HAL Id: jpa-00226267
https://hal.archives-ouvertes.fr/jpa-00226267
Submitted on 1 Jan 1987
HAL
is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire
HAL, estdestinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
DIELECTRIC PROPERTIES OF STRAINED ICE. II : EFFECT OF SAMPLE PREPARATION METHOD
K. Itagaki, G. Lemieux
To cite this version:
K. Itagaki, G. Lemieux. DIELECTRIC PROPERTIES OF STRAINED ICE. II : EFFECT OF SAM- PLE PREPARATION METHOD. Journal de Physique Colloques, 1987, 48 (C1), pp.C1-149-C1-153.
�10.1051/jphyscol:1987122�. �jpa-00226267�
JOURNAL DE PHYSIQUE
Colloque C1, supplhment au n o 3, Tome 48, mars 1987
DIELECTRIC PROPERTIES OF STRAINED ICE. I1 : EFFECT OF SAMPLE PREPARATION METHOD
K. ITAGAKI and G.E. LEMIEUX
U.S. A r m y Cold Regions Research and Engineering Laboratory, Hanover, NH 03755-1290, U.S.A.
R&urn& L a pr6paration dl&chantillons d e g l a c e u t i l i s b dans l a mesure d e s propriCt6s d i 8 e c - triques induit t r o p souvent d'importants e f f e t s micaniques. C e s e f f e t s o n t &t& &tudi&s par com- paraison a v e c des mesures r&alis&es a v e c d e s &hantillons obtenus s a n s e f f e t s mgcaniques im- portants.
Abstract. Since most commonly used sample preparation methods f o r i c e dielectric studies in- volve r a t h e r heavy mechanical straining, t h e e f f e c t s of straining were studied and compared with more strain-free sample preparation methods.
1. Introduction
As discussed in a n accompanying paper, t h e dielectric properties of ice c a n be modified drastic- ally by straining, probably through t h e modification of dislocation structures. Most methods customarily used in sample preparation f o r dielectric studies involve mechanical c u t t i n g or grind- ing, o r freezing of t h e samples between t h e electrodes. I t is obvious t h a t such t r e a t m e n t c a n distort t h e crystals, making t h e dielectric properties different from t h e as-grown s t a t e . In t h e case of samples grown between electrodes, mechanical s t r a i n could a r i s e due t o t h e difference in t h e r m a l expansion coefficients. Auty and Cole (1) reported cracking between electrodes and ice upon cooling t o below -40°C.
During t h e course of our dielectric studies we also encountered several anomalous phenomena which may have been caused by mechanical straining in t h e crystals. In this paper we describe t h e e f f e c t s of various sample preparation methods.
2. Sample Preparation Methods and Results Freezing in situ between electrodes
Freezing of ice samples between electrodes was t h e preparation method used in most early stud- ies (I). Electrode System C (described in P a r t I) was slightly modified t o allow freezing of in situ crystals. A s e t of spacers w e r e inserted in t h e guard ring a r e a between t h e high and low electrodes. The electrodes were wrapped in polyurethane film, and distilled and deionized water was introduced between them. The ice growth methods included very slow i c e growth a t a cold- room t e m p e r a t u r e of - l ° C a s well a s rapid growth induced by filling t h e g a p between t h e elec- t r o d e s and a seed crystal with distilled and deionized w a t e r at -lO°C.
For t h e slowly grown ice, very strong negative conductance disappeared in 4 days at -2g°C and never reappeared. An interesting observation is t h a t relaxation a t t h e higher frequency range was very close t o t h e Debye type. Its closeness can b e judged by t h e straightness of t h e capaci- t a n c e vs conductance plot and also by t h e r a t i o of relaxation time, t h e "Debye index," shown on Line 2 of t h e tabulated d a t a on t h e results s h e e t in t h e accompanying paper. This ratio, which should be 1 f o r t h e c a s e of pure Debye relaxation, was 0.88 o r higher over more than o n e decade and remained this way. This was t h e closest to t h e Debye t y p e during t h e course of o u r studies.
The sample was lost when i t was being dismounted from t h e electrodes s o t h a t we were unable t o determine t h e orientation. But t h e lower frequency "tail" of this particular sample was very short, suggesting t h a t t h e major p a r t of t h e c r y s t a l was oriented with considerable deviation from C-axis parallel.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1987122
C1-150 JOURNAL DE PHYSIQUE
Compared with this slowly grown sample, rapidly grown samples show considerable deviation from Debye relaxation. Samples with s e e d s having C-axes oriented parallel t o t h e e l e c t r i c field
Fig. 1. R e a l p a r t of dielectric constant becomes negative but returns t o positive when t h e t e m p e r a t u r e drops.
0
had large lower frequency tails, a n d t h e transition between high and low frequency relaxation was sometimes obscured. In c o n t r a s t , crystals grown with C-axes perpendicular t o t h e e l e c t r i c field had very small low-frequency tails and relaxation was closer t o t h e Debye type.
Samples frozen onto e l e c t r o d e s
One side of a 45-degree-cut crystal slab sample was slightly melted with a warm m e t a l surface, and quickly frozen o n t o a cold electrode s u r f a c e at -lO°C. The o t h e r side of t h e sample was t r e a t e d in t h e s a m e way.
A very long t a i l was observed on t h e low-frequency side of a Cole-Cole plot made immediately a f t e r sample preparation, a s shown in Figure 1 (designated -9.310 hr). On warming u p t o - l . l ° C , t h e t a i l shifted toward t h e negative side of K ' (designated -1.1114 hr) but shifted back t o t h e positive side following slight cooling (-4.3128 hr). On slight warming of 0.7OC, t h e slope again shifted t o negative (-3.61122 hr). A t a much lower t e m p e r a t u r e of -22.8OC t h e Cole-Cole plot appears perfectly normal (-22.81165 hr) but i t returned t o negative by warming up t o -1.4OC.
As annealing proceeded, a stronger negative trend seemed t o e m e r g e (-31254 hr). The GR 1615 c a p a c i t a n c e bridge is capable of switching both t h e conductance standard and capacitance t o t h e sample side of t h e bridge. Balancing t h e bridge by connecting t h e internal c a p a c i t a n c e par- allel t o t h e sample means negative c a p a c i t a n c e of t h e sample. Since only t h e t e m p e r a t u r e s of t h e sample, t h e e l e c t r o d e and a short length of cable w e r e varied, t h e source of such negative c a p a c i t a n c e must be within this system. Quite possibly t h e ice sample is t h e source of t h i s anom- aly.
During studies of t h e internal field by charged dislocation using a computer model, we observed negative capacitance behaving in a similar manner. Beyond a certain limit of dislocation den- sity, t h e internal field e f f e c t c a n feed back t o t h e dislocation motion, causing catastrophic be- havior similar t o t h e Mossotti catastrophe. With distributed relaxation t i m e such catastrophic behavior appears in t h e lower frequency range a s observed here.
Warm die-drawn samples
All samples used in t h e straining experiments were prepared by warm die-drawing methods.
Also, one series of experiments were conducted without straining. This method is preferred t o t h e others, since it easily produces samples of uniform size with a minimum of induced strain.
Stacking faults may be generated in t h e course of t h e drawing, but will be eliminated in about 12 days at -20°C s o t h a t a few days of annealing near O°C will suffice t o c l e a r t h e stacking faults (2).
The general f e a t u r e s of t h e results using warm die-drawn crystals lie between those of slow freez- ing, showing neaily pure Debye type relaxation, and those of crushed ice, having a widely dis- tributed relaxation time. Samples oriented in a n e l e c t r i c field a t 90 and 45 degrees with respect t o t h e C-axis tended t o have shorter low-frequency tails. Longer tails were shrunk quickly t o a minimum by annealing. Contrarily, crystals c u t s o a s t o have their C-axes parallel t o t h e elec- t r i c field showed very large low-frequency tails.
c
Fig. 2. Strong negative conductance graduallydecreased and became positive.
- 200
10 20 30 40 50
k'
One of t h e C 45-degree c u t samples showed strong negative conductance, extending t h e Cole- Cole plot t o t h e negative side of t h e K'-K" plane, a s shown in Figure 2. This negative conduc- t a n c e gradually diminished a s t h e straining progressed and t h e sample eventually showed normal behavior. A t t e m p t s t o reproduce such persistent negative conductance samples have failed, though short-lasting negative tails have been observed frequently.
Alcohol-polished samples
Both sides of a C 45-degree c u t sample were polished using a polishing cloth impregnated with a 40% alcohol, 60% water solution (Buehler Microcloth) and a holder similar t o t h a t described by Takei and Maeno (3). About 1 mm was removed from e a c h side of t h e bandsaw-cut sample.
Interestingly, t h e Debye index remained relatively low (0.7 t o 0.9). Relaxation strength s t a r t e d low (40) and went up t o 8 0 a f t e r slight straining. Lower relaxation strength was expected for this sample since dislocation density would be low with l i t t l e straining.
Sanded samples
Sanding was used t o make f l a t surfaces in studies by vonHippel e t al. (4). In our studies a sam- ple was oriented and c u t by a bandsaw s o i t s C-axis was p e r p e n d i ~ u l a r t o t h e e l e c t r i c field. The sample was f i r s t microtomed t o form a slab about 6 mm thick and then both surfaces were sand- ed with #220 sandpaper. The Cole-Cole plot was rather flat, but t h e Debye index had a relative- ly high value of 0.85 t o 0.9. Relaxation strength was generally 90 t o 100, considerably g r e a t e r
JOURNAL DE PHYSIQUE
than t h e strain-free samples prepared using t h e alcohol polishing or die-drawing methods but comparable t o t h e plastically strained samples discussed in t h e accompanying paper.
Crushed single-crystal samples
A piece of single-crystal i c e was crushed between 4.05-cm-diameter anvils at 15 MPa in a vac- uum. A c e n t r a l piece was c u t and again compressed between t h e anvils at 21 MPa t o make a 4.8-mm-thick disk. The sample was chilled t o -6Z°C in a cold box prior t o preparation, and im- mediately a f t e r i t was made and mounted between electrodes i t was placed back in t h e cold box t o prevent recrystallization.
We s t a r t e d with very pure i c e and crushed i t in a vacuum t o make a solid column of ice. So al- though t h e sample produced was very fine-grained i t s extensive grain boundary surface ran l i t t l e risk of being exposed t o contamination. Also, we s t a r t e d at a lower t e m p e r a t u r e s o t h a t residu- a l s t r e s s would remain high. A large deviation f r o m Debye relaxation, a s would b e expected, was observed even a f t e r 2 weeks of annealing above -S°C. Most of t h e s t r e s s should have been relaxed, and considerable grain growth took place within this annealing time; grain boundary conduction would be t h e major source of deviation a f t e r annealing.
Dislocation-free sample
The dielectric properties of dislocation-free a r e a s of hoarfrost crystals were measured using electrode System A (micro-mercury electrodes). As discussed previously, very small dielectric relaxation was observed a t about one-tenth of t h e normally observed relaxation time. The relax- ation s t r e n g t h and t i m e w e r e changed t o n e a r normal by scratching t h e crystal s u r f a c e with t h e t i p of a hypodermic needle. A detailed discussion of this has already been published (5); no fur- t h e r discussion will be given.
3. Discussion
Complex behavior was observed in connection with t h e various single-crystal sample preparation methods. Though dispersion seemed close t o t h e Debye type, t h e observed Debye index [ratio of relaxation t i m e s R t ( ~ " x w)/Rt(~"/wj] was generally low (0.7-0.9), and t h e frequency range required t o produce a r a t i o larger t h a n 0.8 seldom covered m o r e t h a n one decade.
Relaxation closest t o t h e Debye t y p e was observed in t h e samples frozen slowly between t h e electrodes, e v e n though lower frequency conductance showed a strong negative trend at t h e beginning of t h e series of measurements. This was not a single c r y s t a l but a coarse-grained poly- crystal sample, however. When t h e seed crystal filled t h e s p a c e between t h e electrodes, t h e results we observed constituted mostly t h e seed crystal, but air g a p e f f e c t s should b e absent.
Polycrystalline ice made by crushing single crystals between anvils produced a heavily skewed Cole-Cole semicircle and a broad relaxation peak a t -40°C. A f t e r prolonged annealing above -S°C f o r over 10 days, a slightly f l a t t e r Cole-Cole semicircle with a large t a i l developed. How- ever, t h e relaxation peak remained broad. Considerable flattening of t h e Cole-Cole semicircle a t lower t e m p e r a t u r e was observed in t h e second cooling-warming cycle a f t e r annealing.
Computer model studies of a i r gap and film m a t e r i a l e f f e c t s on purely Debye dispersion mater- ials indicated t h a t t h e Debye index was larger than 0.85 and covered at least one decade of fre- quency range unless t h e a i r g a p was very wide (0.5 mm) on t h e model. Also, t h e K' vs K " x w plot on t h e computer model has a long, straight portion, unlike t h e real crystal. On t h e real c r y s t a l t h e K' vs K " x w plot was never straight, indicating t h a t t h e relaxation s p e c t r a w e r e generally spread, and real Debye relaxation was seldom observed in t h e extended frequency range.
As discussed in t h e accompanying paper, dielectric dispersion c a n be modified considerably by plastic deformation. Moreoever, t h e dielectric dispersion observed o n dislocation-free ice w a s very small, indicating t h e strong e f f e c t s of dislocations on t h e dielectric properties of ice.
P a r t of t h e observed behavior could b e explained by point d e f e c t theory. An increase in relaxas tion strength could b e caused by a n increase in point defects. However, t h e s a m e theory would have to explain t h e increase in relaxation time. And i t should b e able t o describe t h e distribut- e d dispersion and negative conductance a n d capacitance. (Or, alternatively, a s e p a r a t e theory would have t o be developed.)
Many of t h e observed properties, including t h e linear relationship between dielectric relaxation t i m e and strength and t h e distribution of relaxation time, c a n be easily explained by charged dislocation theory. Also, some anomalies observed during this study, such a s negative conduc- tance, negative capacitance, and occasional sudden disturbances in bridge balancing, might pos- sibly be explained by charged dislocation theory.
References
(1) Auty, R.P. and R.H. Cole, J. Chem. Phys. 20 (1952) 1309-1314.
(2) Hondo, T., T. Itoh and A. Higashi, Jap. J. Appl. Phys. 20 (1981) L737-L740.
(3) Takei, I. and N. Maeno, Low Temp. Sci. 44 (1985) 177-181.
(4) vonHippe1, A., D.B. Knoll and W.B. Westphal, J. Chem. Phys. 54 11971) 134-144.
(5) Itagaki, K., J. Glaciol. 21 (1978) 207-217.
COMMENTS
G.W. G R O S S
Have you c o n s i d e r e d e l e c t r o d e p o l a r i z a t i o n a s a p o s s i b l e r e a s o n f o r f ' n e g a t i v e e l e c t r i c a l c o n d u c t i v i t y f f ?
Answer :
I had n o t t a k e n i n t o a c c o u n t any e l e c t r o d e p o l a r i z a t i o n h e r e , b u t I f e e l it may a p p e a r o u t s i d e o f my measurements.
