducedto zero whenthe barriers are impermeable, but approaches an asymptotic value DI., whenthe barriers maybe crossed.

'111

Do
Do

2A

al

-\ 1

Dal 2 D) Dal at 2.%

as 2 Figure 3.12 Restrictocl aitfusion L9 Permiable ana-impopmtalle membinnof Unction of aelfa.

by these trom derived plots be A , ailie tor the characteristic barrier seFarationa'. -,an where the diffusion Delta' point time the at usiq

/2 (D'O_, + Dlm),

by Tanneris given by: The-haracteristic length proposed

Delta'll [2(D'o - Dlm).

3.66

direction in the the NME to the of translational properties motional measurements relate Spatial field gradient G. By changingthe orientation of G, the apparentself-ditfusion coetticient D' may be tested for anisotropy.

It was found for example that for water in keratin fibres'"' 2: 1 between the diftusivi I t tes parallel

there existed a ratio of axis. The angular

and perpendicular to the fibrillar coefficient

dependence of the observed apparent diffusion

can be written as;

61

D'

=A

D4sin-'(0) + D. Lcos2 (9) 1

3.67

where i is the angle betweenthe field gradient and the normal to the fibrillar

axis of the

D(O)/D(g)

or

jr,

V2-

Degrees,0

aepenaence Fi7ure313 Orientationai of the aiffusion coefficientof water in keratin fibres. D(O) = cosz0Djj+ c()sz(9@-0)Dj keratin. andD andDJto the are the diffusion coefficients parallel andperpendicular fibrillar axis. (Fi; ure 3.13)

Thephysicalsignificanceof an observed valuefor the diffusion coefficient D' for the be in heterogeneous molecules moving requirescareful interpretation, since it may systems that D' is the Intrinsic valueof the coefficient, or that D' reflects the geometry of, and interactions with, the biological system in which the molecules reside.

References

1. 2. 3.

' "ThePrinci ,pies of Nuclear Abragam A. Press,Oxford. (1961) Magnetism. " Clarendon Slichter'C.P. 'Principles of Magnetic * Harper Resonance. New York. (1963) andRow, Farrar T.C., Becke*r E.D. "PulseandFourierTransform NMR.Introductionto theory " Academic andmethods. Press,New York (1971) Bloembergen " N., Purcell E.M., Pound 'V. R. Phys.Rev.Vol 73, p679. (1948)

62

Belton F. S., Ratcliffe

in Biological tissues. " R.G. WIR and Camparmentation

Frog.

in NMR Spectroscopy. Vol 17, pp 241-279.1 (1985) Steiskal, E.O., Tanner J. E. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Tanner J. E.. Stejskal, Bloch F. Hahn E.L. J. Chem.Phys. 'jol 42. p 2^88. ( 1965155

Rev. Sci. Instrum. Vol 36,. p 1086. (1965) J. Chem.Phys. Vol, 49, p 1768. (1969) -

E.O., Tanner J. E.

Fhys. Rev. Vol 70, P 460. (1946) Phys. Rev. Vol 80, P 5580. (1950)

Carr H.Y., Purcell E.M. Phys. Rev. Vol 94, p 630. Meiboom S., Gill D. Rev. Sci. Instrum.

(1954)

Vol 29, P 688. (1958) (1954)

Kubo and Tomita J. Physic. Sao. Japan. Vol 73. p U& Zimmerman J. R., Britten V. E. Resing H.A. Resing H.A.

J. Fhys. Chem. Vol 61, p 13& 28- (1957) (1965)

J. Chem.Phys. Vol 43, p 669.

J. Chem.Phys. Vol SO. p 186. (1976) (1961)

goessner D.E. J. Cheri. Phys. Vol 35, pl. Odailma A.

Progr. Theoret. Phys. (Kyoto) Suppl. Vol 10, p 142. (1959)

Miyake, J. Polym. Science. Vol 28, p 477. (1958) McCall, Douglass and Anderson, 1. Chem.Physics. Solomon1. Fhys. Rev. Vol 99, p 559. (1955) Vol 30, p 1272. (1959)

Campbell I. D., FreemanR. Kalk A.. BerendsonH.J. C. Brooks A.A. et al.

J. Main. Reson. Vol 11, p 143. (1973) J. Magn. Reson. Vol 24, p 343. (1976)

J. Chem.Phys. Vol 49, p 1571. (1968)

25. 26. 27. 28. 29. 3o. 31.

EdzesH.T., SamulskiE.T., Nature (London) Vol Z65, p 521. (1977) Woessner D.E. Mal. Phys. Vol 34, p 899. (1977) Mal. Phys. Vol 33, p 857. (1977)

Ahmed S.B., PackerK.J., Ramsden J. M.

FungB.M. Science. Vol 190, p 800. (1975) Fick A. Fogg. Ann. Vol 94, p 59. (1855) CrankJ. "The mathematics of diffusion. ' Clarendon Press, Oxford. (1956)

McCallD.W., Douglass D.C., Anderson E.W. Ber. Bunsenges. Physik. Chem. Vol 67, p

63

336. (1963) 32. 33.

d4.

Robertson E. C.H. Neuman

Phys. R ftev. Vol 151. p Z'64. (1966) J. Chem. Phys. Vol 60, p 4508. (1974) J. Chem. Phys. Vol 49, p 1768. (1968) J. Poly. 11.Sci. Symp. Vol 49, p 43. (1975)

TannerJ. E., Steiskal E.O. LynchL. J., 'JebsterD.S.

35.

I:

64

CHAPTER THENMR OF IJATER INBI OLOG i CALSYSTEMS

-II

4.1 Introduction.

-II.

I,

I--,

I.

-- 1 II

Sialojicai

Tissues are complicated cellular

systems in wnich there is a high degree ot ot water and other metabolites in these

comDartmentation and heterogeneity.

The distribution

systems 3monestdifferent

environments leacs to a numoerof populations vith different

R NM.

properties and cnaracteristics. teen intracelluiar

At an elementary levej it may be possible to distinguish bemetabolites, or even between one intracellular component may,itself com-

and extra-cellular

ponent ano another, since the intracellular, seieral ; oculations or intracellular

be partitioned between

environments. In leaf tissue for examplea-distinction, space,, the cytosol-, tne 'Jacuole

MiZht be irawn Detweenpopulations existing in extra-celiular and the -cmaropiast, inere the only phvsical partiticninq ternal and external components"'. internal partitioning

in this system is between the indoes not necessarily correspond to

the membrane-bound subcellular compart2entation that is observable by microscopy.

Theability to distinguish one population from another is determined by the resolving to experimental power of, the-technique, which is subject and Instrumental considerations. * In (imaging) experiments spin mapping for example wherethe heterogeneity is on a cellular level, spatial resolution is important in order to distinguish one type of cell from another.

In most biological systemsthe temporalresolution is also an important concernas different populations are often in exchange two populawith one another. A distinction between Lions can only be made If on the time scale of the experimentthe exchange rate is slow, or if the dimensions of the compartments are such that the different populations are established
without interference by diffusion.

65

The.NMR, int IIuI I race an at, eveI, is based approacn to,,water,,in heterogeneous on systems upon the usual resonance properties ot signal intensity, whicn reflect frequency. and relaxation behaviour.

the aynamic properties ofthe water molecules. The state of water in biological of biological : acremolecules is, ii; nift,, 3ntlY different to the

tissues and-in the vicinity

spectroscopy is a state of water in solutions of simple molecules and, in Dure-%ater. NMR powerful tlecnnique for studying in detail the structure. mobilit'l, and extent of ordering of

water molecules in various biological systems. If these properties vary from one population ot iater to a, nother then it should in principle be possible to distinguish betweenthem.

-This

chapter begins with a description of the characteristics

folhyaration water, of systems. The application

lowed by 3 general review covering NMR stuaies of water in cellular

is then interacting walls NVIR the cell to with cellulose and,wood of study of water molecules reviewed in detail.

4. ' Hiaration Water.

'Jater and biological These interactions

interact with one another via electrostatic macromolecules

forces..

have two main sources;

(1) (ii)

The high dipole moment of, the water molecule. Extensive hydrogen bonding arising from the large numberof potentially proton donor and proton acceptor sites-potentially available

available an the macromolecule,

66

These we3k macromolecular-water motions of the water molecules, extent ot ordering and

interactions rise

restrict

the rotational'

and translational and

-give

to a hydration

layer in which the mobility from the iast

of the water molecules are distinctly

different

and random macro-

motions ot water molecules in the bulk. These effects MOleQUIeSLnteract with %4ater in both solutions

ccme into play when biological systems.

and biological

in the bulk phase a water molecule will ditiuse.

rotate and exchangeprotons with characin any molecule oi a water phases exchange

teflStiC fate$ Td. Tr and ie respectively. Moreover, the lijetime one zhaSe: an be char3cterised by a time 'W". iith one another.

since molecules of difierent

The

of restricted Jegree

is layer hydration motion of water molecules throughout the be longer these no can molecules and the aynamicmodesof 0i correlation

ncn-initorm ana anisotropic,

distribution but time require a dE:,:rlbea completely by a sinije correlation times. Diffusion in these layers is also anisotropic.

and the hydrogen bonaing interactions

betieen iater molecules and several proton donor and acceptor groups or lacromolecules live layer, proton hydration the in promotes which hydrogen to bonded rise a continuous path

in the that diffusion is There processes transter along the macromolecular evidence chain. inprotons the of involve hydration layer, bond, while exchange the breakingof one hydrogen vaives the breaking or two"O', if determine to which, be possible and it should therefore

dominates the NMR either of these two processes characteristics or a system.

hydration bulk the below the freezing At temperatures water molecules water iell point of from the NMR for this comes signals of water in remainunfrozen and mobile. Evidence btaiogicaj samples from between arise which only a fraction at temperatures and -604C, -S'C ot the total water content. This fraction is proportional to the macromolecular concentratiOn. -ihich indicates

67

that the water NMRsignals i. e."the hydration layer.,

in frozen samples arise from water associated with macromolecules.

Ti-O types at water molecules contribute water and specific 'irrotationally mediate layer-of tional binding sites

to the hydration layer.

The interactions

between or

an macromolecular chains result in'a specific

hydration

bound' layer"' Irotationally

Between this and the oulk phase there exists an inter.bound'-water molecules arising from the spatial and orientanon-specific hydration'layer contributes to

mismatch between the two regions. -This freezing traction

the non-treeZing

and water molecules residing in this layer-have rates Of

the for than lower for than molecules In the bulk, but, greater molecular nations which are specifically bound layer. AFigure 4.1)

-hiqbw crap" 3.! bauma

ec r&,, v
c. bi),wria watera Fitare 41 Thetht*e ImIs 6f ill wide" "c"'a "lkcl"'locttll'

in liquid Under the condition of rapid exchangebetween the hydration and bulk water is by given the solutions of macromolecules, observed relaxation rate

QIT)obs

Xf(I/Tf)

XMIM)

4.1

68

Although the traction Xh. the term MUM

of free water Xf is muchgreater than the fraction

of hydration water be-

makesan important contribution, to (I.Mobs., Since Q/Th) ),>, (I/Tf) motion of water moiecules in the hydration layer.

cause of the-considerably restricted

rhere have been a numter of studies carried out to investigate the phenomena occurring hydrated systems are cycled above and below Z73K, in order to asses the of 'wnen -quantity 'bound' dater based on the definition that its properties are so modified 3s. to render, it

have the IMIR Most by amountof non-treezable water methods assessmentsof non-freezzable. cnosen the intensity of, the mobile water signal at a specific bein"i 27.31K below as temperature

proportional to the amountofbound water. the aosorFtion signal, or the initial

One can use either the integrated intensity Of

intensity of the transverse relaxation of the mobile

parameter""". onent, as a measurement water com;

in most cases the quantIty of maoile water betow any treezing point has been found to be temperature oependent'11*121. It is questionable that the non-treezable quantity and hydr3, tion water are the same. Conflictinj calorimetric from the comparison betweenNMR and comes evidence

indilagen, NMR 131 denaturation of col studies the where studies carried out on , studies indicate an in-,
related, t.o, at

We a decrease in the quantity of bound water while calorimetric

cre3se. Resing'"I 3n identifiable asserted that non-ireezable

water in muscle is not necessarily alters

phase at room temperature since the act of freezing

the structure

the muscle.,

In some the identifiable hydrato be a close correlation between systemsthere does, seem tion layer and-the non-freezablecomponent"'03,161, For,a study relationship sys_ofthis temsthat can tolerate a wide range-in their level ofhydration are useful, e. g. wool-keratin tibres"". On this systemthree independent to determinethe quanweremade measurements

titY of hydration water within the system;

69

(D (ii)

The observation or signal intensity change across the freezing event. -, The resolution of the mobile water signal above the freezing event into two components and the measurement of the intensity of the non-interacting component.

(iii)

The measurement 727 beiow the freezing event as a function of water content.

In the latter, the spin-spin 'relaxation time increases with water content-intil is reacnea at the fibre saturation point. With C310riletfiC

a plateau

Results of these studies 'Jere generally consistent

and quantitative techniques,", to,. There has been goodcorrelation using

other MR technijues sucn as measuringthe height of, the C.W absorption peak as a function of iater content using high power r. f. fields, in the peak represents the Maxiwhere a maximum water leads to enhanced

hydration point because the further addition of non-interacting mum r. t. saturation due to its long TZ"".

In collagen the 2H coupling constant has been used

to issess the quantity of non-freezable water.

In jenerai there are tour phenomena otservable in temperature scanning experiments'12,20-221:

A Sharp transition cooling

in the intensity

of. the signal

and in relaxation as a result

times during of the stepwise

for samples with higher water contents,, occurring

changes in the amount and nature of-mobile .,uper-cooling.

water in the system. seen to freeze at temperatures

The water in the sample isoften

telow that oi pure water. (ili) UV) A decrease in the mobile water signal Thermal hysteresis at temperatures below the freezing event.

in the amount and nature of thp mobile traction event.

as the-tempera-

ture is cYclad above and below the freezing

70

( i) 3nd t 11) are consistent with the presence at non-interacting water. The temperature dependence 3f the MIR signals arising from the aobile water content have been explained using, the folluing effects: 4

An '3pp3rent the motional

phase tr3nsition'

effect.

whereby molecules progressively lattice state-as a result of

transfer

from de-

_narrowing

state to the

-rigid

_thermal

apillary t.,

effects.

'Jater contained in pores having a range of different

si: es Uill

i)rcgressively freeze according to the Helium equation, 24 251

Free-'ing point a I/Pore Diameter.

the depending freezing lowered is on the point of water A non-specitic effect where the Ib, the subwith structure caused to interaction the water alegfee ol pertuftati-3n $trite ataterLal.

ro explain the NMR upona distribution of behaviour, all of the aboveare dependent states for the interacting leading to a distribution of correlation times. A schematic

by Resing"'I carproposed representation is given in tigure 4.2. Theapparentphasechange from the to a transfer of the moststrongly perturbed or least iontle molecules responds mobile tegions to the lattice state. Thecapillary and non-specific effects correspondto
the least modified water molecules passing through a changeOT Fh3se""-

Calorimetric studies show a diffuse endotherm extending over temperature intervals 273K for samples with low water concentrations 1211

below

As the water content is increased the

endothermsconcentrate close to Z7 3K. The endothermscorrespond to a gradual fusing of some water that is modified by interactions with the substrate to the extent that the freezing

is OePressed P0111t but not Prevented. In terms of a distribution of correlation times this

71

DistJiion o corielitiori ti, wc (J(fl)

'Og, P" houndury pmton

mobile watel% liqW-lattics

-,, F, z. 10

A-I

it

Oe of flon-interioting 273X. Portion a 42(a) lemperature is ahove, water urefrozen. small interactini waterfalls into rigia-tattice state.,

Harrintepicting, frmen water

correlation timesQ0 1)

httice/mobile hwxjal%g proton Interactinir Water

(T) time Loqr Cerrelatian 71 19A Is fraotion troren. of water greater MW-intfracting 42 G) le"Perature It low273X. .4 ralls into interacting rivii-lattioe state. water

tranis there that temperatures lower gradual a so to in the distribution relates to a shift The lattice the narand states. mobile rigid the populationsof sition of moleculesbetween destrucfrom is a temperature the endotherm observedresults rowingof the rangeover which turinj of the substrate leading to a moreopenstructure and a narrowerdistribution of correlation times.

Studies on wool above273Kindicate that the mobile water population increasesabovethe freezing event with temperature and this is also attributed to the apparentapparent phase transition effectill, Z7,211 as water moleculesprogressively attain sufficient mobility to , the rigid lattice limit and therefore to contribute to the population weightedaverage exceed relaxation process instead of the solid.

72

Ramirez, ", Interpreted

the gradual disappearanceof the mobile water an cooling in terms freezing -

of a classical binary solution model. Lowering the temperature below the initial

event causesthe formation of -ice., allowing the resultant more concentrated solution to remain liquid.

As the temperature is further reduced-the quantity of ice increases and the the eutectic tempera-

concentration of the remaining solution becomes more concentrated until ture is reached and the whoie solution freezes.

Hysteresis effects have been explained in a numberof. ways. Morarju and Mills""

ob-

served hysteresis in both spin-spin relaxation behaviour and in the amount, of mobile water adsorbedon silica gel. 'they explained this in terms of a physical shift betweenfrozen and unfrozen, states. in the phase boundPearsonand-

aryrelating to the gradual transition

Derbyshire"11 explained hysteresis effects1n super-heating for water-structurally

terms of an overlap of super-cooling and Barnes"" suggested that the act

ordered within pores.

of water freezing disturbs the sorption equilibrium. and therefore that unfrozen water moves awayfrom the pores to the surface of the substrate where it freezes as a result of ex- ,, periencing a-lower pressure. On re-heatingAhe specimenthe water would then becomeliquid

at'a higher temperature.-, In addition to, these effects one should also consider the pos-, sibility of structural Itself the substrate changes'In as a result of freezing.,

4.3 NMR Relaxation Studies of WaterIn Protons and Biological Systems., II1 1-1 1, I-1 11 11" -I

4.3.1, Possible origins of, non-exponential relaxation behaviour.

73

TheNHIR properties. of water in biological systems have ceen studied extensively in time domainexper, iments becauseof the ease with whic, h sucn experiments can ne carried out and the dramatic differences in NMR sometimes provide inforProperties observed. Such measurements mation atout. the dynamicsof water molecules within the system.. Changesin chemical shifts are proportionally by the increasea peak width in the NMR and are often masked signal.

The most useful NMR parameters in the study, of water are the relaxation times TI, T2 and T_p from which,one can derive correlation t'im, es and diffusion ccoeificients. Vater relaxation _

6 has been studied using, 'H, 2h and 'IQ nucl, ei.,. The use a,, 'K nuclei has,been limited because at their radioactive nature. 1H nuclei are popular because at, their high natural abundance.

by Proton signals are complicated by signals from other components t Chemias as wel, 'howeer, cal exchany between the protons oi vater, and labile protons of proteins. etc., and by cross-reiaxation effects.

from Deuterium does but suffer : toss-relaxation effects or masking not undergoes excnange by signalsirom other components.1_70 suffers_fromnoneof these difficulties intrinsically 'a but has

I -- 5/2. of its nuclear spin juantumnumber multi-exponential decaybecause

high to Ua de; decay 170 a curves ree single nuclear component using single are -Spin-spin ', is multi-expenenti; environment exists, but "j, il,, e5peciallyin caseswhereJ(O) )>, J(Vo). which is generally the case in cellular. systems. Both.110 and -'H-often require some enrichmentdue to their low natural abundance.

Thenature and significance of water relaxation

been, have, reviewedby a,number 'behaviour

and,the general features of,,the.behaviourwhich are described in chapter 3, maybe summarised as:

The spin-lattice relaxation time is greater than the spin-spin relaxation, time, i. e. ' TDT2.

74

Relaxation rates of water in biological systems are faster than in the bulk. (M) temperatures below the leaves at LI temperature the even signal? owering of a residual ireezing point of the bulk. i 11 The magnetisaticn decay curves are often :aulti-exFonential.

spin-lattice ieature observed in proton and ZH relaxation is single component A common multi'TO but exhibits SPin-spin relaxation processes. multi-component relaxation.

features are These, both in behaviour and spin-spin spin-lattice relaxation. exponential
of based between compartments or on exchangeprocesses phases usually explained using models for TZ Of time time the TI, the but scale being scale of on on slow exchange water. I 'rapid much at, 110 however, proceeds in deuterons. nuclei the where relaxation case of protons and faster, rates for both. T1 and TZ, the exchangerate is relatively slow.

forward: been have two put and alternatives This mechanism is not universallY acceptedt

Q)

Based an diffusion-dominatedProPettiesBasedon residual static effects.

I .., III

these the but of origin Both again depend uponsites of varying.relaxation characteristics, differ. bulk the their characteristics the relaxation sites and nature of effect on

biological In the first case it is assumed within that proteins and substrate materials systems act'as patent magneticsinks for the relaxation of water molecules. If a water moleculeis sufficiently close to the surface of the substrate so,that within the lifetime of its ownintrinsic, relaxation it can,samplethe sink, then'its relaxation,will be strongly effected, otherwise it will relax at its ownintrinsic rate. Using this model,a variety of relaxation, times are to. be expectedfor water in cellular systems which depend uponthe size and shapeof the cells.

75
4

by Brownstein and The mathematical approach to such a model has been developed. tar r, is due heterogeneous to in that I systems proton re axati on of wat2r, who assumed

in diffuse must to molecules wnich magnetic sinks contained within active volumesor surfaces, order to relax. The general solution to this problem has the form:

E l(n) exp(-t/Tin) . n=O

4.2

i. e. the observed magnetisation M(t) at time t is the sua of an infinite ponent3. l(n) is the intensity

numberof com-

(j--1,2. p). Tjn, the time nth component of with relaxation and

Calculations based on models containing only active surfaces showedthat the intensities

the system dimensions the the of the functions times geometry of system, of relaxation are

the (M), resonant the diffusion of and coefficients (a), the strength of the magneticsinks by the three parameter identified 0). Brownstein regions characterised Tarr nuclei, and Ma/D;

Q)

Ma/D< 1. This is the region of fast diffusion, whereall the nuclei samplethe sink. in the discrete multi-phase It is equivalent to the region of fast exchange

intenhas 4.1 term first In significant this of equation any model. region only the sity, i. e. I(n The is cor0) ZZ predicted. 0, and single exponential relaxation

V is the V/Ms, To is by decay where time lowest for the given mode responding sample and s is the active surface area. volume I Ma/D ( 10. In this region a larger proportion of the resonantnuclei samplethe

magneticsinks due to their intermediate diffusion rates. The lowest mode still dominates but n sity. 1.2 modes nowcontribute a few percent to the total relaii've inten-

*for The decaytime of the lowest mode is greater than waspredicted the region

of fast diffusion, and Is about one or two orders of magnitude greater than the decay times of the first two higher modes.

76

10 ( Mail).

In this region few nuclei sample the magnetic sink as a result of their rates. The majority of the relative intensity is still in the lowest for

slow diffusion

intensities the' 60%-80%, depending the and the system, upon geometry of' mode; terms with n)Z are still negligible. The relaxation times of the n

0,1,2 com-

the decay times the of in the Intermediate than region, with ponents-are greater first two higher modesnowbeing only about one order of magnitude less than that of

the lo'we'stmode.

in for water It was suggested that multi-exponential relaxation may be observed from maga 30um In the I the to away heterogeneoussystems'when range of water molecules are netic sink, "'. second. 10' I between distance To di f fuse this a water molecule"Wouldrequite andI

biological of real Thesuccessof this model'in predicting relaxation characteristics Ic, I Tarr that the acBrownstein ould model sho and wed is In their systems mixed. case study, it in behaviour muscle observed rat lastrocnemius count well for the transverse relaxation in Since'thls'value'was the cells 88P. expected be to the diameterof the, about wasassumed this then due biopolymers it the However, sinks are range the modelwasthought to hold.' fibrils the between longer holds because distance the modelno 2-3pm, is and nearer of muscle

I It fibrfls th"e between Moreav he the distance Ibe'tween, 10-30nm sp'aic er the' f Ibr'es is around the To fs'fifl-ed wit'hsarcoplasmic determine only whether reticulum and mitochondria. ,, d-iole in thes'esystems,tests werecarrie out on giant squid embrane mI-I`I-I, plais a relaxing , in'dia zeter and contain very little axont311 The'cells of giant squid axonare aroundSOOpm inter naI str ucture Iements"carriedout an extrudedaxoncontaining virtually no Measur
TI

membrane showthat the relaxation times of the internal medium are independent of the boundary, which suggested Is caused that the multi-exponential relaxation observed by protein-' 'and "0. " ,n, inter*actions, I " t water t . t For"such the'descriptlo inte no membra ne wa er system's ractio ns'. .I ' relaxation'c'haracteristics in terms of tIhe diffu .sion'M untenab-l'e'. ode appears

77

Systemsto which it maybe applicable are those-in -inich compartmentsare of the order of micrometersacross, with little or no internal structure, i. e. only localised relaxation Even in systems where the

sinks such as cell walls such as the vacuoles of plants.

Brownstein and Tarr model is not expected to work. there is support for the role of diffusion in ordered systeiss's".

As a, second alternative very different

to the discrete multi-phase model, Fung-andPuon 1,40, suggested a based on the. f,act that

origin of multi-exponential transverse relaxation,

SH OH-3nd Mi. protons and deuterons are in constant chemical exchange, with species, such as on the protein surface. They point out that it is customary to, analyse the-decay interms of

be fraction can to a type of proton. -Four Iroups sumsof, exponentiais and then assign each identified from their characteristic relaxation times:

(D
(ID

Z Z44s. This signal originates from protons in proteins and other macromolecules. MTZ z 0.4ms, madeup from irrotationally bound water and highly mobile protons in

large molecules. T2 ZZ 40-60ms. Typically this is assignedto the Intracellular water. 12 Z' 150-400ms. Extra-cellular water.

in their work, Fungand Puonstudied T2 decaysin skeletal muscleand found that after a modelfor period of longer than IOms the decaycurve could be describedby a two component both proton and deuteronspin-echodecays,which they identified with components (M) and

Qv) 3bove. However, the assignment of QiD, to intracellular and Uv)to extra-cellula"'r, water wasquestionedon the groundsthat:

(a)

In deuteratedsamples,the slowly decayingpart zonstituted a larger fraction than for the proton decay, whereas It the long component weredue to extra-cellular water it should haveaccounted for the same fraction as in the non-deuterated samples.

78

(b)

In glycerinated muscles, whereall the membranes were disrupted, the proton spinecnoes were similar to those from intact muscles.

(c) (d)

Non-exponentiality increased as rigor mortis set in. The lifetime teristics of intracellular for the in characaccount water exchangewould not

of the two resolvable components observed.

-n-g I Fu Basedon these observations, and Puon rejected the idea that the slowly relaxing component was due to Exi'ra-celllar the TN's proposed'that and the ,unusually short water, and

I -II I in hydrogen due protein to fun exchaqe non expon ential character are with ctio na groups filaments. splitting"or NH2and GHgroups are the c3use of a dipolar The extremely slow motions of SH, -3 .1 . in ,0, I I traction In inor the the tar-split ting deuterium, If case of m and quadroup I I spin- spin dipolar or quadrupolar

to due '!, 'J ith has chemical shift, excnangewI pr otons a shift interaction. then non'-expone-ntialbehaviour is observed if

W/10 Z 1/(TZb + TO

109W-

4.3

gher4 rb is the I if etime'of the' 'protonson the proteins, ' and T2b Is the spin-spin relaxation two protons Thedipole interaction between time of protons attached to' the macromolecules. le separatedby distance r, and'at ang is g6en'by

3aammal. hz.(1-3 cos-I/Zrz)

4.4

but this value is reducedby the rotation of the functional groups. Theauthors estimated that, for the NHI groups, 94 z 10" - 101and T2b 10"s, and from studies an the protonation of methylamine Tb is given by

,rb -- 2.7 x 101.[H+l

4.5

79

whereCHt] is the hydrogen ion concentration in molidm.. - This model predicts thereiore that non-exponential behaviour should be observed for, pH 10. Glycerinated rabbit muscle snows, for PH =7 and PH

isAinear g. the ti-ex; the curve when PH out onential relaxation whereas --, .

5. until 95%or 98%of the magnetisation has decayed away. The model can also account for anisotropic relaxaticn, 9W. dependence be by the of explained -whichcan angular

This model is contested by several authors. Aakes, III

iaI,, Ie exponent observed sing

in Also., relaxation in bovine serum albuminjwater mixtures with up to 70%protein content. -

exponendecay transverse the single from frozen the of samples muscle, unfrozencomponent-is this Since ls, tial whereas observed.., abovethe freezing point muiti-exponential relaxation between the is intimately the proteins., and exchanjejate assiciated with unfrozenwater
to See have to expect is would changed, one very much, not expected water and protein groups

freezing the behaviour below point using this model. multi-exponential at

In the case of protons, the modeldependsupon the static

dipolar interaction

arising the 'When

from proton pairs, and exchangewith other protons will modify this interaction. modulation rate is of the sameorder of the splitting will appiy. since the splitting action strength. This will

'N, the condition of motional narrowing

interto the turned be off at a on and rate comparable will

in to is 10" 10" pH a s-1, corresponding occur whenthe rate not

the range 7.4 -, 8.4, which is close to that in muscle. The quadroupolar Interaction will

It does deuterated Is for be destroyedby exchange tenable muscle, the model and thus, while not appearto apply in the caseof protonatedmuscle. .,: i,,,,, 'I,

In Theorigin of non-exponentialrelation then, has many sources. addition to possible the compartmentation there is goodevidencethat protein-water interacof, water.molecules,, II tions are important. The exchange dif-, of whole water molecules,or just protons,between

1- .

1, 11 "?,

,, -,

I,

-"

80

ferent water: populations, or betweenwater the'relaxation complicates species. and other

be--

haviour observed in compartmented and heterogeneoussystems. These processes may be active jointly or separatelyIn any one system, and careful interpretation of: data is required. --

4.3.2 Applications to unicellular

systems.

In simple unicellular

data relaxation suspensions human such as systems of erythrocytest in terms of'two components"21, i. e. 'intracellular water

has been interpreted satisfactorily and extra-cellular

tImeAnboth Usually have components relaxation water. the same, observed, but coping with paramagnetic impurities

dicating rapid exchangebetweenthe two-fractions,

for ihich the: cell membrane Is impermeable.e. l., MnZ+In concentrations-)1004mole. enhances become the : elaxation rate of the external water soAhat the two components resolvable. Hiiher concentrations-enhance the relaxation-rate to such an extent that the signal from the

for internal leads in to the is water longer value a change and external water no observable, which maybe used to determine the lifetime of molecules inside the cells from the relationship -

4.5

,ra

Tb

Ti

TI Is'its whereTb is the relaxation time for, the Intracellular, watevin the doped-system,

-'

intrinsic relaxation time, and ra is the lifetime of water moleculesIn the cell. The back- flux of the magnetically labelled water Into the, cell mustalso be considered'in caseswhere the dopingof extra-cellular water doesnot reducethe relaxation time sufficiently. ,,

Both spin-spin and spin-lattice relaxation times havebeenused'"-'"

.A complication

arising from the use of spin-lattice times Is that there is an additional signal present from the 2itochondria (TI Z 1004s) which mustbe subtracted from the water signal, whilst with sPin-spin relaxation there is a possibility that the true rate of one might not be measuring

81

transverse relaxation due to chemical shift effects,, whicn arise whenlarge amountsof doping are used. Insufficient doping leads 0 problemswith the backflux term however. Further com-

plications with TI 3easurements unaccompanied arise from spin diffusion through the membrane by 33terill transport.

An alternative approachto the study of water transFort in erythrocytes has been to use 110 whien has intrinsically spin-lattice low relaxation times and gives double-exponential decays in ..

relaxation studies"",

which again can be attributed to the internal and exteris time which has a relaxation water

nal calluiar water. Here it is seen that extra-celluiat

once a; ain a comreaucedas, a result at exchangewith the fast relaxing internal component. parisn of its intrinsic the that leads of to with rate observed measure a relaxation

lifetime of water molecules in the cell. A Value Oi Ta z 9-Ims at 37'r. was founa usin; this method,which Is in load agreement with other-methods.

is double In tore coaplex cells, e. g. relaxation proton exponential even cells, -Lettree with concentrations of doping agent as low as 20 gmales"". The relaxation time of extra-

to whenit is not. Moreover, cellular water is reduced whenIn contact with cells compared An dopedsamples greater than the the relaxation rates of the intracel lular component -is rates of the extra-cellular In found the in studying -results with component, contrast

erythrocyte suspensions. The explanation formulated to explain these observations is Ahat the loosely bound i. the three solution, forms different manganese e. system, within resides as -In to the cell wall, and tightly boundto the cell wall. Vater in the vicinity of the cell walls

therelore relax faster and intracellular the moretijhtly boundMn2+.

interactions because fast is then of relaxation with

82

Treat2ent of Lettree cells with Sendai virus increases the permeability of cells to Mn2+ ions and this leads to tri-exponential behaviour, i. e. two types of intracellular water of

equal populations become apparent. Microscopy does not show two compartmentsof equal volume, however. and it is thought that one component originates from the nucleus, mitochondria, and endoplasai-- reticulum, which bind manganese differently to other components.

In between the simple erythrocytes and complex lettree cell systems are the lymphocytes. In viable cells 110 relaxation is bi-exponential, but, in non-viable cells is single exponential"" The results are ascribed to the Intraceilular . ,

the as compartmentation

nucleoplasmand cytoplasm have volume ratios of 2: 1, which is the'same as'the relative populations observed by 1111R.

in behaviour Theexamples the celorigins of multi-exponential given abovedemonstrate lular systems, i. e. physical partitioning between environments with different relaxation characteristics, and interactions between water moleculesand intracellulir With components.

in these systems,two'factors' have, respect to the values of the relaxation times observed beenconsidered, namelythe conformation'of the substrate material and the total water,con-'' tent of the system, both of which play'a role in'the dynamic'behaviour of water in the-sys--' tem. Beall et all", correlated the Ti of water protons with thetotal water content of Hela"

cells during their life cycle. It wasfound that the relaxation time increasedwith the in the macroby conformationalchanges amount of water in the cells, and wasalso determined moleculesof the cells during growthand division. In cancerous cells It has been demonstrated that TI, T2 and the'diffusion constant are'greater*than in normaltissues-m. It is thought that at temperatures abovethe freezing pointthe observed relaxation rates are the weightedaverageof tree and hydration water as given by equation4*19'and indeed'studies on the non-treezable component in cancer, cells Indicate a ieduction In'the amount of water' occupyingthe hydration layer.

83

4.3.3 Applications to Muscle and other Tissues.

In the case of muscle and similar tissues data has usually been interpretej compartmentationbetweenextra-cellular components,intracellular

in terms of

water' in myofibrils and 'In

sarcoplasmic reticulum. and water associated with proteins ana phosph'olipids"I'll' general both the relaxation times and the diffusion

these in systems are constant of water

the for from For found to be reduced the values observed Dure water., spin-spin relaxation sequence the Cart-Purcell-Meiboom-Gill pulse decay the using curves observed exact shape of 'and 'T"11 passes the depend to found rate spacing the average on pulse relaxation are . of the ZVC. indicate presence to These in temperature the through a maximum range -5 results the the components that of kind theretare observed populations of exchangeprocess, and some cannot be taken as accurately reflecting the numDer of spins in each environment.

The motions at water molecules in muscle cannot be described satisfactorily

by a stille

discrete -inio'a of usually resolved number relaxation time, and the observed decay curves are the been to has assigned The extra-ceilular components. slowest relaxing'component the grounds that it is likely also because its relative water on

' interfaces'; from and proteins and"other to be the most remote

from that extra-cellular to of is expected studies close population

spacing.

Neville-and White433' measuredthe'extra-cellular

leg frog in muscies by an various space

ion efflux method in of the slow ccmponent and found goodcorrelation with the magnitude doesnot contain resonance studies. 'Moreoyerin single cells of the giant barnacie-which extra-cellular spacein the normalsensebut doescontain clefts on the outer surface Wch contain about 6%of the total water volume,Fosterts" obser4ed'aslow relaxing component with a relative'intensity ' Sim'ilar'results wereobtainedby Packer' of around3%. and

Belton,331'whoshowed that by iddinj hypertonic ringer solution which could openor shrink

84

the cletts,

the signal intensity

of the, slow, relaxing

could be increased or _component

jecreased. lt, theretore seemslikely cellular water.

is due to extra-, that the slow relaxing cozzpoikent

Pintar" Negai and 5 solution there was still

Ringer D20 that with in observed whenwater muscle was exchanged, the that, origin therefore, IH a 3low relaxing componentani suggested demonstrited Puon"01, Fung However, 3nd mobile organic species.

ot this resonancewas hijhlj

that in the proton NMR spectrum of deuterated samples there was a peak at the appropriate for, the slo'd but, for t that, the other, any so origin not suggested shif water chemical and was from an incomplete exchangeof protons. relaxing component

The fast relaxing component can be fitted time constants. It seemslikely

two or by one the equally well use of either, demonstrate-3 deuterium which of studies

from thejesults

that part time the some of orier of milliseconds. fraction non-exchangeable with a relaxation It is from suggested non-water species. arises fast component of the signal of the relaxing is a-composite of protons from water, and therefore that thefast resolving, component ,"' proteins. Hazelwood, ZVC in at observed muscle the curves skeletal decomposed spin-echo

following decaying the ratios; into two fast decaying components component with slow one and

(D

8% having T2 ( Smsecs; Wer associatedwith macromolecules ,

,,,

2% 84. TZ 45msecs; AD, the z with myoplass with water, associated ,, -Intracellular (iii) Extra-cellular water, T2 ' 196msecs; 10%-

Hazelwood believed that water in these three fractions did not exchange rapidly, and considered the possibility that each traction maybe composed sub-fractions. of fast exchanging Belton found similar results with T2's 2250 msecs.Both Foster"" of 9,40 and, and Fu411,6'

havesuggested that-the 8%fraction with the fastest relaxation arises from non-rigid protons of macromolecules rather than bound-water.

85

Not-all muscle and such tissue exhibit multi-exponential relaxation behaviour. " In postmortemstudies on mousemuscle it was observed that the slow relaxing componentwas not present' immediately post-martem.'but developedafter 15-60 minutes""'. Similar results have

been founa in experiments with frog muscle, 31-, where the slow relaxing componentwas increaied by immersion in Ringer solution which was matchedwith an increase in extra-cellular space, as measuredby the ion efflux method. -in porcine muscle TI-is single exponential. while TZ' is initially two componentafter several hours-"''. single componentbut, becomes

this be inducea bi-exponential behaviour To freeze-thaw can explain using 1his cycles. behaviouvit was postulated that two compartments exist in slow exchangewith eacn other, both consisting of bulk vater and water interacting with proteins"". tera: ting water remains f ixed but there Is a shift The population of in-

in the population of bulk water from cne '

region to the other inducing changesin the relaxation times in acoardancewith equation 4.1.

indicate a distribution Experiments performedat dif ierent, frequenciesand temperatures of correlation times and a dispersion of relaxation times for water in fibrous materials""
beI.
-

Finch and Homer"" reported TI, T2, andTp values for ftog musclewater protons at different temperatures 'a'distribuaboveO'C and over a wide rangeof'frequencies; This revealed tion of correlation'times ranging from 10*5 to 10*11seconds. - Results were Interpreted In termsof an'exchange two fractions; one with a distribution of of water moleculesbetween different degreesof motional freedom.and the other with unrestricted motion like'ordinary water, having a'relative population of 97%.

NMR experiments on frozen samples of rat jastrocnemiusmuscleover the frequencyrange 4.5 - 6OMHZ and temperaturerange -70 - +27T indicate that the relaxation times of the unfrozen fraction are frequencydependent.In samples at subfreezingtemperaturesthe unfrozen

86'

traction exhibits a distribution of correlation times while above-8'C a single-Ti, wasob-, Basedon these results Fung to renderit trequencyindependent. served4hich wasshort-enough 621 supporteda two phase andMc-3auiny, modelconsisting of an untrozenphasewith a distribution of correlation times and a phase with a single relaxation time.-

Several authors have shownthat the water content of a system-exe ts a determining role Sr on the relaxation behaviour of water nuclei. Cook and Vein-" partiall7 T2 TI of varicus and studied the decreased as times the to protein proportional

hydrated muscle proteins, and found that the relaxaticn

ratio of water to protein decreased. NT was found to be directly

concentration.

Fung-11 measuredthe TI of water protons for dehydrated mousemuscle at three different frelow three to down frequencies; 5, '30, all 100 MHz, at very water and contents,, and j quencies T1 decreased as the water content fell when until, it reached a value of less than 71%

it began to show an increase. This phenomenon mayhave occurred because of a change,in the structure of the hydration layer at low water contents. -"II-III

Belton and PackWlsl, correlated the stepwisedehydrationof, musclewith a decreaseIn the relaxation times. Dehydration followed by rehydration wasalso investigated, and it was in the that the amount the was same as shown (hydration) samples rehydrated water of unfrozen different. in normal This low, temperatures but the-relaxation behaviour quite was at, muscle, the that the amount depends in concentration of suggested, hydration system upon, of water a in the system, whereas differences in relaxation times arise because macromolecules of a-,. in the structure and distribution of the water amongst change different fractions, which Is determined by the structure of the macromolecular system,
-fl -

87

Further evidence of the importance of the macromolecularstructure on the relaxation behaviour comesfrom studies on infected, dead and dying tissues. and in studies on contracted muscleV61-61-b". Healthy tissues- generally have shorter relaxation times than unhealthy or have decreased, to the is in tissues, considered post-rigor which amountof ordering of water leading to an increase in its molecular mobility. In contracted muscles where there Is no

in the in from the reduction that observed relaxed state also suggests change water content the spin-lattice relaxation time is brought on by a change in water structure and mobility,

as a result of changes in the structure of the macromolecular substrate.

4.3.4 Application to plant cells and tissues.

in for behaviours found have plant been water A variety of characteristic relaxation in thylakoids"', time observed relaxation tissues, from the single cells and ranging to

ivy bark leaves, and systems in such as multi-exponential behaviourseen morecomplicated - assigned ' InAeaf: tissues the different spin-lattice relaxation times, wheatcell crowns&691. were found to be anisotropic"'I". to water moleculesresiding in different 'compartments, This

from highly thylakoid membranes dependence, ordered it resulted arientational* suggested, was containing manganese.

data NMR The very, relaxation 'and -interpretation of structure .1 size of plant cells makes difficult. Cells within a 'single piece of tissue mayhavea large rangeof sizes; from 1 to

3 for example in zea maizetoot tips, 7"-., Plant cells contain vacuoleswhich maybe 90% JOOpm of the cell volumebut which contain very little in the wayof proteinacousstructure. This

diversity, of size and absence plant material-fit closely to the of Internal structure makes conditions describedby Lillford et al'391, and In caseswhereexternal relaxation reagents havebeenadded, 'the conditions fit almost'exactly the assumptions of the BrownsteinTarr"&-'? l model, with the added complication of a distribution of, cell dimensions.

88

In light of the above. it maywell be incorrect to analyse multi-component relaxation in plant tissues in terms of discrete populations. even in cases where discrete componentsdo exist.

4.3.5 Applications to Porous Media

to a Study at A good deal of attention has of late been centred on the application oi NMR geometrical properties of porous rocks, where the assumptions of Ahe Brownstein-Tarr model appear to hold"" Paramagneticimpurities at the surface of pores within the rocks act as .,

diftuthe by behaviour, in determined is this relaxation magnetic sinks, and results which sion of-mainetisation by the factor Ma/D). (i. the the a. dimensions pores, of the and structure and pores within

diffusion between Cohen that, when adjacent pores Is suffiand Mendelson"', asserted: throughoutthe sampleis uniform and decaysexponenciently fast, the nuclear magnetisation tially, -and the relaxation rate is linearly related to the surface to volume ratios of the , entire pore spaceof the sample. As the diffusion between pores is reduced,the magnetisation maydiffer from pore to pore, resulting In non-exponential the pores are decay. Where weaklycoupled, the relaxation function for the nuclear magnetisationIs the Laplacetransform of the probability distribution of relaxation rates. whichare linearly related to the surface to volumeratios of the Individual pores.

In other studies, Banavaret

noted a good correlation

between the Darcy per-

meability K, -and the function (QIT12)L.13 for a large numberof sandstones and carbonates. (QIPOrosity). They analysed the main lattice decay in terms of a stretched exponential which

they suggested as a consequence the of randomnature of the system:

89

MM = M(O)exp(-t/Tl)l

4.7

Fitting

their data in this way compared favourably in 60%of the cases to analysing in

terms of a numberof discrete components, and proauced a single value for the spin lattice relaxation time of the system. - This was then related to the dimensions of the rocks. using a similar line of thought to that introduced by Brownstein and Tarr. Since both porosity and

by iined 'a ity I I deter the geometry of the 'rocks. they were,able to establish are so permeabi the relationship above.

In an attempt to quantify the parameterwhich characterlses the surface lagnetisation relaxation effects, N, Lipsicas et all? "" performed Fulsed field gradient diffusion measure-

the ;, DELTA in ) that, limit 9 the DELTA al/D, They dolomite >> and ments on noted samples. diffusion is bouncedand the dephasingof spins becomesindependent of DELTA,

2l tM h(o). exp(-gammaIG294115 -

4.8

whereT2v is the spin lattice reiaxatin time'Of the bulk.

'"'Under these condition's the dephasing by twoIndependent of spins is determined the field by is be to mechanisms.In large'pores the dominant caused expected mechanism gradients, and the dependence of the spin echoan G9 in this regimeis given as:

In M(t) - 'I i2i K(O)

I ') YmZtzgammazG-S. ' )tjs 20

4.9

-;

is the strength'of the 'paramagnetic ' sinks on the pore surface.

90

In smaller pares. more molecules sample the paramagneticsinks an the surface. th dependence of e sival interaction the is this relaxation and predominantly via

The on G9 is

iiven oy:

-IIiII

-In MM ZI+4.10 TZ H(O) ,5 in their experiG9 term the the two and. tor Thepowerdepenoence of regimeswasnoted dolomite the ftr that, 2/3 G^,, dependence MM tound Lipsicas of on suggesting a et ai ments, they fast in the diffusion the From' results. region. at ie3st, the systemwas samples. be could it technique for that 6.7 the 10same was suggested cm/s x a of IM. ceduced 4alue 3Fpliid to a variety of porcusor ceilular systems.

Cellulose 'Jood. in 'Jater and Resonance Nuclear Magnetic )f 4.4

4.4.1

Studies oi water absorbed on cellulOse-

The including wood. systems, common Cellulose is an important constituent of many the hydration by of the degree of to a certain extent properties oi cellulose are determined the interto the of nature been conducted establish systcm,and a number of NMR studies have act ions bet-*; een the polymer and water molecules.

by Child"", A modelfor water sorption by cellulose wasproposed

whostudies spin-

derived from a variety of sourcesfor water conattice and spin-spin relaxation in samples low the tents up to 25%. The temperature 14MR dependence the signal at moisture proton of contents indicated a strong interaction between cellulose and water, together with the Presence oi a distribution of binding energies for the water molecules.The molecularaction Or the water moleculeswasfound to depend on the degreeof crystallinity of the cellulose.

91
I!

1he TliTZ ratio.

which should be near to unity according to EF? theory, was found to be three

orders of magnitude greater than expected wnich Child suggested was caused by a d'istriiut'ion ot correlation times for the tightly bound'dater. There was evidence of proton exchange in

the water-cellulose sYstem! rom the oehaviour of the spin-spin relaxation time with varia-

tions in temperature.

An anomaly in the-ansorption behaviour was observed in the TI data as a function of moisture contentat moisture contents of betweenS and-10%, where a minimum is noted -A tial mechanism Ot water uptake by Cellulose Wasput forward whereby the in,. hydroxyl groups. This causes a breakdownin, the weak'inter-chain stronjer mooliitY. , nain tonds. sOfPtiOn Occurs at

hydrogen bonds in favour of

This is reflected in a reduction in macromaieculat in'chains

As further hydration takes place the length of the bridges betweenthe, thei eventually Oreak. and the molecular motion increases.

creases until

Froix and lieI son,III studiea the,spin I attice and spin-spin reiaxat ion times,in cei IuI ose

two types identified as a tunction ot moisture of bound waterwhichthey termed and content
'primary' and-'secondary'. Theprimary bound water wasobservedbelowwater contents of 9% hydration 'mechanism. to the bridging water in Child' s. proposed As the water and corresponds content increasedaboveIE%,, three sourcesof proton 11MR, signal became apparentand could be associatedwith the cellulose itself, the tightly boundwater phase,,, and the mobile friction,

correspondinj to free layers of capillary water. A limit, to the amount of boundwater,of, 15 gig zan-theretore be inferred. which is in agreement with that, reported by Caries and ocalIan"

Sasaki"", et at studied the spin-spin and spin-lattice relaxation times for water absorbedon cellulose as a function of water content and temperature,in order to establish
whethet the correiation time of Brownian motion measuredby NMR Is consistent with that

measured by dielectric relaxation.

Using samples'withwater contents between 12 and 148%,

92

they found that the spin-latt. ice relaxation was non-exponential and two time constants were required to describe the decay of the magnetisation,, whereasthe spin-spin relaxation was
adequately Cescrited by a single time constant. The benaviour of the two TI's as a function

or, iatar content is. described by figure 4.3.

,.;

Two ComponentSpin-lattice relaxation observed in cellalose

Ti Short

(a see) -4 ter V.14a

Lonrj 11

(A SeO 200
X water 9 lea

Figure 4,3 Spin-lattice relaxation for water in ciluiose v water content79)

The multi-component nature oi the spin-13ttice

relaxation

led them to postulate the co-

existence of two oater phases.with the rapidly relaxing phasebeing associatedwith tightly OH free bound to bonded groupson the cellulose, and the phase molecules. which are probably fraction. with the longer TI originating from a moremobile unbound Thecontribution to the

total signal amplitudeof the fast relaxing component dependwasfound to havea temperature ence. .,. bethe two populations, By assuming between sting protons exchange slow exchange

z they found the tweenthe boundand unbound two the for times phasesfrom correlation phase S, Kuboand Tomita's revises SFPtheory to be 10'10 and 10-11seconds for the fast and slow relaxing components respectively, and these are consistent with times observedin dielectric relaxation. in contradiction to Odailmals"t' earlier findings.

93

', Hsi et'all"I

found similar

relaxation

characteristics as a function

in their

study on the behaviour of It was found that

TI and '12 in frozen suspensions of cellulose there existed-a significant

of temperature.

amount of water in the systems whin

does not freeze 3t low'temrelaxation70% of the proportion that the rate

peratures. -I Once again; two constants were required while the spin-spin signal -was'attributed relaxation

to describe the spin-lattice exponential-function.

was described by a single

to the component with the longest relaxation

time. and this

was round to be independent or temperature below the freezing or exonange between the two populations

event. indicating

had, teen reduced beyond detection.

between There rhe larger TI exhibited a minimum difference oi: 230K. ' was a considerable W ' and the Tils at the minimum the values of T. which. accordine to Resing. is indicative of a distribution data The could be fitted. times. at cor: alation using a jaussian distribution

in'terms the the by to results author, who nose expiain, of in(T); Dut'this, iaea -jasrefuted

or,cioss-relaxatior,

`-FfeeZInj

possibility 111n1mizieg-the

Of chemical exchange, and ail the unirozen water must

be associated with tne surface in someway. To explain the multiple behaviour, ot, the spinlattice Sasaki both in the transverse, and plane, relaxation alongside exponential relaxation interaction relaxation of a cross betweenprotons on the solid

Hsi considered the possibility

and on the liquid. - These eifects are expected to contribute solely to th,,-pin-lattice . relaxation. Using samples Aatural,vd with Deuterated water, a clear doublet was evident below

the treezing event that das not observed above. This is consistent with the inducement of a preterential orientation by the cellulose surface.

ItAs ditficult

by Hsi and those observedby to say whetherthe two populations observed

Sasakiare the same, as the two experimentswereconducted on opposite sides of the freezing event. I, -1 ,,I" 11 1, I,,

94

4.4.2 NMR of water in wood.

Nuclear Magnetic Resonance provides a non-destructive methodfor water in wood. -studying The most simple application has been to measurethe water contents of wood by observing the amplitude of the decay following a single 30-degree pulsel,111. Provided that the applied r. f. field does not overload the receiver for more than a few microseconds, the transient response will consist of 3 signal due to the solid content oi the woodsuperimposedupon the signal from water protons. The amplitude ot each componentto, the total signal is PfOPOr-

tional to the numoerof, protons contributing to it and,so. by resolving the two fractions ana to matter making someassumptions about the proton density in eacft. it becomes a simple derive the moisture content of samples.

"I by lcn Caries Z used was and r3ther method -cal motesophisticated -A
amountof bound water resident in woodand cellulose gels.

to estimate the

They measuredthe line width in

high 300% from these contents as moisture as and, steady state measurements with on samples measurements,they derived a value for the T2 of the vater. A two-state model for water in the system was evoked with water molecules residing in a bound layer and in an equilibrium mixture of structured and unstructured water. Rapid,exchangebetween these two fractions

leads to an observed relaxation time which is the weighted average of the relaxation times of the separate relaxation times. This is represented mathematically as:

BF
4.8 T TIb Tf

where8 and F are the probabilities but, at a jiven, time. a proton is in the bound. or tree state.

-The three orders ot-magnitudethat separateTb and, Tf leads to the simplified, equation:

95

Tobs

Tbr x" [H240161'H2401b

V-

I ation'P and the population of water molecules in where 1,4-'Olt and IH201bare the total popu the bound phase respectively. From this it can -be seen that. provided the bound water fr3c-

tion remains tixed, the observed relaxation time dependsupon the relaxation time of the boundfraction. 'Olb. ina on [H,

Caries and Scallan found a linear depencence of Tobs on [HZO]t for moisture contants of aoove From the linear portion of the grapn they estimated-the value TUCH201b.and

thus T, b. , they is intercept thejob from the obtained, an upper limit.,, to. the value ot,,, s ax on deviates IHA'Olb. the the The from curve value of point at which giving a maximum estimate, from linearity 'the the the fibre samples. saturation point within was taken as a measureof

the to all to upon appeared correspond a monolayer amount of boundwater within the samples accessible surface area.

Riggin focus by the, et in TransverseNMR were oi a wood times water oi relaxation 'study
aI high A 176%. eral,,,, as value of sev. contents as moisture using spruce sapwoods with with In-,

microseconds was found for the T2.at protons in wood, and this increased slightly creasing moisture content.

increased Below30% to the and FID exponential, water was moisture content the attributed 0.9 from,approximately80 microseconds t, c. about of, % mil-_ at,, moisture content,, o, a,value liseconds at the, the higher of At relaxation Magnetisamoisture contents contentwater 310%

tion due tolthe water content is the sumof two exponentials., Thebehaviour the two time 'of in figure 4.41-as Constantsis shown function of the moisture content. Three phasesof a , hydrogen, nuclei wereusedto explain the nature of these curves,.which corresponded to the . protons residing in wood for which Nanassy, 1131_ claimedthat were exchangeable with, '38%,

protons in the water, and two water phasesthat to the boundand non-bound water corresponded

96

components. The exchangebetweenprotons in woodand water is expected to be relatively to enaoie the two signals to be resolved. The exchangeis restricted

slow

to the bound fraction

of water. The,bouna,and mobile tractions are also expected to be in exchangewith each other on a time scale which is slow comparedto the observed relaxation times of each component.

71fast,
72 slolo
. Fk see) Na

, we cookponent spin-Irin relaxation of vater in Spruce SAPvQoA

ax

Plow

ao lt4 120 moisture content W ft is ture contenk:W I 63) Figure 4.4 tpirrivin relaxation in Semco sai'mooav. Aoirture coutunt! a

ia

data. 'the 'we I II for USI n? tIhis model', Riggin et' &I could account'

They pointed out that

of . . for families times the three the observed*v alues of the po'pulatic'ns and relaxation I 'because be the exchange tween fractions. of protons were'-ot the true values It was also

the NMR lines than the that predicted in of greater width were noted steady state experiments by the theory of'chemical exchange.'leading the authors to suggest that the relaxation times reported actually system. Is IrI Ifferent d ted average ove many eI nsemble represented complica phases of the

.d I Ia and' , by Riggin Hsi, Hossfel Bryant"" to that dertook performed and un similar study Id that in foun general, both TI and T2'were'double-exionentlaffor moisture contents greater 'than 30%. ' Belowthis' value 'the t,ra'nsverie"re Ion wascharacterlsed by a single I axa-t parameter', but the longitudinal relaxation still required two time constants. Twotypes of

behaviourwere observed for the signal intensity as a function of temperature: below0.38 9 of water per gramof wood no freezing event wasevident, whereas a clear freezing event was

97

observedin samples of higher moisture content.

A slight decrease in the FID amplitude with Below the freezing event. and as the fibre relaxa-

decreasing temperature could be seen below the. freezing event.

for samplescontaining moisture contents of above and below the 30%identified

saturation point. the transverse relaxation was exponential, while the longitudinal tion required two time constants. 30%to the total signal amplitude.

In both cases the rapidly relaxing fraction contributed

by The appear3nceof a second explained transverse of component adequately : elaxaticn was V andt below 30% the the is that surface all content water assuming moisture associated with above 30%. the additional water Is not absorbed by, the cell wall matter. it was postulated

that the non-exponential longitudinal behaviour arises from coupling between the solid wood protcnshavinj a long T1 but short TZ and water protons with a shorter TI.,

the transfound from the times of Evidence ratio was of a distribution of correlation bY have been TZ depression The caused also longitudinal, of may, verse,to relaxation times. in the liquid those state. between and chemicalexchange solid protons I Thedata of Hsi,et al oas usedby Brownstein"" to support his theory that, as the moisture content is varied. the behaviourof the two transverse relaxation times could.be exthat water could plained using a-modeibasedon the diffusion equation. Themodelassumed reside in a, layer within or boundto the woodwherethe ditfusivity layer wherethe diffusivity severely reduced,or in a non-absorbed, of the moleculesis is close to that observed ,

for water.in the bulk. The relaxation of the proton spin magnetisation of the water

moleculeswasconsideredto be due to a combinationof diffusion and magneticsinks on the surtace of the cell wall.,

98

The solution to the Brownstein and Tarr equations in section 4.3. subject to the boundary conditions that evolve from the Brownsteinmodel. is a sumof exponential$ for which the relaxation times Tn are given by:

(b/a - 1) 1. tan f, =14.10

where I is defined by

T. M) a2ilD --

4.11

water, D is the diffusivity a is a measureof quantity of unbound layer, and b represenrs the thickness of the non-absorbed layer.

of water in the absorbed

Using n_-- 0,1 the model ajequatelY predicted the values of the two transverse relaxation dec3y,times observed by Hsi et al as a function ot moisture content.

woodis a highly structured and fibrous material, and anisotropy of the proton NMR spectrum and T1 relaxation times was found by Vriesengall" et al who mademeasurements as a Their steady state exhumidity

function. of the angle between the woodfibre and the mainatic field.

relative periments implied that three types of water existed in the samples at 100% and the resonant frequency of each was a function of the orientation the nature of this orientation

From the samples. of

dependence,the authors concluded that the vast majority of types three the that of the and wall cell region of symmetryrelative to the fibre axis.

Sorbed. water exists in the inter-tibril water-filled regions exist,

each possessing a different

This symmetrymust persist throughout the be to Its for in observed. effect sample order

In de-II Inif led samp I es, function of the samp Ie the resonance positions as a peaks of -the orientation weremarkedlydifferent from those observed in the fresh timbers. The authors suggested that the orientational dependence ter wasnot a result of direct align'ment'of w3 moleculeswith the celluloses in the cell did not appearas the resonances walls because

99

dounlets. Instead they believed that the symaetry by a magnetically anisotropic wascaused
environmentwithin the inter-fibrillar regions that originated from ordered magnetic comThe ordering in the lignin could be caused by

ponents of the lignin such as phenolic rings.

its interaction with the cellulose via the hemicelluioses. '-

The effective W"' ,

self diffusion

coefficients of water in woodwere measuredby ".acGregor et

from the the field technique. The pulsed gradient reduced using were values obtained coefficient iradient. of distilled'water upon the Orientation and depended IonInAhe was the greatest

value for, the self diffusion of the fibres to the field gitudinal direction,

In generak diffusion

i. e. along the direction of' the fibres.

The depression of the dUfusion coefficient

in the woodwas expiained by assumingthat

the water moleculeswere'bounded by the cell walls.' This leads to restricted diffusion, and diffusion from the be the the dimensions coetthe boundaries of estimated magnitude of could f icients. For diffusion perpendicularto the direction of the fibres, the dimensions agreed In the longitudinal direction, however. To ex-

reasonablywell with the known width of the cells.

the distance between' boundariesis too great to influence the diffusion coefficient.

I I the the diffusion treat to the as observed it coefficient plain reducedvalue, ar-y wasnecess I weightedaverageof coeffi-cients for water residing in the cell lumenthat Is not bounded, by the spiral thickenings that and water residing in*the pits in cell walls'and'bounded I into lumen, the protrude suted rate of diffusion. which lower the mea
these regions is slow, then two componentditfusion reported however.

If exchange between

decay would have been observed. Non was

I Themajor conclusio'ns'tobe derived from

tudies are that the NMR properties of such sI Nevertheless,they provide useful

water in woodare complex and needcareful interpretation.

and probably unique insights into the moleculardynamics of water within'the system"".

too

Reterences:

Belton F. Z., S Ratcliffe

" Biological tissues. in E.G. ' "NMR COI'Fartmentation and

Frog.

Spectroscopy.- VoI17, pp/Z41-279. UMR Packer K. J. "The dynamicsof water in heterogeneous systems." Fhil, Trans. R. Zoe.

3-" Val 178, pp -59-87. (1377) Michelson C.. BerendsenH.J. C. "Proton excnange and moleclar orientation of water (1973) " J. -Chem. Phys. ' fibres. 296-30S. Vol 59, hyarateo in collagen pp 4. Mathur-De Vr; R., 'Bertinchamps A.J., 'Berendsen M. C. ' "The effect Of gamma-ridlation ana polynucieotides. on the hydration of DNA solutions. " ' Vol 68, ' ppI97-214. Radiat. Res. 1. An OR study of frozen H20 and DzO (1976)

5.,

R. Trans. Fhil. " in FranksF. "Solvation interactions oi proteins solution. Vol Z78, pp 69-96. ' ' (1977)

6.

Katayama

' Vol v "i r jiwar a S. j. Fhys. *Chem.

84

,, * 11"0. (1960) ' ,p

7. 8. 9. 10. 11.
12. 13. 14. 15. i6. 17. 18.

', Fijiwara S. J. Am.Chem. Soc. Vol 101, p 4465.- (1979) Katayama'S. i7i, Sci. Vol 1855. (1973) Caries J. E., Scallan'A.M.'-J 'App.Polymer p . 'Soc. Vol'97, Hsi E. Bryant R.G.-' J. Am.Chem. ,
p'3220. ,1 (19-15)

"Resonance, Vol 6, p 244. (1972) Derbyshire'W., 'Parsons'J.L. J. Mag. Vol'63, (1979) 2.38. Sc i. ! LynchL. J., -Webster, -, , Interface Coll. J D.S. p
Lynch L. J. ', I Webster'DA. Hreviishviii Fuel Val 58, p 429. (1979)

G.M., SharimanovY.G. Biophysics Vol 23, p 242. (1978)

Resing H.A., Foster K.R., GarrowayA.N. Science Vol 198, P 1181; (1977) Golton IX., Gellatly BJ., Finney J. L. Studia Biophysica Vol 84, p 5. Val 44, p 1212. ' (1979) (1961)

Leung,H.K., Steinberg M.P. J. Food Sci.

Haly A.R., Snaith J. W. Biopolymers. Vol 7, p 459. '(1969) '' Ashpole D.K. 'Proc. Royal Sac. ' London. 'Vol, A212, p 112-- (1952)

101

19.

Mousseri J., ateinberg M.P., Nelson A. I., Vei. L. S. (1974)

_J.

Food Sci.

Vol 39. p 114.

20.

oodhouseD.R., Derbyshire W., Lillford

P.J.

J. Magn. Res. Vol 19, p 67. Vol 8, -7! p 379., (1980) Vol 399, p 191.

Nareb3kaA., Streich W. Call. Polym. Sci, 22.

Fung 8.11., DurhamID. L., Uassil D.A. Biochim. Biophys. Acta. (1975), Resing H.A. J. Chem.Phys. Vol 43, p 669. (jr-65)

24. 25. 46.

Morariu V. V., Chiricuta I. C. Oncologia Vol 16, p 291. (1977) OakesJ., J. Chem.Soc.. Farad. Soc. Trans. 1. Vo1 72,2 '116. (1976) _p Lynch L. J. 'Vater relaxation in heterogeneousand biological Magnetic " systems.

Res. (2) Chapter5. (1983) Haly,A.R., -naith J. 'J. Vol 10, P 1681. (1971) -Biopolyiners. , Dehl R.E. Science. Vol 170, p 7318. (1970) Ramirez I. E., Cavanaujh J. R., Furcell J-M. J. Fhys. Chem. Vol 78, p 807. (1974) Morariu V.V., Mills R. J. Coll. Sci. Vol 2 406. (197d. ) p WIS. -Interface

249. 30. 31. 32. 33. 34.

FearsonR.T., DerbyshireW. J. Coll. Interface Sci. Vol 46, p 232. (1974) BarnesG.J. Z. AnjewMath. FhYs- V01,13,p 1.33. (1962) FackerK.J. "I'luclear spin relaxation studies of molecues adsorbed on surfaces." Prog. in NMR Spectroscopy. Vol 3, pp 87-128. (1967) Cooke " Ann. R. R., Kuntz I. D. "The peopertiesof water in biological systems. Biophys.,(3) article 9035, pp 95-126. (1974), _

35. 36.

PackerK.J. Phil. Trans. R. Soc. London Vol 278, p 59. (1977) & BrownsteinK.R., 'Tarr C.E. "Spin-lattice relaxation in a systemgoverned by

diffusion. * J, Mag., Reson. Vol,, 26, pp 17-24 (1977) 37. Brownstein. K.R., Tat'r C.E. "Importanceof classical', diffusion in, NMR,, Studieslof, water in biological cells. ", 38. 39. Phys. Rev. Vol 19(6), p 2446. (1979) _A.

Chang. D.C., Hazlewood C.F. Biochim. Biopys. Acta. Vol 630, p 31. (1980) ' Lillfard P.J., Clark A.M., JonesD.V. Set. \vol 1^47, A.C,S. Symp. (1980) p177 ,

102

40. 41. 42. 43. 44. 45. 46.

Fung B.M., Puon P. S. Biophys J. -Vol 33, p (1981) ' P 127 OakesJ.

'"

J. Chem-'-SocFaradayTrans. , Vol 31, p 207 (1976) 354. '(1972) Biochim. Biophys. Acta. ; Vol p -188. Biophys. J. '101 I:Z' P 1101. (1975) Biophys. J., Vol 25, P 389. (1979)

Conlon T.. ' Outhred R. '

Fabry M.E., Eisenstadt M.

Firkle J. L.. Ashley D.L., Goldstein J. H. Shporer M., Civan'M.M.

Biochim. Biophys. Acta.

Vol 773, p /Z19. (1984)

Biol. membr. Getz,D.. Gibson J. F., SheppardR.N. C. A.. Michelin K. J:, Pasternale 'J. -,, Vol50, p 311. (1979) 9iophys. p 601. (1976)

47. 48. 49. 50.

Shporer M. Haas II., Civan M.M.

C.F.,- Roa P.N. Beall P.T., Hazzlewood

Science NY. Vol 192. pp.904-907. -(1976)

Damadian C. '- Science NY. Vol 171, pp 1151-3. (1971) Belton P. S., Jackson R.R., Packer KJ, ' Biophim Biophys. Acta., Vol ^_t6, P 16. (1972)--

51. 52.,

Diei-31 J. G-, Pinter MA

J. -Biophys. 'Vol 15-,P d55-" (1974c) Blophys.-J. Vol 14,p 583.

D.E. Hazlewood C.E., Chang D.C., Nichols S.L., Voessner. (1974)

53. 54. 55. 56. S7.

Neville M. C.,, Vhite S.

J. Physiol., Vol,/288, *p 71 (1979)

Faster K.R., -ResingH.A., Garroway AA Science. Vol 194, p 324. (1976) Belton P.S., Packer, K.J. Iiophys. ' Acta, '101,354,, p 305. (1974)Siochim.

B.M. Fung Biophys. J. 'Vol 18,, p 315. (1977) Pearson R.T., Duff I. D., DerbyshireW., Slanshard JAV. Vol 362, p 188. (1974) Biocim. Biophys. Acta.

58. 59" 60. 61., 62. 63.,

OuthredR.K., George E.P. OuthredR.K., George E.P.

Biophys; J. Vol 13, pp 83-96. (1973) (1973) Biophys.,J., VoI13, pp 97-103. J. Mag.Res. Vol 14, pp 44-51. (1974) "' -, 1

Knispel-R.R. Thompson R. T., Pintar MA 9 Finch E.D., Romer L.D. 'Biophys.-k FungBA,

Vol 14, pp 907-921. ' (1975)

McGaughy, T.W. Biochim. ' Biophys. Acta. Vol 343, pp 663-673. (1974) (1971) pp 10022-1017.

CookeI., Wein'R. '-,, Biophys.

103
4

64. 65. '66.

Fung B.M.

Ai0chim.

Biopnys. Acta.

' 317-13.2'2'. (1977)' Vol'497, pp

Vol 354. pp 305-.314. (1974)

Belton P. S.. Packer k. J.

Biochim. Biophys. Acta.

Chang'D. C.. HazlewoodC.F., 'JoessnerD.E. Z-58. (1976)

Bicchim. Biophys.' Acta. 'Vot

PP25.

67.

Mazlewood"C. F.. Nichols B.L.. ChangD.C., Biown'B. Johns Hopkin's Med. J. ppll7-131. - (1971)

Vol 1Z

68.

MenkeW.

"Biochemistry of Chloraplasts. 'lot I.`

Ed T. J.' '43oodijin.

Academic

Press. London.'(1966) 69. 70. Stout D.G.. SteponkusP.Id Cotts U McCainD.C., Setil'ToC., p 748. 71. 72.. (1984) 'L JuniFer B.E. (196 "Plant Cells. " Chapter 1.81 , -well, Oxford k8 . " H. For-, Chemistry "Physi.., media. us of s and Plant Physioi. Vol 6Z. pp USA. (19731

Govidjee, Markley J. L.

Proc. Nati. Acad.

Clowes'F.A.I.. Banavar JAI.

Koplik J.. "Jinkler K.W.

AIP Conference proceedings 154.,Am. Inst. Phys. NY 1987. 73. CohenM.H., MendelsonK.S., "Nucleai Magnetic Relaxation and the internal geometry of sedimentary rocks. " J. Appl.Phys. 53(2), February 1982 74. Banavar J. R.. Schwartz L. H. "Magnetic Resonance as a probe of permeability-in media." 75. Phys.Rey.Lett. 58(14) April 1987. interfacial porous

Schmidt E. J., Valesco K.K., and Amos M. "Quantifying solid fluid in porous rocks with proton nuclear resonance." phenomena 15 April 1986.

J. Appl. Phys 59(8)t Nmr.

76.

Lipsicas H., Banavar J.,R, 'Jillemsen J.

"Surface relaxation and pore sizes in rock3 -

a nuclear magnetic resonance analysis. " App.Phys.Lett. 48(22) June 1986. 77. Child T. F. cellulose. " 78. "Pulsed NMR study of molecular motion and environment of sorbed water onPolymer Vol 13 June 1972 of water with cellulose from NIMR relaxation

Froix M.F., Nelson R. "The interaction times. "

Macromolecules8(6) Nov-Dee1975.

79.

Carles J. E., Scallan A.M. J. App. Folys. Sci. Vol 17, pp 1855-1865. (1973)

104

SO.

Sasaki,M.-, Kawai T., Hirai-, A., Odajisa'A. -"A study of sorbed water on cellulose by technique. ' Pulsed NMR 1. Phys. Soc. Japan 15(9) Sept 1960.

St.

Oda ji ma, A. 2aterials. "

*Nuclear Magnetic Resonance studies of water sorbec an fibrous J. Fhys. Soc. Japan 14(3) March 1959. J.,

622.

Hsi E., Vogt G2. J., Bryant R.G. 'NMRstudy of water absorbed on cellulose. "Coll. and Interface Sci. '70(2), June 1979.

83., 84.

Sharp A.R.. Rielyin T., Kaiser R.. 1978. M. Echnsider 1OCZ) Fibre. M. H., good .0 and Riggin M.T., Sharp A.R.. Kaiser, R. " in "Transverse NMR wcod. relaxation of water

J. Appl. Folym. Sci. -Vol *; 3, pp 3147-3154. (1972) 85r. 86. 87. NanassyA. Vood Sci. Vol 5, P 181'. (1979. )

Hsi E.. Hosstield R.. Bryant R.G. J. Coll. Interface Eci. -Vol E-7. ? Z89- (1977) Brcwnstein U. " sorbed on wood. "Diffusion as an explanation of observed NMR behaviour of water abJ. Mag. Res. Vol 40, pp SOS-E010.tj"60) 'Froton NMR of adsorbed water in wood." pp 911-921. (1983) "Anisotropic diffusion of

88.

S.. Luner F. 'Iriesenga J. R., Chandrasakaran J. App. Folym.

37. Symposium Folym. App. 4-ci. -

89.

MacGregorR.F., Feecoeler H., Schneider M.H., Sharp A.R, " water in wood.

37. J. App. Folym. Sci. : App. FolYm. Symposium J. Electrochem. Soc. Vol 133. p 631. (1986)

pp 901-909. (1983)

90.

MendelsonK.S.

105

Chapter5.

Experimental Detail and SamplePreparation.

intrmuction.

No experizentai systems were used during the course oi this work. and a description oi each is given here. A numberof changeswere madeto both systems. The changesincluded

modifications to the hardware, and a revision of the data handling techniques. For convenience the two systems were called the BZ-25 system and the A140 system. Ancillary equipto both systems is described in section 5.4. ment common

In section 5.7, a description Is given of the methods-usedin preparing samples'of both polymer solutions, and the xylem itself.

5.2 The B"225 Spectrometer.

5.^..1 TheMagnet.

has The An A.E. I. M12 the field Bo. magnet static water-cooledelectromagnet provided nine inch pole pieces and using a polegap of up to 5 cmwascapableof producinga field of 1 1.4 Tesla., Theoriginal, A.E.J. powersupply waschanged, to a solid state supply which lead to a long term stability in the magneticfield of better than I part in 101, although the, less than,this.

short term,stability wassignificantly

5.2-. 2.The Spectrometer.

106

The basis of the spectrometer was machine designed to a secondhand Bruker 8-KR-w^-02 operate at 14MHz and 6OMHz. It incorporated a homemadepulse programmerfrom which logic level pulses were amplified and fed Into the r. f. gates of the Bruker transmitter. The power

amplifier could deliver approximately 400 volts peak to peak into an impedance of 50 Ohms. (See figure S. D.

Brukep

Figure 5.1 The KXL B225 SYedrometM

Pulse programerl', Ostillatop Power Amp 69MHz pWe shifter Signal Outpat Signal, iverager DL290 sensitive phase Probe Ht

-FDati: LtIP_D

altector Polaron jYmmics Amp

Thereliability

built wasreplaced by a home of this systemwaspoor, and the B-KR-2502

frequencysynthesiser with a frespectrometer,for which the r. f. sourcewasa Schlumberler 2 and 40 Mliz. Theoutput from1his was split into three channels, quency range of between two of which-formedAhe r. t. signal channels,and the third wasused as a reference channel. Ther. f. channelswerephaseshifted and gated. The output from the gates werethen combined and the-signal frequencywasdoubledbefore being fed Into a Bruker high powerr. f. amplifier requiring an Input of, just I volt peak to peak. (Seefigure S. -2)

107

'the r. f. pulse lengths set using the pulse programmer were continuously variable up to 10 ms. and 90' or 180' r. f. pulses were easily obtainable by the observation of the nuclear sigin the signal. nal upon adjustment of the pulse length to give a maximum minimum or phasesof the first The

and subsequentpulses could be shifted independently so that the shifts

required by the CPMG sequenceswere simple to produce.

ulse PiploplammFer

Frequeneg synthesiser

Af A shif ter hi t splittert

gate

comb

inkerj

Powert aimp

shif teR::: Mate

aoubler

1 shil e 1 dotth er
Phase sensitive detector
I

yolar-on amp

yol&yon pre-amp

Probe

e'>
mopt.

averneri U-Spectrum signal DL200

11

SPectrometer.M12. Figure 5.2 TheB225

f. 90' for the length Measurements 60 MHz, r. a this where pulse made on wereat system pulse wasapproximately3 ps.

Theprobe assembly is describedin section 5.4. The nuclear signal wasamplified using a Folaron receiving systemconsisting of pre-amplifier and amplifier units. commercial The

Outputfrom this wasfed into a home built phase-sensitivedetector, for which the reference signal was derived from the oscillator in the spectrometerthat wasfed into a 60 MHz tuned amplifier and then phaseshifted. Theoverall deadtime of the receiver mechanism wasquoted

as 40 PS.

108

A oiode detection system provided an alternative causeof the non-linearities preferred.

to the phase sensitive detector, but be-

in responseat low signal strengths phase detection was

ZI

DataHandling.

a using Signal averagingcn the B2_5 TZ in system was achieved measurements the case of ZMS DataLab DLZOO linked to the paper-tape punchof a Data Dynamics type 1133teleprinter. For

P.F.G. diffusion measuresignal averagingof transients suchas those in T1 measurements. to Biomation a Interfaced 600 freezing transient a recorder ment3 and curve experiments, Sinclair ZX Spectrum wasused.

Data processingwascarried out on either the departmentalDataGeneralCarporationNova 12220 minicomputer,or directly on the .7XSpectrum.

5.3 The A140 Spectrometer.

5.3.1 The Magnet.

here. 5 BZ". An A.E. 1 water-cooled'magnet in the employed to was that system used similar Thepowersupply wasfound to be far morereliable however,and a stability part in 104 as easily maintainedover a period of several hours. of better than I

1 5.3-2

The Spectrometer.

109

The spectrometer was a homebuilt broad-band system, operating over a frequency range of hard-wire to 30 MH:. It designed be to pulse modified 4 was controlled using either a programmer,or-by the use of a software controlled pulse programmer.

The output of a Fluka 6160 AM frequency synthesiser was split presented at the inputs of a wide band phase shifter. -shifters-

into four and each was Two fixed phase

(See figure 5.3).

' 900,18 manually 0', as 0'. '. 170" two available of either or were shifts roduced and a and into fed signal a were

variable channels. The outputs of selected phase shifters reierence"zultiplexer.

a of these by the software the of was-determined output and either

the Hazeltine manual, operation of hard-wired puise pio"grammer. via or unit, programmable A

channelswere Theoutputfrom. the signal channelmultiplexer wasspiit into two. These they following f. were the which gates. and r. t. through passed a stage of r. amplification he'j"i The first accomplished was stage amplification. of further two"juj stages combined a'nd'entered -I by the a out second stage was carried Polaron and RF 460-BAZ amplifier, using a wide-band 50 into 500 to a home volt peak peaksignal built Mullard design amplifier whichsupplied a Ohm impedance.

The reference channel from the multiplexer wasfed into a variable phaseshiiter and then to the detector.

Thepre-amplifier of the receiving systemwasattached to the same as the probe mounting assembly, and wasdesignedarounda PlesseySL 610 r. f. amplifier with automaticgain control. input stage wasa conventionally tuned F.E.T. amplifier. Theoverall gain of the pre-amp

was4OdB. The output of the pre-amp wide bandamplifier with was fed Into a Hewlett Packard IUt from this went directly Into the signal channelof switched2OdB or 40dBgain, and the outp the detector.

110

===a . ==

. == Z. Z.

2mc

. 7-

-1-6 -C.

t=,

k.

LA-1 . Mao

LA .A 0=4=5 9==x EZ" LA-A . zr-= L. "

Am=
16AA

., S

IJA

30= LLJ

L-JLA

&-" 2

L. A. J mmoc &A-A 9-

L. "
L. Lj

Sze-

'k--o

16LL. A.. A Qdc-, mm LAJ -=

L-=M s-"

-= 310C

. -f

a-

,I

&-=j

ME

Lr;

t., n
C=
Q-.. LA. A
L-A-A

w LA-A L ;a
--mr __j -; A= lQs-.

A. Lj

I-&-A

6-4

6-4

1(. 21 x: cEi
8-. -x ,

Thedetector was designed around a Minicircuits The local oscillator

SRA-I passive double-balanced modulator. of

Input to the modulator was arranged to be connected to the output

the reference phase shifter tion.

for phase sensitive detection, or to the earth for diode detec-

This switching was done using a coaxial relay.

The output from the detector was amplified and a simple L-C filter provided a time constant for the signal. proximately 25 ps.

on the output of this

The overall dead time for the system was ap-

Thesoftware controlled pulse programmer and a wasdesigned, arounda Z80microprocessor handHazeltine terminal., Theprogrammer one Of 3 second Operated on a menu systemwhereby of could be selected., Variables suchas the number numoer of pre-deflned pulse sequences Hazelthe delays the between sequences via repetitions and specific were entered on request tine keyboard.

For the hardware and parameters were set manuallyusing pole systemthe pulse sequences carrying out experiments switches, and wasfound to be moreconvenient wherea single when f In 1801-r-90' the r. parameterin a sequence such as value of T a wasconstantly changed, Thesystem pulse sequence also incorporated an nth point sampling used in T1 measurements. facility thereby so that r could be maintainedat a constant value during T2 measurements,

eliminating the effects of varying r in systems wherediffusion of the resonantnuclei takes, place.

Measurements on this system were madeat 16MH., - ISMHz.

3.3 Data Handlini. ..

ill

Signa I averagingI on the A140 systemwas achieved using a Data Lab DL4000signal averager, Interfaced with an Apple 11 microcomputerwith two 'disc -drives for data storage. MX-80printer. of results could be madeusing an Epsom A hard copy

5.4 Ancillary Equipment.

5.4.1 The Probe Assembly.

, two Th coil'design conventional a was obe assemol y pr e

wound coaxially on' a 10 mmglass former. This then fitted

coils receiver and transmitter with

inside Jacket enaan evacuated

suffiformer was be The inner the to diameter used. of bling a gas-flow temperature system be tuned ' that Separate inserts tube. could 7 mm ' sample probe ciently large to accommodate'a' " Fine 16MHz l. 60 MHz the syste. to Iset-frequencies were employedon the' spectrometer and tuning 'was madepossible in the receiver coil b meansof a homemadeseries npacitor tem., sys-

4.4. 2 TheTemperature Controller. .,

2349 EA temI . instruments oxford as Temperatu're"'contro'l iam achievedusing an of the s pie

IS'C For temperatures of and perature control unit in conjunction with a gas flow mechanism. above,'air was used irom-thi'laborat'ory this nitrogen was below for temperatures supply, and

"' dewe the being from boiled off used as rI vessel. cryogenic, a

Ip *d I' heater Control of, ihe gas temperature'was ''in the ga'sI in'e' by ace achieved m ns of a in front of the probi. A'copper-cons'tantan thirmoC'Ouple sensor, situated just below'thi r-fColl. fed the iignal back to the temperature controller and regulated the heater current.

112-

Thecontrol I er was caI ibrated againstagI sampletube containing glucose solution. the temperature,range 1-10K to 370K.

ass thermometerwith 'its

buIb 'immersed in'a

The system was accurate to within 0.7 degrees over

5.4.3 The Bruker Minispec.

On loan to the Mary RoseTrust was a Bruker CP-Z Minispec.

This is a device which to

operates at' a frequency of 10 Mliz,'using a pole temperature of 40*C. It has the ability measuresolid: liquid ratios, ano single componentTI and T2s.

by the delivery of a single 90' pulse to the Thesolid: liquid ratios weremeasured for TOPS the lips two The and FID at is then points. chosen at sampled sample. resulting to: the CRO to be minispec, this using a attached The FID purposes of work. can monitored detection. which there is a choice of diode or phase

5.5

TheDiffusion Equipment

5.5.1 ThePulsedField Gradient'Coils.

StandardBruker field gradient coils wereused throughoutthis work. Thesewere a Helmholtzarrangement wherethe'coils were electrically in series, but magnetically opposed.

The coils each consisted of four pancake in a resin , layers of copperwire coils embedded block, fixed to'low eddycurrent plates. Theyweremounted on the side plates of the probe headso that the samplelay in the region of the most linear field gradient, and so that the gradient was in'thp direction of1o. '

113

Thelargest region of uniform magnetic is found to field gradient in such an arrangement be when the radij, ot the coils are 3-% Aimestheir separation. and although this condition waspot met by the Bruker coils the field wasexpectedto by approximatelylinear over the small volume at the sample.

To ensure that the diffusion able the coils were'ealibrated. of the fact

coefficients

determined usIng this configuration were relimakesuse

This is done using one of two methods. The first

for &, in spin'echo a of a sample a field the'shape

by in determined is G t gradia For a cylinder of

both the geometry of the sample, and the magnitude of the f! eld Ig raIdient'. radius r. the height of the echo at time t is dominated by the first given by' II;, - .111.

junction Bessei order

(2,1tri G(2T-t)r gamma.

by denoted is the If the time between side of either echo centre the first two zeros on is by; 9t, then the field gradient when given protons observing

gausscm*l St

The shapeof the Spin echo wasobserved for a cylindrical water sampleas the steady current through the field gradient coils was increased. Thevalues of the field gradient were then estimated from the separation of the null points and using the expressiongiven above. The results are shown in figure 5.4. The gradient was found to be independent of the diameter of the sample within experimentalerror, indicating that at the site of the sample the region of linearity wassufficiently large.

Thecalls could also be calibrated by assuming a value of 2.3 x 10*lcm2s-I for the self diffusion coefficient of distilled water at 251C.Steadyfield gradient and pulsed field

114

calibration

132 G cW14'I r

int gra 71
CWI

Fiel

3-I,. -,

'shape of ", spin echo

YrRul'cea 6 G
4 34 Ar

G=2,865 x 10Current through field gradient coils (Amps) rit Figure 5.4 Calibration of the field graclient coils.
fielo Z5*C from for the the were made of gradient measurements water at which strength gradients, could be Interred. coils was measureddirectly In the steady field digital a using the through the current gradient experiment

9.9 9.1 9.2 ' 9.3 9.4 9.5 9

,1

ammeter, while for the PFGexperiment the current

through the gradient coils was determined from the voltage of the pulses viewed across a 0.4. is 5.1, in table Ohm these'results and areconsisshown monitor resistor. -A comparisonof tent, to within 10%.

Source

Field gradient. G cm-1 13.2 t .7 13.5 t .5 12.4 i .8

Table 5.1

Calibration of the Field It adlent coils.

Echo width FFGexperiment SFGexperiment

Having est ablished that the apparatus was functioning adequately, the daily calibration of the apparatus was performed by measuring the self diffusion at ZVC using either the PFGor SFGtechniques, and diffusion relative to this value. coefficient coefficients of distilled water-

were measured

US

Tests were carried out previously'21 to ensure that the field

gradient, coiis did not field Bo as this ,-

produceany field gradients in the direction perpendicular to the static

would cause motions other than those along So to contribute to the attenuation of the spin0 echo amplitude, and lead to an over-estimate of the component of diffusion The methodemployedconsisted of placing two fine capillary sampletube. and separated by a few millimetres. the two tubes would then result in different parallel to So.

tubes containing water inside a

in field in the experienced Any difference

beat a cause resonant frequencies. which would No such beat pattern was observed when the fieWin -, the SFG, experiment.

pattern to appear superimposedon the spin-echo.

common axis of the tubes was perpendicular to the static

ThePulsediieid GradientUnit.

Theunit wasdesignedin the laboratory to producetwo identical field gradient pulses of triggered first the The 9. with of two was pulses variable separation delta, and duration reference to the initial by the ,,, produced go# pulse in the 90*-1-180*pulse sequence

904 be the triggered pulse, either using spectrometer. The secondfield gradient pulse could the 180' pulse, or the first field gradient pulse as a reierence..:.The delays of the leading edgesof the field gradient pulses were set relative to their respective triggers.

Thewidth and delay of eachpulse weredigitally controlled and could be varied over the to finely be adjusted range10psto 8 seconds. Thewidth of the second match pulse could that of -the first. In general the 4xl2, volt rechargeable batteries which provided the source of-the DC'fleld gradient pulses did not have sufficient time to recover between pulses In order to deliver a second pulse of the same amplitudeas the first. especially whenthe time

interval between for this the lenith of the second pulses wassmall. In order to compensate field gradient pulse was by means extended of the fine adjustmentfacility, to ensure,that

the 'area', GS(Seefigure 3.11), for each pulse was identical. and that the height of the,

116

resulting spin echo was maximised. Failure to match the pulses in this way leads to an incomplete reiocussing of the spins and therefore to an over-estimate of the diffusion coeifi-

cient.

5.6 Data Analysis.

The decay of the magnetisation following a series of r. f. pulses in both the longitudinal direction and in the transverse plane, are in general complex functions that maybe expressed in terms ai a summationoi exponential decays. In the 'FMGsequencethe height of the . magnetisation's spin-echo envelope at a time t may be expressed as;

MR)

n, Z Mol x exp (-t/T21

initial T21 imes Mol relaxation and the where and are apparent spin-spin te that total for number of components ith represents n the magnetisations component,and contribute to the signal. Brownstein theory the of and Tarr'31, although n to According

should take an infinite value, In practice OMYthe first

few terms make a significant con*

tribution to the decaycurve and orders of n greater than 3 or 4 maybe ignored. If one asTarr, Brownstein by forward and that the decayis not determined the put sumes mechanics via then but rather that eachcomponent n should different of spins, population a represents decay into to However, trying a different equal the total number resolve more of populations. than 3 or 4 components generally leads to ambiguous results. maximum of n4 has beenassumed. in this work, a Consequently

Using a Newton-Raphson least squaresfitting

routine. the transverse decaysfor n

1,2,3

and 4 were resolved using as many as 4000data points for eachdecayenvelope. The ability of the programme to resolve the decayinto these components is determinedIn part by the mat,

117

nituces or the Mots. and hy the ditterences

i the Waxation

rates of the 3epar3te com-

ponents.A'sa measure of the re! ia.,iiity of the fit eacn the standardjeviation to noise f3l ! owingeachiteration of the fit, rat. o '*as -31CUlated an Aeal fit. ';hich should be close to unity for

In the'case

at

"il me3surements

lising

a 180*-, T-90# pulse

sequence

the magnetisation

decay

isxprs

as;

n Malx

2.expt-t/T,,

Siri,:e eacn oint desc: tbing the decay has to be derive, t'rom a complete pulse sequence.Which shouic be repeated several times in order to improve on the signal to noise ratio, tewer

fiftY defined by is between Of these The thirty decay and usually measurements are made. The inaccurate. the separate f itting computer method rather points. and this makes using
dec3y's in "I'l 'measure'ment'swere thereiGre determined using irapnical longer relaxation times are successively from the curve. away peeled '-ane: eby the methos. Once again the method

reiies on sufficiently

large valuesof Mot, and a sipiticant

difference in successiveTit

vaiues.

Results of sin'i'le 'c'omponent are generally thought to t,Its for both TI andTZ measurements 4sOere be accurateto a'round in . '" t 5%. Sy incre3se this case's non'value stem atic error may

' , the to de Due ,p, t 10%. decays exponential tre to greeot are ated as single"com onent, around . .nIg . diff,, culty in sepa I by fits in graphical rati the dif ferent comp, onents multi-eiponential It the inaccuracies is methods for the Independent times vary. may usually simple relaxation "a value' -for the I. I to deri'e to within tS%,since'the'tilure canbe onge st relaxing component
'Ofthe fast relax resolved from the part of the dacayfollowing the disappearance no 'The 'fast ielaxing . ponent. is then'derived with a'inowledgeof the slow relaxation component t rate. and Is therefore subject to cumulative errors of aroundt10%. In the case wherethe

1iB

tit

is by computer, the errors in the tit are indicated by the r. m.3 : noise ratio. -

For a

goodfit.

this value is close to unity, and errors in the relaxation times are aDout 5%. is not so good, the r. m.s. : noise ratio may increase to 3-4, ana errors in

Wherethe fit

relaxation times are then accurate to 10-15%.

The basis-for the measurement of self diffusion

has PFG 3FG methods the and rates usin;

by the two equations;, been discussed in chapter 3 and maybe summarised

for SFG,

InA

(G) )l A(GzO

2 gammaZG2Drl 3

for FFG.

In

TG--O)

A(G)

gammazG-'92D(Delta-9/3)

in G these 'r or equations one can vary either To derive D experimentallyusing either of SFG the Under favourable PFG the case. conditions the SFG case, and either G. Delta or 9 in diffusion different within the self coefficients derivation lends itself of to the experiment multi-component systems.

for PEGiVater mixtures Attemptsto resolve the different values of Dp. and tv ... r 11I1 .1 1 'I " ,. In the two the In signals resolving provedunsuccessful, however,due to the difficulties for the diffusion coefficient the field the presenceof self gradient. Themeasurement of polymerin solution wasmade possible by using heavywater as a solvent.

Errors incurred while. measuring the diffusion rates are relatively high at about tIS% due the the needto closely matchthe size of the pulses In the caseof the PFG experiment, and due to the low signal: noise ratio in the case of the SFG experiment.

119

5.7 Sample Freparation.

. 7.1 Polyethylene Glycol ioilitions.,

Polyethylene Glycols were obtained from the British nameIreox. The different stored in air-tight

Petroleum Company under the trade

to PEG6000. They were grades supplied varied in size from PEG300

containers and were used without any other treatments.

NO. to basis or Solutions were made up on a weight weight with either distilled water The solutions were gently-agitated and-keptin air-tignt homogenisell-11-, days to three containers ior two to,

5.7.24 Xylem

Rose foruse by Mary in this the kindly Samples work provided of water-loggedwood, were the in bottom the the Trust. TheMary, of tested silts Rosesank in action In 1545,, at and in time During late this began the 1970s. work solent.until -the restoration and reclamation the degradation wood. of the become and has water-logged, underwater wood severely naturally depending has taken place tovarying degreesthroughoutthe wreck, and on the type of wood,, its exposure to the elements.,,, - .1 -1

from a variety of Divers whoexcavatedthe wreckhaveretrieved samples that weremade that werecommon in Europeduring the fifteenth and sixteenth centurY,,suchas oak, woods In treatment, ash, poplar, birch and so on. Thesesamples to prior a large casstored, were cade systemof baths, wherefresh tap water could flow aroundthe samples,flushing awaythe salt water and silts from the wood. Following this they werestored In large tanks of water, to which a small amount of fungicide wasaddedto reducebacterial growth.

120

3mall,blocks of wood were cut from planks treated had been in this manner, that beams and and were sealed in plastic laboratory. to maintain their moisture ccntents before being sent to the about gcc,. and the species of woodwere on Most of the in

The size ofl the blocks was typically the different cellular

identifledfrom

patterns. which,-, dere clearly visible

samples. On arrival their plastic

at the laboratory, the woodswere piaced into fresh water, tank still

Jackets.

", I 5mm., in diameter and cores, of wood. mmin, length. were carefully zmal

taken from these

blocxs so 'that the grain was as near parallel as possible to the long a'xis of the cylinder. The Samples checked for obvious defects in the rejuiar cell patterns, and were then lwere thoroughly washedwith distilled before each to remove any water resioual soluble material. the days; four for of upto, water periods

was placed into 3 separate container of distilled

distilled water being changed daily"'.

by carefully dabbingthe Following the soaking period, excesssurface water wasremoved Into inserted then The weighed and tissue jently. were wood with absorbent cylinders paper. Insert teflon diameter, the baseof a sampletube with a 7mm glass with and a sealed outside fitting tape wasused to plul the tube near to the snugly, sample.thereby reducing the space,

This from the froa the was to wood. around, sample a minimum seaping water preventing and 3 , wasraised aboveroom particularly important In caseswherethe temperature of the sample be The, tube. temperature,, could from the sample wood. at which water wasprone to evaporate Ion further sealed either hermetically, or by means tape. After tef and more of a stopper had beentaken, the samplewasre-weighed eachmeasurement to test for loss of moisture.

121

I Between the sampleswere allowed to dry. measurements this is a natural process once the woodis exposedto air,

With fully

water-logged samples

but in drier samples the evacua-

tion at 4ater could be accelerated by the gentle* application of heat using a 60 Watt lightbulb. After a period of a 'time the sampleswere reweighed and re-sealed. lowed to equilibrate They were then al-

for a period of' twenty-four''o hours before further measurements were made.

the the to relaxation samples on ests were carried out establjsh'te' on effects , 'many ferromagnetic by impurities. In casesthe Wood properties of wood caused para- and broughtup from the remainsof sea-faring vessels havebeencontaminated with iron compounds that haveleachedfrom the iron workon the boats, such as cannons' etc can Theie'^. ompounds

-with EDTA Tetra-acetzite)SQlu(Ethylenediamine be extracted to a large degreeby-treatment by further washing. tion. oich forms various comp! qxeswith the iron that cz_ibe' removed treated between for NMR decays differences the samples to be no measurable therp nztared -dith EVA, and those that werenot.

PEG by Impregnation of was submerging grades accomplished the Yarious of samples using in concentratedsolutions that wereperiodically the samples for periods of uDto 18 months that In increase by the degree The to weight indicated slight refreshed. a uptake of PEG was the samples during this time.", q) underwent

Theabsolute moisture contents of the samples at the endof eachexperiweredetermined for six to twelve hours, or until mentby completely drying the samples In an ovenat 1050C the mass had become of the sample constant.

122

Re fer'ence s:,,.

II

McCall D.'J.. Douglass D.C.. AndersonE.'J. 336. (1963) oassall S.R. PhD. Thesis. Nottingham

Bar. Bunsenjes. Physik. Chem. Vo1 -637. p

3.

Brownstein K. R, Tarr C.E. Physical Rev. A. -rporation. U'nicn Carbide Co (1972).

jol 19(6) P 'Z446. (19,19)

Tech. Bullet 11 in F-2933. wpolyox water' soluble resins. "

S. t.

I . Union Carbide Corporat ion. ' Carbowax catalog -u es, Murray H. "Conservation of artifacts from the Mary Rose. ' Proc. ICOM water-log-jed

Ottowa. (1981) woodworking group conference. 7. a. 9. -nshin A-%. DeZaeuw Pa C. "Textbook of Vood Technology.' Stamm'A. J. Stamm A.J. For'. Frod. J. For. Frod. J. Vol 3p 27. (1957). McGr3w-Hifl- NY. 1980.

, 61) jo'l 10, p 41 V _4.15

123

CHAFTER 6.

A Pulsed IIMRStudy oi Water in Water-logged 'wood.

I.

6.1

introduction-,

I.

in The 1-4MR 4. previous studies properties oi water in woodhave been reviewed in, c, -,apter _ the moisture contents of woodhas beenmonitored under equilibrium conditions at relative humidities of 100% madeon measurements. and below. This chapter presents the results of NMR samplesof water-logged woodhaving moisture contents far in excess of those normally encountered. and In which the celLvall structures have been broken cown by long degradation

processes.

in deiree In general the NMR by the of wood water are of ccmplicated properties

.-

and heterogeneitywithin the samples. Thenature of interactions bet-deen the water molecules degree in by determined of the and the tissues of the wood part structure are cell walls NMR degradationof the wood.This is reflected in the behaviourof measurable as a parameters function of moisture contents and temperature,suchas the signal amplitudes, the relaxation of the magnetisationin transverseand longitudinal directions. and the self diffusion coefficients. Thesehavebeenused in order to characterise the moleculardynamics. of the,water

in water-loggedwood.

In The samples they fragile a saturated conare maintained are and unstable, and unless dition strong capillary forces to tissues The the collapse. into woody cause come play which shrinkagecan be.arrested by the action of water-soluble polymersimpregnatingthe system, whichnot only consolidate fragile cellular components, but also stemthe flow of water from the wood as,a result of their hygroscopicnature. The molecularaynaalcof the'Wate r nd -a POlYmer within. the wood the mo. lecular weight of are modified in different waysdepending, -upon the polymerused. Different grades penetrate different parts of the xylemcells, and the effects on the -water'in the samples maybe monitoredby NMR techniques.

124

Broadly speaking this work has addressed Ahe following questions: -,

Howdoes water in water-logged wooddiffer-from

bulk, in the or from water, that,

found in normal xylem tissues. and are the NMR properties of the water in waterlogged wood Indicative of the degree of degradation of the samples? Does water in water-logged woodexhibit similar multi-exponential behaviour to that

found in other systems, and if so can these be related to the macroscopic structural properties of the woodsuch as cellular the Are perhaps composition and pore sizes?

NIS properties of the system consistent with the existence of 'bound' and 'free' phases. and is there any evidence of exchangebetween the two' 'ihat are the effects of the impregnation of polymers on the nature of water within the wood, and how do these effects vary with grade and concentration of polymer used? factors effect the self-diffusion does the how the and system, of water within

impregnation of polymers modify this behaviour? (v) Can NMR be effectively the for the treatment in rejuired parameters assessing used

successful preservation of water-logged woodenartefacts.?

Jn an,attempt to answerthese questionst three types of NMR experimentswereperformed: in treated 1) Freezing curve analysis. 2) and untreated of samples of water relaxation INMR water-loggedwood.3) PulsedField Gradient measurements of self-diffusion coefficients.

6.2 -,

Freezing Curves,

Previous have that water in proximity to a macromolecular studies shown -systemexists in ,, a number of fractions that can be classified as 'bound' and 'free"". In fresh everyday

samples of wood'free' water is only in evidenceabovethe fibre saturation point of about 0-38 gramsof water per gramof solid'21. The 'bound' phaseIs the water of hydration of the

US

sYstem which is intimately

related to the constituents of the woodcell wall. Tests were

arried out to establish the presence of boundwater in water-logged wom and to see how it : relates to the hydration water found in fresh timbers. In other systems it has been found I that the quantity of hydration water is related to the amount of substrate material that is accessible to water"'. binding sites'to The effect of drying the sample could cause the numberof accessible surface area

diminish as the pores in the sample collapse and the effective

within the sample is reduced. If there is no change in the numberof available surface sites. and the bound water bound water should remain fixed during the drying process until all of the 'free' 'been free The water bound to the removed. ratio, of phase"has of quantities a

followinj D. I. F. te freezinj found the the curves, whereby the amplitude of can with aid of K degree pulse is measuredas a function of'the temperature of the sample.

is D.,, I iant_nuclei F. The immediate reso, oj, proportional. the number ampiitudeof., ., _to 'the
within the sample that are capable of makinga contribution direction the transverse In the solid state relaxation time to the signal. For protons in the takes place on a time Sc3le of

for the time dead the of many receiver to mechanism is Several microseconds, which close spectrometers. If the first tens Is in D. I. the F. times the of order made at observation of

have decayed awayand will not will then from the of microseconds, the signal solid component be Included in the signal, the,,only contribution to the amplitudebeing from protons residing in mobile moleculeswith longer relaxation times.

liquid from the to transition the therefore Freezingcurves are capableof monitoring solid states and vice-versa across ireezing events. Figures 6.1a - 6.1d showthe freezing curves for water in water-loggedpoplar as the water content of the sampleis reduced.The temperaturewascycled between 230Kand 300K,and the amplitudeof the signal wassampled at 70 microseconds,. Corrections to allow for the change In the Bolt: mann factor with temperature. and for the increase in sensitivity have of the receiver unit at lower temperatures beenapplied to the curves.

126

LA rL 0,1
temperature Decreasing Increasing temperatare

:1

0. e1 ignal Amplitttde

0: w 08 12
4' Impepaturi W for voplar. Samvle irampts fil (a) Frenimcurve Figure weight 9191 242

Amp li tadt 0.5 0.4

F 0.6 Sipat ,
IF

220

231

E42

E E65 2? 254 W 'Temperature

2S8

29

310

6.1G) Figure

BJ55 Heigit gpam'4so : apita Partiany sampie,

266K for Onthe downward about samplescon at cycle a clear-freezing event wasobserved taining high moisture contents. This indicates that the bulk of water within the samplebe-havesin a similar wayto ordinary water. The depressionin the temperaturefrom the freezing point of distilled by supercoolingof water within the sample. Belowthe water is caused

127

am L. k3 .
019

0.8 0.7 0'-6 Signal Amplitade INr, W. W 0.4 3.3

temperature 0- Decreasing
Increasintwerature

0.
011

ti

16

0.3

4E

41 Temperaiure W

. 174

13

TiSure61 (.

Foeezing weight= ISO gramAs cameaf tv tupther h-jing. SaAple

1.3 a.9

Silnal Amplituce 0,3 0.4 0.3 0.2 0.1 0.0

Decreasing temFerature te, Incivasing 4peratut

231

242

27E Z!426 W Temperature

Figure 035; 64 (a) Finalfreezin curve.Sample gramNs : weight

freezing event there exists a significant quantity of water which is left unfrozen. This is the bound water that remainsin a mobile state through int'aractions with the substrate. The amplitudeof this signal decreases steadily with temperature.suggestingthat a smoothdis. i. tribution of interaction strengths exists, so that there is a gradual freezing of the bound Onthe upward component. cycle the frozen water thawsat approximately273Kas expected.

128

As the sampte. aries the signal bel. amplitude at event

and the

tota,

number of water above freezing little attenuation.

molecules is ii2inishea. it 6.2 is

within

the sample

is reduced, the signal water_tat is

temperatures

The height, -of the, tree

the rrsezing

shows very

therefore

is zrimarily plottel and free

removed from the weight

the wood-on drying. of the sample. Me shown. It

In figure,

the total to the total

signal signal

amplitude

against

Contributicn that

t. rom the bound water is

COMPOneDtSare also the sam; le. before of Ahe

douid appear

nearly

all

of the free Mo indicates. 6.2

removed from driea

any-of-the,

the bouna, water Js or the broken line

lost. in

the Oven the is

weight

sample. the free It

The intercept

figure

indicates

point that

at wnich

all

water

in the sample has been rem6v oved and the remaining gives 0.4 a measure of 1/1. the fibre saturation error point

water

in the bound phase.

thererare

of the

water-logced

wood 4nicn

is 3; proximately

and within

experimental

is equivalent

to tne value found in fresh 'wood'".

silnal

freezing eyent am'plituk lboye

01 f,

Zr
.0

freezing WOW event I- amplitude

0- aifference watep non-interacting . 0.4 02 0:1 0 25 MO I .001.

lei* lea 125 150 175 200

( milligpamms S114ple weight

Figure 6.2 Signal interacting components amplitudes ana rorrinteractini of , in waler-logya weight. of sample Nooa asa tuncticn
-The observedsignal has beenrelated to the total water content of the wood.However It Is ccnceivablethat the signal, is enhanced by a contribution from the moremobile protons residing in the solid content of the wood,or, reducedbecause water moleculesthat are tightly boundto the substrate relax at sucha rate that they no longer producean observable

129

itude. grows for 3 samole of -oven-dried : ontribution. Fi,, ure 6.3 showshowthe signal amol poplar as it is re-hydratel. FOr -Omparisonit is plotted alongside-the signal amplitude ob. water. Ulthin exzerimental error the two curves

served trom different -quantities of distilled

are the same, but displaced by-37 mailgramsalong the X-axis. This displacement corresponds to the ven-dried weight of the sample, and one mayconclude that the observed signal arises

from the -dater alone, with no contribution from the protons in the wood.

signai C500 rmplituie EOM IM 1000

%0

aCo
a

25

'15

ice

122

130

17 5

zoo

sampleweight ( milligraAms, ),

fcr function Figure 6.3 Signal amplitade sample weight re-hgaratea poplar. of is a to thoseobtaineafor iistilled oater. Fesultsare coNparea

that drying the s5 gs et T su Re fixed quantity of bound period the over sample water within the shrinkageof the sampledoesnot significantly effected the number of-accessible binding been the removed., the of water-has sites within matrixof the substrate material until most

the is openstructure degree to of, a mark binding the, water of accessibility of sites ihe

ToAest, for exchange Of-thesYstem. between, the free and boundphases,and between protons-,,t 3ttached to the xylemand in the water experiments werecarried out on deuterated samplesi Figure b. 4 represent,the freezing curve derived from a, sample of poplar-that wassoakedfor eighteenzonths In D20.-During this time one wouldexpect,the NO moleculesto establish an equilibrium with the-system,, so that deuteriumresides-in the bulk, and on all sites that are accessible to water molecules.Theproton NMR signal. amplitudeobservedin, fully water-logged

130

sampieswas reduced. but the residual signal, is consistent with fully

protonated samples.

with a ratio of bound to free protons of 6: 1, suggesting that the deuterium nao penetrated both free and bound phases to the same degree. As the samples dry the free water molecules are preferentially til removedtrom the wood. ieaving the tne majority of bound water in situ un-

it is 'Cheonly water remaining.

M Signal amplitude 0.4

swle weight 9.4213 1 9,0679 g

40 C; . 1; .0

NO

2. "1 0 Es .dd

ItmPerature (N deNdration aeuterated durini yoplar Fivre 6A Freeziag of olserNed curves

in be detected, then vial a Thesample NKR scaKed could and dried signal was. out until no figure in free., ing is The shown days. deuterated curve resulting of waterfor a periodof- 10 6-5. Theprotonsignal amplitudes Indicatethat the,8:1 ratio of protonsIn the free and in detected the bound the re-swollen has been The signal of phases amplitude re-established. before low in contents increased at moisture sample fromthat-observed samples significantly by that to the sample is dried. assuming prior fully This in increase explained was amplitude drying the exchange had deuterium the an equireached between in the and protons xylem fixed librium. Theconcentration throughout the in the remained the wood of protons waterof drying process.so that at lowmoisture of waterprotonsapproaches contentsthe number zero. When further D20is added to the system to the xylem onceagainexchange protonsattached with the deuterons, so that there is an increaseIn the number of protonsin the liquid phase, in the protonsignal amplitude.Themass andhence attainedby the re-deuterated samples wasless thanthe mass despitethe lengthof the reof the original sample

131

aeuteration period. This Indicated that the ability

le to hold water had been the samr; of of the

reducedby the extensive drying, probably due to somet:ollapse ot the uittastfuctute cell walls. P

i.
I3

1 7' Sample weight"

I 1

IIII

Signal 0.6 -amplitude

0.4

40

2!0 Z60 ltnperaire

a702.

02.9

300

" ' Fivure 6.5 Freezing for Proton re-deuteratea saisple, signal, curves

in that From exists wood conclude the results performed water water-logged can one above is low This fraction in at least twostates, oneof which tesceratures. at unfrozen remains fresh Umbers, in identified with the waterof hydration and is less readily removed observed in Is fromthe sample which The one wood open drying. a very upon structureof water-logged there are no. fractions two The with each inaccessible exchange undergo observable regions. to-the between and xylem other, andthere Is some attached protons chemical evident exchange thosewithin the water. Irreversible denaturation place. until of the cell walls doesnotAake, from, been has the, all-of the non-interacting,and, interacting removed the wood. water much of

132

tluclear Reiaxation at -i-Ater in 1.1irar-logged'; Cods. ; iiagnetic Resonance

6.2.1 Sine Measurements I e-Component oi 11 and -.2.

lhe spin-spin and spin-lattlice

reiaxation. times of water in wooda: re sensitive to the

temDeratureat which tfey are measured,ano on the water : ontent of the wood. 'I 7heYdiffer from the relaxation times observed for water in the bulk. being reauced by typically 2, ratios of bletween Z and of ten or mote, and with Tl: T. 3 factor

In the case where bounaand free

Idw. by is phasesexist in rapid exchange3 single relaxation time is observe cetermined nicn he bound ; hase. and the intrinsic Ve Uantity oi jater in +. ticns. These qu3ntities ffaC7 two the times ci relaxation

determined by the zacroscopic and micrOsclopic themselves are

Is difterent, in tne the of motional responses Celluiar system, water molecuies. which provoke
it possible then- to deri'veyny iniormaticn from the, values_ot TI _3nc TI atout the, wood/vater

interaction, and in turn about tte ceilular compositionof the woodin relation to the degree of degradation?

Minispec on samples of -doodwhich were waterMe3surements, were madeusing a ErukerPC-Z logged and degraded to different hether a two phase mooel is viable degrees, to establish w. This device

and if so to see what addit'ional Information"could subsequently be inferred.

which operates at a, constant, probe tempe'rature. 40 degrees centigrade can be,used to make of__ Simpiified measurements of sol id: I iquid ratios',, single, componentspin-lattice times. and'singie componentspin-spin'telaxation times. ''' relaxation

Cylindrical

samples of water-logged wood fros a variety of species were cut from blocks

1. found on the Mary Rose. The , samples were placed into, the Minispec and allowed to re ach a

thermal equilibrium with the probe. The spin-lattice,

spin-spin, and solid: 11quidrRt! Gi'were

133

-Ato
.......................

..............

11&T2 ( msec
,0

.....................
v............ ..............

............

..................

...............

..............

........... .....................................................

..............

............

.......

til L02

. (0) anaSpin-Lattice(1) relaxation timesfor water.in,water7loggea Figure 6.6 Spin-Spin Wicker,Jnitial water content 7.6591g,

I H, It I MO ( g/g

II-,,

-00
........... .............. .............

ISO ........................

......

.........

..........

..............

..............

msee

-Lo ...........................................

........

..............

Ile:

.....

......

......

..............

..............

.IIIIIII.

I.

in ter for Figure6,7 Spin-Spin times W (0) andSpin-Lattice relaxation water-logged wa. 7.37 Ashas & functionof,witer content.Initial watep = q/g content

I H20ItJ MO ( 1/1

measured as a function of the weight loss of. the samples as they dried out. The sampleweight immediatelyafter eachset of measurements wasmeasured hadbeentaken, and the oven-dried, weight 'wasdetermined at the end of the experiment.

134

The results of relaxation measurements madein this way are shown in f1pres 6.6 - 6.9.

Thetimoerschosenincluded three highly delr3ded of poplar. asn, and wicker. and a samples
PieCeof water-logged birch neartwood. for which, the degree of aegradation was less severe.

....................

...................

..................

Ti, 12,

leo

............

..........

................

...................

..............

..........

....................

............

...................

.............

0,
0

...................

........

..........

3.0 4.0 H20It/ MO ( q/q )

313

Figure6,8 Spin-Spin (*)relaxation inwater-loggea times. forwater Spin-Lattice (0)ana Initial 5.40 Poplar function content. water : water content Y/q. of asa

x, Z

F T'

240
.........................................................................

.....................................

.........

T2
ra0 60 0 0.0

. .........
......................... ...................

................
................................... .......................

.........................

0-1 .......

............

0.0 1.6 1.2 2.0 I H20It/ MO Vq) Figure 6.9 Spin-Spin (0),& Spin-Lattice N relaxation for water in water-loggea times BirchHeartwooa Initial water asa function of water content. content : 1.79 g/g.

0.4

135

An indication of the extent to which-1egradattan,has taken place' is given by the Of -ratio the, initial weight, of the sample to the oven-dried weight. In all cases where the moisture

content is greater than't). 149 Of water per-,,ram of, woodboth the spin-spin and spin-iattice relaxation times decrease monotonicly with water content. -,here is a considerable difference bet-ieenthe values of T11 and TZ for all of the samples. with T1:TZ ratios varying from about 2 to S.

Fc., the three heavily degraded samplesthe relationship relaxation times is a linear one at high water contents. two : nase model used by Caries and Scallan"If

between the water : ontent and the This linearity is predicted by the

Assuming tnat one fraction of 'dater exists ir-

a boundrnase. and a second exists as a, free componentin excnangewith the bound fr3ction. then using equation 4.9 of section 4.4.2, the gradient snould ; ive a zeasure of the value T^d T1 T Tib or of the bouna traction, represents either lb/NzOlb, where represents and CH301b all of the free

to .e constant until the ; uantity of bound water present, wnich is assumed water has been removed.

It the conditions exist underwhichequation4.9 are satisfied, then one would expect the intercept of the graph to be at the origin. As can be seenfrom table 6.1, however,there is a broaddistribution of values for the intercept times, and while this condition appearsto In the hold goodfor the sample for the the three. samples other it does other not of poplar, lowered. Is bend-towards the points on the content the water crigin'as graph

Equation4.9 is a Special case of equation 4.8,, which can be re-written as,

CH201t Tobs

IHIO]b (111101t-114201b) ....... . ................ Tf ...

136

IR201t and'JHIO)b are the total and bound water contents in the wood. Tabs. T...... 'where are the observea relaxation time, and the relaxation times ofthe phasesrespectively. Umes. -

and

bouna ana tree water n

The equation equation holds for both spin-spin and spin-reiax3ti;

Sample Poplar

intercept of T1 urve 3t5 'Isec

Intercept of T2 curve 1 msec

Ash
'Jicker Eirch

I +-!: sec
33 tm "C '9t5

4t4

msec
msec'

85 t 20 msec

12 !: 10 msec

fits to T1 andT2 Table6.1 Intercepttimesfor, single component data as samples are dried. Ejuation. 6.1 may be simplified to give.,

Tobs--

Tf... (HIOR x To ... ,x ------------------------------------------Too., OIt [H, CH201b x # x (T,,.. -T.... d) +

NOR

IHzOIb 6.2

by to that used the then P. above equation reduces is rovided that To small sufficiently ... g, Caries and Scallan's'.

Thesolid lines drawnin figures 6.6-6.1 are plots of, Tobsas a function'Of [HIO)t using '. T..... T,,.. 6.42. lines the three The, to and variables equation of the are sensitive 'form .r 11401b.. In eachcase the quantity oi boundwater wasiound to be close to 0.141 of water pe 'O&iw'a`ra"61 found bound by favourably This gramoi wood. water compares with the quantities of in wood by Boesen"' for-cellulose in cotton (0.2g/1), and suggested 1, Pulps inconsistent, with the value found for, samples of, fresh1imbers"'. _. b'ut'is

137

Table 6. presents the, fitting _

'parametersused for the curves. For the spin-lattice consistent. with Tlt ... in the order of 0.7 seconds.

"It'im'es' 'are' relaxation these reiativeiY and TI......

z9 milliseconds. ' In ash-the value of T2,,..

appears to be close to the value of

suggisting that' it has properties like those oi bulk water. where as in 'jicker and in papiar there the T1f : TR ratio is approximately 8: 1. 'within experimental error the two-phase
model iccounts weil for the' data in degraded samples at high water contents.

Samp ie
Vicker

101iflo
7.6-5

Tit,.
700

Oda TIN

TZ,
90

a 3

Ash Poplar Birch 1

Table 6.2 Fitting

7.37 5.04 1.79

700 1000

9 9

EOO 120 -10 1 9

using

1 11 variable varaible

lines in figure, jor solid parameters 0.14g/g. EH201b 6.2. = equation

At .low moisture contents the signal to noise ratio, does not permit accurate measurement

I for, Values TI Minispec. ash with the of obtained were 0f the spin-spin relaxation times using from is to believed In0.14J/I, the originate signal ' where moisture contents'that are lower tha through TI, a miniIs passes that bound reduced content-is to the the As water moisture wood. for by samplesof 0-3 0.29/9. been authors has of A similar minima mum reported a number at pulp. Child"' cellulose derived from-a variety of sources, including wood made a detailled

'As the moisture content is reducedthe mobility of the water in the study of this phenomena. the is the interaction decreas'ed because Increase in samples macromolecules. with rate of an , Abovethe minimum bulking the system, and Imfree'water exists between the macromolecules, parting a degreeof mobility to the macromolecules.. free the is amount of water reduced -As the boundwater molecules are no longer separated, and are expectedto, form btidges, between the macromolecules bondingwith hydroxyl groups.This causesa reduction in the via hydrogen mobility of'the'macramolecules,This is reflected In the lower values of T1. As the water

I- 138

torm that to is further are the bound able ontent reduced. number of ,: water molecules bridges betweenmacromoiecuiesis'diminished. The numtevot-briages tails. and there is-

therefore in increase in the mobility of the macromolecules.and hence of the bound water,

causingthe TI to increase.

An 3[ternative. interpretation,. of,, the-minima.in t-he TI.

-data

invoke be spinto may

have the, bound' a mobility In the would exchange. absence of spin-exchange water molecules in the region of their T7Iminima. in contrast the cellulose protons placing them somewhere 'r-i-gid"'I"it, . -T1 , ti-ce 1. have 3 mobility putting I them under conditio ns, e. 'Tl-minimum, an '12 ZZ Thus the spin, As the tCtal

10 - 20 oec. 'Spin-excnanje occurs in 10 - 30 psecs, and is very efficient.

lattice relaxation time observed is dominated by the value for the bound water.

'water in althe iystem bound the proportions is amount varies the of water reduced amount'of T1. influence, increasing the observed ter. and the solid protons exert a greater

break to theory degraded"the two-phase appears less is In birch heartwood wherethe wood I initially Is TZ There a In the of also. case down,certainly in the case of TI. andprobably is increase arrested -This the in the contents-Increase. times water sharp rise as relaxation Is there a point which at 0.4g woodof gram at moisture, contents of around of water per increase. to begin plateau.,At higher moisture contents the relaxation times oncemore

Thepoint at which the behaviourof the relaxation times appearsto change,i. e. 0.4 coincides _with in Northern 1,41 Hsi by all et the measurement of the fibre saturation point

function times as a In 6.10 figure the measured white-cedar chips. of spin-lattice relaxation water content havebeenreplotted. The solid line has beencalculated using equation 6.2 by that the values of the spin-lattice relaxation times for the free and boundphases, assuming and the quantity of boundwater, are functions of the moisture content. Thenature of this curve is relatively insensitive to the value of T1,,.., provided that it is large enough.It Is motesensitive to the values of IH:O]t and TIk... 4. Thevalues used to calculate the curve

139

"40 U. 0 -I 120
...........

TI msec

................

..........................

.............
......................

.......................
..........................................

......................................

..........................................

0.3 H20 It/

112 Mo ( g/sr)

2.3

......................

.............................................................

iE 11 Bouna
24

..........
...............................

..........

..............

..........

...................................

...............

....................................

................................

L-

0.0

EH20 It/

Ma (9/0

4 ...................

................... I

...............

I H201t ( g/g )
I ml -. -, . 11 111.. .......... .................

11

!I

____j

0.4

0.3

1.2

1.6

2.0

(01' birch (A) Figure 6.10 7heoretic-al Elt0lb/lItOlt (m) in bound TI and changes . W.
heartwood, used to predict nature of Ti observed

are shown in the figure. Thevalue of Tib,,,,

increasesalmost linearly with the water con-

tent. probably as a result of the Increase in mobility of the boundphaseas the structure of the woodbecomes moreopen. Theproportion of the total water content that constitutes the

140

bour-aphase is initiaily,

nigh. but droi)s dramatically as the water content of the wood tends ' of total watet held in the the proQrtio pn

towards 0.11.4p 1. Between0.'14 gi q and 1.4 j/j, boundphase is relatively

declines. it 50%. beyond steadily constant at around which

Both trends were obtained theoretically experimentally by Hsi et all"'

''nds tre Tobserved were same e via equation 6.2. ,, that were ob-

T1, TIb the although and magnitude of ... , ... ,

the watertheoretical be by lower is Hsi to 3s than the values. expected -This served are logged woodconstitutes a far more open and less' 'rigid" structure, contain far -nore mobile water molecules. likely Is therefore and to

dlitribution lends the CH: Olb/EH201t T support behaviour model of moisture of 1he Ne I Fro by ii son' proposed and four uptake water of existence of stages the

based n

In the dry

interby held together state the system exists as a jetwork of macromoleculatchains , mol ecular hydrogen bonding. Uoon the initial hydrogen nondinj ocof water. strong addition structure. be may which

the on c eilular sites between surface cuts water molecules and active

In the a exists wood hydroxyl groups. During this stage the majority of water within 'Primary' bound phase.

further ' water but of the addition 0.14 g/g the surface sites are wholly occupied, Above further revealing causesthe networkof macromolecules to swell, openingup the structure and the this During proportion region bound by 'Secondary' water. active sites which ate occupied of. the water boundwater In the woodis relatively'stable, swelling is proportional-to-the water content. degree that the of suggesting

At around 1.4 g/g the swelling of the structure is slowed'down, and fewer active sites are'revealed by the addition of morewater.' At this point the proportion of water in the boundphasebegins its decline, causing the free water In the woodto dominatethe behaviour of the spin-lattice relaxation time. It would appearthat the secondary boundphasedoesnot

141 -

exist in severely water-101ged wood. suggesting that the severely water-iogged wooddoes not experience a swelling'phase. The quantity of bound water per gram of wood at the point of

saturation for the secondary bound water, is 0.335g/q. This is, consistent with the fibre saturation point for fresh timbers. The deterioration in the structure of the cell walls of

degradedtimcer lowers the fibre saturation point in the water-logged sampl s. _e

varies ana -allan"I

suggestel that the bound phase exists aS3 monolayer of,,destruc, A_

tured 'dater on the surface of macromolecules,and the quantity of bound water istheref0te

measureof the acc'essible area of the wood. The area zoverea by a single water moiecule was calculated is where 18 is the molecular weight of 'he -water Is N molecule,

Avagadro's number, and 1" is the specific volume of the destructured water which is taken as i; nlY 0.9 cc/g. Each gr3ai of 'bound water therefore covers a surface area of 3000 mz. In 6, degradedtimbers the accessible surface are is approximately 440 mzig, and in the heartwood 50 timbers the accessible surface area. following this line of thought As approximateiy- 10.
m2/g, but it Is possible that the water in these samples resides exist in multiple layers that are

he by supported capillary

system that may still

in well preserved wood.,,

6.2.2

Multi-component Relaxation of Vater in Vater-logged Vood.

The results of section 6.2.1 ate analysed in terms of. 3, two-component system'. forwhich the observable Tis and T2s are treated as the weighted averages of thesp components.. The values of TI"a'nd T2 were found to be functions of the degree of degradation and the total water content of the sample. However, - to analyse relaxation in water-logged wood In this way is to adopt &'simplistic approach as in general the behaviour of the mainetisation following

a sequenceof r. f. pulses Is complicated and cannot be adequately, described In terms of single componentexponentiat'decays.

142

The results -of multi-exponential

analysis ate presented in this section.

In studies of

water in fresh timbers, two components of 'relaxation have been observed. and related to the ' Here we-question whether the two componentfit molecules in the bouna and free phases. sufficient is in

to describe the relaxation curves, and If so whether this assignment Is valid. systems where multi-exponential

other water-logged cellular

behaviour has been observed the

lations -of -water that, are res ident In -dIfdata has of tan been-interpreted in terms of Dopu

'in wood' different enferent -. the there' sample. water-logged numerous are, ompartments within in be relaxation different lead seen populations that that of water could could vironments %to distribuTI T2 size Perhaps different the the pore and are related to values of experiments. Perhaps the suiti-component of compartmen" tion Athin the wood^ decaysare not a consequence It mayalso be possible that multitation or to the existence of bound and free phases? interpretation is itself incorrect. component

irea resonant using made Measurements of water-loggedwoodwere on a number of samples the , degree degradation Ie were of wood of and , content 17ft tempe th ture, quency water, of as ra , It In two was into general components. or more The varied. magnetisationcurveswereresolved ' takes fit in the improvement Some is satisfactory. found, that the fit using two components Wa and Is likely to of consequence is three this but ' slight, place with or, morecomponents, the'greater-number of parameters used In'the'fit. - 11 9 1

180'-1-90' the pulse Themeasurements usinj times made were of spin-lattice relaxation Carr-Purcell-Helboom-Gill the pulse sesequence and the spin-spin decayswereobserved using is less quence.The resolution of the longitudinal magnetisationinto two or morecomponents reliable than in the case of the transverse relaxation experiment,as accurate accumulation heavily on the reliability of data is time consuming and depends of the apparatusover that

period of time. Since the magneticfield and amplifier gains are prone to change over a period of hours, results are often ambiguous.

143

Thecompcnents, of spin-spin relaxation, can be extracted with moreconfidence, as upto

4000sampling points can be,observedduring the decayfollowing a single pulse sequence, in the data quality and reliability. Transversedecay times have giving a vast improvement therercre beenthe prime focal point of this work.

IiIIIii

M9' ; M(O)a exy(- t/12a)+

WO)b exy ( -t/12h)

Hahn SpinEcho 40 20
0

N, "20

' i, 'Q 10, 610 0"1,-,15 200' -'. - 7ce -'00

( Asec Time

Jecaq transverse of relaxation figure 6-11 Non-exponential spin-spin observealn ash. water in water-loggea
in for directions longitudinal water Themagnetisationdecaycurves In the transverseand the log-normal The 6. observed lz^.. of plats 6.11 in figures water-loggedwood and are shown The nature. their multi-exponential magnetisation against time are non-linear, indicating decayis expressed by the function; magnetisation n Z MI(O).exp(-t/Ti) 1=1

NO

6.3

deviation by is the fit to the s. wheren=1.2 or 3. A measure r. m. given of goodness of noise ration. For an ideal fit this value should equal unity. In water-loggedwoodthere is a distinct improvement in the fit for nz2 over n: 1, but the further improvement in the fit

144

wherenz-3Is marginal. - The spin-spin and spin lattice

have therefore been times relaxation

analysed in terms of two and sometimesthree components., The value of Mi(O) is a measureof the number of nuclei within the 'phase' relaxing at the rate I/TI.

height Pttlse following

* III iI\

r.f. x1se
-.1

L sequence, '0
%. A" 11. M0

4CO tNO SOO INno 1 0,111 Time( Nsec) in JecaU longitudinal for Figure Spin-lattice 6.12 Mon-exponential relaiation waterloggetl 'ash.

.%1

00

1SCO

Voodas a function of 6.2.3 NMR Relaxation Timesof Vater in Vater-logged Moisture Content and Temperature. -

of 300K Figure 13jashowsthe behaviourof the spin-spin relaxation'times at a temperature The intIs for water in poplar, as the amount reduced. steadily of water within"the sample tial water content of the poplar is given by

(t1I,t.

IM.,. A

"

145

T the sample is relatively

well preserved in terms of the amountof solid matter still

residing within the wooc. The data has been resolved into two components. 'the shorter reI3xation time is usually attributed strong interactions with the to. nuclei within molecules that are motionally restricted by

'substrate,

wnile the componenthaving a slow rate of relaxation

is assigned to a more mobile fraction4.

I.
.......... .............

A,
................. I ...... ..........

............

- IN

...........
.............. .......

0 3-hort comporens =MXrer-', ............... .............................................. ..............

..............................
.............................. ...............
A4

........... ............ .........................

1.Iq 11.2 (M-Mo)IMO (g/g)

4 "E.

runction in times Fijupe13a. Spin-spin a as content. poplap of moistupe relaxation 3.9 off. initial fit. wateP content Two comment Moo
4

BGBO

.......................

..................

.............

rt c&xrent mponent Long a : 6660 1., .................. .......................

''' 4440........................ ...............

.............................

...................

..........

..................... ......................................... ..................

................ ............... ....................... ...... .............

Ulu U.W

FigureC-13h Relative in-spinrelaxition in joplar S? amplituhsfor twoconponent hving aninitial watercontent 3 of 9/g.

N- No )/ No (1/1)

146

The figure also shows the amplitudes in arbitrary units of the two populations (figure 13b). The most striking "fe3ture of this' graph is that the 'population of the rapidij relaxing

is muchgreater, than that'of component

the-slowly relaxing i:omponent.From the analysi's of bound fraction to contribute ap-

the'treezing curve data one'iould have expected the tightly

proximately" 15%of the total signal. The two fractions observed In the T22 measuresents'do not thereftre correspond to'the 'bound and 'free' curve experiments. fractions that were observed in the freezing

In addition the relaxation time of the long componefitis reduced from that

is considered that both componentscontain bound and free water. of free water.,, Hence ; it

Abovea water content LE jig the most readily removed-dater is that which resides in the 'mobile' fraction. At 2.5 g/j, however, there is an increase in the rate at which water Is

to harder becomes the in fraction the from the-limmobilel fraction. and water moblie removed time by transition In of the relaxation from This a the is spin-spin extract matched wood. the Immobilefraction which f3ils at an increasedrate, with water,content belowthis point.

Therelaxation time of the mobile fraction maintains a constantyalue of around215 millow The 0.6 between 0.35 Ilsecondsuntil the water content'reaches I/g. I/j and a value of defined be to actransition more the data point does number the of points on not allow iraph
curate I y. Once3gain the transit Ion coInc Ides 'with'a' change'ln -the' mannerthat the two amplitudes behave as a function of sample dehydration, and the rate, at which the mobile fraction of'water Is removedis again reduced in favour, of the removal of water from the less mobile fraction. At the transition fraction for the T2 the mobile point value of falls

rapidly as the, sample is further dehydrated.

At water contents of 2.5 g/g and above, a third componentis resolvable from the decay,

which has an intermediate relaxation rate. The low,number of data points doesnot allow this fraction to be well characterised, 'with inled out' on saiples on wood and tests, were-6airr creasedinitial water content In order to study this component.

147

FI-Pf9s popiar naving an initial ponents at all

Show the-SPin-spin "atef

relaxation

times and amPiitudes-ior

water in. into two cc

content oi 4.2 gig. The aecay has neen resolved and could be resolved into

water ccntents.

three components at water contents 3 third component using

?reater than, 2.6 gig. Below this , -I the 'Newton-Raphsonmethod.

value the computer could not resolve

300

short

240
T2

lor-g-miporent

ccwwant
..................... ................... -6 . ....... .................... ............................................... ................................................

60

............... 019

........ 1.3

....... ................ .... 14 ....................... 3.6 4.5

0.0

times rov initial , 6141 Uapaltion in two Pophr, relaxation 110istare fivre component gurin; Irving. Unction 4.2 weiqht of a as content q1q,

-6000 3000

I lorg comporent 0 short cmvonent

S0

4000 Signal Applitades 3000 2000 1000 0

I'S ,12.7 N- MO )I No ( s/q Figure614b Signal anlituJesof two-coNponent relaxationin A4.5

iniffil water content : 4.1

witer-ionea

poplar.

148

e ............. ........................ ..................... T2 msea

.

........................ ......................... ... ... ..................

......................... ........................ ............. ... . ........ .........

-1 ....................... 1: ........................
. .......................

to I

................

A--, A . ....................... I......... 1. .......

11

1-

Vi

Le

(M-Mo )/MO - g/Y J'aying 111ree relaxation poplar component-spin-spin water-logged ,n initial water content of 4.2g/g.

onent Slort comp;

A Intermeaiale component Coal NEO,

A ....................................... .......... .................... .................

...................................... ..........................................................

...........

1140........... ...........

.......................

in Fiaure6.155'Signal'amplifitaes for three-component relaxation water-loggea 4.2g/g. poplar. Initial watercontent

2.7 M-No MO Y/g

3.6

4.5

The initial

water content of the woodindicates that degradationIs moreadvanced in this contains a considerable amount of solid content. The charweris-

sample,although it still

tics of the two-component fit are similar to those observedIn the previous sampie, and two transition points can be identified at water contents of Z.7cl/I and 0.4 - 0.8 1/g. The

149

reduced scatter

in the jata

points

marking the Ioni 7: values allows for 3 trend to be obinitial TZI.., 7alue of 4. -' gig to the transition locai ! 70 ainimum of arouna a reacnes

served: as the water : ontent point at TZI,,,

is reduced trom its

Z. 7 gij At Cteasas. "?

increase until asecs. It then shows a / -rteao, Paint. dhen it ialls rapidly.

the water content reaches the second transition

Ideally

ene would like

tc assign the oifferent

components of relaxation copulations

to -nown PoPula-

tion3 of water residing are not cnsistant layer.

In the wood. and yet the relative in a hYar3tion

for the t-, o tomponents to this

with water residinj

layer.

and water external

Looking zloser

at the macros: cPic structure

of the samples one notes that F6PJar, Ci

like many ct the timoers found on the Mary Rose is a angiosperm., and as such is =00sed two main types ot cells, given in below; "'. fibres. the and vessels namely 'typical dimensions ! 0: these are

Parameters observed

fresh poplar in

Lengthof vessels Diameterof vessels

Percentage of Voody tissue in vessels 40 23

cm: 100-300 AM

Length of fibres Diameter of fibres Percentage of woodytissue in fibres

900-1100 pm 23--,6 p

56-79 %

Thickness of cell walls Total solid content of wood

2.8-4.3 pm 28-40 %

1S0

'by k volume).

In water-logged timbers a 1, ood deal of the solid content of the woodwill

have been re-

placed by water molecules. and it can be estimated from freezing curves that 14--^, 0% of the water within the wood resides in close proximity to suostr3te materials, and will, constitute , a "bound' phase.

FC r fresh -water at amaient temperature. the distance diffused during the spin relaxation time, is give by equation 3.53, and using D zz2xio-s cm2sec-I -and T", , tance is of the order of I!), ' meters. It is therefore likely witnin the -dcoi wili diffuse to regions seconds. this dis-

that the vast majority Of water

'bound' water exists. Oncethere it will 'In-

der; o exchangewith that fraction.

and an avsr3ge r'elaxation rate will be observe-C.

An obvious divide which ",rOvides the correct initial

; opulations observed in Doplar In

the two componentcase is to associate the fast relaxing component with water held within the fibres, This tenlong time in the the the relaxation with water vessels. and with component and the

tative assignment can explain the relaxation data for the two componentfits. validity of this model is discussed below.

Vessels are the main conducting elements 'Of -the wood. They form the primary route at access for water molecules moving between'the exterior and the Interior. The Initial drying can

therefore be related to a reduction in the population of water within these ducts which is seen as the fall in the population of the slowly relaxing component. As the water content of

the woodfalls, the water molecules within the fibres leave the coils, and their from passage I wood is directed Primarily via the vessieli. "Wer therefore passes from the fibres into 'the, the vessels, producing an increase in the rate at which water'leaves the quickly relaxing

151

The rate at which the population of the slowly relaxing componentpopulation falls component. is reducedby the influx of molecules from the quickly relaxing -component, corresponding to the first transition noted in figure 6.1a and 6.14a.

ater in both vessels and fibres is expected to consist of 'bound' and 'tree'

phases. In

in and the to cell walls, reside an the surface of this zooel the 'bound. phase is thought the voids created by the removal of hydrolysed cell wall materials. figures 13 and 14 are fitted The solid lines In

time observed 6. Z the relaxation using equation which predicts

the in Making The in with exchange each other. parameters used two rapid are populations when fits are listed below;

I rar3meter

Short Component,

Long Component,

msecsto asecs EH-'Olb, 0.7 919

200 msecs Z SOmsecs 0.002g/l

the by is ly that the restricted motional I H4 'Olb represents the amount sample of water within the total solid bound of per gram water in for of wood cell walls eachphaseexpressed grams is fast component and for the.slow component , content within the sample. The sumof EH201b Initially bound the, total amount the of sample. water within in the the wood 19% water Of all

fteezing from findings curve exthe bound In is In the resides with phase. which agreement Justified be by can asthe One therefore of-[HZO]b values periments. question whether might suminj that the two components relate to water In fibres and vessels.

to dimensions before had those, degradation, listed that the similar poplar of sample -Given above,one can calculate the fraction of the total volumeof the, samplethat was taken up by
L

1:

the cell walls of the fibres and vessels in the fresh state. Thevolumeof each fibre is divided fairly evenly between the cell wall substance and the lumen.In vessels the cell wall

152

only contributes between 4.6 and 5.9% of the volume. If -one assumesthat in water-logged wood the solid content of Ahe cell wal I has been replaced by water, and that this water acts as a bulking agent for the remaining skeleton of xylem, then the motional, properties of this fraction will be modified and it will constitute the bound phase. If bulking Is extensive then

the volumeof the cell wall prior to degradation can be equated to the volume of the bound phase in the water-logged woo&

It all oi the volume previously associated with the cell walls is occuoied by bound water 2.0 gig for the fibres, then this would lead to values of [PZOIbZZ sels. It is not likely the for ves0.08 9/9 and

It 3nd bV water that all of the cell wail material has been replaced

to IHZO)b is plausible that the remaining solid content in the fibre cell walls reduces however for CH2.01b to the vessels. 0.3 A of amendment the value similar gig o0served. around leads to 0.022gig, which isone order of magnitude ireater than observed. Onewould expect.

is efficient the : ore walls that cell of preservation deiradation is too severe, provided not in the fibres than in the vessels. If this Is so, then the case iay 3rise whereby the cell to such an exIn

he2icellulose and cellulose by the'loss the of walls of vessels are weakened

bound leading fall to water full of a tent that they can no longer sustain a population

the value of EHZO]b. The network of hydrogenbonds giving rise to the bound phase loses its integrity, the become leading to this more mobile, phase and the remaining molecules within

higher meantransverse relaxation time.

Tb, Tf. and IHZQ]b that the parameters In Htting the data by this method it wasassumed do not changeduring the drying process. This assumption Is likely to be false at low water In cellular structures contents. whenthe gradual shrinkageof the wood changes produces whichcausea redistribution of water moleculesthroughoutthe system, and a modification of their motional properties. Thesudden fall in the value of T2,.. j maytherefore be a conse-

153

juence ot both the two-componentnature of the water within vessels. and of this change in
stfucturS. To achieve the lit ot the curve at a water content oi auout 0. Z 10 msecs. g! g the calcuia-

tion has teen %ade usiq

the behaviour for the be ient torward to of notions. put above may suifi,. account Ihe ior in is degr3dation not 3dvanced. spin-spin : elaxation times water-logged wood the which The appearanceof a third component at hither moisture : cntents nowever, leads to Speculation
nout the validity of this model.

Above3 datetcontentof 2.7 jig the to compon ent fitAs fit. : omponent ratio.

three the tive as. not as etfe,: .

'two-fold 'improvement 'a IiInI the stanjard deviation : noise for -which there is

The third componentappears is a population with an intermediate relaxation rate. and

be it for resolvel. can contents the water of wnicn nas 3 constant amplitjoe over range

More inrormation about the , hjrd -omponenthas been obtained from Studies on hilhly oegradedtimbeis for ohich the initial See figures 7.3 is ( gig 3pproximateiy water content

the decay from the even when 16a-17b) The third component be magnetisation could resolved
The low 1.6 to amplitudes 91g. 35 values as water content withinthe sample had been reduced for the two componentfit exhibit characteristics that are similar to these observed In less

, The transition'at aegraded timbers.

low"moisture' contents'has not been observed"as the data This is because the, signal to noise

Points do not extend to low enoughwater contents.

ratios were unfavourable at'these va'lues,"and this has led to the scatter in values oi

that can be seen in figure 6.17. It is possible that the secondtransition would not be visible in an, y event, as the behaviourof TIL.k1j, is different in this sample,and decre3ses j rom the offset. Thesudden drop in the values of T2-.. ljj, at low water contents

doesnot oc-. ur, and so the transition maynot exist.

154

zoo

..........

......... ................ .................

12 LIM.......................

.................

......... ........... ................. ................... .......................

si
3
,tj 11010

............... ....................... ........................ ........................ if 1.1-1

1

I&A

( MO -v luto (A) 7.43 in lar, iritial rielaxation po? 4itjl water content Filart 616a Iwo colkponent loss Iving dyging, function of oeight i is g/g 1

.1m

hm : vfgara .: + int t xmFonEnt lon; rent ......... compo .2 Sivul ............... .............
................. .......................

60 V ...........

Asplituae NO
20
Ll I-I

.......... ....... ........................

.......................

....................

++
116
Figart 6.16b

1 312

411

S. 0

iflifial for fit Signil Nocomponent poplar, for the amPlituaes

watep content : 7.439/g.

The transition observedin the two component fit at 'a water content of 4.9 g/g coincides with the point at which the amplitudeof the third begins to decrease. It correspondsto the 2.6 ill same transition observed at water.contents of. around, In timber,that has beenbetter

preserved-The notions put forwardabovemaybe sufficient to accountfor the behaviour of the

Iss,

sPln-sDin jelaxation

times in water-logged woodior which tne aegradation 'is not advanced.

rhe ippearance ji a third componentat higher moisture ccntents leads to 'speculatic'n about the validity of this model.

tic 240 ............ To Asec 0 IS

............ .

....................... ......... ......................... ........................ .............................. .......... ............ ........................................... ............ .........................
I-

4 ti -,,:: -

6(l I I,,. . % kA I'N '. , """

t,

No No
initial * 111fee in in times FiTjr! 4,17a' rilaxition with poplar spin-s? component Aincticni aaring Jrqinj, Icss 7.42 weight a if as slS, waterr: ontmtof

a0

.0

scrt comporent + interm't cotApment 4M ,,,a Ion; com-pnent .................... 11.1 ......................... ................................

300....................... ................. .......

................... ... ............... 1100
lco
0 00 0i

..........

++ . ................ ....................... ..............

cc

L-

++ "00 ,. r.............. .......... ....................... I.......... I: a .0 OC130

OU ..........................

ri on

4.*3 3.2 N- MO Ma ( g/g

Figurv6.17-hnree compenent lar, Inithl watercontent signalaitplitabs for po? : 7.43q/q.

156

In'well preservedtimbers the amplitudeof the intermediatecomponent remains! ndevendent 'two-comPonent. of the total watei content of the sample at which^point until the data becomes
it ac-. ounts for 460% oi the total 'signal amplitude. The amplitude of the short componentalso -- , acCountsfor 60%of the, total signal amplitude whenthe data tirst"beccmes two component. These; oints. together with the fact that the value of ^ nituce as T, ...... is of the sameorder of maj-

in the case-of two Components ieads +' the postulate that the intermediate fibres.

iraclcn ^.:: responds'to watei within'the

' The then be "he large and short components can assigned to water within the vess'9is. the TT-t.., 120 wlih is 'which -aintains a constant value of around, consistent value 3r 3secs, ',,.. value .,t Ti. 'observed in tne vessels, and to 2-, -from falls that Msecs has a value The'chanze, from three to two .,A. Mponentbein the'amount'of 'water within the vessels leading

lamsecs, whi. -h is not inconsistent- with haviour may possibly be caused by the fall to alass of resolution

two component Theidea that the water in the vessels becomes at higher water constants the relative one considers until to the the relaxation appears explain spin-spin character of the above, long to the presented According arguments the amplitudesof and short components. bound Is indeed It the if less phase. 6% short -omponent the or should contribute signal of tenuous. If one caiTheactual contribution is nearer 60%.which is wherethe modelbecomes fit becomes valid using the culates the value of TZI,. g at the point wherethe two component 4 fit, one arfrom the three component for the vessels as measured amplitudesand values of T22 rives at a value of 18 milliseconds, which Is clearly not consistent. Yhether this evidence i.s suificient to discount the idea of compartmentation between vessels and fibres in not clear as existence of the third component Is Itself in doubt.

157

Results similar to those observed during this work were found in samples of muscle tissues,.'- I. The data was interpreted in terms oi three components of water, and the aecay . times and their associated amplitudes match those in wood. The cells in muscle ate designed, like those in wood. from fibrillar units. and the similar relaxation behaviour may not-there-

fore be unexpected. Differences in the dehydration dynamics are noted however: As the mois-, , ture content of the muscles vas reduced the amplitude of the long and intermediate decays fell while the amplitude of the short component maintained its Initial intensity until the

te fast had frcm a the 7he the r fract,. st cother comporents samples. vanished on of water with ot decay ias associated with that ihich is most closely related to the protein jithin 3uscle. 3na It was suggested-that extracellular ponent. Brownstein and Tarr"" the

T2 long com6, t, be the e water might at origin

used the relaxation, in frog gastronemusmuscle 3s 3 test case

for their wo. of rei3xation with the longest relaxation time 'Jas con-k in which the component sidered to originate from extracellvi3r relaxation of the intraceilular water, separate to the intracellular component.The equation. treat-

diffusion the using evaluated was component

ing, tne Pometry of the cells as having a cylindrical tion is then given by;

symmetry. The decay of the magnetisa-

MM z

M(O) ......

x eip 6,4

+E Mi(O) x exp (-t/Tiit izo

Mi(O) and Ti are found using equationsthat are similar to equation6.3. This modelsuccessfully predicted the behaviourof the transverse relaxation in the muscle.

In wood the Intercellular

voids are packed with lignin

in the form of the middle lamella,

Thevolumeoccupiedby the middle lamella contributes only a few percent to the total volume of the wood,and It is therefore unlikely to give rise to T2,..,. The assignment of T21, j to the extracellular component is also negatedby studies of the spin-lattice relaxation describedbelow.

158

Etc-nstain's ipprcach to the U311SVerse iias to assume that tvo relaxation in fresh wood phases, I -iiater resided in the wood'"'. The first is the water of hydration within the wood

, ghereaitfusicn is inhibited. Abovethe saturation value tot the wood a secondlayer or, water exists for -thich the diffusion is similar to that observedin bulk water. The relaxation of the spins wasconsideredto. be due to a combination of, diffusion and magneticsinks on the surface of the wood,with the density m(r.t) of the proton magneticmoment obeyinga modified diffusion e-%3tionas described in section 0.1.

The 3-otution-to, the equation. st;bject to boundary conditions, yields 11 appropriate, -the, sumof ax; cnentials. Brownstein" decived a. transcendental equation for the nth relaxation

time Ti., ;

(b/ a where aZ/(Drl)

6.5

6.6

in the Is b represents the amount the layer, water in quantity of the a second of water hydration layer, and D, is the diffusion coefficient of the water in this layer.

by the single adjusTheoverall time scale of-the relaxation times,, T,.,, is determined table parameteral/D1, which doesnot effect the form in which the Tt, ) vary with b/3. tsee figure 6.18). The theory wassuccessful in predicting the behaviourof spin-spin relaxation times of water in fresh timbers with moisture contents upto 1.4 g/g. By choosinga suitable value of az/D1 for the first mode (n=O), the values of the second (nzI) can theorettmode cally be predicted.

r.

1251 Ow
Msee 00 0 00 0 Figure618 Erownstein's analysisof T2aata for Eahn's Northernwhite-cedu

vooa,
12fast
0

0000000

012 04,96
grams

O's 110 1
wooa grams

1.1 Is

Samp Ie Ash Birch Elm Oak Pine Poplar

(Min-Mo)/Mo S. 7.5 5.3 3.3 6.0 4.3 19 420 436 502. 257 6.5 60 32 145 85

%TILGIII 87 63 62 62 so 60

'different degradation. degrees of Table 6.3 Variation in TI for and samples TIL,. for tIO% Tlok., (Valuesaccurate to t5% for j) t, and The theory breaksdownin the case of water-loggedtimbers. Equations6.5 and 6.6 predict that for the first mode Tio, Increases linearly with water content, and this maybe Idenin times tified by its dependence the two coathe bla two the relaxation with on shorter of ponent fit. Is required for a2/D,. The value In order to fit the curve a value of 3.38 msecs

usedby Brownstein"31 in fresh timbers was25.8 msecs,leading to an upper boundfor the thickness a of the hydration layer of 0.09pm.Using a value of 3.38msecs to makethe fit for

160

"he snort -omzonentsnould produce a value of T.,, that is approximately 0., 74 milliseconas. not the value of ZOO asecs observed. Using this model does not therefore predict the correct values of the decay times iIn water-logged woodat the water contents studied. sibiA. however, to attribute 7 conditions that were applied. It may be pos-

the failure of this model to the ad-hoc boundary and initial

Spin-iattice

decay in water-iogged woodis non-exponential above and below the freezing the variety of longitudinal be that Can behaviour relaxation

event. Figure 6.19 illustrates expected in different

I 6.3 Table the time shows samples. constants and re lative populations decays in the figure. The character isation of water jithin

used to describe the different woodis difficult

Iowfo'. the but because of the inaccuracies involved. using TI measurements

ing observations can be made;

(a)

The decay of the longitudinal

that in wood of malnetisation -dater-logged component

into be tuo comtypically degradation can resolved has a medium to high degreeof
Tilt.. '400-330 Of "' -'and'2.0-40% zz msecs. relaxinj with 60-80%'of'the signal ponents with I

the signal relaxing with TI. h..., zz40-120msecs. (b) Both TIl.., 355K. ,, (C) the slower rate at the In heavily degraded relaxing signal timbers the proportion of increases, and both TI II.. 'The exact values higher of Tl, values. attain j and h,, t 275K between the and temperature of range independent Tl,.., of and t are

functions both the Of TI, complicated are and h., -t-and-the relative populations species and moisture content of the wood. (d) As the water content of the woodIs reduced,both T11, and Tl, h,, t fall. and the

populations of both components are diminishedwithout preference to either. In woods Tij.., wherethe degradationIs not advanced the drying processhas started. (e) Belowthe freezing event both relaxation times fall with temperature. can be as low as 14.5 msecs even before

161

AS IndiC3tea in , a). the spin-lattice

reiaxation times for the two componentsare populations

sl Wntiy ireater than the spin-spin -times. Rather more surprisingly. - the relative

of tte two fractions do not correspond to those seen in T2 measurements.and the majority of water relaxes at a slower rate in contrast to the observations madein transverse relaxation studies. This is an extremely impcrtant result. It meansthat the two values of T1, and the populations of water.

two values of '12 cannot be related to two identifiable

The fact that the re[3xaton times are Inaependent of temperature t%e Rjure 19a). maybe explained using the diffusion model. Assumirg that the water Molecules are movingwith corHowever,as the

relation times that are low. T1 should be seen to increase with temperature. temperature increases, the diffusion rates of the molecules are likely Interaction rate between water and macromOlecules will offsetting changes due to temperature.

ind the Increase, to

increase, 1016r4ni the 71 values, and

400

300 11 long, TI short (Oksee) 200

* *

$ * *

100

O O

o a.,q9

I,

0iI

2.78

fl'short Fig. 6,19a. 71 long (*) and (o) as a function of temperature for witerlog7ed poplar above the freezing event.

3.21 CX") logo / Temy

'42 3

3.64

162

C3

-Z cc gen C>

(o

C3

><

C>

I.

P fli j
c

C3 91-0

><

IN4 of ca

C3

><

C3

C3<

<>

-6-

c2c >C3 C3 <> >< <>

man -I.. -I-*-4-

C,

c5t

KI

95

cc0

c? -, I DI W.", c 14. W3 " CD

4 " '95

heavily decay '&n the aegradej timbers even is spin-spin component that the-three observed at iow water contents maybe an indication that the t,o-phase 2odel is incorrect. Three over-

lapping : esonanceswere observed in the proton spectra of NorwaySpruce and Sugar Maple that had been saturated in distilied water"". The position of eacn resonance dependedupon the the difthat the it of origin thought and was

orientation of the wood in the magnetic field, ferent signals was from different

morphological regions within the wood. The nature of these

regions -jas not defined.

It Seems likelY that a three component-fit for the datals a two componentfit,

a better representation

than

despite the tact that the third, componentis not always cefined. If so Even assu2inj be the three average 3 weighted components. oi will is -unlikely that any one traction rep-

then the-two-component fit

that the three-component model is correct, however, it

Is fractions coma but that, three the of rather eacn resents the non-freeZ3ble water alone, differs destructuring for degree the of bination of exchanging bound and tree water which .

from one population to another. -

with each are in exchange Evidencethat the different types of water within !the sampies is the times temperature sample of from as-the other comes a study of, the T2 relaxation for 3.2.1 of an ensemble VEK BPP between theory section to-the of 350K. cycled Accordinj and nuclei wnose motion is describedby a single correlation time, the spin-spin relaxation of the to a fall In the static component should monotoniclyincrease with temperaturedue dephasing field. betweenfracHowever, the discrete multiphasemodelpredicts that exchange

times', tions within a sample lead in to the elaxation : and since excan observed a reduction Is a thermally activated process the rate of exchange change will increase as the ;temperature of the sampleis raised. Underthe appropriate conditions, therefore, it Is possible for, the effects of exchanging to swamp the effects of temperatureon the dephasing nuclei on T'41. -, field. in which case the observedrelaxation times will be seento decreasewith rising temperature.

163

I 'te'm'pera Figure i. 20 shows the 'e fIfeI the two-compon ent fit ture oInI cts of popiar. As the temperature is increased T4 ..... initially

for water in, ',-

increases. in accordance with the

BFIII prediction. The rate at which T-')...,, rises with temperature steadily declines. until at
is reached, and the curve takes a downwardturn. - From 315K to 355K abcut 31EKa maximum T2" continues to decline, indicating that an exchangebetween fractions is in effect.

in the case of the slowly relaxing component. the relaxation

time decreases with Increas-

inz temperature over the entire range studied. and'exchange Process'esdominate the behaviour of As the temperature of the system is increased there is a slight shift in the

popuiations of the two components.with nuclei in the immobile phase moving Into the mobile fractior. "this behaviour persists to varying degrees in woods at all moisture contents.

As the temperature is cycled the curves for the upward and downwardlegs follow the same

the tor temperatures higher the account Some would which is at paths. evaporation expected apparentnon-teversibility observed.

I An Interesting behaviour In the temperature during which experiments observationwasmade the function conwater of times as a the two-component transverse studied were of relaxation tent of the wood.Th -esample had an Initial Qsed The 4.0 samplewas g/j. of water-content

A , Following dried then rehydration to partially 1.3 rehydrated. a moist g/g, and, ure content of 0.9 the s, 1/g. dried, approximately time this to of amplewasagain a water content

ihe long component of the transverse relaxation time fell with increasing temperaturein the fully saturated timber. and the short component T2 exhibited a maximum at 315Kas above.
As the water contentwas lowered the form of the T2. but the magcurve was preserved, h., t

nitude,. T2...,, -of

fell In a waythat wasconsistent with studies made at constant tempera-

164

Figure6.2OVARIATIONOF SPIN-SPIN RELAXATIGN'WITH TEMPERATURE WATER CONTENT, OF FORPOPLAR AS A FUNCTION

Mass of Samples a

D-y 2536. NO. g 0 . 0.19849

go-D. 08219 -0.1493. g

. 200

En 9 0 %..
Co

100 go 80 70

cl x 0 0-4 Co

so so

tn

CL

40

CL tn

3c

2.7

2.8

2.9 , 3.13 3.1

3.2 A. 3

3.4

3.5

3.6

3.7

3.13

I/Temparatum

(xio73)

I ture. The value 3nd behaviour with temperature of the long componentT2 was not initially fected by the loss of water ttOM the WoOd. onsistent with earlier Again this was e., ef-

measure-

3ents.

As the water content was further reduced T&^...,, continued to decrease. The value of the long component relaxation time suddenly began to fall, became apparent at point saddle and a

6.13 figures by T21.., and lowering in 315K. The is this of approximately way predicted
6.14.

After rehydration the spin-spin relaxation times relained their previous values and characteristics, the drying indicated that process had not had a significant. which long-

lasting iifect,

the 'that is to large that this a provided extent reversible process and

drying is not complete.

As the water content of the samplewas reduced T2, k,, t fell

In the usual way, and 72. tsa,

In for that a system which exchange temperature dependence typical of exhibited a TZ,.. behaviour the with of g/j about 11.9 of predominates. However, content at a moisture the was the temperature sample of decreasing Instead temperature dramatically. as changed of in the low contents T21,, increased. transition moisture the raised, This that at suigests j the between populations value of TZI.., is causedby a'diminished exchange nuclei rate of that contribute to the slowly relaxing component.It should be pointed out that this result wasnot reproducable,but is worthy of note, at it mayindicate an end to exchange processes at low water contents.

In light of the earlier discussion one can obviously question the validity of analysinj the temperaturescanningdata using two instead of three. At higher moisture concomponents tents a three-component fit is generally better than a two component fit. Similarly the d3t3

165

at hilher

temperatures can be described more effectively

using 3 three

component fit,

but at

temperatures of 3110K and below the third componentis jifficWt

to resolve. ana may irceed no

longer ce cresent.

In the cases wriere three components can be rescived certain trends can

be seen:

'the rapidly relaxing ccxpcnent has a value T2,.,,,

t :j temperature that is indapande", typijeCre3Sesq

"or any given water content. As the water content is reducedTZ... 9 msecs. to around cally from -20msecs The intermediate component hasa value of about thrse times

3vir the

range of temperat-ures and -ater contents used. The intermediate relaxation time 315K. uhroujh passes it about a maximum
(c) The relaxation time of the slowly relaxing component. TZI.., # decreases with increasinj 1Iles has a ral. le of '43 temperatures.

between 100 and 300 msecs. T2,..

(d)

4ater is most readily removed from the third traction. As the water -.ontent of the sampleis reducedfrom saturatedvalues the population of the intermediate phase Instant until a significantportion of the fast and slow phases havebeen remainsco 60% lost. The intermediate traction contributes approximately of the total slinal, and the slow component about 30%.,

(e)

As the temperature, increases'there is a slight increase,in the population of the in, both the fast'and 3'lowrelaxi'ng Populations. termediate phase'at the expese'of

Similar resultsio

those describedabovefor the case of poplar were found in other

species of woodremoved from the MaryRosesuch -asa'shandbeech.

160

It &s suz; estad . ere th; t' the two or three component-fits convenient representations tion decay'curve or a continuous distribution

to 'the --Jat3 are no Pore' than times. The magnetisa-

of relaxation

is then expressed as.

1(t)

M(0)

P(T2). exp(-tYT2). dT:

6.7

for the relaxation

in the 'transverse direction. dire-. tion? (Tj)

and a similar

expression

exists

for relaxa" of

tion in the loniituainal relaxation

and P(T2) represent

the normalised ai3tributions fOr

times for T, and T2. These two distributions relaxing spin within In figure the system. 'it

owevit, need not be the same. . T2, and the two distributions dist: ibutions

any particuiar consistent might five

must be

with this rise

fact.

6.2.1 two hypothetical

are shown whicn the sample. The which best

to the two -omponent, f it

that nas been used to charaoterise

blocks drawn over the distribution desc, ribe the distribution

fits two the three and component represent T, Tz and of are used. values

when discrete

to times in waing Figure 621--Hypothetical aistribution wooa of relaxation Jecaus., two otservation of component

P(T2:

72 (msec)

429

167

The distribution

of T, and T2 values is shownby the Brownstein, and Taw"

model to be considered the

a consequence of the sizze ot Pores within the sample. Cohenand Mendel'son"', relaxation of water in porous systems as being a comoination of relaxation

in the bulk and

diffusion processes that carry the spins in the bulk to tne surface of the pores in rocks, where they exchangewith rapidly relaxing spins of the hydration layer. 'Jithin a pore of radius r, a uniform magnetisation is established across the pore In a time,

-J

-(r2/6D)(SI/VJ.

3.3

where D is the self 6iffusion

Is I is "jV and to the sur'ace pores. rate, ratio of volume 'the size. which In porous rocks 4as time is observed

the thickness oi the adsorbed layer. Below a certain critical taken to be r<r, = Emm,a single averaje relaxation T...

i.

(T.,

'1 : /7) Tf ..

(SIM T,
pore.

dependsupon the surface to volume4atic Of each ; articular and T, ...

by the distribution of pore he shapeof the P(TI) andP(T2) distributions are determined sizes in the sample.and the. relaxation times of Ahe boundand free phasesof water within the system.Since TI neednot equal T2 for the boundphase,or possibly for the free phase. the two distributions can be expectedto differ, -thus giving feasibility to the two distribu-

but be, tions suggested M, would The, P(Tj) specified, and 2), cannot above. precise nature of be expectedto, v3ry from sampleto sample,and within the same sampleunder different condi-, tions of temperature and concentration. Thepossible existence of these continuousdistributions as a result of different pore sizes is a moregeneralapproachthan assigning different Populationsto specific cell types.,

168

Ihe fact that the TI values were independent of temperature can be explained by, the fact that aithough normally when one is in the region of motional -narrowing the TI, would be expected to increase with temperature, the rate of relaxation is enhancedin porous substances
by the increased r3ta 3t which diffusion of tne hydration layer. processes reed the spins in the bulk to the region ,

Ihe incrs3se

in the 3mplitude af, the slowly rei3xing

711component in degraced s3apies c3n

ce expiainedin terms of a distribution of T, values,in terms,ot an shift In the populatiOns

causedby either a more open and poroui system,- which undeniably exists within iater-lciged -wooa. or avincrease in-Tilbound, and is most likely to be a comoination of the two.,

!n spherical pores the ratio of SIV ; 3)r.,, Usinj the uta

from freezing curve excetimentThereicre SWUX

the boundwater within the samplecontributes to about ! 0% of the total.

31ir '. 13.1. According, to. Brownstain' s calculatlens, 13, the thickness of the hydration ! a'Yer,I in iresh timbers, is about 0.09p. leadint tn an avt-raje pore radius r... -- 2.7m which is a

fact: r ui 5 down on the average cell width. Af

it is assumedthat water within 4xist at all

levels within the cell, wall, and that all of this water is exchangeable with the water, within cell lumens, then IZ 3p., and 90,1n. Given that this analysis is based on a spherical

pore,- whereas,in reality, the-cel Is in woodare tubular, this value seemsa reasonable one for the, pore size as seen,by an average water molecule diffusing T1 TZ the and system. within -The

data is therefore. consistent with a heavy,presence of water residing in the voids within the call walls. that Is in exchangewith the large -juantities of-water In the cell lumens.,Can the therefore reflect-the relaxation phenomena, nature of the woodcell wall?

Thedehydration of fibrous muscletissue leads to a denaturation of the protein structure that is reflected in the signal amplitudeof the non-freezinj water. After rehydration the amplitudeof the non-ireezablecomponent wasseen to decrease with decreasingtemperature$ whereas prior to dehydrationthe amount of mobile water is constant below the freezing event.

169

In water-logged wood the'dehydrat i on-rehydrat ion cycle produces no such effect

in spjteof

the ract that the TZ values of the woodate greater'in rehydrated samples than in samples, that have never been dried, wnere the long Tz is increased from the typical value of 210

100 msecs. Evidence that somedenaturation takes place xay comefrom studies msecs, to about . of the non-freezable component:

60% Belcw2.73K the transverse relaxation of the non-freezing water is two component. oi the magnetisationdecays'with a relaxation time of about,, 4 msecs'at 270K.and 4% with a' ^_;f th2'f3st As the water content of the sampieIsreduced', relaxation time of 0.17 msecs., A increases the to time'of the -mponent. relaxing component match sucdenly other , relaxation Ch single decayis therefore observableat moisturecontents of' le-esthan about Z..: ; /it Vh, at is the transition point at which the transversedecaycharges! r: m three to two : ^, mDonents in Is Indicative This the morphoicy the ireezing'point. temperatures change event 3 above .; is in Upon the fast component agaii evidence. the of wood. relaxing rehydratiOn

The'two component'nature of the non-ireez3ble phase I eads'to the assumption that two types of bound water exist within the woodat higher water contents resulting In the ap-

if the then two temper3tures. the, third effects so time'at pearance, of ambient relaxation Is the bound to content the moisture disappearance of the phase of could live rise water as reduced. The water may be lost from the system during drying, or its motional properties may ,e. As be modified by changes in the structure of the wood leading to a transformation of phas the bound water in question has the fastest*decay one can relate It to water that is most Intimately associated with the'constituents WWI be the easiest to remove. Is it unlikely such as and of-the cell walls, that

170

If the jater Icst to the syste%it is probablybecause the e cells that regions within 61. , this wateroccupiesare disturbed by the drying process. it Mid te for exampiethat this elesentary fibrils. wateroccupiesthe fissures between
this water ! mFarts a mobilitty of hydrogen Do-ndedbridges to the fibrils

and that at higher moisture contents

upcn drying , Jue to the formation 1 111

which is lost

of water molecules between the fibriis.

'L

A more plausible explanation is perhaps that the r3Pidly decaying signal originates from an extremely mobile fraction of protons attached to the xylem itteif.. If this iere the case.

is lowered beco disappearance q and then the 20 the content ar the water as ana structure t33 more rigid is predict3ole. this is not reflected Houever. in the amplituie of 'he frozen comThe r3turg oi this

bY decrease be 'wnicn to about 40%at the transition. ponent expected would

at secondcomponent,and its possible connection with the appearance of a third component high water contents above the freezing Foint remain undetermined.

is it data conabove relaxation presented the studies of single andmulti-compOnent From the character of the magnetisationdecay. ciuded that water In the hydration layer dominates Single-component samplesIs data indicates that the fibre saturation point of highly degraced equivalent to the value found in solutions of celluloses, but in iess'degradedsamples-this In fresh timbers. is is increased to the 0.40 ooserved value as same about g/g, which-is,

- Thedecaysin both the transverse and longitudinal directions ire non-exponential.They bound In between phaseand water are consistent with theories of molecular exchange a water in an unbound from to chemical exthe also result phase. contribution, -Some relaxation may between change protons in the water and protons attached to the xylem. Theboundand unbound phases between the two fractions Is rapid. cannot be resolved, - Indicating that the exchange Direct evidenceof exchange from studies of relaxation behaviouras anfu cprocesses comes tion of temperature.

171

The dif ferent- componentsof the magnetisation decay cannot be related to water in different , ompartmentswithin the syste3i. The two and three componertfit3 result from a distribution are more likely to of

of Dore sizes that in turn lead to a continuous distribution

relaxation ti3ies. The bound phase occupies the , oids in the woodcell wall caused by the hydrolysis of the xylem, and is far less mobile than the water in the cell, lumens. The nature of the distributions for transverse and longitudinal decays have not been established.

Dehydration of the sampie is reversible provided that it is not advanced. In cases where the --oundwater has been lett in tact there Is evidence of a reversible transition by the caused variation material cell wall morphologyof wood. Further investigation is necessary to palify in the

the of the total of water content

the origin i this transition. .,

Polyethylene-Glycol impregrated Uood 6.3. !IMR with 'A'ater-i-ogged studies on Solutions.,

by impregnation Thepreservation of water-logged with polyethylene-glycol (PEG) wood the Impregnate cell the to the the prevent ability of polymer systemand solutions relies on wall collapse by bulking voids within the structure of the wood.Different degreesof degradation, require that different gradesof polymerbe used in order to infiltrate port different regions within the wood. and sup-

NVIR ocsamplesof, water"loggedwood, werecarried out-on a number relaxation measurements that, were Impregnated using a variety of water-polymer mixtures to establish whether the water and polymercould be resolved from the transverse decaycurves, and if so what deductiOns could be made about the behaviourof the two molecularspecies once resident in the --., wood.Different trades of FEG havebeenused to determineif possible the different types of

172

WAVOur provoked within the woodby "eir 'n tne Eystem. 3in, 3 va, utaollsn Iiii'Ers

dYnamics molecular of the water the on presence

FEG imptegnatel samples the of it possible, now aeparation

itcm U11treated-dater-logged wood.

As an aid to te, 'Interpretation

ai results preliminary experiments were carried out on

the PEEGwater system. to establish typical relaxation rates for polymer and water mixtures. -the rates oi relaxation for the water'and polymer molecules are discussed In terms of the

Liu, by Ullman described 16, and reiaxation aacnanisms

6.3-

The Polyethyiene-Iglycolill'ater

System.

14 re carried Measurements we.

PEG determine to the effect Of s3mples on out

on th e spin-spin

and spin-lattice

temperature. and molecular : oncentration times weight, varying of reiaxation

dlreclongitudinal, in, both the, the transveme dec3y. the and As expected -of-the. magnetisation

tions is multi-exponential. Twotime constantsare required to describe the shapeof the . the in be the water, and one relaxation of with protons associated may which of curve: one the relaxing protons,in moleculesof PEG.

for both times dependence temperature the of the spin-lattice and spin-spin rel xation figures in Oe fast and slowly relaxing components Illustrated in 50% solutions are im In general, the moremobile a population of protons is tne longer the relaxation teb provided that the correlation times are in the region of motional narrowing, i. e. 'r, < when many of the local fields experiencedby static or slow movingmoleculesare averaged-to zero in fast movingpopulations..Thusthe! 1ongrelaxation times havebeenassigned to the, water Protons and the short relaxation time to the less mcbile PEG protons. This is supported -aecom es

173

JQQW

101

------

4667 -0

T2IAnj

11 lantj a733 -

0+

TI Lcrv, M_-Long

)2000 1867 933 + + +

3.10 '1090/ Tomp

4.10, -3 N-1) v

Figure 622a Spin-spin against Spin-lattice relaxation times a!t a function of ienyerature in 59 2 solation of FEG 1548.

IMM

iIIIIIi

U Sh4vt 12 Sh4l"t 1067 -

1333 13 13

13T1 Short X T2 Short

XX

Goo -x 533 267

2.50 2.70

13

"a

3.10 3.30 4.10 3. M 3.90 3.70 Ig-3R IWO / Temp -1) ,, -( Figurle 6.22h Spin-spin apinst svin-lattice volamatian tiw as a function of te"perature in 58 V.solution of PIG 1540. 2.90

4400 12Long 33M lisoc

0 Ti Lzrv + TE LDrg

2.97 3.16 3.35 3.54 3-91 4. 1600'/ femperature -( 10-30) Figure 622o CIPin-slin alainst 31pin-lattloce *s as a tun0tion of temperabnv relaxat ion t 5' In 517. solution of PIG4M.,,

2.60

2.79

41 A 3.72

174

4A

13 L120 -XaX TI Short

0 TI Short T2 Short

T2Short 940 -

me)

560

13 X

X I" V[ m 4.10 2.79 2.97,3.16 2.60 3.54_ 3.72.3.91 -, -3.35 . 1290/lemperatupe (19-31-1) figure, 6-Ma WmPin againstsPirrIattice relaxation tines as a functionof,tenperature in 50V.solutim of PIG4009.

13 U Shortl 940 T2Shopt (mime) Z60 280

I

13TI Short XM short 13 13 AX 13

XX
2.79

11
91 MI

3.16 3.35 4.10 2.97 3. !43.72 3.91 logo / Temperatart 622# Spin-"in airainstSpin-lattice relaxation tims as a function of teloeratum Figurke 28,091. in 56 Z solutionof PEG

' IIIIXt92.60

;m

In

-. 4 1

4500 0 4000 MOO-0+ * Long 3000 * Long 2500 CPseo

0 T1 Lora T2 Long 0 0

+4++00 LOBO 500 2.97 3.72 3.54 3.16 3.35 3.91 4.113 1896/ Tomptibatum CIg-31-1) figure 622t Spin-svinagainst opin-lattice relaxation timesas a function of too4yeratu" IR39 7.SOlUtiDn 28,069. Ot PEG 2.60 2.79

175

by the measurement of the relative

which are close to the amplitudes of the two components

'the "ie"Su'l calculated values. - The tim'pe'ratu're was-'varied 6eti-een"27OGid' 360K, -'and tiaie characterised by the following points;

(1)

Using solutions of low molecular weights the short componentattributed

to the

protons in the PEGmolecules has TI z T2 over the temperature range studied. Using PEGsof high mo I ecuI ar 'weight, -T2,, Is Iess than TI, suggesting possibIy that -the -, molecular motion of the larger PEGmolecule's-i'i-"restricted play a role In'the determination of the' relaxation times. (ii) The long componentwhich is attributed to the lel., has TI zz In the water moecus and that viscosity effects

'protons T2 at low temperatures but as the temperature Is increased TI becomes progressively longer than T2., This difference becomes more pronouncedas the molecular weight is

increased.

Figures 6.23a-d showthe relaxation data for the 225% solutions of PEG with two different solutions the-low molecularweight solutions haveTI molecularweights. As with the 50% T2

components, I molecularweights'.The 'ana,TI greater than T2 for . for the short''ielaxing higher similar trends as those for the 50% water relaxation shows solutions, but the point at which 6-apiratir'e-s-'. -TI deviates f romTZ occurs"at'm6- lo'we'r'
A It fAM

71Lont

4WO 3100

3000 12 LDni 25M EQDZ 1500 Im Soo

2.60 e. vq 'E. 97 3.16 3.35 1989/ Temperature 3.54 (Ig-3X-1) 3.72 3.91 4AO

Fhruzle 6.23a Spin-sPin atainst Spin-lattice relaxation in 23 % solution of PEG29, QN.

times as a function of temyoratur, @

176

5
? IQ

f2f%r% VJKAV

IiIi

IIII'I' ' ,-,,

(3

IZ16 U short 1333 M Short =X

13TI Short X, r6 Short 13

Cmseo

8B9 444

X X 13 X XR

-' '1 3.72 3.54 19-3 X-L 3.91

'4. N

2.60

'2.79

2.97 3.16 3.35 18901 7emmature

Figure 6231 Spin-spin against Spirriattice relaxation times as a functign of temperature in 25 V,solution of PEG 20,800.

WOW

X
IZ16 TI Short 1333

0 TI Short
X T2 Short

12 Short

667 -X 444 222 3.35 3.16 3.54 4.10 2.97 3.91 3.72 low / 11"Pepatm 1,9 -3 X-1) timesis a functi(m of temperature r1gart 6,23c Svin-spinaninst svin-lattice Pelaxaton in 25%solution or IEC IM. 2.60 2.79

7500 ''6667 -0 3833 -0+ 3000 -0 U Long 72Long 4167 -++0 3333 -+ E500 -, 1667

0 TI Long T2 Long m -

2.60 Figure 6MI

2.79

2.97

3.16

3.35

0-0 -I+ 3.54 3-72

-,3.91,4.10

Spin-spin ajaWt Spin-lattict relaxation tims as a function of tel4popaturg in 23 1.solution d PEG 1990.

177

Thegraph in figure 6.24 comp. in 25% ares the longitudinal relaxation times of PEG soiution and50%, solutions of'PEGsof different molecularweights. It can be seen that the solutions containing only 25% PEG haverelaxation times whichare longer than those containing 50% PEG, but these times are independent of the molecularweight of the PEG used. Theconis morepronounced centration dependence at lower temperatures,again indicating the in.1 I I I 'in-i6l ution -proton's fluence of viscosity effects. __The tra ,n, SY erse reI'a'xition' times of "the' PEG are Illustrated in figure 6.25. Again there is a concentration dependence which is stronger at low temperatures,but the relaxation times are also influenced by the molecularweight of the polymerused. The relaxation times of the protons In water for the 25% solutions and 50% are Illustrated in figure 6.26. The spin-lattice relaxation times are again Independent of the PEG molecular weight used, and are'lower In solutions of high concentration. Theeffects Thespin-spin relaxation times decreaseIn of concentrationire' greater at low temperatures. high concentration solutions and showsome dependence on molecularweights of the PEG as the temperatureis raised.

08 yj

4C

40 40 0,

4-

10

+0 11Short . 0 ) msec 4.0 3.0 2.0 1.0 0.0

Figure It
I-`--.
. 4 ,-,.,

is 40

4 +0 +0
--

fo

+0
+0%

254 solttions
4

5w.Solution 0+0+
+0

14

3.80 3.20 3.50 4.10 LOWI TeNperature ( 19,3K-i) Spii-spin relaxation of 27/,ani W, salittions of PEG LOBO (a), ana PEG 2MQ(4)1 415 a turlctiOm Of inytrst tl""rat. LLrt,SDrt (PIG)COMPORent,

2.90

178

T2Short 5.0 4.0 3.0 2.0 1.0 4.10 3.80 3.20 MO I= I Teneraturt ( 10'30) (0), 4090 W, 22/. LOW PEG Eq-/. figurl 25. Spin-spin ana relaxationof sailtiong ot (PEG) functios W, Short inverse component 20,606 timperature. a as at aria 2.90 13. in 2.60

40 40

It

Ln
IM

7.0 C11sto)l
6.0

a. -0

J+ 0+ ov
0+ 4

25i.solutioes, OF
OF
Q+

.411, 04 Sov. solutims

4.0 r 4.1, 3.3 3.8, 3.0 --3.6 X-1) IM / Tompepature16-3 W PEG (O)s IM in my, st/. PEG spin-uttice 26a 6 Jigure ana soiutian ot reimtion NAM (4) asa functom Lavg (H2O) compowt. of te%perature. 2.8 3.0 2.5

72 Lon 7.0 -+ Orm

Ln : 8.0 -+t+++
4

237-solutions r4 4

5.0 4.0 -

+ 50% Solations +

3.6 3.0 3.3 3.8 4.1 AGOO / Temperature j@-3x-l) 5W.Solutifts Figure6.261SPilv'svin 1009 goo(j), relaxation in 257. arwl at PEG M (w)as a functiof,of temperature. Log Q 20) ccmponent. ana29, 2.5

3.0

2.8

179

Finally,

the graphs in figure 6.27 showthe effect of varying the concentration of PEGin

solutions at a temperature of 300K. Both TI and T2 of PEG decrease with increasing concentration, and within experimental error the influence of the molecular weight on the PEGrelaxation times Is slight., As the concentration is reduced the spin-lattice relaxation time begins

to exceed the spin-spin relaxation time. TI and T2 of the solute decrease with increasing concentration and are independent of the molecular weight of the PEG.

Thespin-lattice'relazation

times weredependent the uponthe resonant frequencyused. -As

frequencyincreasedTI became longer. This behavioursuggeststhe presence of a distribution of correlation times describing the motionof protons along the polymerchain.

In dilute aqueous Measurements by Liu and of TI and T2 for FEGs solutions weremade Ullman"". The results aboveare consistent with their measurements and findings. Theprin-

chains Is aresult of the dipole-dipole and. ciple mode of relaxation for, protons In. the PEG hydrogen spin rotation Interaction between nuclei. The strength of this interaction depends the spin pairs and so relaxation In uponthe inverse sixth powerof of the distance between the polymeris dominated^by Interactions between near neilhbours. Liu and Ullmanconsidered that the effects of the,hydrogen nuclei other thanIthose attached to the same unit monomeric could be nellected. -The effect of the solvent on the'relaxation times wasdetermined exthat proton perimentally by Heatley and Walton'M using deuteratedsamples.Theyconcluded relaxation times are effected very little by deuteration of either solvent or polymer; and

'Int'ir-ictlo'n's-. ' are therefore determined'by-intra.; mojecular

The structure of these monomeric by the repulsive,and attractive , units Is determined forces associated with the molecular sterlochemistry and is similar in the melt and soluble state to that In the crystalline state proposedby Tadokoro et allM . TheC-Hbondsare

180

600 500
ox

400 Short ", -TI --.

Nsec

200
-100

0x

40 20 30 IITZ H20x PEG pams, pan,

60

5m
400 ) 300 ( Asee ,

>

100

10

20

1101, H PEG grams grams 10-x fiction in (? E0 component, 627aShort, Figure of polUmier a as concentration relaxation W. 29,80 (o) (A), ON PEG AN, ana solutions of , "I,

10 w

40

60

181

melm

Long 11 ) 1200 ( mseo

800

A0

Xo

400 1

20 30 40 PEG / IN H20x 7. paxs gras

50

60

0 ( 0
72Long, Nsec

AX
X

XA
400-X 50 40 20 30 60 H20 x 'ill PIG / grams grams PIG in Figure times M, Long (H20) solutions asaf un coNponent relaxation ction of W in (o), 4NO 28,899 PEG 1699 (A). concentration ana polper - -- 'A

10

-f

182

tetrahedralandapproximately1.10 Angstroms in length.,Eachmonomeric, unit can adopt one of three conformations;trans,, gauche, the describe theltwo position gauche', of which -and oxygen atomsin the O-C-C-0 unit. Theseare Illustrated below;

I-

/0

H
'Now

(1), in O-C-C-0 bonain Figure 628 Relatiye oxvgen of atoms gauche positions (3) trans (2), codormations, ana gauche'
Themotion of the hydrogen atomscan take one of two forms. The first Is the rotational

diffusion of the entire monomeric unit with yespect to the external applied magneticfield. The second-involvestotameric.jumpsbetween the different, conformationsof, the monomer unit by by rotation about the C-Cbonds. The results of. relaiation measurements caused performed Liu and U112an"11wereconsistent,with the idea that the principle contribution to the spin between the gauche-gauche' relaxation In the polymersIs thejumps conformations.

The rate at which the structure of monomeric units transform is consideredto be independentof the molecular, welght of the polymer, and independent, albeit to a lesser extent, of the concentration of the solution, as the effects are not long range. On the other hand, diffusional rotation of the monomeric depends heavily on-the amount segments of chain entanglementwhich increaseswith both molecular,, weight and concentratIon.-The spin-spin relaxation Is effected by the, l,ow frequency component of the local magneticfield and depends

183

to a greater degreeon the diffusional than the spin-lattice relaxarotation of the monomers tion time. This leads to the increasedrate of transverserelaxation at higher molecular
weights and concentrations.

The relaxation times of the water In PEG/watersystems are depressed from those observed In distilled water. At low temperatures they are Independent of molecular weight but vary

with the concentration of polymer.Theycan be explained In terms of a two phasemodel the observedrelaxation time is the usual weightedaverageof bound whereby and unbound phases.If h represents the traction of polymerwithin the solution, and K Is the cothen; ordination number

Kh, Tb

(I-Kh) Tu

6.10

whereTb and Tu are the relaxation times of the boundand unbound phases.Manipulationof this equation leads tolM
ML

Ia

Ci

6.11

Tbi Where the subscript I

K Ior2, and relates to the spin-lattice and spin-spin relaxation.

Using the data of figure 6.27, Mi z 3.57, C, 22 0.305, M2 3.25, and C2 0,463. There Is of sourcesthat two water moleculesassociate with eachmonomeric unit evidencefrom a number In the polymer.,In the measurements in this work It wouldappearthat equation6.11 is made to a monomer no longer valid for polymerconcentrationsgreater than I g/ml. This corresponds to water moleculeratio of about 1:2. Setting K2 therefore, Tbj 478 msecs,Tb3 478

ising result Indicates that TIh msecs.This rather surpr, polymerconcentrationsof less than Ig/ml.

T2b for the boundphase. using

The weakness Is that the constants C, equateto Tul the relaxation of this approach ti. mesof the unbound phase.This leads to values of Tul 3.28 seesand Tu, 2.16 sees, water at

which are equivelent within experimentalerror to the relaxation times of distilled

184

the sametemperature., Moreover, one would have expected to see a difference

in the two

relaxation times of the water in the bound phase becauseot a decrease In the mobility of the boundphase, and because of an increase in spin diffusion between the polymer and water -', molecules. The fact that this change is not observed maybe a reflection the polymer.itself, of, the mobility of,

or It may indicate the weaknessof the, polymer-water interaction.

Thedeviation of T2 from TI for water at higher temperatures Raising unexplained. remains the temperatureIs expectedto enhance the mobility of the water and polymermolecules,, , '
thereby the liquid nature of the solutions. This should lead to similar T1 and T2 -increasing values. Although the exchangeof nuclei betweenthe polymer and water molecules may lead to the low T2 values it does not, explain the fact that the separation of TI and T2 takes place at, lower temperatures in lower, concentrations. It Is-thought that, the difference in Mand T2

contamination of the samples,which werenot oVthe water-content is the result of oxygen vacuum sealed after preparation.

6.3.2 A PulsedNMR Study of Water-logged Impregnated Wood with Polyethylene-Glycol Solutions.-

Thespin lattice relaxation times observedin degraded woodIndicate and water-logged that a large proportion of water resides within a boundlayer of water that most probably oc. cupies voids created in the cell walls of the wood.A cosbination of diffusion and exchange processesgive rise. to relaxation times in these samples by the distribu. that are determined tion of cell and pore sizes within the sample,with larger poresgiving rise to longer ,,, -

relaxation jimes. 1he uptake of large PEG this systemmaypreferen. moleculesimpregnating tially. occupy. the large pores, or they may, get-trapped within the smaller,voids, In the wood, thereby filling 'These these gapsand displacing the water,whichwouldotherwise occupythem.,

effects could be expected, to changethe-relaxation behaviourof water within the woodby dis-

185

turbing the initial

distribution

of water at different

' 'and by the sites within the samples,

direct'inf luence on the dynamic properties of water molecules by'the macromolecules". " T2 behaviour in samples of water-logged woodhave therefore been monitored as a function of the uptake of PEG, to establish whether they 'reflect ' or the redistribution In the wood, the modified dynamic behaviour of -the water

of polymer'within the sample.

Severalsections of water-loggedpoplar werecut front a block of moderately well wood(initial' preserved water content 5.7 g/g) and'soakedin solutions of polyethyline-glycol

havinga variety of molecular weightsand concentrations. The effects on the spin-sPinmagnetisation characteristics of the initial 3400wasmonitoredo'ver'a period of uptakeof PEG

the twentydaysusing solutions with concentratio 5 and 50%.In freshly cut samples ns between transversedecaywasat least two component, with 70% of the signal relaxinl'-With a characterlstic time of 49.8 msecs,and the remainder with a time 'of 170t 20 msecs.

Thetransverse decayof the poplar saturated In 5%-20% solution wasresolved Into two The relative populationsof the fast and slow relating component components. wereconstant 'slowly Inday the twenty time ' the the period, whilst over relaxation of relaxing'cosponent 72 of the fast relaxing component creasedto a value of 240 t 18 msecs,, stayed relatively' constant'at 51 t4 msecs.,
xll

'

Onewouldnot expect'a'polymerof, this size to havepenetrated'the cell walls. 'The polymerinitially the mostaccessible regions of the wood permeates which are the vessels,

the cell lumen and middle lamella. Onecan therefore concludethat If the two CO"Ponent. fit Is correct at'least a proportion of the nu I to the slowly re I al Ing 'Component clei contributing reside In one at moreof these'accessible regions. The-sIlght Increase In the spln-Spln`,,, relaxation time of the slowly relaxing component maybe explained It one assumes that the relaxation rate of this compon ent Is greater than that observedfor protons ln'distllled, water because of Its interactions with water boundto the wood.Theseinteractions reducethe

186

average Brownian mobility of the unboundwater molecules and the relaxation ther enhancedby paramagnetic centres present, in'the cell'walls'of

be furrates may,

the wood.The'PEG

molecules within the samples would be expected to reside close to the cell walls occupying hydrogenbonding sites in preference to the water molecules so blocking these exchangeinteractions, and leading to a-greater value'of T2.

The. 50 asecsfor T:.,.,, constant value of approximately

mayindicate that there are

3400 regions within the woodthat havenot beenpenetratedby the polymer. Molecules of PEG are relativeli 'large, and to test whetherany areas of the wood wherelarge bodiesof water imthe for could accumulate were Inaccessible to the polymer, freezing curve s wereproduced by the solutions the water freezing point is depressed samples. In normalPEG/water pregnated in, PEG 245K 235 between to molecules.Typically the water turns to ice at temperatures of PEG solutions. As the concentrationof the solution Is loweredthe freezing concentrated range so that there is a steadydecline In the smeared over a wider temperature event becomes signal amplitude with temperature.

Twosamples for eighteen monthsin 40% PEG of poplar weresoaked with degreesof -nos'ill`n-6f'attenua'show the slinaCamp'lltu"de'spolymerisation of 1500and 4000. In-b66 ca'ses' tion eyen,, haveperat temperatures as low as 230K This"suilesis't'hat the molecules of PEG 'that bj'water is freezable in normalwater-logged aeatedall regions occupied wood,even -y 4000molecules.However it doesnot excludethe possibility using the relatively large PEG that regions exist that are inaccessible to PEG wherenon-freezablewater exists.

In samplesthat weresoaked-insolutions of 50% 3400the transverse relaxation times PEG for M,. t-and 147aseesfor T2.1'; are reducedto 26 msecs ';.' This Increasedrate of relaxation Is consistent with the T2 of water In the PEG-water systemat higher concentrations, ' ;wherethe greater 'viscosity of the solution results In lower T21s. That the PEG effects both

187

fast and slow componentsagain suggests that if the multi-component nature of the system is a consequence of compartmentation within the sample tIhen each compartmentis Impregnated by the large PEGmolecules to somedegree. -.

I li
"I, "-, ,, "IIo, 0,0, X,

0+x0++++0

0:

0,

,.

Ia

spinin Relaxation Tims (vaec) (LogritImic scale)

o0 f+00 x ox
x x

o+

-0.3

0.4

0.?

0.9

0.9

Amplitudeof T2 slow component

0.5

0X09

+X

0-3
Fisrure 6m

9.4

0.9" 0.8 '0.5 'V0.7 -Sample, weight weight/ initial sample

1.0"

Twocompmwntspin-Win relaxation in watop-loggea Pollar as a function PIEG 390 (o), loss arging. auring poplap with or weight was pre-tPeatea PC 4M (+), ana PEC3490 W.

be distinguished The signal arising from the protons attached to the PEG cannot molecules from the decay. At low water contents the bondingof large PEG moleculesto the cell walls broadensthe NMR detection., The result of resolving the decaysaboveinto signal beyond three components, that, there Is a shift.. In the relative, populations of each, suggests, component, as Impregnationtakes place, with an increase in. the quantity of tightly boundprotons,at the expense In the populations of the intermediate and mobile phases.Given that the number of,,,

188

the prime mechanismfor relaxation

is a result the diffusion of molecules to the region of

the 'bound' water -layer or Ao paramagnetic centres on the xylem surfaces, this suggests that there is a shift before interacting in the distribution of'diffusion lengths which water molecules must cover

with spins of a shorter T2, to lower values.

The preservation of certain'tImbers

requires the use of smaller PEGmotecules,which can

PEGs penetrate the woodcell, wall and support'heavily degraded xylem.,! of-the skeletal'areas Figure with degreesof polymerisation of 300 and 400are common ly'used'in' such circumstances., 6.29 contrasts the spin-spin relaxation behaviourobservedIn samplesof wood heavily impregthat vary in molecular the from 300 340O. impregnation nated with PEGs, to weight -Following samples wereallowed to dry. Thedecayshavebeenresolved into two components andare shown as a function of weight/initial isweight of the sample.It should be noted that the samples

to the same pregnated with PEG could not be dried evenat oventemperatures extent as the un20 and 40% treated sample.This indicates that between of the Initial of the'sampleIs mass

- The following observationscan be made due to the PEGs by the wood. that have beenabsorbed the behaviour'of TZ In these samples; regarding,

, The spin-spin relaxation times for freIsh water-loggedwoodand treated with samples for which'the'degreePEGs as water Is of "polymerisation'is'greaterthan 1000decrease from the wood. removed

QD

For"' I es'trerated'with PEGs less' than 1000the that havedegreesof polym'eris'ation' samp T2 values are independent of the water content of the sample.

Qii)

In untreated samples,and in samples treated using PEG with DPsgreater than 1000, the bulk of the signal relaxes with the slow relaxation, while in impregnated timbers wherethe size of the PEG moleculesused Is less than 1000, this fraction is reduced to between 40 and 50% of the total.

189

FEGs with degrees of polymerisation less than 1000 are liquid in the bulk at room temperature, which may account for the dif f erent, relaxation characteristics magnetisation-Af of the transverse

the PEGsremain In a viscous liquid state within the wood then one might the amountof liquid within the

not, expect to see any change in the relaxation times until

The increased viscosity of the'f luid would reduce the number`of; is much reduced. very sample interactions between the molecules within the solution and the, woodceli, walls, thus main-, taining the values of T2 at'some intrinsic the water may lead to the precipitation level. Using PEGsof, higher grade, the removal Of

of large PEGmolecules which could enhancethe rate

of 'decay of the remaining fluids via spin diffusion and chemical exchange.

a1

0-0 x++

0+080++ x

Spin-spm Relaxation Tims(msec) (laffritImic, scale)

x 0,0

00++0 x x

0x000++0+
0.3 9.4 0.3 0.6 017 0.8

019

1.0

Sample weight I initial saxple weight Figure 6,39 Tya*ec014POnelt sPil-oin relaxation in water-loggeaPOPlaras a function of weight loss auriag arging. PoplapwasPre-tPeAtea 3M (o), with PEG PEG 496 (+), and PEG 34M W, I -,,, -,,, -,; , ,-,,

Ii,,,

It is possible that the different components of the relaxation curve relate to similar spin species In the sample that maybe physically partitioned, attached to different molecular species, or attached to similar molecularspecies with different modes of motion.

ISO

It mayalwbe

that the decay curve represents a complicated aggregate of all, these ef-

fects, which to a good approximation happensto be describable using a double exponential function. A three componentfit represents the data equally well and the results are shownin to draw from these multi-exponential repre-

figure 6.30. Meaningful conclusions are difficult

sentations It an aggregate Is indeed what Is being seen. It seamsreasonable to associate the mobility of-protons within the samples to the value of T2. At higher watevcontents there is a small fraction of very mobile protons evident, in samples of untreated woods, and woods polymers. This, is almost certainly the during that water molecules

treated with 'solid-like'

length of the T2 experiment do not interact strongly with higher molecules. In samples treated with liquid-like polymers this non-interacting fraction of water is no longer ap-

parent. This may be due to the-fluidity,

the/ because are of, the smaller PEG molecules which,

moremobile, can escape,the close attentions of the cell matrix, and better occupy the voids

in, Interaction, the Increasing the the thus, rate the water wood and wood, with mixing within between watevand macromolecules.,

In all samplesIs reduced.An the samples free low the water amount of water contents, .,-At treated with. 'solid-likel causethe fall-in FEGs the interactions are,morefrequent and of such strength as to

TZ.An samples treated with, lliquid7like', PEGs the. increase In the fre-

and strength of the Interactions is offset by the mobility of the macromolecules which quency can Impart-mobility to, the remainingwater and sustain the relatively long T2.

dependence The, temperature In that of water untreated wood'Indicates water7logged ex-, , change play a stinificant, rale In the determinationof relaxation characteristics. mechanisms Figure 6.31 showsthe temperature dependence fit for T2 In samples of the two component of PEG Impregnated timbers. Onesample was impregnated using a 40% solution of PEG 3400, and the other using 100% 600. Both curves follow the same PEG pattern i. e. the spin-spin relaxation times for both decrease components with Inverse temperatureIn accordance with BPPtheory. decisive and very important result Is that the exchange that were evident in nonprocesses

191

treated samples,are no longer dominant In the T2 temperature dependence, indicating

that

whilst the motional properties giving rise to relaxation In these samples on a local level are not substantial ly"effected, -translational the exchangeeffects maybe suppressed by a fall In the

freedom of the water molecules.

At around273Kthere is a change in the rate at whichTZ falls off with inverse temperature. Although there Is no evlencefrom freezing curve experiments In state at of a change the freezing point of water It appearsthat the shift towardsthe rigid lattice condition Is acceleratedbelow 273K.

It Is perhapsnot surprising that the study of NMR relaxation times in Impregnated PEGO the interactions between water-loggedwooddoes not reveal moreInformation regarding* ComIndividually the PEG-water systems complicated. water; are very and and wood-water xylem in which the ensemble bining the two leads to a highly Involvedarrangement, of spins undergo a wide range of activities that lead to distributions of correlation times that cannotbe

of the large PEG easily characterised. Yhat is clear Is that the presence moleculeshas a of water In the samples,mostnotably steamingthe rate of profoundeffect on the dynamics different types of water in the wood.Thesites occupiedby the different between exchange molecularweights cannot be specified using the NMR data presentedabove,only the more that the moleculesinitially generalcomment occupysites within. the largest pores in the

for the case of large PEG high Using can be made sample concentrasufficiently molecules. PEG,regardless of molecularsize, appearsto occupysites previously tions at least some oc. cupied by free water, so that the amount of non-interacting water within the sample. s disappears.

192

Figure6.31.,
WITH TEMPERATURE VARIATION OF SPIN-SPIN RELAXATION WITH 40% PEG 3400 SOLUTIONv FOR POLPARIMPREGNATEO &100% PEG 600c,

600Sao_
400-

3013-

vi S ci ra

1013go8070. so.

4030, a" a

20. c
CL
I

I ll

- __ 1

-1

:,,

V'

:a

I IVtoo

43.

2.7 2.

ig";

'a 3. 3.1 3.2 3.3 3.4 3.5

3.6 3.7 3.9

3.9

4.0

I/TempQrature (xlG-3 )K

-1

6.4

A Study of Self-diffusion

for Water in Untreated and Treated Vater-logged Woods Using

Field Gradient Techniques. 14MR

I -.

The relaxation data in both treated and untreated samples suggest that the bulk of water within the wood behaves-like pure liquid water until layer following randomdiffusion it comesinto contact with the hydration coefficient D,

processes. The value of the self diffusion

can be monitoreddirectly using NMR spin-echotechniques, which maybe used to assessthe nature of this diffusion. Measurements of the self diffusion coefficient of protons In wood how D havebein undertakento establish whetherD Is consistent with the two-phase and model,
Is effected by the structure of the wood, and the water. content of the wood.-

In samplesof water-logged wood'treated with PEGthe relaxation data suggests-that there-is

the water and macromolecular species. This should also be 1reater interaction between reflected In the diffusion properties of the moleculeswithin the wood;an evendistribution proton diffusion coefficient, D'. Onthe of PEGs within the sample should reducethe mean are resident on the surfaces of the cellular structure then other hand, if the PEG molecules DOshouldbe largely unaffected by the impregnationof PEG.

6.4.1 Self Diffusion In UntreatedVater-loggedWood.

Theself diffusion coefficient of water in water-loggedwood using the wasmeasured described in section 3.5.3. Thevalue of the self diffusion pulsed field gradient method coefficient for distilled water in the bulk, D., has beenfound to be 2.5 x 10-3 cm3sec*l at 2311C.In freshly' cut"saaples,of .wat er-logged I woodthis -vIalue II. s reduced'bytypica"lly 10-20% to between 2.0 and 2.3 x 10"'cmlsec-' Effects leading to the fall In D weMconsidered by

Wang'34. ) whosuggested two possible mechanisms, namelythe direct hydration effect and the obstfuction effect.

193

The direct hydration effect leads to lower values of D becausewater in the boundphase has,a reduced translational velocity. Twophasesthat are in rapid exchange give rise to an

observed value of D that, is the weighted average of both., In the case of water that Is bound and,free this can be expressed as
4-

D.

b.,

-Pb.

Db

Pf. Df

6.12

D. and Df are the seIf, dIf f usion coefficients of water in the boundand free statest where andPt and P, represent the probabilities of moleculesbeing in the hydration layer or otherthen Db will be several orders of magnitude wise. If the water moleculesare tightly bound less,than Df, and the first give; II1 to term an the right handside of, equation 6.12 maybe dropped -1 11- 11, .I11

Pt. Df

Df

LH-20lb. Df CH201t

6.13

I at, On is the popu I at ion of, the bound water,phase,, and I H201t is Ahe,tota I. popu whereEH201b of water in the wood.

)AW

1.0

0.7 0.6 0.4 0.6 1.0 L3 IEH ZOR 9/9) Figure 642 Vs tle relative Aitfusion ocettisient top water in water-loggea poplar as a function of 1/[H2Olt, foil the longitedinal lipections Ole 0.3

194

Figure 6.32 shows the behaviour IH201t-I function for a sampleof '/D. of ", D. D1 a' as = of poplar that Is oriented within the probe so that the longitudinal direction aligned with the field of the woodIs

NMR Ig/g the ' Below signal becomes of gradient. water concentrations If equation 6.13 Is correct then as the'sample Is dried the

too weak to measureD' reliably. value of D1 should fall

linearly with'the Inverse water content of the sam'ple',intersecting

Df. ' The'solid line In the the Y axis at D1 -- DI/D., and the gradient should equal EH201b. f Igure, calculated using Df` = D. and IH201b= 0.38 g/g, Is a good representation of the data within experimental error. The significance of this result Is that water external'to the

hydration layer diffuses at a rate equal to water in the bulk, ' and that the hydration layer consists of 0.38g of bound water per gram of wood, which is In good agreementwith other work. Df has been shownto equal D. by MacGregor et all"', of water that was extracted directly cates that the diffusion who measuredthe diffusion rate

from the sample. At the sametime the above result indi-.

rate of water molecules in the bound phase is orders of magnitude

lower than from water in the bulk, and the'amount of water within the bound phase remains relatively constant during drying, in agreementwith freezing curve experiments.

LnDI Long, 0490q Lins'6.1799 LZj 91239 9M69 0' - IVTI

0.54 26

-0.30 2.8

Figare6M In (Y) for vater in poilar at Uterent uter mtents. as a functicmof Alsoshowis thevaluein theraaiaVtangential directien. Inversetemperatare,

2.9 3.0 3.2 loge/ reRperature( is

313

195

Equation 6.12 above,is valid only iU there is a, rapid exchangeof -water molecules between the bound and mobile phases. As the total water content of the system falls diffusion the reduced

that time an average water molecule is in by increase the of rate amount caused an

increase In the interpreted be as an hydration layer, in the turn spendswithin can which strength of the interaction betweenan average water molecule and the cell wall material. then as diffusion the activation if

there Is an Increase in the strength of the interaction tivated process one might expect it to be reflected-in

is a thermally acIn

energy of diffusion.

ice the activation energyIs close to 52 KJ mol-.

the semi-lOg Figure 6.33 shows Plots 01

for wood D' against Inverse temperature aligned longitudinally with the field gradient at a series of different water contents.

Assumingthat

DI

Dlo x exp(-Ea/RT)

'6.14

whereEa is the activation energyfor the diffusion process, R is the universal gas constant, from the temperature the deduced Is the the T Ea have been of sample, absolute values of and Ea figure in 6,4. in the table There the Is In as of and are shown value gradients no change Ea in the woodIs equal to the activation energy the water content Is lowered.Furthermore In pure liquid water which is approximately19.7 KJ mol*'. Theenergyof a single measured of energyrehydrogen bond is about 10.0 KJ mol", and Ea Is roughly equal to the amount to leads break two hydrogen bonds. This to concludethat one quired rather perplexing result by the hydration effect it water In the hydration layer has the same D' can only be reduced activation energyas that of the mobile phase, or that the water In the hydration layer does distance on the time scale of this experiment.This would appearunnot diffuse a measurable likely and one is tempted leading to reducedvalues of to consider other possible mechanisms D'.

196

Samp Ie I 2 3 4 5

Orientation Longitudinal Longitudinal Longitudinal Longitudinal RadlaI/TanjI

I11,1 Table 6.4

Weight (g) 0.226 0.190 0.179 0.123 0.226

Activation energy (KJ mol-I 18.02 18.22 18.54 -_18.56 16.13

11 I, 110 11 in poplar.

Activation energies for diffusion

Theobstructive effect leads to an apparent fall in DI because the translational motion by'the presence'of boundariesand macromolecules. In such of water moleculesIs hindered.. in the Intrinsic diffusion rate of the water whichmust casesthere neednot be a change distances in any specified directravel longer path lengths in order to traverse the same in this work could be varied between 30 and 170 tion. The time scale of the experiments msecs.During this time the rangeof distances travelled by a water moleculewith a rate of self diffusion equal to t 'hat of " 'free, 30 microns. water, would"bebetween-'12"and unbounded

. 0.

The long axes of longitudinal ly- oriented cells are largaIn comparison, the water 'and. diffusion In this direction. moleculesin the lumen of these cells wouldnot exhibit bounded A significant amount of water resides in the medullary rays, however,whichcould could lead to partially restricted diffusion along the longitudinal axis. The resulting value of DI may is then be the weightedaverageof one traction which is bounded that unbounded. and one

Robertson"" has quantified the value of D1 for the one dimensionalcasewherea'population of resonantnuclei are bounded by parallel, Impermeable borders and has found that it is sensitive to the time over which the systemis observedV.

D' W2

Dobs(tl) Din

(2n+l)l nz-O

197

3-4expl-C2n+'1)zyt'i'P)"+' (2n+ I) I al

exP(-(2n-+' )z'Ezt')

6.15

where t' t/td,

is the observation where t, zzal/Din.

timet,

measured In units the Intrinsic

of "diffusion relaxation

time",

t,.

I. e.

to

Din represents

rate of- the water in the the separation of

SYstel, I. e. the self, diffusion the boundaries.

rate, in unbounded samples, and a represents rate.

Dobs Is, the measured diffusion

For water In wood, Din is taken as being in

equal to the diffusion figure 6.34.

rate of pure water Dw. The form of equation 6.15 is illustrated

D'/D

19 .

1.02

Figart 6.34 Solution of lchertsoalsC22) one-aimmional mpatNent Nodel.

t/t a

Figure 6.35 shows the specific self-diffusion coefficient Do, for water diffusing In the longitudinal direction in water-loggedwood as a functlon. of Delta, which represents the time Therange of Delta is scale on which diffusion is measured. by' experimentaland sys-

'limited tematic considerations, but Do is Independent of Delta over the rangestudied, having a con-

O.B. Fromfigure 6.34, this corresponds to a value of to of stant, value of approximately around0.02. Using the maximum value of t 171 msecs,this requires a maximum boundary

separation of approximately140ILs In order to adequatelyaccountfor the data. Since the length of the cells Is moreusually orders of 101 greater than this one can concludethat Do Is not determinedby the compartmentation of the total water population.

198

L. go glee

II

-I

-.

10, Iaa4

a0a00a

x
Im

XX

200 160 120 ac-P) FigUrt 6-m- ThPr'ehhve AMMON cWticient 1", as a fuuti4n of emperimflUi time wait, rap Popiamin iogitmairwai (4)),ant in (x) arvotiom. 80 DIUA

40

doesnot excludethe possibility that the water is bounded Theaboveargument by permewhich could result In a lower averagediffusion rate. able membranes If the depression of Dl

by a fraction of the water being bounded, in the longitudinal direction is caused then an upper value for the separation of the boundariescan be put at a few hundredmicrons, which is beyond the lower limit of the diametersof the cells in the medullary rays. It has beenestimated that as much are marked as 50% of the walls of cell lumen with pits and spiral thickbetween these 8 and 13pa, and* diameter In the pits are MacGregor of enings. popular et al havesuggested that the depressionof Dl for water travelling parallel to the longitudinal by a proportion of the water being bounded axis in poplar maybe caused within these pits. The results of this study are not consistent with this view though, as it Is estimatedthat the asount of water held within these pits is unlikely to account for sore than 4%of the total water population. As pointed out by MacGregor et all"', a contribution from water held

in the medullary rays will Increasethe amount of bounded water, provided that the diameter of the cells within the rays Is less than 1-10p for these samples. MacGregor et al"Ij that the distance between could accountwell for their data by assuming boundariesis around

5-15gm,and the fraction of bounded water wasconsistent with the fractional volumeof transverse elementswithin the woods.Chang et all"', on the other hand, claimed that the effects

199

of compartmentation could accountfor no morethan 10% of the attenuation in Do. observedin their studies on muscletissues. Theypointed out that, in the limiting case when t >> (2/xz)t., equation6.15 reducesto the form :, -, -I t 11

Dobsl,,,. - (az/IOD,, ttz)

6.16

i. e. the observed diffusion rate becomes inversely proportional to the intrinsic dif fusion canwater leading to the fall in Dobs' that Is seenIn wood rate. If the fraction of bounded for 'then' it'is"a'-s'afe holds by"Viarying'Delta', 6.16 that'equation monitored not`be' assumption" that traction.. As,the temperature of the systemis varied, an increase In Dj., shouldbe fibres, decrease in D'obs, No in by a and vice-versa. such effect wasobserved muscle matched SImilarly. for which the ratio of Dobs1/Dj, were effects a constant such maintained value. no j' be wholly during this w ork, which suggeststherefore that compartmentation cannot, observed value of Dobs. responsible for the reduced

Thediffusion of water-in woodis antsotropic. The value of D' in an arbitrary direction perendicular to the long axis of the calls D'N'IT, is Illustrated1n is 6,35 filu 11 and re

lower than D' in the longitudinal direction. In figure 6.35 D'o, is plotted as a function of T Delta. As the value of Delta is increasedthe value of Welt Is steadily attenuated, from a value of about 0.38 at tz 45 msecs,to 0.24 at t -171 3secs. Utilising figure 6.34, this indicates the presence by barriers having a characteristic separaof water that Is bounded tion of a between 26 and 33 microns. -Thediameterof fibres in popular range from 23,to 26 microns, and 20-30% of the total water within the samplemaybe expectedto reside within the lumens of fibres. Onemaytherefore suggestthat the effects observedIn figure 6.35 result from the restricted diffusion in the transverse direction of, water molecules within this Population.

200
flg

Experizents were carried out Wtest, directions. Whilst the identification

for anisotropic effects In the radial-and tangential the, radial

is trivial,, longitudinal, of wood, the axis of

and tangential directions are not so easily distinguished, especially in small samples. Small sampleswere therefore placed with the longitudinal the field-gradient the direction to perpendicular axis axis, -so that of

and.the samplescould be rotated about. the longitudinal

D'e, j could be obsrved as ajunction gradient. The results are illustrated function of sample orientation

field the the between the and axis radial of angle in figure 6.36, which shows the value of Dloi, as a

for a series of water contents.

D' DIDW 1,10

0

IL 0XXXX""x xxXxXXXXx A AA

0x00

a 01

001,,

0A0xaa -x XxXXX0

0'

xx

AAAAAAAA

VIA

'45

90

Ias a 'poplar, D' is shown d aiffusioncoefficientin witerloqgei Figure646 knisotropy infieia rep samples functiosof tle. anqleof sample wighing 9106g(o), Uug, w ana , gMg (-L)

' 135 , 180 ,' 225*- 270 koileof sampleintitia

315

360

Thevalue of D'N,Y is cyclic, and one can concludethat self-diffusion in the radial ,c. Ie I I . I I directio'n prece des at aI differe nt rate to dif fu sion in th tangential dire tion. The peaks in D'n,T are associatedwith the alignmentof the radial axis of the sample with the field gradient and value of DIRit is taken to represent D'a, the self-diffusion coefficient of water in the radial direction. D'j, the value for the self diffusion coefficient In the tangential direction, is equatedto the value of D'et, at the troughs. As the water content of

201

the samplelsreduced both DIN and D't decrease, and the difference In DIN and D'T Is increased. The effects becomemore pronouncedas the water content Is lowered. The behaviour of DI In the -longitudinal, in figure 6.37. directions tangential contrasted are radial and

D/Dw1.0

0.4 0.2 -+ 0.0 0.0 Figm 6M

0 13- Tangential 1L 0.2 0.4 I/EHZO]t 0.6 (glg) (3.8 1.0 I- ;

Diffusion in the iongituainai, airections, asa maial,ana tangentia A function of 1 /-total vater contest of sample,.,

for all water contents within experimentalerrort Theslopes of the graphsare the same It independent Is [H201t*l in decreases the direction D' the which with of and rate at which but the Intercept, which from equation6.13 represents the self diffusion coefis measured, ficlent of the free phase, Is reducedIn the radial direction, and again In the tangential -" -I '.%, I. I, "- :" ' 'I it r''I''''' II., '.. ''t, .11, direction. This implies that Dt is lower for these orientations, but also via equation 6.13 that CH201b, the quantity of water in the hydration layer, also has an orlentational dependence. Althoughit Is possible that diffusion takes place within the hydration layer, living in the amount of hydration water from a diffusional view point. rise to an apparentchange This result, together with the fact that the activation energyof self diffusion is independent of the water content throws doubt on the notion that the hydration effect Is responsible for the low value of Win the radial and tangential directions.

202

Theanisotropic nature of the diffusive tensor reflects the structural composition of

wood. The resistance to diffusion are aligned, i. e. the longitudinal is least In the direction In which the majority of cells direction. A large numberof cells orient radially direction, this for observed and

this leads to the intermediate rate of diffusion are no major cellular

whereas there giving

componentsthat traverse the wood in the tangential direction

rise to the lowest D' value. In view of this one favours the obstruction effect as an ex-

planation of the fall In D'. The fall In D' with reducedwater content is then explained as a the gradual of shrinkagein the woodupondrying. As the water leaves the wood consequence collapse of the larger voids leads to a greater proportion of the water being in a confined is fall in therefore the to diffusion The and a observed shrinkage environment, coefficient. Itself anisotropic, being greatest In the tangential and radial directions, and onemight therefore expect to see a difference In the rate at whichD' falls with [H201t for the Ionwater Situdinal, radial and tangential directions. The fact that it does not maybe because the Into to does falls that has restricted a state and not contribute within a pore collapsed NMR signal.

Inconclusive. In particular the narrowrangeof Theresults describedaboveare soaewhat delta pet21SSibledoesnot allow for distinct characterisation of the boundariesto diffusion In the direction of the three different axes In the wood,leaving a question markover the probablecauseof reduceddiffusion in the longitudinal direction In fully water-loggedtimlikely that the bers, and in all three directions as the water content is reduced. It seems hydration effect plays a significant role on the determinationof D' in these cases, however t he hydration effect doesnot explain the anisotropic nature of the diffusion coefficient. on the other handthe obstruction effect seems to play an important role in reducing DI in the radial and tangential direction, but doesnot adequatelydescribe the behaviourof DI with water content. The causeof the fall in D' In woodIs mostprobably therefore a combination of both effects.

203

6.4.2. Diffusion In Treated Water-logged Woods.

Therelaxation experiments I es there is describedearlier Indicate that in untreated,samp a considerablenumber of protons in different regions of the cells of -the woodthat undergo It wassuggested that these exchange either chemicalor molecularexchange. processesare attenuatedby the presence molecules,which should be reflected by a depressionIn of large PEG the self-diffusion coefficient of protons within the sample. ,, I

been have Theself-diffusion coefficients of hydrogen in that nuclei water-loggedwoods for some time in solutions of polyethylene-glycols are reducedby an order of beimmersed, to untreated samples.As a preliminary experimentthree samples of tween10 and 100compared PEG 300,100%PEG 400, and 50% PEG 3500. After six poplar were soakedIn solutions of 100% had increasedby 2-3%, implying that some hoursthe weight of all three samples at least some PEG had entered the samples In eachcase. Prior to soaking, the self diffusion coefficients for the water in the wood weretypically around2.0 x 10"scm2see".After the period of six hours there wasa dramatic reduction in the value of D in each sample.

Solln Sample ", IIII 300 PEG PEG 400 PEG 3500

DI before soaking 0.82 0.79 0.78

D' after soaking 0.084 t 012 . 0.067 t 018 .

0.06 t 05 .

TheValues of DI weremeasured several times each, and the large error in the value of D, for the samplesoaked in 50% PEG 3500reflects the poor quality of the data. The measurements weretaken at a temperature of 297K.

204

The low values of DI match values obtained in studies of PEG/D20 mixtures. DI for PEG-300 molecules In D20 are shown In figure 6.38 for a*range of temperatures and concentrations. To attain a value of D1 a 0.084 at a temperature of 297K requires the concentration of polymer to be greater than 60%by weight, '-and the smalI increase in sampleweight' IndicatWthat' concentration of PEGwithin the wood is far lessAhan this value. the

Lnk9) C#MiIIiI
1.4? -000000 _ _A,
_"" ,aA. 0 0""

0. B3 0.25

(3 0 13 0 xk10 13 13 13

00

-0.33 -xXXq

13
x0

2. ea
figure

2. Im

3.00 3.10 1000/renperature (10

3020 1

3.30

3.40

6.39 La (DII against invem twPoratuPt to', 19% (O)S M. (A), sez (1), 9jr/. (X) solutions of PEG308.

(0), W. (0) SQY-

300 and PEG In samples that have beenwell saturated in PEG 400 Solutions there Is a diffusion in the the dependence of orientational coefficient, and a steady reducreduction PEG In tion in the activation energyfor diffusion as to 3.8Kcal/molon the sample soaked than lower 300, and to MKcal/mal in the sample PEG These in 400 are values solution. soaked that for the value observedIn pure liquid water (4.7Kcal/mol). X-ray diffraction has shown hydrogen by between there is bulk the in the weak order caused short range molecules water bonds.This structure is rapidly forming and reforming as the moleculestumble, but in solu. that this short range structure is disrupted by the presenceof it wouldseem tIons of PEG Theactivation the macromolecules, and the Inter-molecular interactions are weakened. for different concentrationsof polymersolution, and the results are energies were measured In figure 6.39. As polymeris added to D20there Is an initial shown energy. This initial fall in the activation

decreaseis caused by the destructuring effect describedabove, and Is

205

by a slight Increase in the diffusion matched addition of polymer the activation

the In system. With the further rates observed

energy begins to rise to, the value observed in distilled If the lowering

water, and the diffusion of the polymers Is slowed down. (see figure 6.39). of the activation energy of diffusion Is causedby the destructuring effect,

for water in water-logged woodby the addition of PEG then the amountof PEG within the'wood must be such

that within the samples the concentration of the PEG/watersolutions is less than about 15%.

ktiyation awrgies 16 -a. for adfusionin PEG 380 solas. (X8314 ,

N

1
-

I-

o.
"

28
--0.34
M -O.
I I I I !* !I I

50 60 bv veigh 1.09 i. PEC

70

80

90

100

MD risupe649 Actiyation of ?EC enerliesin solutions

20.

Theobstruction effect described in section 6.4.1 can take one of,,two forms. First there is the casewheremembranes to regions of the wood, within the systemcan tonfine molecules. and this Is believed to be a,contributing factor to the depressed value of D'. There Is also by the presenceof macthe casewherethe translational motionof water moleculesIs Impeded be Isotropic kind in locally. homogeneous, This obstruction would. of a romolecules. solution. Assuming that the concentrationof polyzer,, within the woodis low, then to depress D, to Such a great extent wouldrequire that the motion of the polymersis very greatly reduced, perhaps evenhalted by,their proximity to the cell walls of the wood.This neednot impedethe rotational diffusion of water molecules,and the effect on the relaxation times of the water in treated samples wouldby negligible. The reducedtranslational motionof the water molecules

206

would howeyereffect the rate of chemical exchangebetweenprotons InAhe dif f erent, regions of the wood, which the leads, to, the modification of the temperature dependenceof the spinspin relaxation times of water. in treated samples.

6.5 Conclusions.

In water-loggedwoodthere exists at least two populationsof water moleculesthat are distinguished by the fact that at low temperatures,Le. below the freezing point of ordinary water, one phaseretains a high degreeof mobility and its hydrogen atomsstill the proton NMR signal. contribute to

It is believed that this phase to a population of corresponds

that Interact strongly with the matter in the walls of the xylemcells, and is molecules the below humidities in fresh timbers the found to hydration equivalent at relative water of f lbre saturation point. Freezing curves indicate that the quantity of , boundwater within the fibre, In the is approximately0.38 -grams fresh timbers solid material. wood of gram per of water per gramof solid. saturation point for manytimbers occursat around0.38 grams Therelaxation times of water molecules within this phaseare muchlonger than for equivalent the In fresh timbers, Indicates, that as the degree degradation molecules water as well and of neighbouringwater In the rest of the systemreducethe averagestrength of the Interaction the xylemand water molecules.This Interacting water Is called 'bound' or 'nonbetween freezablel water, and is commonly hydratedsystems. found in many

the The molecules water within system are relatively remaining ,

free from interactions to

with the molecules of the cell walls and undergo the usual change of phase from liquid

solid as the temperature Is lowered. As the wood is dried this phase Is removedalmost exclusively from the sample, and only whenall of this 'unbound' or 'freezablel fraction has

been removeddoes the bound water begin Its evacuation., As Ung as the quantity of bound ,

207

water has not been depleted during the drying process the sample of wood may be rehydrated to

Its original form. Theunbound fraction is consideredto havemotional properties that are similar to pure water In the bulk.

The'two populations of water within water-logged woodconstantly exchangemolecules, and

deuterationstudies Indicate that the structure of degraded and water-loggedwoodis very openandaccessible. Thereare no regions of the wood wherethe boundfraction cannotmix and exchange molecules with the unbound of-water molecules population. In addition to exchange between the two phasesthe protons attached to molecules undergo of the xylem itself may chemicalexchange with protons of water molecules.Theexchange of protons from oneenvironmentto another is clearly demonstrated by the temperature dependence of the transverse relaxation times, which seem to indicate that there is a distribution of-exchange rates.

Valuesof single component are less than relaxation times for water In water-loggedwood for fresh water but greater than for newtimbers. The times of both,T1 and TZ in degraded both showa proportionality to the amount of water within the wood samples provided that noneof the boundwater has beenremoved.This fact Is consistent with a, two-phase model, but Indicate that. the amount from of water heldAn the lboundl,phaseis less than that found, freezing curve experiments,having a value of about 0.14 grams boundwater per gramof of solid. In relatively well preservedtimbers suchas some found on the Mary ofAhe heartwoods

Rose,however,the two component T1 modelcan only explain the behaviourof single component that the values of T1 and T2 for the boundwater fraction are themand T2 values by assuming selves functions of the water content of the woodand,this leads to the Postulation that there exists two kinds of boundwater; a 'primary' and a 'secondary'. Evidence the of existence of secondary boundwater also comes-from the multi-exponential nature of the non-7 freezable component.Secondary bound water exists at water contents over 0.141/g, and.the boundwater-that, the systemcan sustain reachesa saturation quantity of secondary value at water contents of about 0.4 g/g.. It wouldappearthen that with freezlg curve analysis one

208
H

observes both the primary and secondary bound water, whereasfor single componentrelaxation measurements only the primary bound water has a sufficiently tinguished from the freezable fraction. high relaxation rate to be dis-

The value of 0.14 g/g of primary bound water within the system is consistent with measurements of bound water found in samplesof hydrated cellulose from a of the amount, variety of sources. Below water contents of 0.14 g/S there Is evidence of the formation of

bridge-like structures between water adjacent molecules of xylembuild from hydrogen-bonded molecules. ,,, vI, , -III-, 11-

Examination decaycurvesJor, spin- I attice, and,spin-spin experiments of the magnetisation behaviour, and at least two time constantsare required shows clearly their non-exponential indito explain the data. In dehydrationexperiments the, two-component spin-spin behaviour. 0.4 and cates two transition points that occur at-water,contents of about 2.7 C/Sand between 0.8 0/g. Therelative populations of the two components vary, and the easewith which the different components in the from the wood also changes,which promptchanges can be removed behaviourof the spin-spin relaxation times of. the two components.The relative populations , of the two components are such that the moleculescontributing to the signal of eachcomRather ponentcannot be identified as the moleculeswithin the bound layers., and unbound eachcomponent wouldappearto consist of a blend of both boundand free water In different proportions. -,, 1 ", ""I. 1,

Thenature of, the NMR relaxation in water-loggedwood-is not unique, similar results have beenobservedIn other fibrous materials such as meat In the past the multi-tissue. decayshavebeenanalysedin terms of different populations of water that component are compartmentallsed In some waywithin the samples. An attempt to explain the two component nature of the transverse NMR behaviourin terms of the physical nature of the wood wasmade by assuming that populationsof water in the fibres behavedifferently and vessels of the wood

209

and consist of different

is bound model successful until free proportions of water. -The and

behaviour in more degradedsampleswhere a third component one considers the transverse NMR decay for these samples suggest can be resolved. Furthermore studies of the longitudinal NMR the Invalidation of this model. II., .

distribuis It seems In there likely that a continuous more samples wood of water-logged tion of relaxation times brought about by a distribution in pore sizes and relaxation times for water In both boundand free phases,as well as in the self diffusion rates of water bound by different the of the and water held within eachpore. proportions within sample, Thedistribution and magnitude relaxation times are thought to be established of the observed by studydescribed by Brownstein Tarr"" workers expanded upon and and the mechanisms via , 11 A two In the this component as rocks. sense Inj the NMR such of systems porous characteristics In this fits to that the approximadata merely are work presented are three component and tions whichdescribethe envelopeof the distribution. function. ,III

"trailht S

forwardmeasurements of water7logged samples of TI andT2 on polymerimpregnated,

into information the Its that has do amount any of about polymer way provide made wood not in solution can be resolved into a comthe wood. Theproton NMR signal of PEG molecules, The the from ponentoriginating from the macromolecules water. and a component originating by conformationalchanges relaxation of polymerprotons Is governed along the chain of the of Ahe water In PEG molecule.Therelaxation mechanism solutions is complicatedby the formation of hydration layers along the polymerchain. At high concentrationsof polymer, visinto play to restrict the motionof both polymerand water molecules. cosity effects come Vhenwoodis Immersed in a PEG solution the contribution to the, signal due to protons In. the PEG 'absorbed, into the dIsmoleculesis no longer resolvable. - Instead this signal becomes tribution of relaxation times observedIn untreated timbers.,

210

The impregnation of PEGinto water-logged wooddoes not restrict ,

the rotational

motion of

the water molecules in the wood, and the relaxation rates and multi-component nature of the relaxation of the wood-water-PEG compositedo not differ significantlY from those of samples

that have not been Impregnated by the larger molecules. The presence of PEGwithin the systea, -however, has a profound effect on the translational freedom of the water molecules.

This is reflected In the reducedvalue for the self diffusion coefficient of the water in the dependence wood,the temperature of the transverserelaxation rates which no longer exhibits the spin-spin relaxation times and the water exchange effcts, and the relationship between content of impregnated samples.

Thedifferent gradesof PEG, upto molecularweightsof 4000, all penetrate the waterts loggedareas of wood degree, thoughprolongedsoaking, wherefree water exists to the same the degreeof polymerisationof -the PEG Is increased. Whether required when or not the lighter gradespenetrate the infra structure of the cell walls better than heavier grades hasn't beenestablished through this work. TheNMR doessugproperties of Impregnated woods gest that the lighter pades allow a greater degreeof motional freedomfor the water molecules. This degreeof mobility is maintained eyenat lower water,contents, whereas using high gradesof PEG the mobility of the water moleculesis greatest at, high water contents.

At low water contents the water mobility is restricted, probably as a result of the precipitation of the polymeronto the xylemcell walls and the strong interaction of the water and solid-like PEG molecules. -,, -, -, t

Theself diffusion coefficient of water-in water-logged woods are depressedfrom the values observedin pure distilled water. This depressioncould be the result of either an obstruction mechanism, wherebythe effective path-lengths for diffusion are Increasedby obstacles In the way, or it result from hydration effects the protons held within the whereby

211

hydration layer lead to a reduced average rate of diffusion., diffusion coefficient, .0 reality

The anisotropic

nature of the

lead to the conclusion that in and Its water-content dependence is probably a combination of both these effects.

the lover rate of self diffusion

The rate of translational

diffusion

in PEGimpregnatedwater-logged woods is reduced by a energy for

factor of 10. There is a reduction in the anisotropy of D', and the activation diffusion

Is-lowered In Impregnated samples, leading one to suggest that the obstruction efin D'.

feet is the primary cause of the fall

the five this out In the Introductory section of chapter questions were posedwhich set The first this work. almsof to a certain degreein this chapter, four havebeenanswered

treatment the NMR in the to It of application assessing regarding and remains say something, parameters, required for the preservationof yater-logged timbers.

Simple, straight forward NKR that can be made a conwithout first making measurements siderable Initial capital outlay are the water contents of woods,the spin-lattice relaxation gives any detailed, and saturated form, and even

times, and the spin-spin relaxation times. Of these only the first information about the state of the woodin its Initial unambiguous

this measurement is less straiiht, forward than it at first appears. Eachtime the water conbe for to tent of a sampleis measured NMR, have estabexample, using certain parameters lished suchas the relaxation rate of the signal originating from the solid content, which for the deadtime of the spectrometers mustbe knownIn order to be able to compensate Given that wa contents can be measured receiver mechanism. with far moreaccuracyand cer:te'r tainty using the relatively cheapoven-dryprocess, the use of NMR seems an unnecessary overkill.

212

The relaxation of PEG molecules within woodcannot be resolved from the rest of the signal. PEG is The fact that the proton NMR from the of effectively signal stemming molecules has Impregnated the system meansthat assessing the degree of impregonce the PEG Nor can TI and T2 be realistically

Invisible

Is not possible. nation via single TI and TZ measurements

used to establish degrees of degradation in water-logged woodsince the values of these entities are determined by extremely complicated processeswhich appear to vary from one

to another. specimen

inhefty . Makinj detailed NMR capital studies on samples of water-logged woodrequires a be to justified in the likely terms Information is ascerthat of cannot one vestmentwhich tain from suchInvestigations. Theresults of the temperature and water content studies performedin this work, whilst useful in determiningthe nature of the system,cannotunfortunately, be readily applied to the preservationprocedure.

References:

1. 2. 3.

FranksF. "Water-A comprehensive treatise. '

York. (1972) Press, New Pub. -Plenum

SkaarC.' 'Water In Wood' Syracuse University Press- Syracuse. (1972) DerbyshireW. 'Water -A comprehensive treatise. ' (1980) Val 7. Plenum Press, New York.

4. S. 6. 7. s.

U. Stamm

"Woodand Cellulose Science. " Ronald Press, NewYork.

(1964)

Carles J. E. Scallan A.M. J. App. Polym. Sci. ,

Vol 17, ppIB55-1865. (1973)

Ogiwara Y., Kubota H. J. App. Polym. Sci. Vol 13, p1689. (1969) BoesenC.E. Call. Chem.Tech. Vol 4, p149. (1970) Neal J. L., Goring D.AJ. J. polya Sci. C. Vol 2.8, p103. (1969)

9.

Child T.F. Polymer. Vol 13, p259. (1972)

213

10. 11. 12.

Hsi E., Hossfield R., Bryant R.G. J. Coll. Interface Sci. Froix M.F., Nelson R. Panshin A.J., (1980)

Vol 62(3), p389. (1977)

Macromolecules. Vol 8, p726. (1975) McGraw-Hill, NewYork.

de ZeeuwC. "Textbook of Wood ' TechnolotY.

13. 14.

Brownstein K.R. J. Magn. Resonance. Vol 40, ppSO5-510. (1980) Vriesenja J. R., 37. Symposium ChandrasekaranS., Luner P. J. App. Polym. Sci. : App. Polymer pp9ll-921. (1983)

is. 16. 17. 18. 19.

H. MendelsonK. S. J. App. Phys. Vol 53(2), p1127. (1982) CohenM. Lui K.J., Ullman R. Heatley F., J. Chem.Phys. Vol 48(3), p1158. (1968)

Walton 1. Polymer. Vol 17, p1019. (1976)

TadakoroS., Murahashi S. Makromol.Chem. Vol 74p p109. (1964) MacGregor R.P., Peemoeller H., Schneider M. H., Sharp A.R. j. App, Polym. Sci. Polymer Symposium 37. pp901-909. (1983) App.

20. 21.

Robertson B.

Phys. Rev. Vol 151, p273. (1966) 39, Annala NY AcadamY

ChangD.C., RorschachH.E., Nichols B.L., Hazlewood C.F. p432. (1971)

22.

Belton P.S., Jackson R.R., Packer K.J. (1974)

Biochim. Biophys. Acta.

Vol 354, p305-

23.

Brownstein K.R., Tarr C.E. *Importance of classical diffusion water in biological

in NMR studies of.

cells. " Phys. Rev. A. ' Vol 19(6), p2446. (079) Vol 76, p4755. (1954)

24.

J. H. Wang

J. Am. Chem.Soc.

214