!

DIELECTRIC MATERIALSMEASUREMENTS

Network Measurements Division1400 Fountaingrove ParkwaySanta Rosa, California 95403

AUTHORS:David Blackham

Frank DavidDavid Engelder

RF & MicrowaveMeasurementsSymposiumandExhibition

F/in- HEWLETT~~ PACKARD© 1990 Hewlett-Packard Company

www.HPARCHIVE.com

ABSTRACTThe electro-magnetic properties of materials must be accurately measured before these materials can be skillfully applied.This paper covers the basics of permittivity (dielectric constant) and permeability at high frequencies, includingterminology and data formats used in the field

Next, it surveys a variety of measurement methods based on RF/microwave network analyzers, and discusses the strengthsand limitations of each. This includes recent enhancements to the popular "S-parameter method" (HP Product Note 8510-3),plus a new coaxial dielectric probe for making permittivity measurements easier and more convenient.

Finally, the presentation addresses some applications, which require accurate dielectric measurements.

AUTHORS

David Blackham is an R&D Development Engineer at HP's Network Measurements Division in Santa Rosa, California. Hereceived his BSEE from Brigham Young University and MSEM from Stanford University. David joined HP in 1979 andworked on the HP 8340A Synthesized Sweeper, scalar detectors and bridges, and microwave vector network analyzers. Heis currently working on material characterization.

Frank David is an R&D Project Manager at HP's Network Measurements Division in Santa Rosa, California. He earned hisBSEE from the University of California at Berkeley, and his MSEE from Oregon State University. After joining HP in 1969,Frank designed microwave components for the HP 8755 Frequency Response 'lest Set and HP 8566A Spectrum Analyzer.He later managed the millimeter-wave mixer and HP 8720 Network Analyzer projects.

David Engelder received a BSEE from the University of California, Santa Barbara, in 1979. He joined HP's NetworkMeasurements Division in Santa Rosa, California, in 1980 - working first as an engineer and later as a manager in theProduct Support area of Marketing. Dave is currently in Product Marketing, promoting the HP 8720 and applyingmicrowave network analyzers to dielectric materials measurements.

www.HPARCHIVE.com

,

Dielectric Materials Measurements

Slide 8971

rL~ HEWLETT~~ PACKARD

Slide 7861S

3

HIGH FREQUENCY

DIELECTRIC MATERIALSMEASUREMENTS

--_.,--.....­.....W---.leDWt __woo,... .,.......... c. ....

JV4 __~...­-­..-liilJ=

p.,..,

ELECTRO-MAGNETIC SPECTRUM

10' 10' 10' 10' 10" 10" 10"

III11I111111111111 I f~IF WICt'Cw",. JR uv~ ~I!i.:.:.:.:.l!l~

~~ I~RF mm-w.... V X-Ray

Swept Frequency Stimulus-Reponse

--------------m~~This presentation concerns the electro-magnetic propertiesof materials at high frequencies; specifically:• Pennittivity (or dielectric constant); and• Permeability

It is important to understand and measure these propertiesto skillfully apply materials in a given application. In thetraditional electronic industries, this information isobviously needed for solid design and quality control. Inaddition, this paper will also relate these properties toindustrial high-power microwave heating and drying.

Slide 8972A

OUTLINE

• BASICSElectro-Magnetic Principles

Common Examples

Terminology

Data Formats

" MEASUREMENT METHODS

e APPLICATIONS

--------------m =:.:;';--'

The laws of "electro-magnetics" describe the behavior ofelectric fields and magnetic fields. In the time-varying case(e.g. a sinusoid), both kinds of fields appear together. Thisradiation can propogate through free space, or throughmaterials. As frequency changes across the spectrum, theelectro-magnetic radiation appears in many different forms.However, the same basic laws apply.

Today, we are concerned with the RF and microwave part ofthe spectrum.

Slide 8973

iNTERACTIONSBetween EM-Fields and Materials

Electric -- -.~MagnetiCFields~ ~ Fields

Permittivity Permeability

f; = f; - If;'L-J----STOfiAGE

1_.---- LOSS

Dielectric Constant K" PIl.TM7J

"'---------------M=~ ~The presentation has three parts:• Basics• Measurement Methods• Applications

Let's begin by reviewing the basic principles and termsused in this field.

---------------m~--'

When electric and magnetic fields pass through a material,each can interact with that material in two ways:

• Storage: Energy may be exchanged between the field andthe material, in a bi-directional (lossless) manner.

• Loss: Energy may be permanently lost from the field, andabsorbed in the material (usually as heat).

The electric interactions are quantified by permittivity(£r*), also called dielectric constant (K*). The magneticproperties are described by permeability (~*). These arecomplex numbers with two parts:

• Real Part: Represents storage term; denoted with '.

• Imaginary Part: Represents loss term; denoted with .'.

This paper focuses on permittivity, since many commonmaterials are completely non-magnetic.

www.HPARCHIVE.com

W13 ~EWLETT_4__-=..="?'....=....=......:;,A..:;:C:.;KA=.:..::R:D:....- ...- ..;;;D;;.,;i:.:e:.::l..::,ec=tr:..:i:.::c~M=a:.:t;.::e..::,r.::.:ia=l=s:..:M=e::..:a::.:su::..=.:r:..:e::..:m:::.:::e:.:;n:..:t::::....s

__L..ltul__

100,.--

AJu...!!!.."o!----+------iI----+--cA"'....-::....::;---j

Slide 8751S

"LOSS TANGENT"

ta n6 ~ D (Dlaalpaflon Fact""

,.i Gnyarae Quality Fact""

ex Ena'VY Loat par Cycla

Enargy Storad par Cycle

..""',---------------m~

Complex permittivity can be drawn as a simple vectordiagram. The real and imaginary parts form a 90-degreeangle to each other. The vector sum forms an angle 0 withthej-axis.

The "lossiness" of a material is the ratio of energy lost toenergy stored, or £"f£'. This ratio is the tangent of the angle0- so people call it:a tan 0 ("tan delta")• loss tangent~ tangent loss

Tan 0 is directly related to D (dissipation factor) and Q(quality factor). -------------Slide 8752

DIELECTRIC PROPERTIES OF

COMMON MATERIA S__HlgItLOu-_

.w ...,...f---~~---+_--_+----+--_'____I

$1_.. I---'-I---+_--_+----+-------i

E; ..• 1---,_,........,...+---;.:PC::.cBo=..'1---.--!I----_--+-~::--!

eWytar k., I---T~on::----+---=.:'-I-...,o~,,~··:::.···+--'''''------;

AIr

Slide 8749A

DIELECTRIC MECHANISMS

lc(~G G 1: r ~~t G 1

III -I

~~lc(.J 010 l'- IDipolar ionic Atomic Electronic

fOr1antatlon) (SpaceCBonda)

Charge) ------------------m~

At the microscopic level, several dielectric mechanisms cancontribute to dielectric behavior:

Dipole orientation (and ionic conduction) interact stronglyat microwave frequencies. Water molecules, for example,are permanent dipoles, which rotate to follow analternating electric field. These mechanisms are quite lossy- which explains why food heats in a microwave oven.

Atomic and electronic mechanisms are relatively weak, andusually constant over the microwave region.

Slide 8786S

ACROSS THE FUll SPECTRUM

(if

.....---------------IH~This graph has £' (storage) along the vertical axis, and tan 0(loss) on the horizontal axis (both logarithmic). The valuesfor several common materials are shown as dots (at a singlefrequency and temperature).

"Low-loss" materials, such as Teflon, have small values oftan o. They are commonly used in electronic applicationssuch as: insulators (e.g. for cables), substrates, anddielectric resonators.

"High-loss" materials include water, food, and many naturalmaterials. These materials quickly absorb microwaveenergy, and so are not used for electronic components.However, they are important to:• Understanding microwave radiation in the "real world"• Material analysis (e.g. moisture content)• High-power microwave processing (heating and drying)

t...... ..,.. .001 ...01

tan6 =~E', ..",., .....---------------m~

Each dielectric mechanism has a characteristic "cutofffrequency." As frequency increases, the slow mechanismsdrop out in turn, leaving the faster ones to contribute to £'.

The magnitude and "cutoff frequency" of each mechanism isunique for different materials.

Water has a strong dipolar effect at low frequencies - butits dielectric constant rolls off dramatically around 20 CHz.Teflon, on the other hand, has no dipolar mechanisms - soits permittivity is remarkably constant well into themillimeter-wave region.

www.HPARCHIVE.com

Dielectric Materials MeasurementsHEWLETTPACKARD 5

Slide 8754 Slide 8756S

RELAXATION TIME T REFLECTION

"""""""""~~~~~~~~'~~~~~~~"""'~"""'""""" ~"""""""" 1"'''''''''""""""""""""""""""""""""""""""-~"""""""H' ~""'"" "~ tA~""

" "" U'lilI""" ...." ""'"'" """""" .... '""""",... ~ """"""""""""""""""""""""""""""""""""""""""""""

oWlter It 20 CMaximum Ablorptlon:

f.'" 22GHz1

Q.- '= 15tln6

• ......

~~;'JO,.

V.. ......./

10

100

f

'Time required for ~ of a perlurbed (aligned) IYllem10 relurn 10 equilibrium (random).'

1 GHz to GHz 100 GMJ' BOUNDARY

-----------------m~For dipolar dielectrics (such as water), a "relaxationconstant" 1: describes the time required for dipoles tobecome oriented in an electric field. (Or the time needed forthermal agitation to disorient the dipoles after the electricfield is removed.) At low frequencies, the dipoles can"follow" the field and E' will be high.

At high frequencies, the dipoles can't follow the rapidlychanging field - and E' falls off.

The loss factor E" peaks at the frequency 1!'t. Here, energy istransferred into the material at the fastest possible rate.

'----------------m~Consider a plane boundary formed by two materials (e.g. adielectric slab in air). The electro-magnetic impedance of amedium depends on E* and Il*, so the boundary has animpedance mismatch. Incident radiation will be partlyreflected, and partly transmitted into the dielectric.

E'=2.04 for Teflon, which is fairly close to air (E'=I). Thereflection coefficient at this boundary is roughly 0.17 (areturn loss of 15 dB).

For water, E'=80. The large mismatch reflects about 80% ofthe field back into air (return loss of only 2 dB). Forheating/drying applications, boundary reflections are animportant part of the overall efficiency.

Slide 8755Slide 8974

-----------------m~Once inside a material, energy is "lost" (absorbed as heat)at a rate dependent on E*. The field strength follows anexponential decay, related to distance d by an attenuationfactor a. One can also define a "penetration depth" at whichthe field strength or power has dropped by 112 or lie.

Teflon is very low loss - just 0.0006 dB/em. The fieldsdecay very slowly over a large distance.

Water, on the other hand, loses 3.9 dB/cm (at 3 GHz).Energy is transferred into the water over a very shortdistance.

d2

-2 DC dP (d) = PT e

-2 DC dE (ell = ET e

DC =: 10.5 f.JT: (tan 6l (t In GHz)

ATTENUAr~ON/PIENIETRATIONE

Boundary /0

COLE-COLE PLOTS(',..",

20·C40

:so "",--~ ...a...; .20

/ eo.c 2s.1 , , ,1.14 ~

10 \ \1_4.e3~

0.51(0

0 10 20 .0 .., 50 eO 70 10 00 IR..,PATI~

----------------m~

For high-loss materials, both E' and r." change dramaticallywith frequency. A Cole-Cole plot (similar to a Smith chart)is often used to plot the "frequency response" of materials.Simple lossy materials (e.g. water) scribe a semi-circle on aCole-Cole plot. More complex materials may form anellipse, or an arc with bumps on it.

The two traces demonstrate that E* (for water) changesdramatically with temperature, as well as frequency.

•

www.HPARCHIVE.com

6rcw HEWLETTIL..,...... PACKARD Dielectric Materials Measurements

Slide 8975 Slide 8760S

Function of Fr..,uIJncy, TtJmperlllurtJ, Etc.

SUMMARY

DFIXTURE

"~';'* 0.!",,-~.!. .s,,~

U.ECTlIIC....tth..y

1Dl.-c1rio Cefl••• 1'l1 K)

(. (- "

' .....1

..n'.~.D ...!-t, Q

_ Ion

Att t1.,

-----------------m~These are the key points - the properties that describe theinteraction between electro-magnetic fields and materials.Of course, these properties are the foundation for applyingmaterials. But they also form the basis for the variousmethods to measure these properties.

Slide 8972B

.....--------------m~HP's network analyzers make a good foundation for allswept high-frequency stimulus-response measurements. 'Ibmeasure dielectric materials, two additional elements areneeded:• Fixture - contains material; applies electro-magneticfields in a predictable way; allows connection to networkanalyzer.• Software - converts S-parameter data from networkanalyzer into £* and/or 1.1.* (depends on fixture).

Each method (and fixture) has strengths and weaknesses.

OUTLINESlide 8772S

www.HPARCHIVE.com

o BASICS

• MEASUREMENT METHODSResonant CavityTransmls sion-LineCoaxial Probe

o APPLICATIONS

-----------------m~The presentation now turns to three methods used tomeasure e* and/or 1.1.* over frequency.

(Note: All three approaches are based on RF or microwavenetwork analyzers. Other methods exist, but are notaddressed in this paper.)

CAVITY METHODS

+.~,

'.

"".ms

-----------------m~

Cavities are high-Q resonant structures. A sample of thematerial-under-test (MDT), inside the cavity, affects itscenter frequency and Q (quality factor). From these twoparameters, the complex permittivity of the sample - bothe' and e" - can be calculated.

Two approaches can be used:• Perturbation: measure fc and Q of empty cavity, then findshifts caused by small sample in cavity.• Absolute: directly calculate relation from fc plus Q to e*(requires precise understanding of fields in cavity).

Dielectric Materials MeasurementsrI3 HEWLETT~PACKARD 7

Slide 8773S Slide 8766

CAVITY METHODS TRANSMISSION - UNES-Parameter Method

Strength.

'ftry ..nonl.. to low tan 6

(to 10"')

Weakne••e.

Sing" "~uency per ca'llty

Complex anolyo"

PrK:l•• ••mp~ .hap.

(d..tructlve)

Few fixture. avaUabie

Tranaml••lon(S%I)

..,.........---------------m~

Cavity methods are very accurate, and can measure tan /) ofvery low-loss materials with high precision.

However, most cavities resonate (and yield E*) at discretefrequencies. The analysis can be very complex, especiallyfor the absolute approach. The MUT must be destructivelysampled, and formed into a precise shape. And few fIxturesare commercially available - most users design andmanufacture their own.

......---------------m~The measurement can be based on the reflection coefficient(S11), transmission coefficient (S21), or both.

The popular "S-parameter" approach (Nicolson-Ross orWeir) uses both S11 and S21 to calculate both E* and ~*. HPdescribed this technique, adapted for the HP 8510, inProduct Note 8510-3 (Aug 85).

Note that S11 and Szl are composed of multiple-reflectionsfrom both boundaries.

Slide 8764 Slide

z

IIw

• e..l., to .r.p.

• 8.~ '-.0-".2 -12." GH»

• Bre.dband •.0. DC - 11 GH.

• "or- dlffleutt til '''-IM

TRANSMISSION - LINE METHODS

"----------------m~Transmission-line methods involve putting the MUT insidea portion of an enclosed transmission line. The line isusually rectangular waveguide or a coaxial airline.

·'~.-,-,......l~"",,"""'~f:,.;J-'-'-~,!-:-,-'-'~""--f-:!,-"""'-o;-:IlJnICOUDCT (048"

.....---------------m~These traces are E' versus frequency for Teflon, measured ina broadband coaxial transmission line with an HP 8510.(Results courtesy ofNIST.)

The blue trace uses the traditional Nicolson-Rossalgorithm. The periodic drop-outs occur at frequencieswhere the sample is nA/2 wavelengths thick. There, thereflections from the two boundaries exactly cancel eachother, and S11 drops to zero. Researchers usually avoid thisproblem by keeping the sample thickness below A/2 at thehighest frequency.

The red trace is actual data, using the same equipment andS-parameter data as before. However, new algorithms ­developed by Baker-Jarvis at NIST - have solved the nA/2dropout problem. This method now works well withsamples much longer than A/2 - yielding better results fortan Ii on low-loss materials. (Note: New NIST algorithmsapply to non-magnetic materials.)

www.HPARCHIVE.com

8~HEWLETT

~~ PACKARD Dielectric Materials Measurements

Slide 8768 Slide 8981

SiMP IflED MODel fOR COAX PROBETRANSMISSION

I Strength. IYlolda (' and 1"

Broad fr04loncy ranllOIto 110 GHJl

LINE METHODS

I Weaknesses IUmked low-Io.. roaoluUon

f > 500 MHz!banded In wavoguldo) 11001 I(~) .... Ie: J Radlafloft

Simple fh,Iu,oa...ay tocuotomlzo

Adoptabl. to "fr• .-apaeo"

Proeloe aamplo ahape

(deatruetlvo)

'----------,.. f

Slide 8769

'----------------m=:;:;!;

f; MAPPED ON $11

Slide 8982

---------------- m =:;:;!;Ultimately, the model relates Su of the probe-pIus-MDT(measured by a network analyzer) to E* for the material.Here, a grid of E' and tan 0 is mapped over a polar chart ofS» for a single frequency.

The network analyzer's uncertainty in measuring S» is asmall uniform circle on the polar chart. This map, then, alsoindicates the resolution in measuring E* due to the Slluncertainty. Where the E* lines are widely spaced, the smallS1I uncertainty causes little error in measuring E*. Whereclosely spaced, the E* resolution is not as good.

......

'---------------m~The geometry can be modeled with a simple circuit, wherethe total admittance Y is composed of:• A pure capacitance (phase change), related to E' (storageinMUT).• A conductance (magnitude change), related to E" (losses inMUT).• An additional conductance, represented radiation losses.

This model yields a single-pole frequency response. Theprobe's sensitivity in measuring E' is related to the slope ofthe phase response. Hence, the accuracy in E' of this methodis optimum over a 2-decade frequency range, depending onthe MUT. Radiation losses vary with E', E", and frequency­the model must be quite complex to accurately account forthem. These losses limit the measurement sensitivity in E"(or tan 0), especially at higher frequencies. This is becauseGr is in parallel with GE" - as Gr grows more conductive, itbecomes more difficult to determine GE" from Y.

Solida Uqulda

5" ~w~tO.I "",,.,

--Reflection

(5" I

Coax Open-EndedPROBE METHOD

I. hlA••The third method is simply an open-ended coaxial line. TheMDT is measured by simply touching the probe to a flatface of a solid, or immersing its end into a liquid. The fieldsat the probe end "fringe" into the material, causing areflection (S11) that can be related to E*.

'----------------m~Transmission-line methods have advantages over cavitymethods: They yield swept-frequency results over a widerange of frequencies. The Nicolson-Ross algorithms provideboth E* and ~*. And transmission-line fixtures are easier tobuild and analyze.

However, the transmission-line methods are not as accurateor sensitive to low tan o. Sample preparation is just asdifficult, and is usually destructive.

(Note: As of July 1990, HP offers the software to performthese calculations as the HP 85071A MaterialsMeasurement Software. HP also supplies waveguide andcoax accessories to serve as fixtures.)

www.HPARCHIVE.com

Dielectric Materials Measurements~ HEWLETT~~ PACKARD 9

Slide 8983A Slide 8984B

Slide 8985

2...-to•

"----------------m~These plots demonstrate the probe's resolution versusfrequency, for various values of £' and tan o. The verticalaxis represents both %-uncertainty in £' and absoluteuncertainty in tan o. Obviously, the probe's accuracydepends on:• £' of the MUT• £" or tan 0 of the MUT• Frequency

These graphs assume that 8 11 uncertainty is 0.05 at allfrequencies, and plot the corresponding maximum error in£*. In fact, the specified 8 11 uncertainty is often smaller,and the actual 8 11 accuracy is typically much smaller still.However, there may be other sources of error - such ascalibration quality or contact gaps between the probe tipand a solid MUT.

----------------m~

f; MAPPED ON 5 11

..

........

These plots show how the map of £' and tan 0 can changewith frequency. At this low frequency, the £* lines areclosely spaced for virtually all values.

Slide 8983B

PROBE CALIBRATION

"----------------m=These plots map £' and tan 0 at a high frequency. Thenetwork analyzer's uncertainty in 8 11 would be negligiblewhen £' is low, and would become more significant as £'

increased.

Simil.r 10 I-port (3-tlirm) c.1

for vliclor nlitlt'Ork .n.lyzlir

(DirliClivity, Tr.cking, Sourcli M.lch)

1. Un model to predict 5" of 3 standarda::> Air *,C· '0o ShOf1(RC--1)

:) U.aor-e.f1Md standard (.-G. Wti.ri

2. Measure standards

3. Calculate difference arrays and apply

'correction" to later measurements

----------------m~For best results, the probe (with network analyzer) shouldbe "calibrated" by measuring known standards. The modelis used to predict 8 11 for various standard materials, suchas air. The difference between this predicted value and theactual measured 8 11 is used to correct subsequentmeasurements.

www.HPARCHIVE.com

10rcw HEWLETT~PACKARD Dielectric Materials Measurements

Slide 8986C Slide 8987

MEASUREMENT RESULTSWaler & Alcohols n.ossy liquidS> - Rear

COMPARING METHODS

a._

••.• ,...._.-r_-.,.'--.:.:Th::;:•.::Ory~_...,...--=;:=-.:M:;.::a.::au::;r..t=-....,....~

.... l=:::t::::tl--..-t-;w~a±.,::-+---+-+---t-+-l

.... t--+-+--+-+---f--::::::"-l-~-+~L---l-.... t--+-+--+-+---+--l----J--+~-~_____=I....... 101 than I

.... P\~r-f-"'od--__-'--"'-'+'--+-+----+--I------+--+----l

I:-Elt1~a~olE~±=E3:=E3=E~~~e.• I:..nGIV

Flexibility / Convenience

r,....-_"

'--------------m~~Here are some actual measurement results, using theprobe:

For Teflon (not shown), the values of E' are consistent withthe literature across the entire frequency range. Theprevious slides predicted the poor resolution at lowfrequencies, since E' is so low. (The E" data is not useful forsuch a low-loss material.)

For water, E' again agrees with published values. Since E' ishigh, there is no noise at low frequencies - including 2.45,GHz. E" also measures as expected.

'--------------m~---ILet's summarize the three methods, while comparing theirkey strengths and weakness. This chart plots accuracyversus convenience.

Cavity methods offer the highest accuracy, especially forlow tan o. But they are narrowband, complicated, andrequire destructive sample preparation.

Transmission-line methods provide the widest possiblefrequency range - and can determine both E* and ll-*.However, they are not as accurate as cavities, and stillrequire destructive sample preparation.

The probe method is by far the most convenient: it yieldswide-band results, and requires no sample preparation(non-destructive). The probe is especially convenient forliquids and semi-solids (e.g, food and biological materials).Generally, however, the probe offers less precision thanother methods.

Weakneaaea

PROBE METHOD

IStrength.

Slide 8771

Convenient.•••y-to-u..- No aampl. pr.p.atlon

- lJaua"y non-<loatructlv.- qulda, food., organic.

Brosd frequency rang.

t200 101Hz - 20 GHz.dspendlng on £;)

£; only (non-magnetic)

Slide 8972C

OUTLINEo BASICS

o MEASUREMENT METHODS

e APPUCATIONS

"'--------------IiJ~~The key advantage of the probe method is its convenience:for most materials, no sample preparation is required andthe method is non-destructive. The probe works well forliquids and semi-solids - making it popular for food andbiological researchers.

While the probe provides reasonable results over a 2-decadebandwidth, 't is not as accurate as other methods andcannot resolve E or tan Ii for low-loss materials.

(Note: As of July 1990, a probe - with its dedicatedsoftware and accessories - are offered as HP 85070ADielectric Probe.)

Electronics

Industrial Heating/Drying

---------------IiJ~~Let's see how these measurements can help in applyingmaterials and microwaves in practice. We11 touch briefly ontheir use in electronic components - then explore world ofcommercial microwave industrial processing.

www.HPARCHIVE.com

Dielectric Materials Measurementsr1~ HEWLETT.d~ PACKARD 11

Slide 8752 Slide 8989

f; II) f----"'....."'M1'"'''..../---+--___iI---+-:-:..Ieo.,..,..,.,'''''::---l

2.<450 '!: 0.050 GHz

5.800 '!: 0.075 GHz

2<4.125 + 0.125 GHz

61.250 : 0.250 GHz

122.500 : 0.500 GHz

245.000: 1.000 GHz

INDUSTlRlAl / SCiENTIFiC / MEDiCAL(WISM W) FIREQUENCIES

6.780 :! 0.015 MHz

13.560 '!: 0.007 MHz

27.120 + 0.163 MHz

<40.680 :! 0.020 MHz

(<433.920 :! 0.870 MHz)

915.000 + 13.000 MHz

""""

.t ,.0001 .001 ••01

lan6 =~£;

.."

• f-----:----:-+-.;I'C-=""::r:..:;-_____if---+----=.o-:-::-::___i

, f-_-_rizT.ii=o_..,.,_ar_+_--'...:..:o.___i'-~-:':'..::::--_t·-··;Io"'·~"----··-..___iT ~0Il ...

DIELECTRIC PROPERTIES OF

COMMON MATERIALS.. r--__- ....-=--;:_L-_L_.._.::;===-_---,_-_-_-_"'..:.'g/I.LN=-·~_- -fit-..,

"'~ ...... t-__T1...:OZ'+-__---l -+-__-+_o__o"':0,"----"

St_,. f----+---f---+-----j-.:.:.:..:-----l

•

•

""-------------M~~Here again is the map of materials.

Low-loss materials are commonly used in electronics (alongwith lossy absorbers or shielding). When used in cables, E'affects the impedance and E" the loss. In substrates, E' setsthe impedance/capacitance of lines. Used in dielectricresonators, E' sets the frequency for a given geometry, whileE" determines the Q and spectral purity. Materials are alsoused in ferrites, antenna lenses, windows, radomes, andwaveguides - £* is always the critical design parameter.

High-loss materials are rarely used in electronics. However,the same microwave expertise and technology (historicallydeveloped for defense applications) is required to applyhigh-power microwave energy to (commercial) industrialprocessing. (The following topics will reinforce the theoryjust presented, regardless of application.)

""--------------M~~The "ISM" frequencies are those allocated for industrialprocessing in the US. (Most other nations allocate asimilar list).

However, the vast majority of today's industrialprocessing (and home microwave ovens) use only the2.45 GHz band. Might other approved frequencies bebetter choices?

Yes! Let's see why ...

Slide 8991

DIFF'Eli1IEN1IAl HEATING V5 fMicrowave Energy "Seekso Higher Loss Material

.....---------------M~One reason microwave processing can be more "effective" isdifferential heating.

Consider a mixture of two materials, where the loss factors£1" and ~" are different. Microwave fields will couple morestrongly with the higher-loss material, and heat itselectively.

In making paper, this effect is used to "l.evel" the moisture(make it uniform across the web). In another example,selective heating removed high-loss contaminants (sulfercompounds) from low-loss coal.

Obviously, measuring how £1" and ~" vary with frequencyhelps to choose the frequency that optimizes differentialheating.

Slide 8988

MICROWAVE PROCIESSi~G

EXAMPLES ADVANTAGES

• T..III.. • L.ew.f' fMt-ri.1 l""Po

Drying: '''- • Gr...... ,.n.tratiOfl ~th.- • ..-.tu,. ,...111'1'

• Pu'. • ,..1... ,,...-11'10

• C..I • c..t11'1WUa 'rec.... lno

fMd.....t~"G • er..t.r ~""t1en "Itl

".....lno: • 8.Ml'I .....lng• &...wIot __1.-1_1 _.....

• f,..u .....)l'dntlfH'l • Cdnwu. ,,....1"0

• Blanchll'lG' • ,..1... prec••.tnCl

...- • Au..., wlMnlau.n • Gr...., P«t.tr.11on ""'hPi_tic

• Ptntlre curing• L...,., aw.e. .....p.

PT.......g: • F••ttor proc_all\;

c...... • Oryll'l'• Fa"" proc...lno• Gr.....,. P«'l.tr.tMHt ~!'I

..,......Ing: • 11ft_In, • Cen.. In.... r.maln c.oI PIl.T""

""-------------m~~This is a partial list of applications where industrialmicrowave processing has been used successfully. Becauseof its unique advantages, microwave processing hasdisplaced conventional heating in many applications ­even though microwave energy is more expensive on a cost­per-Joule basis.

We'll see now that a solid understanding of dielectricbehavior, and the ability to measure £* over frequency andtemperature, enable us to confidently and analyticallydesign microwave processing systems.

f:~;C

• \ nil),.Dry Water from Paper

Moisture Levelling

Heat Sulfer from Coal

PATt99\

www.HPARCHIVE.com

12..... HEWLETT~~ PACKARD Dielectric Materials Measurements

Slide 8995

PENETRATION DEPTH vs fOptimize for Uniform Heating

Slide 8990

EFFICIENCY vs f

Water: 2,45 GHz V$ 24.125 GHz

APPROXIMATEFREQUENCY HALF-POWER DEPTH (mm)

Water Ice Rubber Wood

(2<fCl (-12"Cl (~C) (2~C)

0.915 GHz 116 557 26 7902.-450 GHz 9 8,000 15 360

24.125 GHz 0.25 -- 4 32

E;II

•••

Frequency (GHz)

1k ,.~.... ......

Temperature ( C)

•

,---------------IiJ~Another reason microwave heating can be more effective isits ability to penetrate materials and unifonnly heat theentire volume of an object. (Conventional approaches, bycontrast, heat the outside surface and rely on conduction totransmit heat to the core.)

A good example is vulcanizing or molding rubber (e.g. tires).Since rubber has high thermal insulation, the heatconduction from surface to core is slow and time-consuming.Higher surface temperatures can't be used, since thesurface layer of the product would be degraded. Instead,microwave heating raises the temperature quickly andunifonnly, thoughout the object's volume. For best results,the penetration depth must be optmized for the E* of thematerial and the physical dimensions of the object. Aknowledge of E* versus frequency is crucial. In many cases,the operating frequency is the only parameter in which onehas any choice.

Slide 8996

EFFICIENCY vs fPowor Tronaforrod from E-Fleld to Moterlol

ike?;FttqJMlcy

-----------------IiJ~For a given electric field strength, the power dissipated in amaterial is proportional to E" and f. In turn, E" is a functionoffrequency (with a typical shape, as shown). The top curveis power dissipation - the product of (' and f. If we assumea fixed field strength, the efficiency of power transfer intothe material can be maximized by choosing the rightfrequency.

,---------------IiJ~

In practice, other concerns - such as penetration depth ­may be more important than simple efficiency. But considerthe problem of drying thin sheets of material (paper, fabric,wood veneer). In this case, maximum power absorption isessential. The left plot shows E" for water across frequency,at several temperatures.

Taken at two ISM frequencies - 2.45 GHz and 24.125 GHz- this same data can be plotted versus temperature. Notethat En is much higher at 24.125 GHz, for all temperaturesbetween freezing and boiling.

Slide 8997

EFFICIENCY vs fWater: 2,45 GHz V$ 24.125 GHz

2

Po DC f E: I E I

T (·C) f: f P D

20 3x 10x 30x50 6x 10x 60x75 9x 10x 90x

-----------------IiJ=This table compares the power absorption factors forheating water at 2.45 and 24.125 GHz.

En at 24.125 GHz is 3,6, or even 9 times higher than at 2.45GHz. The frequency itself, of course, is about 10 timesgreater. The overall efficiency of power dissipation, then, isEn times f - which is up to 90 times higher at 24.125 GHzthan at 2.45 GHz! (Ironically, this frequency is very rarelyused.)

•

www.HPARCHIVE.com

Dielectric Materials MeasurementsrI~ HEWLETT~~ PACKARD 13

Slide 8992

SAFER PROCESSING

~u dT~" PI = H (T-To ) + mC­

dt

•

""""Time

__---PWED

-e::;...------P LO

......---------------18=There is one final reason to characterize £* beforeprocessing with microwaves: safety.

The top equation describes the energy balance formicrowave heating. The microwave energy absorbed in amaterial (£"P) must equal the energy lost by convection(H term) plus the energy raising the material'stemperature (mC term).

But we know that £" is a function of temperature. Apolynomial can approximate this relation, and be used inthe top equation. Solving the differential equation givescurves of temperature versus time. The three curves are forthree levels of incident microwave power. Two power levelsstabilize over time at a certain temperature. But the thirddemonstrates "thermal runaway" - causing the material tomelt, burn, or even explode.

Slide 8994

SUMMARY &. CONC USIONS• BASICS

EM Fields Interact with MaterialsReal (Storage) and Imaginary (loss)

Reflection, Absorption. Penetration

• MEASUREMENT METHODS

Network Analyzer + Fixture + Software

Cavity - Transmission-Une - Probe

• APPUCATIONS

Quantitative Knowledge Gives Better Results

Need to Know f: versus Frequency

----------------m =:;:;!;In summary, we laid the foundation for a goodunderstanding of dielectric behavior, and described severalmethods for measuring complex permittivity £*.

These properties must be known before materials can beused in electronic components. Microwave designers arevery familiar with the impact of £* in their designs.

But we have also demonstrated the need to characterize£* over frequency to optimally apply microwave energy incommercial processing.

www.HPARCHIVE.com

14r13 HEWL:ETT..~ PACKARD Dielectric Materials Measurements

REFERENCESBASICS:

D.K Cheng; "Field and Wave Electronmagnetics"; Addison­Wesley, 2nd ed (1989)

H.M. Altschuler; "Dielectric Constant"; Ch IX of "Handbookof Microwave Measurements"; ed M. Sucher and J. Fox;Wiley (1963)

A.R. von Hippel (ed); "Dielectric Materials andApplications"; MIT Press (1954) - (Over 35 years old, butstill the bible on dielectrics and measurements; goodintroduction to basics)

MEASUREMENT METHODS:

M. Afsar et al; "Measurement of the Properties ofMaterials"; Proceedings of the IEEE vol 74 no 1 (Jan 1986)- (Excellent survey of many methods across widefrequency range; good starting point to choose method; 187references point to further reading)

Murata Mfg Co, Ltd; Int'l Div; 3-26-11 Nichi-rokugo; Ohta­ku; Tokyo 144; Japan - (Manufactures cavity fixtures)

R.G. Geyer; "Electrodynamics of materials for dielectricmeasurement standardization"; Proc oflEEE, IM-TC (1990)

A.M. Nicolson, G.F. Ross; "Measurement of the IntrinsicProperties of Materials by Time Domain Techniques"; IEEEvol IM-19, p377·382, Nov 1970

W.E. Weir; "Automatic Measurement of Complex DielectricConstant and Permeability at Microwave Frequencies";IEEE, vol 62, no 1, p33-36 (Jan 1974)

"Measuring the Dielectric Constant of Solids with the HP8510 Network Analyzer"; Product Note 8510-3, HP PartNumber 5954-1535 (Aug 1985)

J. Baker..Jarvis; (in press)

D.K. Ghodgaonkar et al; "A Free-Space Method forMeasurement of Dielectric Constants and Loss Tangents atMicrowave Frequencies"; IEEE, vol IM-37, no 3 (June 1989)

E.C Burdette, F.L. Cain, J. Seals; "In Vivo ProbeMeasurement Technique for Determining DielectricProperties at VHF Through Microwave Frequencies"; IEEE,vol MTT-28, no 4 (April 1980)

T.W. Athey, M.A. Stuchly, S.S. Stuchly; "Measurement ofRadio Frequency Permittivity of Biological Tissues with anOpen-Ended Coaxial Line"; IEEE (1982)

M.A. Stuchly, S.S. Stuchly; "Coaxial Line ReflectionMethods for Measuring Dielectric Properties of BiologicalSubstances at Radio and Microwave Frequencies - AReview"; IEEE, vol IM-29, no 3 (September 1980)

C. Gabriel, E. H. Grant; "Dielectric Sensors for IndustrialMicrowave Measurement and Control"; Proc of High­Frequency/Microwave Conf at Arnhem (KEMA,Netherlands; September 1989)

APPLICATIONS:

"Tables of Frequency Allocations"; Mnl of Reg and Proc forFederal Radio Frequency Mgt, u.S. Dept of Commerce

MA Stuchly, S.S. Stuchly; "Industrial, schientific, medicaland domestic applications of microwaves"; lEE Proc, vol130, pt A, no 8, November 1983

S.O. Nelson et al; "Frequency Dependence of the DielectricProperties. of Coal"; J Microwave Power, vol 16, no 3&4,p319-326 (Dec 1981)

"Microwave Power in Industry"; EPRI Report, ch 9, p4-5(Aug 1984)

F. Franks; "Water: A Comprehensive Treatise"; vol 1,Plenum Press, London p278-281 (1972)

IT!'; "Reference Data for Radio Engineers"; 4th ed, p70-71

G. Roussy et al; "Temperature Runaway of MicrowaveHeated Materials: Study and Control"; J Microwave Power,vol 20, no 1, p47-51 (1985)

H.D. Kimrey et al; "Initial Results of a High PowerMicrowave Sintering Experiment at ORNL"; J MicrowavePower, vol 21, no 2, p81-82 (1986)

Paper, Wood, Textiles, Coal:

P.L. Jones, J. Lawton, I.M. Parker; "High frequency paperdrying: paper drying in radio and microwave frequencyfields"; Trans Inst Chern Eng, 52, p121-131 (1974)

S.F. Galeano; "The application of electromagnetic radiationin the drying of paper"; J. Microwave Power, 6, p131-140(1971)

N.H. Williams; "Moisture levelling in paper, wood, textiles,and other mixed dielectric sheet"; J. Microwave Power, 1,p73-80 (1966)

H.F. Huang; "A microwave apparatus for rapid heating ofthreadlines"; J. Microwave Power, 4, p283-288 (1969)

M. Baginski et al; "Experimental and numericalcharacterization of radio-frequency drying of textilematerials"; J. Microwave Power & Electromagnetic Energy,24, p14-20 (1989)

J. Wilson; "Radio-frequency drying of wood veneer ­commercial use"; J. Microwave Power & ElectromagneticEnergy, 24, p67-73 (1989)

H. Resch; "Drying of incense cedar pencil slats bymicrowave power"; J. Microwave Power, 2, p45-50 (1967)

N. Standish; "Microwave drying of brown coalagglomerates"; J. Microwave Power & ElectromagneticEnergy, 23, pl71.175 (1988)

D.P. Lindroth; "Microwave drying of fine coal"; USBMReport of Investigations, 9005, Minneapolis (1985)

Food:

D. Bialod et al; "Microwave thawing of food products usingassociated surface cooling"; J. Microwave Power, 13, p269­274 (1978)

A. Priou et al; "Microwave thawing of large pieces of beef';Proc 8th European Microwave Conference, Paris, France,p589-593 (1978)

Anon; "Microwaves cook bacon"; Food Engineering (June1977)

•

•

www.HPARCHIVE.com

..Dielectric Materials Measurements

rl3 HEWLETTL~ PACKARD 15

•

Anon; EPRI EM-3645, Projects 1967-7, 2416-11, FinalReport, Section 2 (August 1984)

YH. Ma, P.R. Peltre; "Freeze dehydration by microwaveenergy"; AICHE Journal, 21, p335-350 (1975)

J.E. Sunderland; "An economic study of microwavefreeze-drying"; J Food Technology, 361, p50-55 (1982)

C. Avisse, P. Varoguaux; "Microwave blanching of peaches";J Microwave Power, 12, p73-77 (1977)

S.C. Chen et al; "Blanching of white potatoes by microwaveenergy followed by boiling water"; J Food Science, 36, p742­743 (1971)

Anon; "Microwaves dry pasta"; Food Engineering, 94, 96(1972)

F.J. Smith; "Microwave hot air drying of pasta, onions, andbacon"; Microwave Newsletter, vol XII, no 6 (1979)

Rubber, Plastics, Polymers:

Anon; "Microwaves in the rubber industry"; RubberJournal, p43-49 (1970)

H.F. Schwarz; "Microwave curing of synthetic rubber"; JMicrowave Power, 8, p303-322 (1973)

R.A. Shute; "Industrial microwave systems for the rubberindustry"; J Microwave Power, 6, p193-206 (1971)

E.O. Forster; "Microwave drying process for syntheticpolymers"; U.S. Patent 3,997,089 (August 1976)

P.J. Minett; "Microwave and RF heating in plastics andrubber processing"; Plastics and Rubber (GB), 1, p197-200(1976)

Ceramics:

R.F. Stengel; "Microwaves speed drying of ceramic blocks";Process Design Ideas (Sept 1974)

L.W. Tobin, Jr; "Cermaic material processing"; U.S. Patent4,292,262 (Sept 1981)

A.J. Berteaud, J.C. Badot; "High temperature heating inrefractory materials"; J Microwave Power, 11, p315-320(1976)

N.W. Schubring; "Microwave sintering of alina spark pl~g

insulators"; paper at American Ceramic Society, ElectrorucsDivision (Sept 1983)

Other:

M.A.V. Ward; "Microwave Stimulated Conbustion"; JMicrowave Power, 15, 2, p81-86 (1980)

K.R. Foster, J.L. Schepps; "Dielectric Properties of Tumorand Normal Tissues at Radio through MicrowaveFrequencies"; J Microwave Power, 16,2, pl07-120 (1981)

R.G. Olsen, W.C. Hammer; "Evidence for Microwave­induced Acoustical Resonances in Biological Material"; JMicrowave Power, 16, 3&4, p263-270 (1982)

M. Hamid; "Microwave Thawing of Frozen Soil"; JMicrowave Power, 17, 3, p167-174 (1982)

M.F. Iskander, J.B. DuBow; "Time- and Frequency-DeomainTechniques for Measuring the Dielectric Properties ofRocks"; J Microwave Power, 18, 1, p55-74 (1983)

R. Chahine et al; "Computer-based permittivitymeasurements and analysis of microwave power absoptionstabilities'; IJ Microwave Power, 19,2, p127.134 (1984)

V.N. 'fran et al; "Dielectric Properties of Selected Vegetablesand Fruits, 0.1 - 10.0 GHz"; J Microwave Power, 19, 4,p251-258 (1984)

R.A. Budd, P. Czerski; "Modulation of MammalianImmunity by Electromagnetic Radiation (EMR)"; JMicrowave Power, 20, 4, p217-232 (1985)

P.C. Vasavada; "Effect of Microwave Energy on Bacteria"; JMicrowave Power, 21, 3, p187-188 (1986)

C. Gibson et al; "Microwave Enhanced Diffusion inPolymeric Materials"; J Microwave Power, 28, 1, p17-28(1988)

Lemelson; "Methods of Forming Synthetic DiamondCoatings on Particles Using Microwaves"; U.S. Patent4,859,493 (August 1989)

Yamazaki; "Microwave-Enhanced CVD Method forDepositing Carbon"; U.S. Patent 4,869,923 (Sept 1989)

Chu; "Crystallization Method Employing MicrowaveRadiation"; U.S. Patent 4,778,666 (Oct 1988)

Wolf; "Use of Microwave Energy in Separating Emulsionsand Dispersions of Hydrocarbons in Water"; U.S. Patent4,582,629 (1987)

Henderson; "Microwave Heated Hair Curler"; U.S. Patent4,538,630 (1985)

Taylor; "Microwave Coagulating Scalpel"; U.S. Patent4,534,347 (1985)

www.HPARCHIVE.com

www.HPARCHIVE.com

~ HEWLETT~~ PACKARD

Copyright © 1990Hewlett-Packard CompanyPrinted in U.S.A. 4/905952-2385 (D)

•