Buschow K.H.J. (Ed.) Concise Encyclopedia of Magnetic and Superconducting Materials

Подождите немного. Документ загружается.

1. LIESST and Related Phenomena

The abbreviation LIESST refers to the phenomenon

of light-induced excited spin state trapping, which

was discovered in 1984 on [Fe(ptz)

](BF

)

(ptz ¼1-

propyl-tetrazole), an iron(II) coordination compound

known to exhibit thermal spin transition between

high-spin (HS,

) and low-spin (LS,

) states (see

Spin Transition Compounds (Decurtins et al. 1984).

The material, as a single-crystal, polycrystalline, or

embedded in a polymer foil, converts from the LS to

the HS state by irradiation with green light into the

1,2

absorption band, followed by two suc-

cessive intersystem crossing (ISC) processes

1,2

to populate the HS state

. Ra-

diative decay

is spin and parity forbidden,

and due to the considerably larger M–L bond dis-

tance in the HS state as compared to the LS state,

which builds up an energy barrier between the po-

tential wells of the two states (see Fig. 1), the lifetime

of the metastable LIESST state can be very long (e.g.,

weeks at 20 K in the case of [Fe(ptz)

](BF

)

) (Gu

tlich

et al. 1994).

Thermal relaxation of the metastable LIESST state

does not occur in a classical manner by passing over

the barrier, but by nonadiabatic tunneling through

the barrier, whereby the observed temperature de-

pendence of the relaxation rate is due to thermal

population of the vibronic levels of the

manifold

(Hauser 1995). Optical switching back from the

LIESST state to the

ground state is also pos-

sible by irradiation with red light for the initial

excitation

E followed by two ISC processes

E-

1,2

(Hauser 1991). It has been found

that LIESST and reverse LIESST may be observed

with all iron(II) compounds exhibiting thermal SC, a

major difference, however, being the lifetime of the

light-induced LIESST state at a given temperature,

which depends on the ligand ﬁeld strength, viz., the

weaker the ligand ﬁeld strength, the smaller the dif-

ference DE

between the lowest vibronic levels of

the HS and LS states, the longer the lifetime of the

LIESST state. LIESST has also been observed in

Langmuir–Blodgett thin ﬁlms (Letard et al. 1999).

More recently, photophysical phenomena in iro-

n(II) SC systems, which are related to LIESST have

been observed. If a sample is continuously irradiated

with green light while the temperature is raised and

then lowered again, a thermal hysteresis may develop;

this phenomenon has been termed light-induced ther-

mal hysteresis (Desaix et al. 1998). In another case

existing thermal hysteresis of an iron(II) SC com-

pound was observed to shift to lower temperatures

under continuous irradiation with green light, and to

higher temperatures with red light (light-perturbed

thermal hysteresis) (Renz et al. 2000). Finally there is

the SOXIESST (soft x-ray induced excited spin state

trapping) which uses a source of soft x rays to initiate

the

transition (Collison et al. 1997), but

the relaxation mechanism should be the same as for

LIESST.

2. NIESST Effect

The acronym NIESST stands for nuclear decay-in-

duced excited spin state trapping. This phenomenon

is closely related to LIESST, as it makes use of the

nuclear decay and its energy release as an intrinsic

molecular excitation source, whereas LIESST is the

result of irradiation with an external visible light

source. Thus the initial step of electronic excitation is

different, but the ﬁnal step of ligand ﬁeld state relax-

ations has been found to be the same for both phe-

nomena.

Fe Mo

ssbauer emission (ME) spectroscopy has

been a most elegant tool for the observation of me-

tastable ligand ﬁeld states at ca. 100 ns after electron

capture decay of

Co(EC)

Fe in transition metals

compounds (Sano et al. 1984). A

Co labeled com-

pound is used as the Mo

ssbauer source at variable

temperatures vs. K

[Fe(CN)

] as a single-line absorb-

er. If

Co is embedded in strong ﬁeld surroundings,

which causes LS behavior in the corresponding

iron(II) compound like the tris-phen or tris-bpy

complexes (phen ¼1,10-phenanthroline; bpy ¼2,2

bipyridine), one observes the nucleogenic

ions

in LS (

) state at room temperature, which is the

ground state in the corresponding synthesized iro-

n(II) compound. Below ca. 200 K, however, one ob-

serves typical

-HS resonances in the ME spectra

with increasing intensity towards lower temperatures

at the expense of the

-LS resonances. In iron(II)

SC systems with weaker ligand ﬁeld strength one ob-

serves no more

-LS resonances in the ME spec-

tra, even at very low temperatures where the

corresponding synthesized iron(II) complex has al-

ready turned to the LS (

) ground state. The

trapped HS state of the nucleogenic

ion was

found to have very similar lifetimes as the corre-

sponding LIESST state under comparable conditions

(Deisenroth et al. 1998). It was, therefore, concluded

that NIESST and LIESST are closely related phe-

nomena. Whereas the LIESST effect bears basically

the potential for practical applications, NIESST is

evidently of more academic interest and has contrib-

uted much to a deeper knowledge of mechanistic

aspects in hot atom chemistry in the solid state.

3. Stilbenoid Complexes

Another elegant way to change spin states of

some transition metal molecular compounds can be

achieved using the well-known cis–trans photo-

isomerization of stilbenoid units incorporated as

ligand molecules. This effect, called ligand driven

light-induced spin change (LD-LISC), was ﬁrst ob-

served on the iron(II) complex [Fe(stpy)

(NCBPh

)

]

1060

Photomagnetism of Molecular Systems

(stpy ¼4-styrylpyridine) (Boillot et al. 1996). This

compound undergoes sharp thermal spin transition

around ca. 190 K under its trans form whereas the cis

form remains permanently HS. This is the conse-

quence of the different ligand ﬁeld strength for these

two forms due to the different conformations of the

coordinated ligand. Irradiation of a thin ﬁlm of trans-

[Fe(stpy)

(NCBPh

)

] with UV light around 140 K,

that is in the LS area, resulted in a LS-HS conver-

sion which was followed by absorption spectroscopy.

This transformation was found to be incomplete,

probably due to the rigidity of the matrix at this

temperature. Thus light irradiation affords a cis–trans

isomerization of the ligand with subsequent spin

transition at the iron(II) center as a consequence of a

change in ligand ﬁeld strength.

It has recently become possible to observe the LD-

LISC effect at room temperature for [Fe(4-methyl-4

trans-styryl-bpy)

(NCS)

] (Boillot et al. 1999) and also

for some iron(III) complexes but only in solution. The

limiting point of this phenomenon is that it has for now

only been observed in thin ﬁlms or in solution, the

reason being the enormous space required for the

photoisomerization to proceed. This weak point

regarding practical applications could be overcome

by implementation of these photosensitive materials in

organized media on the mesoscopic scale, such as

Langmuir–Blodgett ﬁlms, which is actually in progress.

4. Nitrosyl Complexes

The discovery of light-induced long lived metastable

states (41 10

s) in Na

[Fe(CN)

(NO)].2H

O (Ha-

user et al. 1977) represents an important step in view

of possible applications for optical information stor-

age and data recording. Irradiating a single crystal of

sodium nitroprusside (SNP) below typically 100 K

with green light results in the subsequent population

of two metastable states (MS

and MS

) accompa-

nied by a color change. It was also proved that it is

possible to go from MS

to MS

with red light and

more generally, that all possible photoswitching path-

ways between the ground state, MS

and MS

, are

accessible using an appropriate wavelength. Upon

heating, radiationless relaxation is observed around

ca. 200 K (MS

) and 150 K (MS

). The photoswitch-

ing mechanism involves a two-step process: a charge

transfer transition from the metal d orbital to an an-

tibonding p* NO ligand orbital followed by a relax-

ation into the metastable potential minima. It may

also involve a reorientation of the nitrosyl group at-

oms upon photoexcitation. This possibility as well as

the origin of the longevity of the metastable states is

still debated. It should be mentioned that the photo

switching of SNP differs from the LIESST phenom-

enon as the ground state and the metastable states are

diamagnetic and as no change of the lattice constants

is observed during the photoexcitation.

The existence of photo-induced metastable states is

not restricted to SNP. It has also been found for in-

stance for related complexes such as XY[Fe(CN)

(NO)]

.10H

O(X¼Rb or Cs), K

[Ru(CN)

(NO)].

O and for K

[RuCl

(NO)] (Woike et al. 1990);

this latter compound points out the crucial role of the

nitrosyl group for the photoswitching process.

Thus, SNP and its derivatives represent an appeal-

ing class of photochromic compounds, which are

suitable for optical memory devices with a very high

storage capacity and for holographic information

storage (Haussu

hl et al. 1995). It is noteworthy that

the effect is not limited to solids but also occurs

in glassy matrices. Research efforts are currently

underway towards materials with a faster popula-

tion velocity and high decay temperatures. Chemical

modiﬁcation of the SNP system in a controlled

manner might be the appropriate way to reach this

exciting goal.

5. Prussian Blue Analogues

The design of molecular magnets with high Curie

temperature (T

) which can be switched by light rep-

resents a challenging topic in materials science. Par-

ticular attention has been devoted to the Prussian

blue family, which can show spontaneous magneti-

zation above room temperature (Ferlay et al. 1995).

Soon after, the ﬁrst photo-switchable, cobalt–iron

cyanide was reported (Sato et al. 1996). K

0.2

1.4

[Fe(CN)

].6.9H

O possesses an f.c.c. structure in

which cobalt(II) ions are bridged to iron(III) ions

via cyano groups with a few cobalt(III)–iron(II) di-

amagnetic pairs also being present. Irradiation of a

powdered sample with red light at 5 K results in an

increase in the magnetization together with a rise of

from 16 K to 19 K. This light-induced magnetiza-

tion remains trapped at 5 K for several days in the

dark. Switching back to the ground state was

achieved by raising the temperature to 150 K or by

irradiation with blue light.

Before the excitation, the cobalt(III) and iron(II)

ions are in their diamagnetic LS states (S ¼0 for

both). Excitation with red light induces an electron

transfer from the iron to the cobalt leading to co-

balt(II) ions in the HS state (S ¼3/2) and iron(III)

ions in the LS state (S ¼1/2), which couple anti-

ferromagnetically. More interesting is that the reverse

electron transfer is accessible with blue light. The lo-

cal electron transfer switches increases the number of

magnetic neighbors and as a result the ordering tem-

perature (Verdaguer 1996). The key role of this phe-

nomenon is not only the presence of Co

III

–Fe

diamagnetic pairs but also crystal defects which

render the network ﬂexible and allow the extension

of the light-induced electron transfer.

From an application viewpoint, the slow switch-

ing process (a few nanometers), as well as the low

1061

Photomagnetism of Molecular Systems

temperature range of the observation of the phenom-

enon, are the weak points of this system. Anyhow, it

represents the ﬁrst example of ﬁne-tuning of long-

range magnetic ordering by light. Photomagnetic

studies on related materials using different alkali cat-

ions are in progress to further explore this fascinating

phenomenon.

6. Concluding Remarks

The photo-control of the magnetic and optical prop-

erties remains a challenging topic in material science

in view of the possible implementation in optical

switching and memory devices. Various examples be-

longing to this appealing class of materials have been

covered in this article. There are no doubts that a

sizeable number of new compounds, as well as new

photophysical phenomena, may emerge in the near

future from this rapidly evolving ﬁeld. We can, for

instance, mention a novel composite material com-

prising Prussian blue intercalated into photorespon-

sive vesicles based on azobenzene moieties, whose

magnetic properties can be controlled by photoillu-

mination (Einaga et al. 1999) or a nitrosyl derivative,

trans-[Ru(en)

O)(NO)]Cl

(en ¼ethylenediamine)

which show a remarkably long lived metastable state

up to 267 K (Kawano et al. 2000).

Bibliography

Boillot M-L, Chantraine S, Zarembowitch J, Lallemand J-Y,

Prunet J 1999 First ligand-driven, light-induced spin change

at room temperature in a transition-metal molecular com-

pound. New J. Chem., 179–83

Boillot M-L, Roux C, Audie

re J-P, Dausse A, Zarembowitch J

1996 Ligand-driven light-induced spin change in transition-

metal complexes: selection of an appropriate system and ﬁrst

evidence of the effect, in Fe

(4-styrylpyridine)

(NCBPh

)

Inorg. Chem. 35, 3975–80

Collison D, Garner C D, McGrath C M, Mosselmans J F,

Roper M D, Seddon J M W, Sinn E, Young N A 1997 Soft

X-ray induced excited spin state trapping and soft X-ray

photochemistry at the iron L-2, L-3 edge in [Fe(phen)

(NCS)

] and [Fe(phen)

(NCSe)

] (phen ¼1,10-phenanthro-

line). J. Chem. Soc. Dalton Trans. 4371–6

Decurtins S, Gu

tlich P, Ko

hler C P, Spiering H, Hauser A 1984

Light-induced excited spin state trapping in a transition metal

complex. Chem. Phys. Lett. 105, 1–4

Deisenroth S, Spiering H, Nagy D L, Gu

tlich P 1998 Lamb–

ssbauer factor of electronically excited molecular states

measured by time-differential Mo

ssbauer emission spectro-

scopy. Hyperﬁne Interact. 113, 351–5

Desaix A, Roubeau O, Jeftic J, Haasnoot J G, Boukheddaden

K, Codjovi E, Linares M, Nogues M, Varret F 1998 Light-

induced bistability in spin transition solids leading to thermal

and optical hysteresis. Eur. Phys. B 6, 183–93

Einaga Y, Sato O, Iyoda T, Fujishima A, Hashimoto K 1999

Photofunctional vesicles containing Prussian blue and azo-

benzene. J. Am. Chem. Soc. 121, 3745–50

Ferlay S, Mallah T, Ouahes R, Veillet P, Verdaguer M 1995 A

room-temperature organometallic magnet based on Prussian

blue. Nature 378, 701–3

tlich P, Hauser A, Spiering H 1994 Thermal and optical

switching of iron(II) complexes. Angew. Chem. Int. Ed. Engl.

33, 2024–54

Hauser A 1995 Intersystem crossing in iron(II) coordination

compounds: a model process between classical and quantum

mechanical behaviour. Comments Inorg. Chem. 17, 17–40

Hauser U, Oestereich V, Rohrweck H D 1977 On optical

dispersion in transparent molecular systems. Z. Phys. A 280,

17–25

Haussu

hl S, Schetter G, Woike Th 1995 Nitroprussides, a new

group of materials for holographic information storage on

the basis of metastable electronic states. Opt. Comm. 114,

219–22

Kawano M, Ishikawa A, Morioka Y, Tomizawa H, Miki E -i,

Ohashi Y 2000 X-ray diffraction and spectroscopic studies

of the light-induced metastable state of a ethylenediamine

nitrosyl ruthenium complex. J. Chem. Soc. Dalton Trans.

2425–31

Letard J-F, Nguyen O, Soyer H, Mingotaud C, Delhae

sP,

Kahn O 1999 First evidence of the LIESST effect in a

Langmuir–Blodgett ﬁlm. Inorg. Chem. 38, 3020–1

Renz F, Spiering H, Goodwin H A, Gu

tlich P 2000 Light per-

turbed hysteresis in an iron(II) crossover compound observed

by the Mo

ssbauer effect. Hyperﬁne Interact. 126, 155–8

Sano H, Gu

tlich P 1984 Hot atom chemistry in relation to

ssbauer emission spectroscopy. In: Matsuura T (ed.) Hot

Atom Chemistry. Kodanski, Tokyo, pp. 265–302

Sato O, Iyoda T, Fujishima A, Hashimoto K 1996 Photoin-

duced magnetization of a cobalt–iron cyanide. Science 272,

704–5

Verdaguer M 1996 Molecular electronics emerge from molec-

ular magnetism. Science 272, 698–9

Woike Th, Zo

llner H, Krasser W 1990 Raman-spectroscopic

and differential scanning calorimetric studies of the light

induced metastable states in K

[RuCl

NO]. Solid State

Commun. 73, 149–52

P. Gu

tlich and Y. Garcia

Institut fu

r Anorganische Chemie und Analytische

Chemie, Johannes-Gutenberg Universita

t, Mainz

Germany

Pnictides and Chalcogenides: Transition

Metal Compounds

Research on the magnetic properties of 3d transition

metal pnictides and chalcogenides has advanced

enormously. Realistic energy band calculations as

well as experimental techniques such as photoemis-

sion spectroscopy are available and play an impor-

tant role in clarifying the variety of magnetic and

related electronic characteristics. Chalcogens (M)

(pnicogens (M

)), elements from the 6a (5a) group

of the periodic table, have a valence electron conﬁg-

uration ns

(ns

). The former group includes

oxygen, sulfur, selenium, tellurium, and polonium

and the latter nitrogen, phosphorus, arsenic, anti-

mony, and bismuth. Chalcogens and pnicogens form

many compounds with 3 d elements. The total number

1062

Pnictides and Chalcogenides: Transition Metal Compounds

of transition metal (T) pnictides T

and chalco-

genides T

exceeds 150. This article does not

include oxides (see Transition Metal Oxides: Mag-

netism), nitrides, or polonides.

The generally stronger electronegativity of chal-

cogens compared to pnicogens implies two principal

differences between chalcogenides and pnictides.

First, in chalcogenides like Fe

and Fe

order-

ing of vacancies in the crystals often occurs, whereas

in pnictides vacancies are generally absent. A second

important difference is observed in the magnetic and

electronic characteristics. For example, NiS is anti-

ferromagnetic with a nickel magnetic moment of

1.7m

and undergoes a nonmetal to metal transition

at T ¼263K, whereas NiAs is a weakly paramagnetic

metal. The most interesting chalcogenides and pnic-

tides are of the types TM, TM

,TM

, and T

. This

article briefly reviews the magnetic and electronic

properties mainly of the typical compounds men-

tioned above. Further details can be found in the

review by Beckman and Lundgren (1991).

1. Chalcogenides

1.1 Pyrite-type Compounds of TM

A pyrite-type structure is adopted by the TM

com-

pounds with manganese, iron, cobalt, nickel, and

copper. The nearest neighbor sulfur atoms to iron in

FeS

are shown in Fig. 1. The 3d orbitals in the oc-

tahedral crystal ﬁeld are split into the three de (or t

)

and two dg (or e

) orbitals. The distorted octahedron

and the inﬂuence from more distant neighbors cause

further splitting of the de and dg orbitals. The related

energy levels are broadened into energy bands owing

to the mixing with s and p orbitals of the neighboring

atoms. The MnM

compounds have lattice para-

meters several percent larger than the other TM

compounds. Owing to strong electron correlation,

they do not form energy bands and become antif-

erromagnetic with a local manganese magnetic mo-

ment of 5m

in the high-spin state of the de

and dg

conﬁguration.

The Ne

el temperature, T

, is 48K, 47K, and 87K

for MnS

, MnSe

, and MnTe

, respectively (Ishikawa

and Miura 1991). Energy band calculations per-

formed by Bullet (1982) for TS

with T ¼Fe, Co, Ni,

Cu, and Zn are successful in explaining the magnetic

and electric properties such as Pauli paramagnetic

and semiconductive FeS

and ferromagnetic and me-

tallic CoS

, etc. However, the band calculations can-

not explain semiconductive NiS

(antiferromagnetic

with a nickel moment of 1.17m

). Note that the band

description is not relevant for this strongly correlated

electron system.

The band calculations predict NiS

to be a Mott–

Hubbard insulator because of the existence of a

half-ﬁlled 3d band at the Fermi level. Also, NiS

regarded as an insulator of the charge-transfer type

(Fujimori et al. 1996, 1997). Since the nickel atom

appears in the octahedral surrounding as in Fig. 1,

p–d charge transfer is likely to occur. The photo-

emission spectrum of NiS

has been explained by a

cluster model with a charge transfer energy of 1.8eV,

a d–d electron correlation energy, U, of 3.3eV, and a

p–d transfer integral, pds,of1.5eV.

Both CoS

and CoSe

are metallic. The former is

ferromagnetic with T

¼125K and a cobalt magnetic

moment of 0.85m

, and the latter is apparently a

Curie–Weiss-type paramagnet. The solid solutions

Co(S

1x

)

in the composition region 0.12 pxp0.4

show a very interesting metamagnetic transition at

T ¼4.2K (Adachi et al. 1970, 1979). The metamag-

netism was successfully explained based on the energy

band structure (Takahashi and Tano 1982). The

magnetic characteristics of Co(S

1x

)

at ﬁnite

temperature are typical examples for which the self-

consistent renormalization (SCR) theory is applied

(Moriya 1979).

1.2 TM Compounds and Related Types

The TM compounds, which crystallize mainly as a

hexagonal NiAs-type structure (Fig. 2(a)), often con-

tain vacancies at the T site, and the crystal becomes

distorted owing to vacancy ordering. Therefore, the

magnetic and electrical properties of TM compounds

are generally complex. For example, within the Fe–S

system, FeS, Fe

1x

S with xo0.075, and Fe

com-

pounds have been found. The magnetic and electrical

properties of such materials depend on the type of

Figure 1

The iron atom in FeS

with the cubic pyrite-type crystal

structure (space group Pa3) is surrounded octahedrally

by six nearest neighbor sulfur atoms with the same Fe–S

distance of 2.29A

. Two sides of the rectangle are not

equal. If the position parameter, u, is just 0.5, the

octahedron becomes perfect with y ¼0.

1063

Pnictides and Chalcogenides: Transition Metal Compounds

vacancy order, which varies with temperature as well

as with the composition. Other known compounds

with vacancies are Fe

,Fe

,Cr

and Cr

. Generally, the TM systems with T ¼V,

Cr, and M ¼S, Se show crystallographic similarities

with the Cr–Te system.

The antiferromagnetic compound FeS with

¼598K undergoes an interesting transition at

¼420K, which is called the a transition. At T

the

lattice parameters, magnetic susceptibility, and elec-

trical resistivity measured along the c-axis (r

0.2Ocm for TpT

) change very sharply, while the

resistivity perpendicular to the c-axis (r

D10

3

Ocm)

does not change significantly, and r

for TXT

(Wijn 1990). The coefﬁcient of the electronic specific

heat, g, has been found to be zero within experimental

error (Kobayashi et al. 1999), which suggests that

FeS is an insulator at low temperatures. T

of the

1x

S compounds is strongly concentration depend-

ent. The mechanism responsible for the a transition is

still not clear.

NiS shows an insulator to metal transition at

¼263K with increasing temperature. The low-

temperature phase is antiferromagnetic with a nickel

magnetic moment of 1.7m

and the high-temperature

phase is paramagnetic. The insulator phase is known

to be of a charge transfer type and it is considered

to be at the metal–insulator boundary. An energy

gap of 0.1eV has been obtained for NiS (Fujimori

et al. 1996). The paramagnetic state of NiS is not

considered to be of a usual Pauli type (Takahashi and

Kanamori 1991).

2. Pnictides

All the 3d pnictides are metallic with resistivities of

the order of 10

3

–10

4

O cm at 300 K, and their mag-

netic properties are described by the itinerant electron

model. TM

and T

compounds are reviewed here.

The TM

compounds mainly crystallize either as

NiAs-or MnP-type structures (Fig. 2), while the T

compounds crystallize as Cu

Sb- or Fe

P-type struc-

tures (Fruchart et al. 1969). The most interesting are

the TM

compounds. The magnetic moment per

transition metal atom is displayed as a function of the

lattice parameter for T ¼Mn and Cr in Fig. 3. Man-

ganese pnictides are generally ferromagnetic, while

chromium ones are antiferromagnetic except for

Pauli paramagnetic CrP. The B-axis of CrP (the

B-axis corresponds to the a-axis of the hexagonal

NiAs-type structure; see also Fig. 2 and its caption.)

is about 25% smaller than the a-axis of CrSb, while

the c-axis (or the A-axis) of the chromium pnictides

shown in Fig. 3 remains within about 2% variation.

Therefore, m

seems to be a reasonable function of a

(or B).

Somewhat analogous features are observed for

manganese pnictides (Fig. 3). These ﬁndings imply

somewhat the existence of a critical separation where

the spontaneous atomic magnetic moment changes

drastically. The crossover is located around MnAs

Figure 2

Crystal structures of (a) the hexagonal NiAs-type

structure and (b) the orthorhombic MnP-type structure.

The MnP- becomes the NiAs-type structure if C ¼

ﬃﬃﬃ

B=2 and U ¼V ¼X

¼0. The parameter d, which

expresses a degree of the distortion from the hexagonal

to the orthorhombic structure, is deﬁned as

d ¼ðC 

ﬃﬃﬃ

BÞ=

ﬃﬃﬃ

B. The parameters U, V, and d are

generally less than 0.05, and X

is negligibly small in

general. The lattice parameters A, B, and C in (b)

correspond to c, a, and b ( ¼

ﬃﬃﬃ

a=2) in (a), respectively.

Figure 3

Relationship between the magnetic moment per T atom

and the lattice parameter a

, where a

means either the

a-axis in (a) or the B-axis in (b) of Fig. 2. The ﬁlled and

the open circles and triangles are the values for the

NiAs- and the MnP-type structures, respectively. The

data are taken from Wijn (1990) and Kallel et al . (1974).

1064

Pnictides and Chalcogenides: Transition Metal Compounds

and CrAs. This conclusion is corroborated by the fact

that both MnAs and CrAs undergo a ﬁrst-order

transition at T ¼318 K and 265 K, respectively, where

the magnetic order disappears, accompanied by an

abrupt shrinkage of the a-orB-axis. Besides, the

magnetic properties of MnAs and CrAs are also very

sensitive to substitutions of other elements. Typical

examples are summarized in Fig. 4. A substitution of

only 7.5% of phosphorus for arsenic in CrAs yields a

collapse of the double spiral magnetic order (Selte

et al. 1975). The B-axis of CrAs extrapolated to

T ¼0 K is about 6.1% (or 0.21A

) larger than that of

nonmagnetic CrAs

0.9

0.1

(Suzuki and Ido 1993).

The substitution of phosphorus for arsenic in

MnAs also has a strong effect on magnetic proper-

ties. Magnetic properties and the various phase tran-

sitions in these materials have been explained in the

framework of energy band calculations together with

SCR theory (Motizuki and Katoh 1984). A large ﬁeld

effect on the magnetic properties is shown in Fig. 5.

At temperatures in the paramagnetic region of

240 KoToT

, a metamagnetic (ﬁeld-induced) tran-

sition accompanied by a crystallographic transition

from the MnP- to the NiAs-type structure is ob-

served. The mechanism responsible for the various

phase transitions mentioned above has not been fully

clariﬁed.

Sb with the tetragonal Cu

Sb-type structure is

ferrimagnetic with T

¼550 K. This compound is in-

teresting because of the strong effect of substitutions

of other elements such as chromium for manganese.

For example, Mn

1.9

0.1

Sb undergoes a ﬁrst-order

transition from antiferromagnetic to ferrimagnetic

order at T ¼300 K with increasing temperature. The

magnetic structures in the Cu

Sb-type compounds

have been explained on the basis of energy band cal-

culations (Shirai and Motizuki 1992).

The ground state magnetic and electrical proper-

ties of 3d metal chalcogenides and pnictides can be

explained on the basis of the electronic structures.

However, numerous phase transitions induced at ﬁ-

nite temperatures by the external ﬁeld and/or by the

temperature change remain unexplained theoreti-

cally. Detailed band calculations should be extended

to compounds with complex magnetic structures such

as CrAs, etc. SCR theory has been exploited to ex-

plain the magnetic properties at ﬁnite temperatures.

Some magnetic and electrical properties often de-

pend strongly on the crystallographic states, especially

for the chalcogenides. In addition to the application

of new experimental techniques, preparation of single

crystals and polycrystals with well-controlled compo-

sition and the related better experimental data may

enhance progress in this ﬁeld. High-pressure studies in

the area up to about 10GPa may throw new light on

the physics of these materials.

See also: Magnetic Refrigeration at Room Tempera-

ture; Magnetism in Solids: General Introduction;

Transition Metal Oxides: Magnetism

Bibliography

Adachi K, Matuura M, Ohashi M, Kawai M 1979 Further

investigations of magnetic properties of Co(S

1x

)

(0pxp1). J. Phys. Soc. Jpn. 46, 1474–82

Adachi K, Sato K, Matsui M 1970 New type metamagnetism

in Co(S

1x

)

: exchange-compensated paramagnetism.

J. Phys. Soc. Jpn. 29, 323–32

Beckman O, Lundgren L 1991 Compounds of transition ele-

ments with nonmetals. In: Buschow K H J (ed.) Handbook of

Figure 4

Some examples of the temperature dependences of

magnetic and crystallographic properties of the pnictides

(Wijn 1990, Selte et al. 1975).

Figure 5

Temperature dependences of magnetizations at constant

ﬁelds for MnAs and MnAs

0.9

0.1

. See also main text.

1065

Pnictides and Chalcogenides: Transition Metal Compounds

Magnetic Materials. North-Holland, Amsterdam, Vol. 6,

Chap. 3

Bullett D W 1982 Electronic structure of 3d pyrite-and mar-

casite-type sulphides. J. Phys. C15, 6163–74

Fruchart R, Roger A, Senateur J P 1969 Crystallographic and

magnetic properties of solid solutions of the phosphides. J.

Appl. Phys. 40, 1250–7

Fujimori A, Mamiya K, Mizokawa T 1997 Photoemission

spectroscopy of correlated transition-metal compounds in the

charge transfer regime. Physica B237–238, 137–41

Fujimori A, Mamiya K, Mizokawa T, Miyadai T, Sekiguchi T,

Takahashi H, Mori N 1996 Resonant photoemission study of

pyrite-type NiS

, CoS

and FeS

. Phys. Rev. B 54, 16329–32

Ishikawa Y, Miura N 1991 Physics and Engineering Application

of Magnetism, Solid State Science 92. Springer, Berlin

Kallel A, Boller H, Bertaut E F 1974 Helimagnetism in MnP-

type compounds: MnP, FeP, and CrAs

1x

mixed crystals.

J. Phys. Chem. Solids 35, 1139–52

Katoh K, Motizuki K 1984 Role of spin ﬂuctuation in spon-

taneous magneto-volume effect of intermetallic compounds

MnAs and MnAs

1x

. J. Phys. Soc. Jpn. 53, 3166–71

Kobayashi H, Nozue T, Matsumura T, Suzuki T, Kamimura T

1999 The low temperature specific heat of FeS and M

0.875

(M ¼Fe, Co; X ¼S, Se) with a NiAs-like structure. J. Phys.

Condens. Matter 11, 8673–9

Moriya T 1979 Recent progress in the theory of spin ﬂuctua-

tions in itinerant electron magnets. J. Magn. Magn. Mater.

14, 1–46

Motizuki K, Katoh K 1984 Spin ﬂuctuation theory of

intermetallic compound MnAs. J. Phys. Soc. Jpn. 53,

735–46

Selte K, Hjersing H, Kjekshus A, Andresen A F, Fischer P 1975

Magnetic structures and properties of CrAs

1x

. Acta

Chem. Scand. A29, 695–8

Shirai M, Motizuki K 1992 In: Kotani A, Suzuki N (eds.) Re-

cent Advances in Magnetism of Transition Metal Compounds.

World Scientific, Singapore, pp. 67–77

Suzuki T, Ido H 1993 Magnetic–nonmagnetic transition in

CrAs and related compounds. J. Appl. Phys. 73, 5686–8

Takahashi M, Kanamori J 1991 Electron correlation in Ni

compounds. J. Phys. Soc. Jpn. 60, 3154–61

Takahashi Y, Tano M 1982 The metamagnetism and temper-

ature dependence of the magnetic susceptibility of

Co(S

1x

)

. J. Phys. Soc. Jpn. 51, 1792–8

Wijn H P J (ed.) 1990 Landolt-Bo

rnstein New Series. Springer,

Berlin, Group III, Vol. 27a

H. Ido

Tohoku Gakuin University, Tagajo City, Japan

1066

Pnictides and Chalcogenides: Transition Metal Compounds

Radiation and Particle Detectors

Superconducting radiation and particle detectors

measure the energy deposited by photons or parti-

cles that interact with the detector. Superconducting

detectors measure this energy with orders of magni-

tude more accuracy than semiconductor detectors.

Most superconducting photon and particle detectors

operate at temperatures below 1 K. At this low tem-

perature, the energy deposited by the photons and

particles of interest is not obscured by the thermal

energy of the detector. It is the low temperature of

operation that allows superconducting detectors to

measure the energy of each incident photon or par-

ticle with high precision.

Superconducting radiation and particle detectors

are energy dispersive, that is, they provide spectral

information by measuring the energy of particles or

photons absorbed in the active volume of the detec-

tor. Superconducting detectors can often be con-

structed so that almost every incident particle or

photon interacts with the detector. Thus, these de-

tectors can provide a combination of high resolution

and high quantum efﬁciency which is not available in

more conventional detectors.

The quest for new detectors with high resolution

and high efﬁciency has been strongly motivated by

x-ray astrophysics. However, superconducting detec-

tors are now being developed to study astronomical

objects at energies from the infrared to gamma-rays.

Superconducting x-ray detectors are also being de-

veloped for laboratory plasma experiments and the

measurement of nuclear materials. These detectors

are being used to improve x-ray ﬂuorescence analysis

of many types of materials ranging from integrated

circuits to proteins. Superconducting detectors are

also used to search for proposed weakly interacting

massive particles (WIMPs), which may compose the

‘‘dark matter’’ of the universe. They may also be

useful for studying ions and very large molecules.

Various detector designs have been developed

which take advantage of the unique thermal and elec-

trical properties of low-temperature superconductors.

1. Microcalorimeters

A microcalorimeter is a small detector used to meas-

ure the total energy of incident photons or particles.

As illustrated in Fig. 1, the device consists of an ab-

sorber, which is heated by individual photons or

particles, and a thermal sensor, which measures the

resulting change in temperature. The absorber and

thermal sensor are in good thermal contact, while the

assembly is isolated from the cold bath by a weak

thermal link. The temperature rise due to an incident

particle is DT ¼E/C where E is its energy and C is the

heat capacity of the device. A typical microcalori-

meter operates at 0.1 K where its heat capacity can be

so small that even optical photons will raise the tem-

perature by a significant amount.

In addition to having a low heat capacity absorber,

the microcalorimeter must include a sensitive thermal

sensor. Thermistors have provided the best results

to date. The sensitivity of the thermistor is typically

parameterized by a

a 

ð1Þ

If a low-noise ampliﬁer is used such that the dom-

inant noise contributions are from the thermal ﬂuc-

tuations in the detector and the Johnson noise in the

thermistor, then the limit to the accuracy one can

measure the energy is

FWHM

¼ x2:35

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

ð2Þ

where k

is the Boltzmann constant and x is a pre-

factor that depends on a and the power dissipated in

the detector by the bias circuit.

For semiconducting thermistors, aB5 and xB3.

Values of xo1, and therefore better energy resolu-

tion, can be achieved with a superconducting transi-

tion-edge sensor (TES). A TES is a superconducting

ﬁlm operated at the transition between the supercon-

ducting and normal state where the change in resist-

ance versus temperature is very steep and a4100 can

be obtained. The transition temperature of a TES can

be tuned to almost any value with the proper com-

bination of superconducting and normal metals.

When the photon or particle deposits energy in the

microcalorimeter, the temperature quickly increases,

and then returns to the operating temperature. The

Figure 1

Schematic diagram of a microcalorimeter photon or

particle detector.

1067

time constant for a device operating without bias

power is C/G where G is the thermal conductivity of

the weak link. When bias power is applied to the

thermistor, electrothermal feedback can significantly

lengthen or shorten the pulse.

For the TES, a constant voltage bias is usually

used. With a constant voltage bias, V, the detector is

heated above the base temperature by the bias power

P ¼V

/R. When the detector is heated by a photon or

particle, its temperature increases, thereby increasing

the resistance, R, of the TES. This increase in R

decreases the bias power providing negative electro-

thermal feedback. This feedback reduces the pulse

length. For a TES device with a large a, electro-

thermal feedback can shorten the pulse time by more

than an order of magnitude. In this case, the prefactor

in Eqn. (2) is given by xE2:5=

ﬃﬃﬃ

, which can be much

less than one. A high-bandwidth (41 MHz) super-

conducting quantum interference device (SQUID)

current ampliﬁer is typically used to operate the

low-resistance voltage-biased TES detectors.

While the resolution achieved with microcalorim-

eters has been remarkable, the results are not gener-

ally limited by Eqn. (2). For instance, Eqn. (2)

assumes that 100% of the photon or particle energy is

converted to heat in the absorber. In real devices,

however, some of the particle’s energy can either ex-

cite long-lived states in the absorber material, or leave

the detector before being measured by the thermal

sensor. In both cases, the energy resolution is de-

graded. For optimal results, one must therefore ﬁnd

absorber materials with good photon or particle ab-

sorption properties, low specific heat, and good ther-

malization properties. Insulators and semiconductors

can have very low specific heat, but statistical ﬂuc-

tuations in the generation of long-lived electronic

states typically degrades the energy resolution to a

much worse value than Eqn. (2). Good results, how-

ever, have been obtained with high-purity germanium

and silicon when all the trapping centers are ﬁlled by

infrared photon excitation prior to the measurement.

Normal metals have excellent thermalization, but

they also contain much more specific heat. Semimet-

als, bismuth in particular, provide an excellent com-

bination of good thermalization with low specific

heat. Superconductors cooled far below their transi-

tion temperature have low specific heat, and their

small energy gap can minimize statistical ﬂuctuations

from trapped states. Tin appears to work exception-

ally well.

X-ray microcalorimeters are typically o1mm

area, and several microns thick. When cooled to

TB0.1 K, resolutions of about 5 eV full width at half

maximum (FWHM) have been obtained at 6 keV,

and 2 eV at 1.4 keV. The response time of these de-

vices generally limits their count rate to a few 100

counts per second. These results have been achieved

with TES devices and with germanium neutron-

transmutation-doped thermistors attached to tin-foil

absorbers by epoxy. For comparison, a good energy-

dispersive semiconductor detector has a resolution 20

times worse, about 110 eV FWHM at 6 keV. Optical

TES devices have also been developed, where 400 mm

pixels operating at TB80 mK provide single photon

counting with a resolving power of about 15, and

count rates as high as 3.0 10

counts per second.

Another thermal sensor under development uses the

change in magnetization of a paramagnetic material.

A resolution of 13 eV FWHM at 6 keV has been

obtained using an erbium-doped gold sensor and a

0.01 mm

gold absorber.

Larger microcalorimeters are being developed to

improve our understanding of the nonbaryonic mat-

ter of the universe. When a proposed WIMP particle

scatters off a nucleus in a semiconductor crystal, the

energy transferred to the detector is divided between

electron–hole pairs and thermal energy. By measur-

ing both the electron–hole pairs, and the prompt

phonons, dark-matter detectors can differentiate nu-

clear-recoil events from electron-recoil events which

tend to produce more electron–holes for a given en-

ergy input. This allows electron-recoil events due to

background radiation to be excluded from the meas-

urement, increasing the sensitivity to WIMPs by two

orders of magnitude. A similar differentiation has

been achieved by comparing the scintillation light

and thermal energy collected when photons and par-

ticles scatter in calcium tungstate crystals. Despite

their modest size (o1 kg), superconducting WIMP

detectors have achieved more sensitivity than much

more massive detectors operating at higher temper-

atures where it is not possible to reject background

interactions to such a high degree.

2. Superconducting Tunnel Junctions

The superconducting tunnel junction (STJ) detector

typically consists of a superconductor-insulator-

superconductor (SIS) thin-ﬁlm structure where the

insulator is a very thin (B2 nm) native-oxide layer,

usually Al

. The STJ is cooled below 1/10 of the

superconducting transition temperature to condense

almost all its electrons into Cooper pairs. An incident

photon or particle breaks some of these Cooper pairs

and produces electronic excitations called quasipar-

ticles. When the Josephson current is suppressed by a

magnetic ﬁeld and the junction is biased at a nonzero

voltage, these excess quasiparticles tunnel through

the insulating barrier, producing a current pulse (see

Josephson Junctions: Low-T

). The number of quasi-

particles produced is N

¼E/e, where e is the average

energy required to produce a quasiparticle. For most

superconductors, eE1.7 D where D is the supercon-

ducting energy gap (see Electrodynamics of Super-

conductors: Flux Properties; Electrodynamics of

Superconductors: Weakly Coupled). Quasiparticle

can tunnel back and forth through the barrier of an

1068

Radiation and Particle Detectors

STJ detector and this multiple tunneling can increase

the measured signal.

If a low-noise ampliﬁer is used such that the total

noise in the measurement is dominated by the statis-

tical ﬂuctuations in the generation and tunneling of

quasiparticles, then the resolution of the STJ detector

would be limited by

FWHM

¼ 2:35

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

EeðF þ F

ð3Þ

Where F is the so-called Fano factor that parame-

terizes the statistical ﬂuctuations in the generation of

quasiparticles, and F

is a term which describes the

noise associated with the variations in how many

times each quasiparticle tunnels through the barrier.

Monte Carlo simulations ﬁnd that FE0.2 for most

superconductors. The total integrated current meas-

ured with the device is Q ¼nN

where n is the

average number of times a quasiparticle tunnels

through the barrier. Typically n ¼G

tun

life

where

tun

is the tunneling rate, and .t

life

is the quasiparticle

lifetime. For a symmetric STJ detector with multiple

tunneling, F

¼ 1 þ 1=n. Thus the multiple tunneling

of quasiparticles can produce a bigger signal, but

usually increases the statistical noise.

In general, a high tunnel rate, G

tun

, will produce

larger signals and better results. The tunnel rate is

inversely proportional to the barrier thickness and

the thickness of the superconductor. The barrier can

be only so thin before it becomes leaky, allowing re-

sistive current ﬂow and excess noise. A thin super-

conductor will have a high tunnel rate, but poor x-ray

absorption efﬁciency. It is possible, however, to use

two or more superconductors to concentrate the qua-

siparticles from a larger absorbing layer into a thin-

ner layer next to the tunnel barrier. If the

superconducting energy gap of the absorbing layer

(typically niobium or tantalum) is larger than the en-

ergy gap of the layer next to the tunnel barrier (typ-

ically aluminum), then quasiparticles created in the

higher-gap material can emit a phonon while in the

lower-gap material. As illustrated in Fig. 2, the qua-

siparticle is then trapped in the lower gap material

close to the barrier. Quasiparticle trapping can be

used to improve the performance of an STJ x-ray

detector.

In general, STJ photon detectors have not achieved

the resolution described in Eqn. (3), but the discrep-

ancy is mostly understood. STJ detectors are gener-

ally operated with room-temperature current or

charge ampliﬁers. The noise contribution from these

ampliﬁers depends on the resistance of the STJ. This

noise contribution is typically measured during in-

tervals when no photons strike the detector. When

this quiescent noise estimate is subtracted in quad-

rature from the measured resolution, the estimated

intrinsic resolution can approach the limit in Eqn. (3).

Better agreement is obtained with measurements of

the electronic noise contribution when a photon

strikes the detector. The quasiparticles created by the

photon causes a reduction in the STJ’s resistance and

an increase in electronic noise. This increase tends to

bring the measurements into closer agreement with

the limit described in Eqn. (3).

Five-layer STJ detectors of Nb/Al/Al

/Al/Nb

and Ta/Al/Al

/Al/Ta have successfully been imple-

mented as both x-ray and optical detectors. For x-rays

below 1 keV, these detectors have good absorption

efﬁciency and resolution between 2 eV FWHM and

10 eV FWHM, and can operate at temperatures as

warm as 0.5 K. At 6 keV, the resolution achieved is

about 25 eV FWHM, and the efﬁciency of these de-

tectors is not as good. A notable exception is a Al/

/Al STJ coupled to a lead ﬁlm through a thick

insulating barrier. In this case, the photon energy is

converted to phonons in the lead, which pass through

the insulator into the aluminum where the phonons

break Cooper pairs. A resolution of 12 eV FWHM at

6vkeV was obtained with this device, and the lead ﬁlm

provides very high absorption efﬁciency even at 6 keV.

Compared to the calorimeters developed so far, the

STJ detectors can operate at higher count rates. STJ

x-ray detectors have been operated at count rates up

to 2.0 10

counts per second with little degradation

in energy resolution. STJ detectors have also been

successfully implemented to measure individual opti-

cal photons with a resolving power of about 15.

Quasiparticle trapping is a widely used technique to

collect quasiparticles at various locations on a super-

conducting ﬁlm or crystal. When a photon or particle

interacts with the superconductor, the resulting qua-

siparticles will diffuse throughout the material, until

recombining or becoming trapped. If most of the

quasiparticles end up in the trapping centers adjacent

to STJs, then both the energy and the position of the

absorption event can be determined. Position-sensi-

tive detectors have been successfully fabricated in

both one and two-dimensional geometries. An energy

Figure 2

Energy levels in a symmetric superconducting tunnel

junction photon or particle detector.

1069

Radiation and Particle Detectors