[Springer Series in Materials Science] Magneto-Science Volume 89 || Novel Magnetic Field Effects

7 Novel Magnetic Field Effects

Many examples of various MFEs whose broad mechanisms are known are discussed in Chapters 2 through 6. However, there are several MFEs whose mechanisms are not well understood at present. For example, adsorption of gas to solid is significantly affected by an external magnetic field of less than 1 T. The effects cannot be explained by the mechanisms of MFEs discussed in Chapter 1. The novel effects of magnetic fields are important from both basic and applied points of view, since understanding of the phenomena will open a new field of magneto-science and the effects have the potential for further applications in industry.

In this chapter, MFEs on adsorption of gas to solids, optical properties of water, phase transition in diamagnetic materials, photo-induced ultrafine particle formation and related topics are described.

7,1 Magnetic Field Effects on Adsorption of Gas

Serious attention has focused on magnetic field effects on chemical reactions. These have been reviewed extensively by Steiner and Ulrich.'^ On the other hand, there are not many reports on external magnetic field effects on chemical processes in heterogeneous systems such as catalytic reactions on solids. ' ^ Due to cage effects, interfaces should provide more favorable conditions for magnetic field effects than three-dimensional bulk phases. Turro and coworkers^^ reported the cage effects and magnetic field effects on the photodecomposition of several dibenzyl derivertives adsorbed on porous glass, silica gel and reversed-phase silica gel.

Solid surfaces generally having geometrical or energetic heterogeneities which arise from defects, pores and variously indexed surfaces, exert adsorption potentials on the adsorptives. Thus, a phase adsorbed on such solid surfaces may be different from the bulk phase. An example is changes in the magnetic property of adsorptives in adsorbing. Paramagnetic nitrous oxide (NO) may form a diamagnetic condensed phase on solids^^ at a much higher temperature than the boiling and critical

296 7 Novel Magnetic Field Effects

points of its bulk phase. The diamagnetic property was ascribed to the NO dimer. ^ Paramagnetic O2 was also adsorbed as a diamagnetic O2 dimer in zeolite micropores at low temperatures.^^

A typical magnetic interaction in inhomogeneous systems is the magnetocatalytic para/ortho-hydrogen conversion on metals and metal oxides.'^ It is believed that the magnetocatalytic ability arises from paramagnetic interactions between adsorbates and solids, such as paramagnetic centers on surfaces of diamagnetic oxides and ferro-(ferri-)magnetic solids above the Curie temperature. The antiferromagnetic oxides a-Cr203, CoO and MnO have been investigated with respect to their magnetocatalytic properties around their Neel temperature (TN). The chemisorption rate of NO over 5-FeOOH showed an anomaly at their magnetic transition.^^ Selwood showed that during adsorption of molecules onto metals magnetic susceptibility of the metals decreased.' ^

Under external magnetic fields, the catalytic activity of solid surfaces for the para/ortho-hydrogen conversion showed a maximum at the TNI depression in the paramagnetic region {T > T^) and promotion in the antiferromagnetic region (7 < T^)?^ Ippommatsu and coworkers''^ found that an increase in the conductivity of Sn02 thin film by applying a strong magnetic field was observed when a H2-O2 reaction was proceeding on the Sn02 surface in an oxygen atmosphere at 773 K. The rate of increase in the conductivity (ca. 2.3% at 5 T) was proportional to the square of the magnetic field intensity. They explained the magnetocatalytic H2-O2 reaction by changes in the frequency factor for Sn02 surface reactions of H2 and adsorbed O2' by magnetic fields.

From these aspects, we may expect that physisorption as well as chemisorption should be influenced by an external magnetic field as well as the intrinsic magnetic field of solids (magnetisms of a bulk phase and surface). NO has been used extensively as a probe molecule for the survey of active sites on solid surfaces. Therefore, we first used NO for investigation of external magnetic field effects on gas adsorption. The adsorption of paramagnetic NO was affected by steady magnetic fields even at room temperature.^ ' ^ NO adsorption was enhanced on metal oxides by magnetic fields besides FeOOH polymorphs from which NO was desorbed. ^ On microporous materials, such as zeolites and activated carbons, magnetic fields promoted NO micropore filling.'^^ Some causes for magnetic field-induced adsorption and desorption (MAD), such as magnetism, adsorption sites and porosity of solids, have been examined, and some correlation was recognized between them, although the mechanism of MAD is not understood yet.

NO was chemically as well as physically adsorbed at room temperature. Therefore, the magnetoresponses of NO adsorption are very complex because there are many possible NO speices on surfaces and in pores which respond differently to magnetic fields. Thus, O2, which also is

7.1 Magnetic Field Effects on Adsorption of Gas 297

paramagnetic, was used for the approach to the mechanisms of magnetic field effects on adsorption''^ because O2 is physically adsorbed at 77 K and its adsorption state is relatively simple. Moreover, the magnetic properties of adsorbed O2 have been well investigated. O2 is very interesting from many points of view, one of which is the O2 separation from air. We will show the possibility of air separation due to gradient'"' as well as steady magnetic fields,"^ which can lead to the magnetic separation of other gas mixtures.

Generally speaking, magnetic energies are trivial even compared with thermal energy. Therefore, it is difficult to expect magnetoadsorption in paramagnetic systems, much less in any diamagnetic system such as water, besides ferromagnetic systems such as hydrogen absorption alloys, e.g., LaCo5H;c.' ^ There are reports concerning the properties of water on solid surfaces and around solutes in aqueous solutions because water bound to molecules and solids plays very important roles in the activities of solids and the conformation of molecules, and also its own activity. It is well known that hydrated or adsorbed water is different from bulk water in structure and static and dynamic properties. If the magnetic properties of water around substances change via interactions with the surfaces, some phenomena via the changes may be expected. Magnetic effects on adsorption seem to be one of the phenomena based on such magnetic behavior. It was confirmed that diamagnetic water'^' and organics (benzene and alcohols) were affected by external, steady magnetic fields of even less than 1 T.

It is possible to affect gas adsorption if the magnetization of an adsorptive and/or adsorbent changes during adsorption. When the equilibrium pressure (p) of a water/solid adsorption system changes to p + A/7 due to the application of a magnetic field (//) at T K, the change in the free energy of the gas phase (per mole) is dGg = RT{\n(p + A/7) -\n p). On the other hand, the magnetization change (AM) of the system due to the adsorption of 1 mole of molecules (adsorptives) under H causes a change in the magnetic free energy of the system, dGm = AM//. In an adsorption equilibrium under a steady magnetic field, dGg + dGm = 0, and, thus, the relation between A/7 and H may be given by'^'

RT{ ln(p + A/7) - In /7} = -AM// (1)

Under the experimental condition p > A/7, Eq. (2) becomes

Ap/ p = -AMH/ RT (2)

Here R is the gas constant. Eq. (2) demonstrates that whenever the magnetization of the adsorption system changes in adsorbing, the equilibrium pressure in the system should change, e.g., a positive AM can cause decrease in pressure, i.e., magnetoadsorption.


7.1.1 Magnetic Field Effects on Adsorption

Magnetic effects on adsorption were investigated by pressure changes in adsorption systems. "'' '" ^^ The adsorption cell was an ESR tube and a quartz cell with a stop-cock; this was connected to a vacuum line as the adsorption system having a total volume of 50 ml or less.^ ' ^ The adsorption cell was kept at 303 K by circulating temperature-controlled water or at 77 K by dipping in a liquid nitrogen whose level was kept constant during O2 and N2 adsorption using an automatic supplier of liquid nitrogen. H2 adsorption was measured with a cryostat in the range 20 to 303 K. The critical point of H2 is 33.0 K and 1.29 Pa. Pressure of the adsorption system was measured by a Baratron sensor (sensitivity 0.001 Torr) and recorded by a personal computer at intervals of 2 s. Magnetic fields were applied to the cell with an electric magnet (< 1 T) and a liquid helium-free superconducting magnet (< 10 T).

Solid samples (< 200 mg) were pretreated at a proper temperature and 1 mPa for 2 h. A gas or vapor was introduced over adsorbents in the adsorption cell at the temperature. After 30 min, a steady magnetic field was applied to the sample; the pressure change was monitored with time.

Figure 7-1-1 summarizes the schematic patterns of the magnetoresponses of pressure (p) as a function of time t. The equilibrium amount of MAD is denoted by Av, which was calculated from the pressure change (Ap).

Time

Fig. 7-1-1 Schematic patterns of the magnetoresponses of pressure (p) of adsorption systems as a function of time t.


A. Steady Magnetic Field Effects on Paramagnetic Gases a. Nitrous oxide Magnetic field effects on NO on iron oxides were observed at 303 K. ^ Ferrimagnetic magnetite and maghemite showed magnetoadsorption in the external magnetic fields of 0.1-0.8 T. Hematite (parastic ferromagnetism) also showed magnetoadsorptivity. Antiferromagnetic iron oxyhydroxides (a-, p-, and y-FeOOH) showed magnetodesorption. On the other hand, an antiferromagnetic NiO and a ferrimagnetic 5-FeOOH exhibited magnetoad-and magnetodesorptivity, respectively. The magnetoadsorptivity of magnetites depended on the pretreating conditions and that of 5-FeOOH was independent of its magnetic susceptibility. The results suggest that surface sites and porosity, i.e., adsorption state of NO, rather than solid magnetism play an important role in the magnetic interactions between solid surfaces and NO molecules under a magnetic field.

It is known that (N0)2 is formed at low temperatures or in the condensed phase.^^ Usually, the adsorbed phase in micropores and adsorbed multilayers are liquidlike, thus one may expect that NO on solids exist as an (N0)2. In fact, Enault and Larher showed that NO is converted to (NO)2 in the adsorbed layers on flat surface below 90 K.' "" In addition, the magnetic susceptibility measurements showed that most NO molecules in the micropores are dimerized even above room temperature by the stabiUzation (ca. 10 kJ mol"') of (N0)2.^' The dimers seem to be stabilized in 0.5- and 1.0-nm micropores which fit the size of the dimer (0.3 x 0.41 X 0.30 nm^). Since (N0)2 has a boiling point or behaves as a vapor, (NO)2 formation in the micropores can lead to anomalous NO physisorption even at room temperature. The NO dimer has a weak chemical bond which arises from electron pairing between two NO ( 11) molecules, but the coupling between two interacting NO molecules will still be weak and the unpaired electrons will be localized mainly on each NO molecule. ^^^ The ground state of (NO)2 as a two-spin system is singlet

The micropore filling of a supercritical NO onto activated carbons was enhanced by a steady magnetic field (< 1 T), e.g., 130 jUg g ' at 1 T on polyacrylonitril-based activated carbon fibers. This is referred to as magnetomicropore filling.'^^ The magnetomicropore filling of pitch-based activated carbon fiber occurred just after application of a magnetic field. Cellulose- and polyacrylonitril-based activated carbon fibers, a coconut shell-based activated carbon, and a molecular sieving carbon showed markedly a transient magnetomicropore filling: rapid NO adsorption due to a magnetic field and a subsequent, exponential decrease under the magnetic field. The magnetomicropore filling relates closely to slitlike micropores and acid sites. Micropores, especially, seem to be useful for magnetically induced formation of an NO dimer, making micropore filling of a supercritical NO easy. Fig. 7-1-2 demonstrates clearly that 0.5- and


1.0

Pore size, cl I nm

Fig. 7-1-2 Magnetoadsorption (Av) of NO at 303 K and 0.8 T as a function of pore size (diameter for cylinder and width for lamellae) of solids. A\' for the carbons is reduced by multiplying by a factor of 0.27. Adsorbents: solid, zeolites (Z3A, Z4A, Z5A, mordenite, TSZ-500, MS13X); open, activated carbon fibers (AlO, A15, A25, AlO-1173); half-filled, carbon blacks (PC, NPC). [Reproduced from S. Ozeki, J.Colloid Interface Sci., 154( 1), 304, Elsevier (1992)1

1.0-nm micropores can stabilize adsorbed NO under magnetic fields, b. Oxygen O2 pressure in the Z5A and AlO systems decreased reversely by the application of low magnetic fields.'*' This pressure decrease should be the magnetoadsorption, since pressure changes in the adsorbent-free (only O2) systems were at most a few percent of those in the adsorption systems.

Amounts of O2 adsorbed on AlO increased with increasing magnetic field intensity below around 8 T, whereas it decreased over 8 T. The magnetic field Hz (= 8 T) at which no magnetic effect was observed shifted to a lower field with increasing relative pressure (Fig. 7-1-3(A)). In the Z5A system, similar behavior was observed, although the //z value was slightly smaller. The magnetoadsorption at around 3 T for A20 was more significant than that for AlO. In the case of Y and NPC no magnetoadsorption was observed below 10 T, i.e., only magnetodesorption was observed. Since the pore width became larger in the order AlO, A20, Y and NPC (no pores), the magnetoadsorption at least should arise from oxygen in micropores (< 1 nm), which may form a certain domain. The most suitable state and/or domain size of oxygen for the magnetoadsorption are also inferred from the results that the magnetoadsorption of O2 on ACFs was greatest at the relative pressure of p/po = 0.1.

O2 adsorbed on ACFs exist in different states, depending on O2 pressure and pore size. O2 adsorbed on AlO is a mixture of a two-dimensional (paramagnetic) gas and nonmagnetic clusters (random magnetism) in our experimental conditions (77 K and 250 Torr).'^' In the


2 4 6 1 Magnetic field, HII

10 2 4 6 Magnetic field, H11

Fig. 7-1-3 Pressure change due to magnetic field zV? of O: (A) and N: (B) on A10 at 77 K as a function of magnetic field. Relative pressure /?//7o: circle, 0.01: square, 0.1; triangle, 0.4.

Z5A system, a few percent of O2 adsorbed at 78 K exist as diamagnetic linear 04/^ Accordingly, from the viewpoint of the magnetization change during adsorption, the magnetoadsorption of O2 in both cases would presume a certain, more paramagnetic oxygen species (states) than gaseous O2 in the adsorbed phases.

B. Steady Magnetic Field Effects on Diamagnetic Gases a. Nitrogen A diamagnetic N2 was magnetically desorbed under all examined H regions and adsorbents (NPC, A10 and Y) at 77 K, as shown in Fig. 7-1-3(B). " The magnetodesorption was marked in smaller micropore systems. b. Hydrogen H2 changes from a vapor to a supercritical gas at 33.0 K, the critical temperature (Ic) of H2. Generally speaking, it is difficult for supercritical gases to be physically adsorbed on solid surfaces and pores. Thus the adsorptivity of H2 must change near the critical point, reflecting solid-H2 interactions such as magnetization modification. The pressure of the AlO-H2 adsorption equilibrium systems decreased with application of a 10-T magnetic field. The magnetoadsorptivity of H2 onto various carbons including a single wall carbon nanotube depends markedly on temperature, as shown in Fig. 7-1-4(A). H2 vapor was adsorbed much more than the supercritical gas by a magnetic field. In the vapor region, the pressure change A/? depends on r~ . On the other hand, in the supercritical region, ^p depends on (r - Tc)"^^ below 40 K and {T - Tc)~^-^^ in the range 40 -303 K, suggesting two magnetic states of supercritical states in the micropores.

Figure 7-1-4(B) shows that Ap increased with increase in magnetic field. The H dependence of Ap is //^^' for the vapor (below Tc) and for the supercritical state //^^^ below 40 K and //^ ' in the range 40 - 303 K.


30 40 50 60 70

Temperature, 7 / K

2 4 6 Magnetic field intensity, HIT

Fig. 7-1-4 (A) Pressure change due to magnetic field Ap of H: on A10 at 10 T and 11.6 Torr as a function of temperature. (B) Pressure change due to magnetic field A/? of H: on A10 at various temperatures and 11.6 Torr as a function of magnetic field.

10 15 20 Po / Torr 50 100 150 200

v/mgg-'

Fig. 7-1-5 (A) Pressure change (Ap) of chrysotile asbestos/water adsorption systems due to steady magnetic fields (1 and 10 T) as a function of the equilibrium pressure (po) of water at 303 K. (B) 100 Av/v for water on activated carbon fiber (A 10) as a function of v at 303 K. Magnetic field intensity ///T: open, 0.1; half-filled, 0.4; solid, 1.0.

c. Water Water in the first layer of hydrophilic oxide surfaces, such as y-FeOOH, silica and chrysotile asbestos (Fig. 7-1-5(A)), and in ultramicropores of zeolite, did not respond to a magnetic field.' ^ In the case of silica, water in the first layer also responded slightly because the surfaces are partially hydrophobic. 100 Av/v for NPC changed with v through a maximum or minimum, depending on //, suggesting that water in a multilayer on nonporous NPC seems to be less sensitive to magnetic fields. 100 Av/v for A10 decreased stepwise along with increase in v, as shown in Fig. 7-1-


5(B). AlO has discrete, slitlike micropores of about 0.7 and 1.0 nm in width, whose micropore capacity corresponds to 80 and 190 mg H2O g~\ respectively. Thus, the steps in the 100 Av/v - v plot for AlO seem to reflect the difference in the response of micropore filling water in micropores having a discrete size. Also, the steep drop in the Ap - p plot in Fig. 7-1-5(A) is considered to arise from capillary condensed water in the cylindrical mesopores of 7.0 nm in diameter of the chrysotile asbestos, which should occur at around the p of 22 Torr. The lower pressure shift under 10 T suggests that such high magnetic fields should induce certain modifications of condensed water in the mesopores such as surface tension and contact angle.

The magnetic energy of water seems to be too small, less than 1 cal/mol even under 1.0 T, to bring about such a large MAD, considering that the energy for the physical adsorption of water is around 10 kcal m o l ' . In addition, magnetoadsorption is not expected, because diamagnetic water tends to be repelled from a magnetic field. From these viewpoints, the magnetic properties of water (phase) could change through an interaction with surfaces, as suggested by the trends in the ;fapp-// relations of water adsorbed on SiO: and TiO: similar to their MAD profiles.'"^^

The positive ;fapp of water adsorbed on SiO: beyond 0.45 T and the lower diamagnetic water phase on TiO: than bulk water demonstrate that Ap < 0 may be possible because of A/ > 0. On the other hand, the magnetodesorption (Ap > 0) should be ascribed to more diamagnetic water than bulk water from the viewpoint of Eq. (2). If Aj ~ ;t'app - ±10"' cm^ g~', Eq. (2) gives Ap/p = ±(10-' ~ 10' ) in the range 0.1 - LOT, which is comparable with the observed \Ap/p\, (0.3-5) x 10'\ A portion of the adsorbed water phase should behave paramagnetically under static magnetic fields via an interaction with the paramagnetic centers on the surfaces, as indicated by the changes in the e.s.r. signals of the solids due to water adsorption, d. Organics Methanol was adsorbed on NPC at 303 K by steady magnetic fields of less than 0.7 T and desorbed over 0.8 T.' ^ The magnetic field dependence of the MAD of methanol was quite opposite that of water, reflecting structural difference in their adsorbed phase: a zigzag hydrogen-bond chain for methanol and hydrogen-bond network for water. r-Butanol was irreversibly or chemically adsorbed by magnetic fields.'^^ Benzene showed high magnetoadsorptivity for NPC under magnetic fields of up to 1.2 T,' ^ suggesting the magnetic orientation of benzene molecules whose C2 axis tends to be perpendicular to solid surfaces.

C. Thermodynamic Features in Magnetoadsorption The thermodynamic consideration (Eq. (2)) presumes that the


magnetoresponses in pressure A/? should be proportional to the initial pressure po, the inverse of temperaturer ' (or r ^ depending on the temperature dependence of Aj), and the square of magnetic field H^; or to AM (or Ax) if the magnetization or magnetic susceptibility changes with a magnetic field. a. po Dependence of MAD In general, the experimental A/? is denoted by

Ap = ki(p-pi) for/7, </?</7,>i(/ = l,2,3...) (3)

where kt is the magnetoresponse coefficient, which represents the sensitivity for the magnetic response of pressure depending on //, and /?, and Pi + 1 are the lower and upper threshold pressures, respectively, which give the border of the pressure region (or the adsorption phase) having a different magnetic response.

In adsorption systems where the interactions between a molecule and solid surfaces and among molecules in the adsorbed phase are weak, Ap was proportional to p. In a nonporous carbon black system, the Ap - p relations for N2 and O2 at 77 K and H2O at 303 K were linear. In the micropore systems, the relations for N: were linear in A10 as well as Y, but for O2 was nonlinear even in Y. The Ap-p plot for H2O on A10 seemed to comprise two or more linear regions. On the other hand, Ap in the systems of chrysotile asbestos, MS5A and probably Si02 changed linearly with p above a certain threshold pressure /?, of around 3 - 4 Torr. i) Temperature dependence of MAD The ;fapp of the water phase adsorbed on Si02 at 1.0 T above 225 K was proportional to T~\ In this case, the temperature rise should reduce MAD according to Ap DC T'^. In fact, the k2 value estimated gave a linear relation against T~^. ii) H dependence of MAD In the magnetodesorption of N2 and magnetoadsorption of O2 on NPC at 77 K, Ap changed linearly with H^ under constant p and T, whereas, since magnetic interaction becomes significant among molecules in the adsorbed phase and between molecules and solid surfaces, the H^-dependence of Ap disappeared. N2/AIO and O2/Y systems showed H^ dependence, but O2/AIO systems did not because the A10 system has smaller micropores than the Y system.

There seemed to be a variety of magnetically interacting systems in which the magnetization changed during adsorption under magnetic fields: Ax should be a function of H. In such adsorption systems, the H dependence of the MAD of vapor and gas on various solids was classified into four types, as summarized in Table 7-1-1: the magnetodesorption (I; MD) and magnetoadsorption (II; MA) at all magneUc field intensities employed; and the magnetodesorption-to-magnetoadsorption (III; MD-A) and the magnetoadsorption-to-magnetodesorption transition with the magnetic field (IV; MA-D).


Table 7-1-1 Classification of the H dependence of the MAD of vapor and gas on various solids

NO 0 . N. H. Type Pattern 303 K 77 K 77 k 20-303 K

< 1 T <30T <30T <10T

I MD a-FeOOH (a) NPC (d) NPC (d) p-FeOOH(a) Y (d) A 10(d) Y-FeOOH (a) Y (d) 6-FeOOH (f)

H.O MeOH 303 K 303 K <10T < 1 T

a-FeÔ. (pf) Y-Fe:0, (f) ZnO (d) DNA (d)

r-BuOH 303 K < I T

benzene 303 K < 1 T

MA NPC (d) AlO(d) AC(d) a-FezOi (pf) Y-Fe203 (f) Fe304 (f) NiO (a)

NPC(d) A 10(d) TiO^(p) SWCNT

NPC(d) NPC(d)

III MD/MA NPC (d) AlO(d) Z5A (d) SiO: (d) Crysotile asbestos(d) Montmorillonite (d) a-FeOOH (a) Y-FeOOH (a)

IV MD/MA AlO(d) Z5A (d)

Fe304 (f) NPC (d)

I : magnetodesorption (MD) II: magnetoadsorption (MA) III: magnetodesorption-magnetoadsorption transition (MD-A) IV : magnetoadsorption-magnetodesorption transition (MA-D) Magnetism: a, antiferro; f, ferro; d, dia; p, para; pf, parastic ferro

b. Quantum Mechanical Aspects of MAD The thermodynamic aspect seems to be useful for some qualitative understanding of MAD. Notwithstanding, it should be pointed out here that some characteristics of MAD seem to be similar to those of the magnetocatalytic effects, which were interpreted by quantum theory.

There seem to be some similarities to the magnetocatalytic photoreactions of dibenzoyl peroxide (the singlet/triplet {SIT) transition of a radical pair ^ ) and the magnetocatalytic ortho/para hydrogen conversion on paramagnetic as well as diamagnetic solids under magnetic fields. ^ The mechanism for both cases is quite analogous to each other, and an H dependence similar to the four types in Table 7-1-1 has been reported both experimentally and theoretically.

The olp conversion process of H2 can be explained by the inhomogeneous magnetic field of the paramagnetic center, which exerts different influences on the two spins of the H2 molecule during a collision, as shown by the Wigner theory.' ^ The magnetocatalytic activity of metal oxides may be classified into three types with respect to its magnetic field dependence, which seems to be quite similar to that of the four MAD types.~^ The absolute rate constants of the olp hydrogen conversion on paramagnetic oxides followed


a proportionality to the square of the effective magnetic moment, as predicted by the Wigner theory, whereas a consistent and quantitative theoretical explanation for the magnetocatalytic effect under magnetic fields is not so far available, although some other theories are partially successful. ^^

The NO dimer and O: dimer may be regarded as a radical pair system. Since the paramagnetic NO molecules adsorbed on NPC at room temperature did not respond to the magnetic field, magnetic adsorption seems to occur via the formation of a diamagnetic NO dimer having a radical pair in the micropores. Since the magnetic adsorption increased monotonically with increasing magnetic field intensity, the so-called Ag mechanism seems to be plausible: Different magnetic fields at the two-spin sites due to magnetically inhomogeneous surfaces and/or different g factors may lead to an S/TQ transition under applied magnetic field. As in radical pair reactions, the S/To transition due to magnetic fields may lead to NO adsorption. In the case of the Z5A/O2 system, Zeeman stabilization of singlet O4 may contribute to the magnetoadsorption, if a stabilizing energy due to the electric field of Ca~ for O4 adsorbed on Ca* ^ sites in micropores become comparable with the Zeeman energy.'^' Therefore, the MAD of O2 should arise from magnetic stability associated with magnetization changes due to adsorption. The rotational behavior of H2O seems to closely resemble that of H2, if a water molecule is bound with its oxygen to solid surfaces, in which each water molecule has in practice only one rotational degree of freedom left.

7.1.2 Local Magnetic Field Gradient Effects on O2 Adsorption

The magnetic effect on the adsorption equilibrium should be accompanied by changes in the magnetization of adsorptives and/or solids while adsorbing under a steady magnetic field, according to Eq. (2). On the other hand, even when the magnetization of an adsorption system remains unchanged, a magnetic effect on adsorption may kinetically occur if a local magnetic field is produced around and/or in solids by applying a steady magnetic field. When a molecule having the magnedc susceptibility X\^ which is at the distance / from the entrance of a cylindrical pore and on the center axis of the pore having radius /p, stands under a magnetic field gradient Hr(l), the force y; is exerted on a molecule interacting with solids in proportion to a gradient of the total potential energy which the molecule feels. As an example, consider //r(/) produced by the perfect diamagnetism of a spherical superconductor particle (radius: Rs) in contact with the entrance of the pore. In this case, when an external magnetic field Ho is applied to the mixture system comprising the superconductor and the adsorbent, the molecule feels the following magnetic field at the position /:

//.(/) = -//o{l + /?sV2(/ + /?s)-M (4)


After a simple calculation, we can obtain the force exerted on the molecule at the pore entrance (/ = 0)

/ ( 0 ) = -3;f,//o-/R + ;rp£(r*Vrp^-r*'V6rp'^) (5)

where e is the depth of the potential well at the potential minimum and r* is the distance between a molecule and an atom in the solid at the potential minimum. ^^ In Eq. (5), if the second term on the right side is negligible, a paramagnetic molecule {x\ > 0) ^t the pore entrance should be expelled by a magnetic force.

O2 pressure decrease (AÂ) in the Z5A and A10 systems was brought about by magnetic fields lower than 7 T, as described in section 7.1.1A2 above. Mixing YBa2Cu30v with both Z5A and activated carbon fibers led to an increase in oxygen pressure (A/7A/SC) by applying magnetic fields of less than 10 T (Fig. 7-l-6(A)).'"^ The oxide superconductor itself showed no detectable pressure change due to magnetic fields. Therefore, the oxide superconductor must indirectly promote the O2 desorption by AA/7A/SC (= A/7A/SC - A/7A) from Z5A and A10 by the application of magnetic fields of uptoca. 8T(Fig. 7-l-6(B)).

In the high Tc superconductors, the upper critical field Hd will be indefinite because of large superconductive fluctuation, and instead the vortex glass-to-liquid phase transition (//t ; //, < Hd) appears in the mixed state. From the magnetization measurements, //t of our YBa2Cu30v at 77 K seemed to exist at around 8 - 10 T, and x^cHd^ ^^^ maximum at around 1.5 T. Thus, it is inferred that the magnetodesorption of O2 from Z5A and A10 due to coexistence of YBa2Cu30v should be brought about by the perfect diamagnetism of the superconductor. According to Eq. (5), the local magnetic field gradient in micropores tends to transport a paramagnetic low-dimensional gas toward the exterior along micropores.

(B) 0.2

0.15

5000 10000 Time, 11 s

0.05 -

2 4 6 8 Magnetic field, HII

10

Fig. 7-1-6 (A) Example of the time course of O: pressure over a (1 + 1) mixture of A10 and YBa2Cu30, at 77 K with the application of steady magnetic fields up to 10 T. (B) Superconductor effect on Q. adsorption onto a (1 + 1) YBazCu.O^/AlO at 77 K. Relative pressure plpo. circle, 0.05; square, 0.3.


A rough estimation for a cylindrical micropore of a radius rp suggests that oxygen molecules at the entrance of a micropore having a diameter of around 0.52 nm should be free from the adsorption potential by a magnetic force under lOT.

7.1.3 Future Problems

A. Gas Separation Due to Static Magnetic Fields In microporous materials such as zeolites and activated carbons, MAD depended strongly on the kind of solids and gases and magnetic field intensity. Thus, a proper combination of them may lead to selective adsorption from a mixed gas and more critically to gas separation. For example, the separation of O2 and N2 due to high steady magnetic fields seems to be plausible, as seen in Fig. 7-1-3. For this purpose, it will be useful to apply a column system, adsorbent mixtures referring to Table 7-1-1, and a magnetic field swing system.

B. Magneto-photoadsorption The photodesorption of water on TiO: with UV irradiation appeared only under magnetic fields, which was enhanced with increasing a magnetic field of up to 1 T. ' On the contrary, the photoadsorption of water on ZnO, which occurred at zero magnetic field, was almost unchanged with a magnetic field. These examples demonstrate that magnetic fields may affect the photoexcitation of electronic states of these n-type semiconductors in a different manner. Thus, the photodecomposition of water on a Pt-deposited TiO: may also be magnetically controlled. Magnetic fields will affect such heterogeneous photocatalytic processes to lead to novel compounds.

C. Electron Spin Resonance Adsorption The pressure decrease of water vapor equilibrated with a zinc oxide and silica was observed only during the paramagnetic resonance of the solids adsorbing water. This may be referred to the electron spin resonance adsorption (ESR adsorption).''* The ESR adsorption of water was detectable with irradiation of a low power microwave. On the contrary, the desorption of water occurred at higher microwave power probably because of a thermal effect. The phenomenon is the first example of the magnetic control of surface/molecule interactions during electron spin excitation.

7.1.4 Conclusion

We have experimentally confirmed the magnetic field-induced adsorption and desorption (magnetoadsorption and magnetodesorption: MAD) of various gases and vapors on many kinds of solids. MAD may be thermodynamically interpreted, but it was only qualitatively and partially successful. A well-designed magnetic field gradient also was able to


dynamically control adsorption. Practically, this method will be more useful than the thermodynamical control of adsorption due to a homogeneous, steady magnetic field. The magnetic field control of adsorption processes may give rise to the possibility of a gas separation method, such as air separation and desulfuration. The magnetic treatment of spin excitation and photoelectrons in solids will lead to new heterogeneous reactions and novel compounds.

References

1. U. E. Steiner, T. Ulrich, Chem. Rev., 89, 51 (1989). 2. M. Misono, P. W. Selwood, J. Am. Chem. Soc, 90, 2977 (1968); P. W. Selwood, Adv.

CataL, 27, 23 {\97S). 3. H. Sasaki, H. Ohnishi, M. Ippommatsu, / Phys. Chem., 94, 4281 (1990); H. Ohnishi,

H. Sasaki, M. Ippommatsu, S. Marteau, J. Phys. Chem., 96, 372 (1992). 4. B. H. Barretz, N. J. Tuiro, J. Am. Chem. Soc, 105, 1309 (1983); N. J. Turro, C.-C.

Cheng, J. Am. Chem. Soc, 106, 5022 (1984); N. J. Turro, C.-C. Cheng, P. Wan, C.-J. Chung, W. Mahler, J. Phys. Chem., 89. 1567 (1985).

5. W. J. Dulmage, E. A. Meyers, W. N. Lipscomp, Acta Crystallogr., 6, 760 (1953); E. F. Wand, W. R. May, E. L. Lippert, Acta Crystallogr., 14, 1100 (1961); A. Enault, Y. Larher, Surf. Sci., 62, 233 (1977).

6. K. Kaneko, N. Fukuzaki, S. Ozeki, J. Chem. Phys., 87, 776 (1987); K. Kaneko, A. Kobayashi, A. Matsumoto, Y. Hotta, N. Fukuzaki, T. Suzuki, S. Ozeki, Chem. Phys. Lett., 163, 61 (1989); K. Kaneko, N. Fukuzaki, K. Kakei, T. Suzuki, S. Ozeki, Langmuir, S, 960 {\9%9)-

1. T. Takaishi, J. Chem. Soc, Faraday Trans., 93, 1257 (1997). 8. K. Kaneko, K. Inouye, Ad. Sci. Tech., 3, 11 (1986). 9. S. Ozeki, H. Uchiyama, J. Phys. Chem., 92, 6485 (1988); S. Ozeki, H. Uchiyama, K.

Kaneko, J. Phys. Chem., 95, 7805 (1991); S. Ozeki, H. Sato, Encyclopedia of Surface and Colloid Science (A. Habbard, ed.), p.3120, Marcel Dekker, New York (2002).

10. H. Uchiyama, S. Ozeki, K. Kaneko, Chem. Phys. Lett., 166, 531 (1990); H. Uchiyama, K. Kaneko, S. Ozeki, Langmuir, 8, 624 (1992); S. Ozeki, H. Uchiyama, K. Kaneko, J. Colloid Interface Sci., 154, 303 (1992).

11. H. Sato, Y. Matsubara, T. Tazaki, J. Miyamoto, S. Ozeki, in preparation. 12. S. Ozeki, T. Tazaki, Y. Matsubara, J. Miyamoto, H. Sato, Adsorption Science and

Technology (D. D. Do, ed.), p. 492, World Scientific, London (2000); H. Sato, Y. Matsubara, T. Tazaki, J. Miyamoto. S. Ozeki, in preparation.

13. M. Yamaguchi, H. Nomura, I. Yamamoto, T. Ohta, T. Goto, Phys. Lett. A, 126, 133 (1987); M. Yamaguchi, I. Yamamoto, T. Goto, S. Miura, Phys. Lett. A, 134, 504 (1987); I. Yamamoto, M. Yamaguchi, T. Goto, T. Sakakibara, Zeit. Phys. Chem. NF., 163, 671 (1989); M. Yamaguchi, I. Yamamoto, F. Ishikawa, T. Goto, S. Miura, / Alloys Comp., 253-254, 191 (1997).

14. S. Ozeki, C. Wakai, S. Ono, J. Phys. Chem., 95, 10557 (1991); S. Ozeki, J. Miyamoto, T. Watanabe, Langmuir, 12, 2115 (1996); S. Ozeki, J. Miyamoto, S. Ono, C. Wakai, T. Watanabe, / Phys. Chem., 100, 4205 (1996); S. Ozeki, J. Miyamoto, S. Ono, C. Wakai, T. Watanabe, Fundamentals of Adsorption (M. D. LeVan, ed.), p.717, Kluwer Academic Publishers, Boston (1996).

15. a. A. Enault, Y. Larher, Surf. Sci., 62, 233 (1977); b. C. Y. Ng., P. W. Tiedemann, B. H. Mahan, Y. T. Lee, 7. Chem. Phys., 66, 3985 (1977); c. Ph. Brechignac, De. Benedictis,


N. Halberstadt, B. J. Whitaker, S. Avrillier. / Chem. Phys., 83, 2064 (1985). 16. H. Kanoh, K. Kaneko, J. Phys. Chem., 100, 755 (1996). 17. J. Miyamoto, Y. Matsubara, H. Kurashima, T. Tazaki, S. Ozeki, Nippon Kinzoku

Gakkaisi,6h 1300(1997). 18. H. Hayashi, S. Nagakura, Bull. Chem. Soc. Jpn., 51, 2862 (1978); Y. Sakaguti, H.

Hayashi, S. Nagakura, Bull. Chem. Soc. Jpn., 53, 39 (1980). 19. E. Z. Wigner, Physik. Chem., B23, 28 (1933). 20. E. Ilisca, Phys. Rev. Lett., 24, 797 (1970); E. Ilisca, E. Gallais, Phys. Rev., B6, 2858

(1972); E. Ilisca, Phys. Rev. Lett., 40, 1535 (1978). 21. S. Ozeki et al., in preparation.

7.2 Magnetic Field Effect on Optical Properties of Water and Aqueous Electrolyte Solutions

Water, essential for life on Earth, has many peculiar properties, including large heat of vaporization, high boiling and melting temperatures, and high solubility for charged and polar molecules J The distinctive features of liquid water are mainly due to its three-dimensional hydrogen bonding network. Recent works on water have been extended to dynamical structure studied by ultra-fast laser techniques," ^^ theoretical studies," ' ^ and the relaxation dynamics of the interfacial water near the protein surface based on neutron scattering.^' Other recent topics are collected in the literature.^^

While a wealth of studies on water and solutions by light absorption and scattering experiments have been reported, we focus here on the refractive index of water. Recently, several convenient commercial sensing devices based on surface plasmon resonance (SPR) have been developed. ^^^ These devices can determine the refractive indices of liquid samples with very high sensitivity (An/n < 10" ). Recent development of the SPR sensor and its application, especially to the chemical and biological sensor, is reviewed by Homola et al.^' It is also promising and important for the applications to study the basic properties of water itself using this sensitive device. In this section, the effect of high magnetic fields on the refractive indices of water and several aqueous electrolyte solutions is studied in order to obtain insight into the static structure of water.

7.2.1 Sensitive Measurement of the Refractive Index under High Magnetic Fields

The refractive index (n) of water was measured by two methods, SPR and PSD, shown in Fig. 7-2-1 (not to scale), as a function of the magnetic field. All the measurements were carried out at ambient pressure and temperature of 25.0°C stabilized within ±VC. The refractive indices of aqueous electrolyte solutions were measured by the SPR method.

Figure 7-2-l(a) shows a setup using an SPR sensor (Spreeta™

7.2 Magnetic Field Effect on Optical Properties of Water and Aqueous Electrolyte Solutions 311

PTSPR1A170100, Texas Instruments Inc.) which is based on the resonance between the evanescent wave and the surface plasmon.'"^* ' The sensor consists of a light emitting diode (AlGaAs, 840 nm) a molded epoxy waveguide, a sensing area (50 nm gold film) and a photodiode array. The resolution of the refractive index is 5 x 10" . Due to the rapid damping of the evanescent wave, this device is sensitive to the thin layer of the analyte material at the vicinity of the interface. Typical thickness is 400 nm for the water-gold interface. "^ A recommended calibration procedure was carried out by which the refractive index of ultrapure water at 25.0°C was adjusted to 1.333000. The literature value of the refractive index of water at 840 nm and 25.0°C is 1.32796'" ^ so the refractive index measured by the SPR method in this study exhibits a slight constant difference from the literature. This difference, however, should not affect our considerations and conclusions. The sensor was placed at the magnetic center of a superconducting magnet (JMT-10T150, Japan Super Conductor Technology Inc.), which generates magnetic fields of up to 10 T. The sensor was positioned so the gold film was perpendicular to the magnetic field.

Another setup using a He-Ne laser (633 nm) and a position sensitive detector (PSD) obtained from Hamamatsu Photonics K.K. was used as is shown in Fig. 7-2-1(b) in order to measure the refractive index of bulk water and elucidate any surface effects for the SPR measurement. A quartz sample cell with a dimension of 12.5 x 12.5 x 45.0 (mm" ) including the quartz wall thickness of 1.25 mm was placed at the magnetic center with a tilting angle (0) between the cell and the laser beam. The laser beam

(a) (b) A

Sample

150mm

K

1

He-Ne laser

Fig. 7-2-1 Two experimental setups for measuring the refractive indices of water and aqueous solutions (not to scale). Both measurements were carried out under magnetic fields of up to 10 T using a superconducting magnet. (a) A schematic diagram using a commercial surface plasmon resonance (SPR) sensor. The sensor comprises a light emitting diode, a molded epoxy waveguide, a sensing area and a photodiode array. (b) A setup using a He Ne laser, a quartz cell and a position sensitive detector (PSD). [Reproduced from H. Hosoda et al., J. Phys. Chem. A. 108, 1461 (2004)1


passing through the cell is deflected, and the displacement of the optical path is detected by the PSD. The refractive index of water is obtained from the displacement, 0, and the refractive index of water at 633 nm; 1.33158.' ^ The small difference in the optical path due to the quartz cell itself is also taken into account. By varying the tilting angle 6 (70, 73, 75, and 77 degrees), the consistency of the refractive index obtained was checked and the experimental error {An/n) was estimated to be < 1.5 x 10 ^

Ultrapure water was prepared by a commercial water purification system (Direct-Q 5, Millipore Corp., 18.2 MQ cm). NaCl and NiCh (Wako Pure Chemical Industries, Ltd. GR grade) were used without further purification. The refractive index of «-hexane (Wako Pure Chemical Industries, Ltd. 99.5%) was also measured without further purification.

7.2.2 Slight Optical Responses of Water to High Magnetic Fields

The dependence of the refractive index of pure water on the magnetic flux density (B) is shown in Fig. 7-2-2. The refractive indices measured by the SPR setup (triangles, «SPR) and the PSD setup (circles, ^PSD) show increases by L8 X 10"' (0.14%) and 1.3 x 10 ' (0.09%) at 10 T, respectively, from those measured in the absence of the magnetic field. The increase in AZSPR is slightly larger than that in AZPSD. The origin of this discrepancy is unclear at this time, although this may be attributed to the magnetic effect on water at the vicinity of the interface or to the difference in the dielectric constant between the wavelengths of 840 nm (SPR) and 633 nm (PSD). The temperature derivative of the refractive index of water is ~1 x 10"^ deg"' at room temperature in the visible region around 600 nm.'" ' In the current study, the increase in MSPR and npso at 10 T exceeds the temperature fluctuation effects. Thus, the refractive indices of both the vicinity of the interface and the bulk are increased by the magnetic field effect.

A possible explanation for the increase in the refractive index of water is that the hydrogen bond is stabilized under a magnetic field. From a classical electromagnetic point of view, diamagnetism is explained by the anti-parallel magnetization of a molecule to the external magnetic field by electromagnetic induction. It is well known that Pauling explained the diamagnetism of aromatic hydrocarbons by a molecular-size ring current model. ^ Since the diamagnetism of a molecule depends on the extent of electron distribution, the electron delocalization of hydrogen-bonded molecules should increase its paramagnetism. Interactions between material and magnetic fields are described by magnetic susceptibility (jc)-Temperature dependence of j of water is relatively large compared to other materials and is related to the degree of hydrogen bond formation.'^ ' ^ Water's x is expressed by the sum of the diamagnetic and paramagnetic terms: ^ = Xd + XP- Here, Xd originates in the closed-shell electric structure of the molecule. On the other hand, XP is attributed to the deformed


1.3350

^ 1.3340

"- 1.3320

1.3330-

Fig. 7-2-2 Refractive index of pure water plotted as a function of the magnetic field obtained by the position sensitive detector (PSD, circles) setup and the surface plasmon sensor (SPR, triangles) setup. [Reproduced from H. Hosoda et al., J. Phys. Chem. A, 108, 1461 (2004)]

electron cloud of lone-pair electrons that form hydrogen bonds. The experimental temperature dependence of the ratio of the mass susceptibility to the one at 20°C is expressed by x^lx^^'c = 1 + 1.38810 x 10"^ {t - 20) + (higher order terms) for t in °C.' ^ This indicates that water becomes more paramagnetic at lower temperatures. Therefore, water molecules with hydrogen bonds should become more stable under a magnetic field. Iwasaka et al.'''^ found that the frequency of the higher harmonic vibrations of water shifts toward the longer wavelength under 14 T. In comparison to the spectra of water at higher pressures, they suggested the enhancement of the hydrogen bond under high magnetic fields. The enhancement of the hydrogen bond strength should lead to a change in the electronic absorption, which affects the refractive index in the near infrared region. According to the electronic spectra of ice in the vacuum ultraviolet region,'^^ it was observed that the absorption of the first electronic excited state of crystalline hexagonal ice is larger than that of amorphous ice. The increase in the absorption due to the formation of hydrogen bonds should cause the increase in the refractive index via the Kramers-Kronig relation. The present observation strongly indicates that the lifetime of hydrogen bonds is prolonged.

7.2.3 Comparison with a Nonhydrogen-bonded Liquid

In Fig. 7-2-3, the refractive index of A?-hexane ( hexane) is plotted against B (circles) measured by the SPR method. The result for pure water (nwater) by the SPR method is reproduced for clarity (triangles). While water gradually increases with B, nhexane does not change up to 10 T. This difference


1.3350 -z==rz:=Lr ; • Water

# Hexane

^ 1.3340^ 1.3570

1.3330 1.3560

Fig. 7-2-3 Refractive index of /2-hexane plotted against the magnetic field (circles) measured by the SPR method. The refractive index of water measured by the SPR method is reproduced for clarity (triangles). [Reproduced from H. Hosoda et al.. / Phys. Chem. A, 108, 1461 (2004)]

supports the hypothesis that hydrogen bonds are stabihzed under magnetic fields. Moreover, the refractive index of ethanol under a magnetic field (unpublished data) was measured and showed little dependence on B. This may mean that although ethanol forms hydrogen bonds, they are not stabilized significantly under magnetic fields since the number of hydrogen bonds per molecule for ethanol is smaller than that for water.

7.2.4 Magnetic Responses of Hydrated Water of Magnetic and Nonmagnetic Ions

Figure 7-2-4 shows the dependences of the refractive indices of aqueous electrolyte solutions on B as measured by the SPR method. Each mark represents the following electrolyte solutions: NaCl solutions at concentrations of 5.0 M (open squares) and 0.50 M (open circles), and NiCl2 solutions at concentrations of 2.5 M (closed squares) and 0.40 M (closed circles). The refractive index of pure water (Aiwaier) is again reproduced for clarity (triangles). The figure indicates that (1) the refractive indices of electrolyte solutions increase with increase in concentration in the absence of a magnetic field, and (2) the slope of the n-B curves is positive at lower concentrations and negative at higher ones.

Without the magnetic field, the variation in the refractive indices of electrolyte solutions compared with that of pure water increases in the sequence 0.50 M NaCl (0.009) < 0.40 M NiCb (0.011) < 5.0 M NaCl (0.034) < 2.5 M NiCh (0.035). It is intriguing that the increase in the refractive index is dominated by [CI] at higher concentrations. The color of the aqueous NiCh solution is green due to the formation of


1.3680 ^ 3:-^ • NiCl: 2.5M h

1.3670

1.3660

^ O - J X O-^ ^ DNaCTi.OM f

___n^ ._cu TJ~

1.3440:

X -

-a .£

> 1.3430^

\ • U o j i o ______

O

--a--

~o--• NiCl: 0.40M

O NaCl 0.50M

1.3420^

IT 1.3350-

1.3340-

Water [- T"

X 1 . 3 3 3 0 ^ - ^

~0 2 4 ^ 6 8 10

B/T

Fig. 7-2-4 Refractive indices of aqueous solutions of 2.5 M NiCb (solid squares), 5.0 M NaCl (open squares), 0.40 M NiCl: (solid circles), and 0.50 M NaCl (open circles) plotted against the magnetic field, measured by the SPR method. The refractive index of pure water measured by the SPR method is again reproduced for clarity (triangles). [Reproduced from H. Hosoda et al., / Phys. Chem. A, 108, 1461 (2004)]

[Ni(H20)6] ' ,^^^ and therefore its refractive index is expected to be dependent on [Ni ^] due to the Kramers-Kronig relation. The results, however, show that the refractive index depends more strongly on [CI] than on [Ni ] and [Na^] at higher concentrations. The solutions of 2.5 M NiCl2 and 5.0 M NaCl are near saturation; NaCl/aq; 5.6 M, NiClz/aq; 6.0 M. At this concentration, the solution is dominated by ionic atmospheres and the extent of ion-pair formation becomes large.~ ' The refractive index at 840 nm seems to correlate to Cr, possibly due to the change or the appearance of the electronic states of Cr perturbed by ion-pair formation or complexation with cations. It should be noted that in the case of a concentrated NiCl2 solution, the formation of [Ni(H20)5Cl]^ is reported. '

The changes in the refractive indices of the solutions at 10 T from


those without the magnetic field are -0.0006 (2.5 M NiCb), -0.0010 (5.0 M NaCl), 0.0016 (0.40 M NiCh) and 0.0014 (0.50 M NaCl), respectively, as shown in Fig. 7-2-4. The slope of the n-B curves of these electrolyte solutions also seems to be dependent more on [CI] than on [Na" ] or [Ni ""] at higher concentrations. Although the paramagnetism of Ni ^ due to unpaired 3d electrons is important, the refractive index under 10 T shows little difference between the paramagnetic species (Ni^^) and the diamagnetic species (Na^). Therefore, the n-B curves obtained should originate in Cr.

Although the complex magneto-optical behavior of electrolyte solutions cannot be explained easily, we can presume that two species bear different n-B curves, e.g., the one dominant at higher concentrations and the other at lower concentrations. Possible candidates for the higher concentrations are a hydrate complex or an ion pair containing more than one chloride ion, since the absolute value of n and the negative dependence of A2 on ^ is dominated by [CI] at higher concentrations for both NiCh and NaCl solutions. In addition, there are no distinct differences between the two cations. The origin of the n-B curve at lower concentrations should be water itself. According to the three-zone model by Frank and Wen, ^ ions hydrated by water (A zone) are further surrounded by a weakly interacting "destructured" region (B zone) where water molecules are neither oriented to the core ion nor hydrogen-bonded to each other. Around them is bulk water (C zone) in which water molecules are structured by hydrogen bonds. Thus, the n-B curve at lower concentrations is considered to be the one for bulk water (C zone) superimposed by another for the ionic species (A or B zone). For the saturated solutions, the n-B curve should be purely dominated by the hydrated ions or the ion pairs.

The refractive index of aqueous NaCl («Naci) solution is plotted against its concentrations (cNaci) with (circles) and without (squares) the magnetic field of 10 T in Fig. 7-2-5, as measured by the SPR method. The nwaci increases rapidly up to -0.5 M, above which it increases linearly in both the absence and presence of the magnetic field. The AiNaci - CNaci curves at 0 T and 10 T cross at 0.75 M (see inset to Fig. 7-2-5). The figure clearly shows the occurrence of the trade-off between the two bearers of the different magneto-optical behaviors at concentrations of 0.5-0.75 M.

Recently, the Cotton-Mouton (C-M) effect, i.e., a magneto-opfical anisotropy between the axes along and perpendicular to the magnetic fields, was measured for water and aqueous electrolyte solutions. ^^ The authors of the report confirmed the small C-M effect of pure water under 13 T that arises from the anisotropics of the molecular polarizability and the molecular magnetizability of water. This partly explains the increase in the refractive index of water, although the C-M measurement shows only the difference in the components of the refractive index. As for electrolyte solutions, the specific C-M constant for NaCl in the aqueous solution is


1.370

1.360

1.350

1.340

1.330 ^ 1 2 3 4 5

Concentration of NaCl / M

Fig. 7-2-5 Refractive index of the aqueous NaCl solution measured by the SPR method plotted against the concentration in the absence (squares) and the presence (circles) of a magnetic field of 10 T. A magnified view ranging from 0.50 M to 1.0 M is shown in the inset. [Reproduced from H. Hosoda et al., 7. Phys. Chem. A, 108, 1461 (2004)]

six-fold larger than that for KCl, and the sign is opposite. These two cations are closed-shell atomic ions and should not have any intrinsic anisotropic polarizability and magnetizability. The unique difference between Na^ and K" is the hydrate dynamics in aqueous phase. It is well known that K^ and Na^ are termed a structure-making cation and a structure-breaking cation, respectively, according to the Frank and Wen model. ^^ They attributed the difference of the C-M effect between Na^ and K^ to the hydration structure surrounding the cations. Their idea of the magnetism of aqueous electrolyte solution affected by the structure of hydrated water supports the current discussion.

7.2.5 Summary of Recent Magneto-optical Effect Studies on Aqueous Systems

The refractive indices of water and aqueous electrolyte solutions were measured. The refractive index of pure water under 10 T increases by - 0 . 1 % more than that of pure water under zero magnetic field. It is proposed that the hydrogen bond of water is stabilized under a magnetic field. Therefore, the optical properties of the ultraviolet absorption and the refractive index should increase. Aqueous electrolyte solutions at higher concentrations show decrease in the refractive index under 10 T, which may possibly be explained by the formation of a hydrate complex or an ion pair containing more than one chloride ion.

References

1. D. Eisenberg, W. Kauzmann, The Structure and Properties of Water, Oxford


University Press, Oxford (1969), Japanese Translation, Misuzu-Shobou, Tokyo (1983). 2. S. Yeremenko, M. S. Pshenichnikov, D. A. Wiersma, Chem. Phys. Lett., 369, 107

(2003). 3. K. Winkler, J. Lindner, H. Bursing, P. Vohringer, J. Chem. Phys., 113, 4674 (2000). 4. P. Bour, Chem. Phys. Lett., 365, 82 (2002). 5. M. in het Panhuis, P. L. A. Popelier, R. W. Munn, J. G. Angyan, J. Chem. Phys., 114,

7951 (2001). 6. G. S. Tschumper, M. L. Leininger, B. C. Hoffman, E. F. Valeev, H. F. Schaefer III,

M. Quack, J. Chem. Phys., 116, 690 (2002). 7. S. Dellerue, M. -C. Bellissent-Funel, Chem. Phys., 258, 315 (2000). 8. A Special Issue for Water Reserch (A. K. Soper, P. J. Rossky, eds.), Chem. Phys.,

258(2-3), 107 (2000). 9. J. Homola, S. S. Yee, G. Gauglitz, Sensors and Actuators B, 54, 3 (1999).

10. J. Melendez, R. Carr, D. U. Bartholomew, K. Kukanskis, J. Elkind, S. Yee, C. Furlong, R. Woodbury, Sensors and Actuators B, 35-36, 212 (1996).

11. J. Melendez, R. Carr, D. Bartholomew, H. Taneja, S. Yee, C. Jung, C. Furlong, Sensors and Actuators B, 38-39, 375 (1997).

12. J. L. Elkind, D. I. Stimpson, A. A. Strong, D. U. Bartholomew, J. L. Melendez, Sensors and Actuators B, 54, 182 (1999).

13. Handbook of Chemistry and Physics (D. R. Lide, ed.), CRC Press, Boca Raton (2002). 14. L. Pauling, J. Chem. Phys., 4, 673 (1936). 15. R. Cini, M. Torrini, J. Chem. Phys., 15, 2826 (1968). 16. J. S. Philo, W. M. Fairbank, J. Chem. Phys., 11, 4429 (1980). 17. M. Iwasaka, S. Ueno, J. Appl. Phys., 83, 6459 (1998). 18. K. Kobayashi, J. Phys. Chem., 87, 4317 (1983). 19. A. B. P. Lever, Inorganic Electronic Spectroscopy, p.334, and references cited

therein, Elsevier, Amsterdam (1968). 20. H. S. Hamed, B. B. Owen, The Physical Chemistry of Electrolytic Solutions, p.42, and

referenced cited therein, Reinhold, New York (1950). 21. M. Magini, G. Paschina, G. Piccaluga, J. Chem. Phys., 76, 1116 (1982). 22. W.-Y. Wen, Ions and Molecules in Solution (N. Tanaka, H. Ohtani, R. Tamamushi,

eds.), p.45, Elsevier Science, Amsterdam (1983). 23. J. H. Williams, J. Torbet, J. Phys. Chem., 96, 10477 (1992).

7.3 Magnetic Field Effects on Phase Transitions in Diamagnetic Materials

Magnetic field effect on thermodynamic properties at room temperature was reported by Yamaguchi et al. for the change in the equiUbrium vapor pressure in metal-hydrogen systems/^^ In this case, the magnetic effect was explained by adding the magnetic term in the chemical potential of the system. However, there has been almost no report on magnetic effects on the phase transitions of diamagnetic materials. Since the magnetic energy at 10 T, for example, is still weaker than the thermal energy at room temperature and the magnetic effects on the phase transitions of diamagnetic materials are expected to be very small, a highly sensitive apparatus under a strong magnetic field must be used in order to detect the magnetic effects on the phase transitions of diamagnetic materials.

7.3 Magnetic Field Effects on Phase Transitions in Diamagnetic Materials 319

For this purpose, a high resolution and super-sensitive thermal apparatus was developed. A high resolution and super-sensitive differential scanning calorimeter (DSC) working between 120 and 420 K capable of measuring heat as small as on the order of 20 nW with temperature resolution of less than 1 mK has been used to study the magnetic effects on the phase transitions of diamagnetic materials/^^^ The schematic drawing of the DSC is shown in Fig. 7-3-1."^^^ The calorimeter was set in a magnetic bore 100 mm in diameter and was cooled by a refrigerating head, A, which was connected to the calorimeter through copper plates, D. The temperature difference between the test sample and the reference material, produced by the heat absorbed in or released from the sample, was measured by thermoelectric modules, which were made of 18 semiconducting thermoelectric elements connected in a series. The output voltage of the thermoelectric module was about 7.8 mV K"', being about 200 times larger than the conventional thermocouple. The baseline of this

Fig. 7-3-1 Schematic drawing of the high resolution and super-sensitive DSC working in a magnetic bore between 120 and 420 K. A; refrigerating head, B; thermal reservoir, C; thermal insulator, D; copper plates connected to the calorimeter, TS1-TS4; Pt resistance thermometers. [Reproduced from H. Inaba et al., J. Appl. Phys., 96, 6128 (2004)]


DSC was about two orders of magnitude more stable than that of a commercial DSC. The temperature of the sample was measured using a Pt resistance thermometer, TSi. Before measuring the magnetic effect on the phase transitions of diamagnetic materials, the magnetic effect on the Pt thermometer due to the magneto-resistance was measured.^ ^^ The DSC was kept at room temperature without using a temperature control and a magnetic field of 5 T was applied three times to measure the magnetic effect on the Pt thermometer, TSi. The magnetic effect on the Pt thermometer shifted reproducibly was as large as 18.2 mK at 288.65 K. The temperature dependence of the magnetic effect on TSi was also measured by controlling the temperature to be constant using the thermometer, TS4, shown in Fig. 7-3-1, and the magnetic effect on TSi was determined. It decreased with increase in the temperature. The temperature shift of TSi due to the magnetic field was corrected to obtain the magnetic effects on the phase transitions of diamagnetic materials hereafter.

7.3.1 Magnetic Field Effect on the Melting Transition of H2O and D2O

The first example of the magnetic effect on the phase transitions of diamagnetic materials is the melting transition of H2O and D20. ^ Purified H2O and D2O samples of 7.40 and 1.97 mg, respectively, together with quartz powder of about 7 mg were hermetically sealed in an aluminum pan for the DSC measurement. Quartz powder was used with water to make the crystallization of water easier and the measurement reproducible. The D2O sample mixed with quartz powder was analyzed by a mass spectrometer and found to include about 19% of HDO. The inclusion of HDO in the D2O sample may be due to light water adsorbed on the surface of the quartz powder. Heat flux measurements on the melting transition of H2O were conducted three times at a heating rate of 1 mKs' using the DSC. ^ The results of two repeated measurements are shown in Fig. 7-3-2, indicating almost the same results. The reproducibility of the three repeated measurements was ±0.5 mK for the melting temperature. In the cooling run, however, the solidifying temperature was not reproducible due to a large super-cooling effect. The heat flux measurements on the melting transition of H2O under a magnetic field of 6 T were made three times at a heating rate of 1 mKs~^ The DSC curve of the melting transition at 6 T was obtained in a manner similar to that shown in Fig. 7-3-2 within the imprecision of ±0.5 mK except for a small temperature shift of the melting temperature. Since the shift of the transitional peak due to the magnetic field was slightly dependent on the peak position, the temperature shift was averaged over the entire range of the transitional peak. The averaged melting temperature with the magnetic field of 6 T was 5.6 ± 0.7 mK higher than that without the magnetic field. The higher melting temperature obtained under the magnetic field shows that the solid phase becomes relatively stable compared with the liquid phase by the magnetic


^ -2

<

-4 -AOT(lst)

O0T(2nd)

273 273.5 274

r / K

Fig. 7-3-2 DSC curves on the melting transition of H2O without the magnetic field at a heating rate of 1 mKs"'. The results of two repeated measurements are shown. [Reproduced from H. Inaba et al., J. Appl. Phys., 96, 6129 (2004)]

field. Similarly, the shift in the melting temperature due to the magnetic field was measured by changing the magnetic field. The shift in the melting temperature of H2O due to the magnetic field was found to be 4.3 ± 0.8 mK at 5 T and 3.2 ± 0.8 mK at 4 T. The shift in the melting temperature of H2O is plotted against the square of the magnetic field as shown in Fig. 7-3-3, where a linear relationship between the shift in the melting temperature and the square of the magnetic field is seen. The heat flux measurements on the melting transition of D2O with and without the magnetic field were also made three times at a heating rate of 1 mKs"^ The repeated measurements showed almost the same results. The DSC curve showed two endothermic peaks around 277.34 and around 277.04 K due to the melting transition of D2O and HDO, respectively. Similarly, as in the case of H2O, the melting transition of D2O and HDO shifted to the high temperature side by the application of the magnetic field. The shift in the melting temperature of D2O due to a magnetic field of 6 T was found to be 21.9 ± 0.7 mK using the temperature range after the peak, because a temperature lower than 277.34 K includes the effect of melting of HDO. The shift in the melting temperature of HDO due to the magnetic field of 6 T was 17.3 ± 0.8 mK using the temperature range before the peak. The shift in the melting temperature of D2O and HDO due to the magnetic field was measured by changing the magnetic field. The shift in the melting temperature of D2O due to the magnetic field was obtained to be 16.9 ± 1.0 mK at 5 T and 9.6 ± 1.4 mK at 4 T. The shift in the melting temperature of HDO due to the magnetic field was obtained to be 13.9 ± 1.1 mK at 5 T and 7.0 ± 1.2 mK at 4 T. The shift in the melting temperature of D2O and HDO plotted against the square of the magnetic field also showed a linear relationship.

The effects of a magnetic field on the phase transition of diamagnetic


0 10 20 30 B-/V

Fig. 7-3-3 Shift in the melting temperature of H2O against the square of the magnetic field. [Reproduced from H. Inaba et al., J. Appl. Phys., 96, 6129 (2004)]

materials can be discussed on the basis of a simple extension of the Clapeyron equation.^^^ When we consider a diamagnetic substance at temperatures near the phase transition, which has an angle 6 with respect to the magnetic field, the molar Gibbs energy change dG including the magnetic effect is represented by the following equation:

dG = -5dr-(^ /ô)(Zis in-^ + ;^HCOs'0)d^ (1)

where S is the molar entropy, Xi. and X\\ are the diamagnetic susceptibility perpendicular and parallel to the molecular axis, respectively, and B is the magnetic flux. Since x± and X\\ are both negative values, the second term of Eq. (1) becomes positive and then the system becomes unstable by applying the magnetic field. When the phase transition is under way, the molar Gibbs energy at the high temperature phase is equal to the low temperature phase to give dGh = dGi, where subscripts h and 1 mean the high temperature and the low temperature phase, respectively. Then the shift in the transition temperature by the application of the magnetic field, AT, becomes, by integrating the differential of Eq. (1),

(2) AT = -{B^ 12jUo){;fi.h sin' h -Z i i sin' 0\ + X^^-^ cos' 0h -

;fn..cos2^.}/(5h-S,)

Here we consider the melting transition of H2O as an example. Since the diamagnetic susceptibility of ice is -8.11 x 10" along the c axis and -8.07 X 10"^ along the a axis according to Lonsdale,^^ the diamagnetic susceptibility of ice can be regarded to be isotropic as an approximation. Since liquid water is also isotropic, x^ is equal to X\\ (= X) t o h in the solid and the liquid phases. Then AT for H2O and D2O becomes

Ar = (^2/2/io)ai-Zh)/(5h-5,) (3)


Using the value for the diamagnetic susceptibility of H2O ice, -8.09 x 10" , taken as the average value along the a axis and c axis reported by Lonsdale^^ and for liquid H2O, -9.05 x 10" , from the results at 274.15 K by Auer, ^ and the entropy change for the melting of H2O ice: 5h - 5i = 22.00 J K" mor\'^^ ^T for the melting transition of H2O is calculated to be 8.4 juK at 5 T. Similarly, using reference values''^ for the diamagnetic susceptibility of D2O ice and D2O water: -8.10 x 10" and -8.94 x 10" , respectively, and the entropy change for the melting of D2O ice: 5h - 5i = 22.67 J K~ mol"\*^^ Ar for the melting transition of D2O is calculated to be 7.3 juK at 5 T. These facts show that the calculated results based on the simple appUcation of Eq. (3) do not explain the observed results. Therefore, an alternative interpretation is necessary to explain these experimental results. Iwakasa and Ueno' ^ found that a near-infrared spectrum of water shifted to the low frequency side by the application of the magnetic field of 14 T. Hosoda et al. ^ found that the refractive index of water increased by the application of a magnetic field of 10 T. Both effects were considered to be due to the strengthening of hydrogen bonding by the application of the magnetic field. ^ ' ^^ However, it is difficult to explain why the magnetic field strengthens the hydrogen bonding. The diamagnetic energy mag of H2O due to the magnetic field is represented by

^mag=^'K.//(2A/o) (4)

where Vm is the molar volume of H2O. £mag is calculated to be -1.62 mJ mol" at 5 T. On the other hand, the energy of the hydrogen bonding is between 5.4 and 18.8 kJ mol" estimated by various researchers,'^^ being 6 or 7 orders of magnitude larger than the diamagnetic energy of H2O. Therefore, it is difficult to assume that the diamagnetic energy shown in Eq. (4) directly contributes to strengthen the hydrogen bonding. At this point, we would like to consider the idea of dynamic magnetic susceptibility. In order to calculate x in Eq. (2) we have used the data of the magnetic susceptibility which were measured statically, using a microbalance and SQUID. According to Ramsey,' ^ however, the magnetic susceptibility ;|f of a diamagnetic molecule is expressed as

;f = ^static + ;f hf ( 5 )

where the first term is the usual static diamagnetic susceptibility and the second term is the dynamic or high frequency diamagnetic susceptibility due to molecular motions such as rotation and vibration under a strong magnetic field. Ramsey also gave a theoretical equation of x^^ for the rotational part of diatomic molecules.''*^ Since H2O is a polarized molecule, the charge of the nucleus and electrons in the molecule cannot be cancelled completely. Then the thermal motion of the partially charged atoms of H2O in the magnetic field gives rise to the Lorentz force at temperatures near the melting point. The Lorentz force would suppress the thermal motion of


the partially charged atoms, strengthen the hydrogen bonding'^' ^^^ and increase the dynamic magnetic susceptibility. The term "dynamic" means the movement of atoms due to thermal motions such as vibrating and rotating motions at high frequencies. The suppression of the thermal motions by the magnetic field makes the solid phase relatively more stable and makes the dynamic magnetic susceptibility larger. Therefore, we can regard x in Eq. (2) as the dynamic magnetic susceptibility rather than the static magnetic susceptibility. The linear dependencies of AT as a function of the square of the magnetic field for H2O, as shown in Fig. 7-3-3, and for D2O and HDO support this assumption. The reason for the larger temperature shift in the melting transition for D2O than that for H2O is not clear at present but may be related to the difference in the thermal motions between H2O and D2O. The D2O ice has a considerably larger molar heat capacity and a larger linear thermal expansion coefficient than H2O ice in the temperature region higher than 150 K, according to Petrenco and Whitworth.^^^ The vibrational and rotational modes of D2O have lower frequencies than those of H2O due to the heavier mass of hydrogen atoms, so they are more easily excited, giving a larger heat capacity and a larger linear thermal expansion coefficient. Therefore, a larger number of vibrational and rotational modes are considered to be excited at temperatures near the melting point in D2O ice. Such modes in D2O ice would contribute more to increase the dynamic magnetic susceptibility in a strong magnetic field.

7.3.2 Magnetic Field Effect on the Ferroelectric Transition of Single Crystalline KD2PO4

The second example of magnetic field effect on the phase transitions of diamagnetic materials is the ferroelectric transition of single crystalline KD2PO4 (DKDP)^^ using the high resolution and super-sensitive DSC working in a magnetic bore. DKDP belongs to a family of ferroelectric crystals in which the molecular units are linked by hydrogen bonds. Their hydrogen bonds are preferentially oriented in a specific direction in the crystal lattice. In the DKDP crystal, O-H bonds lie in the c plane. Diamagnetic anisotropy was reported for the hydrogen-bonded ferroelectrics due to the anisotropy of spatial atomic distribution. DKDP is known to have a ferroelectric phase transition at about 220 K from a ferroelectric orthorhombic structure to a paraelectric tetragonal structure. ^ The transition temperature of DKDP with a magnetic field of 5 T along the a axis and the c axis was 5.6 ± 0.8 mK and 2.8 ± 0.9 mK, respectively, higher than that without the magnetic field. The temperature shift of the phase transition due to the magnetic field was negative and about three orders of magnitude larger than the calculated one using the magneto-Clapeyron equation shown in Eq. (2), if data of the static diamagnetic susceptibility are used. Therefore, the use of the static diamagnetic


susceptibility does not explain the experimental results. Instead, if it is assumed that the dynamic magnetic susceptibility proposed by Ramsey is predominant, the experimental data are understandable using the same magneto-Clapeyron equation. The thermal motion of constituent ions in DKDP is considered to be suppressed due to the Lorentz force under the magnetic field, making the low temperature phase more stable. However, it is not clear why the shift in the transition temperature along the a axis is higher than that along the c axis.

7.3.3 Magnetic Field Effect on the Liquid Crystal to the Isotropic Liquid Transition of iV-p-ethoxybenzylidene-p'-butylaniline

The third example of magnetic field effect on the phase transitions of diamagnetic materials is the liquid crystal to the isotropic liquid transition of A^-/7-ethoxybenzylidene-/7'-butylaniline (EBBA). ^ EBBA is known to have a liquid crystalline phase near room temperature and have a phase transition from a nematic phase to an isotropic liquid phase at about 352 K.' ^ For the DSC measurement, about 3 mg of sample was enclosed in an aluminum capsule. The heat flux measurements of EBBA were made between 233 and 363 K at a heating and cooling rate of 1 mKs"' using the DSC. ^ The crystal-liquid crystal transition and the liquid crystal-isotropic liquid transition were observed at 309.0 K and at 352.7 K, respectively. The magnetic effect on the liquid crystal-isotropic liquid transition was measured under the same condition. The same measurements were repeated twice, indicating almost the same results. The averaged melting temperature of EBBA with a magnetic field of 5 T was 19.8 ± 1.6 mK, higher than that without the magnetic field, as shown in Fig. 7-3-4. The results measured at the cooling rate of 1 mKs"' showed almost the same results except for a slight change in the transition temperature, showing the reversible nature of the transition. EBBA is a linear molecule and has a diamagnetic anisotropy. Since the diamagnetic susceptibility perpendicular to the molecular axis is larger than that parallel to the molecular axis' ^ because of the large contribution of the benzene ring, EBBA molecules are considered to be oriented along the magnetic field. The temperature shift of the phase transition by the application of the magnetic field of 5 T for EBBA can be calculated using Eq. (2). It was calculated to be 0.25 mK using the reference values for (xw - ;fi)J^^ and assuming that x±, i = X-L^ h, cos ©h = 1/3 and cos 0\ = 0. The calculated value is two orders of magnitude smaller than the observed value, 19.8 mK. In order to explain the larger experimental values, two kinds of mechanisms may be considered. One is the assumption that a partially ordered state during the phase transition determines the entire phase transition. For example, an ordered state even in the liquid phase can be assumed and it can be grown to nuclei of the solid phase due to magnetic orientation under a strong magnetic field. Therefore, the transition temperature under a strong


^Nematic \ \ \ ' "' ^P*^ " "phase \ o " "l"id "

-20- \ ° o phase

=1 A O ^ O

O) - A O

- 4 0 - A O A °

A O A

A

o

O :5T ^ ^"'' o^ A : O T ^

-60 ~ ^ 352.3 ' 352.4 ^ 352.5

r / K

Fig. 7-3-4 Comparison between the averaged DSC curves of EBB A with a magnetic field of 5 T and without magnetic field.

magnetic field can be determined by the far smaller entropy change between a partially ordered state in the liquid phase and the state with nuclei of the solid phase in Eq. (2) than the total entropy change between the liquid crystal and the isotropic liquid. ^ The other mechanism is the effect due to the Lorentz force acting on the partially ionized moving atoms. Since EBBA has polarized components of -CNH and -OC2H5, they may be partially ionized and the thermal motion of the molecule may be suppressed by the magnetic field, rendering the low temperature phase more stable. Then the dynamic magnetic susceptibility is considered to contribute mainly to the magnetic susceptibility in Eq. (2).

7.3.4 Magnetic Field Effect on the Rotator Transition and Melting Transition in C32H66

The fourth example of the magnetic effect on the phase transitions of diamagnetic materials is the rotator transition and melting transition in C32H66. ^ In this case, the transition temperatures due to the rotator transition and the melting transition under a strong magnetic field became higher. The calculated temperature shift using Eq. (2) and static magnetic susceptibility was three orders of magnitude smaller than the observed one. Since C32H66 is also a flexible linear molecule like EBBA, the reason for the magnetic effect may be considered to be similar to that for EBBA.

References

1. M. Yamaguchi, I. Yamamoto, F. Ishikawa, T. Goto, S. Miura, J. Alloys and Comp., 253, 191 (1997).

7.4 Ax of Inorganic Insulators and Detection of Small Ax 327

2. I. Yamamoto, M .Yamaguchi, T. Goto, S. Miura, J. Alloys and Comp., 231, 205 (1995).

3. H. Inaba, T. Saitou, K. Tozaki, H. Hayashi, / AppL Phys., 96, 6127 (2004). 4. H. Hayashi, C. Nonaka, K. Tozaki, H. Inaba, C. Uyeda, Thermochimica Acta, 431,

200 (2005). 5. S. Hosaka , K. Tozaki, H. Hayashi, H. Inaba, Physica R, 337, 138 (2003). 6. H. Inaba, K. Tozaki, H. Hayashi, C. Quan, N. Nemoto, T. Kimura, Physica R, 63,

324 (2002). 7. T. Kimura, Jpn. J. AppL Phys., 40, 6818 (2001). 8 K. Lonsdale, Nature, 164, 101 (1949). 9. H. Auer, Annai der Physik., 8, 595 (1933).

10. D. Eisenberg, W. Kauzmann, The Structure and the Properties of Water, Oxford University Press, Oxford (1969).

11. Yu. V. Ergin, L. I. Kostrova, Zh. Strut. Khim., 11, 481 (1969). 12. M. Iwasaka, S. Ueno, J. AppL Phys., 83, 6459 (1998). 13. H. Hosoda, H. Mori, N. Sogoshi A. Nagasawa, S. Nakabayashi, 7. Phys. Chem. A,

108, 1461 (2004). 14. N. F. Ramsey, Molecular Beams, p. 169, Oxford University Press, New York (1956). 15. V. F. Petrenco, R. W. Whitworth, Physics of Ice, p.44, Oxford University Press, New

York (1999). 16. R. Nelmes, Ferroelectrics, 17, 87 (1987). 17. M. Sorai, T. Nakamura, S. Seki, Bull Chem. Soc. Jpn., 47, 2192 (1974). 18. L. V. Choudary, J. V. Rao, P. Venkatacharyulu, Phase Transitions, 9, 289 (1987).

7.4 Diamagnetic Anisotropy of Inorganic Insulators Deriving from Individual Chemical Bonds and Detection of Small Magnetic Anisotropy Using Micro-gravity

7.4.1 Detection of Magnetic Anisotropy with High Sensitivity

Various magnetic effects have been reported for diamagnetic materials in high magnetic field which are considered to derive from the intrinsic diamagnetic anisotropy (A;f)DiA of the material.'^^ The numerical (A;if)DiA values are essential in investigating the alignment process quantitatively; however, the values are too small to be detected by the conventional torque methods in many cases.^^ The magnetic effects are hence currently recognized for a limited number of materials.^^

Conventional torque methods used to measure magnetic anisotropy are based on a principle proposed by Krishnann and Banergee;'*^ the method is based on a balance between the magnetic anisotropy energy induced in the sample and the restoring force of the fiber suspending the sample in a horizontal field B. The direction of the magnetically stable axis rotates in the horizontal plane. Rotational equation for a solid body having a magnetic anisotropy A;f(emu g"') is described as

I(dê/dt'-) = -(\/2)B'NAxsin2e-iD/f)e d)

Here 9 is the angle between B and the direction of the stable axis. / and N


(a)

^ .

Fine fiber

(or no fiber)

Sample

B

(b)

Stable axis

Ax Measurement

by Magnetic oscillation

5.0

£ 10.0

• AlOOH l:z-y • 2:z-y • 3:z-y O AlOOH l:y-x A 2:y-x O 3:y-x -H^'

1.0 2.0 i / r / x i o - ' K '

Sample I

stage

(0 (g)

:^lField

direction { ^

1 1

Terrestrial gravity

u.

Micro-gravity

Fig. 7-4-1 (a) Principle of measuring magnetic anisotropy using magnetic oscillation of the stable axis of the sample with respect to field direction. {b)Temperature dependence of Ax values measured for AlOOH single crystals (see ref. 14). (c)-(g)Visual images of graphite during micro-gravity experiments taken every 0.07 sec, arranged in order of time from left to right (see ref. 11). (c) was taken just before micro-gravity was applied. Graphite crystal was initially placed on a sample stage with inclination of 45 degrees; c planes of graphite were placed parallel to the inclined slope. The stage was removed from its initial position with high velocity immediately after achievement of micro-gravity to minimize the amount of kinetic energy transferred from the stage to the sample.

are the moment of inertia and the weight of the body, respectively. D and £ denote the tensional rigidity and the length of the fiber, respectively. It is seen that the sensitivity of the measurable Ax value is limited by D and i in the conventional methods.

The term of restoring force energy in Eq. (1) is controlled to be negligible compared to that of magnetic anisotropy energy by the present authors for the purpose of improving sensitivity.^ ^ Direction of the magnetically stable axis shows rotational harmonic oscillation with respect to B in the improved method, as shown in Fig. 7-4-1(a), and Ax is obtained from the period of oscillation r without the use of D and £ as Ax = 4KIN~\BT)~^^ ; Aj is obtained with high precision from the gradient of the proportional relationship observed between r and ^'(see for example Fig. 2 of reference 8). A Ax value as small as 10'^ emu/sample was detected in the field intensity of 5 T using a sub-millimeter size single crystal of a-quartz having a weight of 8 x 10"^ g. ^

It is essential to distinguish the (A;f)DiA value from the anisotropy of

7.4 Ax of Inorganic Insulators and Detection of Small Ax 329

paramagnetic susceptibility (A;f)pARA deriving from the impurity ions. The (A;if)pARA components, which followed the Curie law, were extracted effectively by measuring the temperature dependence of A; ; the Is^-T relationships were measured for several crystals having different paramagnetic concentrations for a single material. The high temperature limits of A;f converged to a single value, which was identical to the intrinsic (A/)PARA value of the material as described in Fig. 7-4-1(b) for diaspore single crystals. ^ Small (Aj;f)DiA values of the order of 10~^ emu g~' were detected by the improvements mentioned above. Accordingly, (A;f)DiA values were newly obtained for 13 basic oxides listed in Table 7-4-1.^'^^ Detection of these (A;f)DiA was possible from a mm-size single crystal containing paramagnetic impurity ions. (A;f)DiA values were previously measured only on large diamagnetic material crystals which can be obtained with high purity. The obtained sensitivity, however, is not high enough to achieve further accumulation of (A;f)DiA values; accumulation is required to clarify the overall characteristics of diamagnetic anisotropy of various diamagnetic materials described in section 7.4.2.

The fiber itself should be deleted from the method described above in order to realize further improvement of sensitivity. A preliminary observation to examine this principle was performed at the Micro-gravity Laboratory of Japan (Toki, Gifu, Japan). A typical example of a visual image of a graphite crystal floated in micro-gravity is shown in Fig. 7.4.1(d)-(g).''^ The sample, in the form of a rectangular prism, is stabilized in micro-gravity. A set of parallel planes of the prism having the widest area, which were identical to the magnetically stable c plane of graphite, showed rotational oscillation with respect to B applied in the vertical direction in the figure. Periods of oscillation x of the samples were measured from the images, r was calculated by deleting the second term of Eq. (1) as

T = 2;r(//AÂ/)-'^'^-'(l + (l/4)sin-(eo/2) + ---) (2)

where o denotes the angle of amplitude which was 45° in the present experiment, ^x is obtained by inserting the measured values of r, B, N, 6Q and / in the above equation. The A; value obtained was consistent with the published value of graphite, ' indicating that factors other than magnetic torque can be neglected as the cause of the observed oscillation. It is expected from Eq. (2) that the sensitivity of A; is improved by increasing B and r; no factor is detected which limits the improvement of sensitivity in the above measurement so far. The sensitivity measured in micro-gravity was improved recently to the order of 10~ emu g"' by increasing the field intensity to 1.3 T. ^ Very small A; may be detected in an orbital laboratory where a long r of range of 1 minute can be measured. Development of the measuring system is now being carried out to achieve the expected sensitivity.


Table 7-4-1 Diamagnetic Anisotropy of Basic Inorganic Oxides

Sample

ADP Apophyllite Brucite Corundum'' Diaspore*"

Gibbsite Gypsum

Muscovite KDP Orthoclase

Petalite

Scapolite Talc

[a-c] : [a - c] [c-a]

'[c-a] [c-a] [a-b] [c-a] Ui-Z : ] [X^-XA [X^-X2] [c-a] [a-c] [X^-Xi] [X^-X2] ^ 3 - ^ 1 ] [X^-X2] [x^-x^] [X2-X^] [a-c] [c-a]

(A;)f)DiA

/xlO'emug-'

11 ± 0.5 3.8 ±0.1 2.6 ±0.2 0.7 ±0.1 4.2 ±0.3 0.9 ±0.2 1.4 ±0.2 9.6 ±0.2 7.0 ±0.1 2.2 ±0.1 11 ±2

8.3 ±0.3 2.1 ±0.1 1.4±0.1 0.6 ±0.1 3.5 ±0.1 2.7 ±0.1 0.8 ±0.1 0.8 ±0.1 220 ± 0.2

Magnetically stable axis

a axis a axis c axis c axis c> a> b

c axis

Xi>X2>X^

c axis a axis

X^>Xi>X2

Xy>X2>Xi

a axis c axis

7.4.2 Origin of Diamagnetic Anisotropy of Inorganic Materials

The origin of anisotropy was explained systematically for the first time on ionic oxide crystals by assigning a constant (AJ)DIA value on individual electron orbitals of the chemical bonds composing the crystal.^^ ^^^ The analysis was based on the (A;f)DiA values of oxides listed in Table 7-4-1. A similar assumption was previously done to explain the (A; )DIA values of organic molecular crystals.^ " Diamagnetic susceptibility of a nonmetallic material is approximately equivalent to the sum of susceptibilities assigned to the individual electron orbitals comprising the material according to Pascal's law. ^ This summation can be described by the 3-dimensional X' tensor of a material, assuming that each orbital possesses a constant uniaxial anisotropy A/BO = jfeon - Xôu where JBOH and ;|fBoi denote the susceptibilities parallel and perpendicular to the bond direction, respectively; the bond direction is identical to the principle axis.'^ '' ^ Anisotropy of a bond is obtained from a field-induced free energy.

UiB) = -(\/2)BHxBoiÂXBoiaW +b'p' ^cY)} (3)

Here the direction cosines of i? is defined as (a,b,c), whereas (a,p,y) denote the direction cosines of the bond direction; the jc, y and z coordinates of the vector components described above are identical to the three magnetic principle axes of the crystal. According to the above model, (A;f)DiA values of the crystal between x-y, y-z and z-x axes should be proportional to, la^ -Ip^, 1(5^ - Xy^ and Xy' - la', respectively. Ia'\ X/J' and ly^ are calculated for all the bonds included in a unit cell of the material; the calculations can be performed for any crystal using the published data of atomic positions. The positive correlations expected in the model were clearly seen for three

7.4 Ax of Inorganic Insulators and Detection of Small A^ 331

types of chemical bonds, as illustrated in Fig. 7-4-2/^^ namely for the T-0 bonds composing the tetrahedral [TO4] units,' ^ for the hydrogen bonds ^ and for the M-O bonds of the [MOe] units/'*^ The diamagnetic anisotropy of a single bond is determined from the gradient of the regression lines as

Hydrogen bond: A/BO = -3.7 X 10" ° emu (4)

T- O bond of a tetrahedral [TO4 ] unit: A; BO = -1.1 X 10" ^ emu (5)

M- O bond of an octahedral [MOe ] unit: A; BO = -0.32 x 10" ^ emu (6)

The negative values of Eqs. (4)-(6) show that the three types of bonds all have easy plane type of anisotropy with their bond directions being the unstable axes. The large differences between the three A;fBo values derive from the difference in the electron density distribution of each orbital. Quantitative analysis of this difference is a subject of theoretical solid state physics.

-0.5 -1.0 Calculated iSx

Fig. 7-4-2 Comparison between measured and calculated values reported for diamagnetic anisotropy of inorganic oxides (see ref. 14). Open, gray and closed symbols indicate the data for T-O, 0-H and M-0 bonds, respectively. Numbers on the symbols denote the analyzed samples, namely a-quartz: 1, orthoclase; 2-4, apophylite: 5, gypsum; 6-8, KDP: 9, ADP: 10, hexagonal ice: 11, free water molecule: 12-14, AUOH)?: 16, and Mg(0H)2: 17. The regressions between experimental and calculated anisotropy were obtained separately for the three chemical bonds as

O- H bond : {^x)^^^ = -2.2 (AI) + 0.06 (X 10^ emu mol') ( /= 0.99), T- O bond : (AxhiA = -0.63(AI) + 0.23 (X 10^ emu mol') ( /= 0.89), M-0 bond : (A; )DIA = -0.19(AI) + 0.03 (X lO^êmumor') (/=0.93),

where AI denotes the differences between l a - , Ip* and ly- described in the text. The correlation factor is denoted as/in the above equations.


The A;fBo of the bonds arranged in various directions cancel out when the [TO4] units or the [MOe] units hold regular symmetry in the crystal structure. However, both units usually show a slight distortion from regular symmetry in an actual crystal structure: This distortion was concluded to be the major cause of the measured (A;f)DiA.' ^ For example, the tetrahedral [TO4] unit is compressed in the direction of a axes in the case of a-quartz; accordingly the bond direction is preferentially aligned toward the c axisJ^^

The three chemical bonds are the major types of bonding orbital that compose inorganic oxides. Most of the unmeasured oxides hence possess finite (A;f)DiA, since the distributions of bond directions are generally not isotropic in a crystal structure with the exception of materials possessing cubic symmetry^ ^ ^ ; these materials have the potential of causing magnetic alignment at low field intensity, as observed in Fig. 5-7-l(b)(c)(d).^^^ Accordingly, most of the oxide crystals have the potential to cause magnetic effects deriving from magnetic anisotropy at finite field intensity. Data accumulation on various types of oxide crystals is required in order to examine the efficiency of the model. Sensitivity obtained at terrestrial gravity is not high enough to detect the (A; )DIA values mentioned in section 7.4.1. Measurement in micro-gravity described in Fig. 7-4-1 may be the breakthrough required for further accumulation of (AJ)DIA data.

7.4.3 Conclusions

1. A new principle to detect magnetic anisotropy ^x îth high sensitivity was established by observing a field-induced rotational-oscillation of the stable axis with respect to the field direction. Sensitivity can be improved by minimizing the restoration force of the fiber suspending the crystal in a horizontal field; the restoration force is the standard in measuring A;f in conventional methods.

2. The fiber itself was deleted from the above method in order to improve sensitivity. Rotational oscillation was observed for a mm-size single crystal of graphite floated in micro-gravity in a low magnetic field of 0.015 T. The amount of (A;t')DiA data is expected to increase considerably by improving this method.

3. (A;|f)DiA data were obtained for various inorganic oxides, i.e., apophylite, corundum, forsterite, orthoclase, KDP, ADP, hexagonal ice, gypsum, muscovite, petalite, scapolite, talc, MgO, Mg(OH)2, Al(OH)3 and AlOOH. The (A;f)DiA value obtained is essential in analyzing magnetic alignments of small particles dispersed in the fluid medium described in section 5.7.

4. Published (AJ)DIA values of inorganic oxides were explained consistently by assigning a constant amount of (A; )DIA on individual chemical bonds. The assigned {^x)^\^ values were 3.7 x 10" ^ emu for a hydrogen bond, 1.1 X 10" ^ emu for a T-O bond composing the tetrahedral [TO4] units and 0.32 x 10'^ emu for a M-O bond of the octahedral [M06] units;

7.5 Magnetic Field Effects on Photo-induced Ultrafine Particle Formation in Gas Phase 333

these bonds are the major types of chemical bonds composing the inorganic oxides.

5. Most of the inorganic oxides have the potential of possessing a finite amount of (A;f)DiA to cause magnetic alignment as well as other magnetic effects due to magnetic anisotropy according to conclusion no. 4 above. The possibility of producing new magnetic devices is expected to increase considerably when the selection of magnetically active materials is expanded to solid nonmagnetic materials in general.

References

1. G. Malet, K. Dransfeld, Topics in Appl Phys., 57, 144 (1985). 2. for example, various papers appearing in Proceedings of the International Symposium

on New Magneto-Science, Jpn. Sci. Tec. Corp., NIMS (1999). 3. R. Gupta, Diamagnetism Landort Bomstein, p.445, Springer-Verlag, Berlin (1983). 4. K. S. Krishnann, S. Banerjee, Philos. Trans. R. Soc. London, A231, 235 (1933). 5. C. Uyeda, Jpn. J. Appl. Phys., 32, 268 (1993). 6. C. Uyeda, A. Tsuchiyama, T. Yamanaka, M. Date, Phys. Chem. Minerals, 20, 82

(1993). 7. C. Uyeda, H. Chihara, K. Okita, Physica B, 246-247, 171 (1998). 8. C. Uyeda, K. Ohtawa, K. Okita, Jpn. J. Appl. Phys., 39, L514 (2(X)0). 9. C. Uyeda, K. Ohtawa, K. Okita, N. Uyeda, Jpn. J. Appl. Phys., 39, L890 (2000).

10. C. Uyeda, K. Ohtawa, K. Okita, J. Phys. Soc. Jpn., 69, 1019 (2000). 11. C. Uyeda, K. Tanaka, R. Takashima, Jpn. J. Appl. Phys., 42, LI226 (2003). 12. C. Uyeda, M. Mamiya, R. Takashima, T. Abe, H. Nagai, T. Okutani, Jpn. J. Appl.

Phys., 45,U24i2006). 13. C. Uyeda, Phys. Chem. Minerals, 20, 77 (1993). 14. C. Uyeda, K. Tanaka, J. Phys. Soc. Jpn., 72, 2334 (2003). 15. C. Uyeda, K. Ohtawa, K. Okita, N. Uyeda, 7. Phys. Soc. Jpn., 70, 2334 (2001). 16. C. Uyeda, K. Tanaka, R. Takashima, Appl. Phys. Lett., 86, 094103 (2005).

7.5 Magnetic Field Effects on Photo-induced Ultrafine Particle Formation in Gas Phase

Several gaseous molecules such as carbon disulfide (CS2) and acrolein (2-propenal) (AC) can produce aerosol particles in the gas phase under ultraviolet (UV) light irradiation with a mercury lamp^^ and a N2 laser. ^ The photochemical reaction to produce aerosol particles was utilized to synthesize composite ultrafine particles from some gaseous mixtures. Considering the rapid development of nanotechnologies in various fields, synthesis of novel ultrafine and nanometer-size particles can contribute to building nanometer-size devices by assembling various kinds of nanoparticles. The photochemical method to synthesize ultrafine particles^^ utilizes photochemical reactions of AC and/or CS2. Radicals and other chemically active species generated from AC and CS2 induce chemical reactions of other gaseous components such as organosilicon compounds and organometal compounds.

The photochemical particle formation processes can be divided into


three phases, i.e., (1) nucleation process of the particle in which several molecules react chemically to form a nucleus, (2) propagation process of the nucleus by reacting chemically with surrounding molecules by collision, and (3) particle growth by coagulation and condensation with other (already grown) particles (accumulation mode). The gas-phase photochemical method has several advantages in the nucleation and propagation processes over particle formation in solution.

These advantages are as follows. (1) Under irradiation with intense laser light, gaseous molecules can be excited to a higher excited state and/or ionic state by two-photon, sometimes by multi-photon processes, followed by unique chemical reactions which differ from the chemical reactions of the first excited state usually observed in solution chemistry." ' ^ (2) Chemical reactivity of the surface of the nucleus in mesoscopic size is largely different from that of the bulk material. This may induce novel chemical reactions and fix chemical products with unique chemical structures in the particles. Furthermore, photochemical reactivity of the surface makes it easy to chemically modify the surface characteristics by exposing the particles to a particular atmosphere of organic molecules.^^ (3) In the gas-phase photochemical method, controlling the particle size can be achieved by regulating the period of photochemical reactions. Usually, particles travel by convection only once inside the irradiation vessel and are captured at the substrate accommodated at the bottom of the irradiation vessel. Hence, in order to obtain smaller particles, it is appropriate to use an irradiation vessel with a smaller diameter.^^ (4) Ultrafine particles synthesized photochemically preserve the photochemical reactivity even after sedimentation. This can be used to immobilize the particles on a substrate and to connect the particles to each other. ^

The above characteristics are favorable for assembling particles into a nanometer-size device. The magnetic field is expected to influence the nucleation and propagation processes in aerosol particle formation. In this section, experimental results on magnetic field effects on some gaseous mixtures are presented and changes in chemical compositions of sedimentary particles induced by the magnetic field are briefly discussed.

7.5.1 Glyoxal/Acrolein Mixture

In contrast to a previous finding that polymerization of AC in the vapor phase resulted in production of a white powder (called laser snow) under UV light irradiation,^^ gaseous AC produced sedimentary spherical aerosol particles of polyacrolein by a two-photon process under N2 laser light irradiation.^^ Improvement of the efficiency of aerosol particle formation has been done by involving glyoxal (GLY) molecules. GLY can initiate chemical reactions by one-photon absorption of visible light (370-460 nm). Actually, under visible light irradiation at 435.8 nm with a mercury lamp, a


gaseous mixture of GLY and AC produced white and spherical aerosol particles with a mean diameter of 0.94 ^m (in a cylindrical cell with inner diameter of 35 mm). ^ The product yield of the sedimentary particles increased almost linearly with increasing partial pressure of GLY (between 1.5 and 3.5 Torr). From analysis of FT-IR spectra of sedimentary particles, polymeric species originating in chemical reactions between AC and GLY were found to be the major components.

The nucleation process in aerosol particle formation can be studied by measuring monitor (He-Ne laser) light intensity scattered by aerosol particles. As for the gaseous mixture of GLY and AC, the measurement showed that with increasing partial pressure of GLY (from 0.3 Torr to 1.6 Torr), (1) the scattered light intensity became stronger, and (2) the induction period to detect scattered light became shorter. Generally, during the induction period, the nucleation reaction is followed by particle growth. Hence, the induction period is a good measure for determining the nucleation and propagation reaction rate in aerosol particle formation. Furthermore, scattered light intensity after reaching maximum value is proportional to the number of aerosol particles of various sizes when a specific particle size distribution is maintained during the entire period of UV light irradiation as in this case. Hence, the increase in the scattered light intensity and the shortening of the induction period in GLY/AC gaseous mixture indicated that electronically excited GLY molecules initiated chemical reactions with AC to produce aerosol particles under Ught irradiation at 435.8 nm.

For the GLY/AC gaseous mixture, magnetic field effect on the nucleation process was detected from the measurement of the scattered light intensity of monitor light. ^ The gaseous sample was placed in an electromagnet and a constant magnetic field of up to 6 kG was applied during light irradiation. In the presence of a magnetic field of 5.3 kG, the induction period to detect scattered light became shorter and the scattered light intensity became stronger until 40 min. In a separate experiment, we observed that the size distribution of the sedimentary particles did not change with the application of a magnetic field. These results indicated that the magnetic field accelerated the nucleation reaction and increased the number of aerosol particles being formed.

Furthermore, in the GLY/AC gaseous mixture, the convection of aerosol particles in the gas phase was considerably influenced by the application of a magnetic field.^^ This phenomenon is related to the change in heat release due to nonradiative processes of excited glyoxal, because the convection of gaseous molecules is induced by the released heat. We can describe chemical processes in aerosol particle formation as follows.

Photoexcited glyoxal in the singlet n-;r* state ('GLY) efficiently intersystem-crosses to the triplet manifold (" GLY), and the triplet state initiated photochemical reactions of GLY:


^^ , ISC , GLY -> 'GLY -^ 'GLY (D

As for the photochemical reactions of GLY, an intermediate complex, M, between ^GLY and GLY in the ground state was postulated.'^^ It was further supported from opto-acoustic measurement that the long-lived complex, M, relaxed only by a bimolecular encounter of the complex to produce photoproducts and heat.''^

'GLY-hGLY->M (2)

M+ M -^ Products + Heat release (3)

Following this scheme and considering the fact that AC in the ground state may form a complex with excited glyoxal, the nucleation reaction of aerosol particles from the gaseous mixture of GLY and AC may be initiated by a complex formation between ^GLY and AC in the ground state accompanying heat release:

^GLY+AC-^M' (4)

M'+ M' -^ Products + Heat release (5)

Acceleration of the nucleation reaction under a magnetic field is accompanied by larger heat release to change the spatial distribution of heat inside the irradiation cell, followed by a change in the convection of gaseous molecules.

7.5.2 Glyoxal/CS2 Mixture

A. Light Irradiation at 435.8 nm Under light irradiation at 435.8 nm, a gaseous mixture of GLY and CS2 produces aerosol particles at the early stage of light irradiation, followed by the deposition of a thin film as the final product.'^^ Under light irradiation at 435.8 nm, pure GLY vapor deposits a thin film but pure CS2 vapor does not produce any deposits, because CS2 molecules do not absorb light longer than 350 nm. Morphological change of the product under visible light irradiation, i.e., formation of aerosol particles from the gaseous mixture, is due to the incorporation of CS2 molecules into chemical reactions. Product yield of the total deposits from the gaseous mixture increases almost linearly with increasing partial pressure of GLY (above L5 Torr to 4.5 Torr), indicating that electronically excited GLY molecules initiate chemical reactions.

Magnetic field effect on the nucleation process in aerosol particle formation was detected for a gaseous mixture of GLY (3.5 Torr) and CS2 (60 Torr). Monitor (He-Ne laser) light scattered by aerosol particles as formed in the irradiation cell was detected only for the first 20-30 min as shown in Fig. 7-5-1, indicating that the aerosol particle formation is a dominant process only in the early stage of photochemical reaction, and the deposited particles contribute to the film formation process. With


0.06 f ^

3 0.04

S 0.02 ^

0.00

20 40 Irradiation time / min

60

Fig. 7-5-1 He-Ne laser light intensity scattered by the aerosol particles produced from a gaseous mixture of GLY (3.5 Torr) and CS2 (60 Torr) under light irradiation at 435.8 nm under a magnetic field of (a) 5.1 kG and (b) 0 kG. [Reproduced from H. Morita et al., Mol. Phys., 101, 2572 (2003)]

the application of a magnetic field of 5.1 kG, the induction period to detect scattered light became shorter (from 80 s to 40 s) and the scattered light intensity became stronger. As in the case of the GLY/AC mixture, magnetic field accelerated the nucleation reaction and particle growth.' ^

With the application of a magnetic field, a change in the sedimentation pattern, i.e., change in convection of aerosol particles in the gas phase, was also observed for the GLY/CS2 gaseous mixture. ^^ The sedimentation pattern observed after 40 min under light irradiation (which is straight along the incident light path without a magnetic field) curved to the right with increasing magnetic field. From the experimental results of both GLY/CS2 and GLY/AC gaseous mixtures where only GLY is excited electronically, it is concluded that chemical reactions originating from ^GLY are accelerated with the application of a magnetic field.

B. Light Irradiation at 313 nm As for the gaseous mixture of GLY and CS2, 313 nm light of mercury lamp can excite predominantly CS2 molecules, which can also initiate the nucleation reaction of aerosol particles. Actually, under light irradiation at 313 nm, a gaseous mixture of GLY (6 Torr) and CS2 (60 Torr) produced only spherical aerosol particles.'^^ Product yield increased with increasing partial pressure of CS2. As shown in Fig. 7-5-2, the monitor (He-Ne laser) light scattered by aerosol particles as formed in the cell was detected during the entire period under light irradiation, confirming that aerosol particle formation is the predominant process. In the electromagnet, a magnetic field of 5 kG was applied to the gaseous mixture. Then, we observed that the induction period to detect scattered light became longer


0.000 30 60 90 120

Irradiation time / min 150 180 210

Fig. 7-5-2 He-Ne laser light intensity scattered by the aerosol particles produced from a gaseous mixture of GLY (6 Torr) and CS: (60 Torr) under light irradiation at 313 nm under a magnetic field of (a) 0, (b) 3 and (c) 5 kG.

(from 14 s to 45 s) and the scattered light intensity became weaker. This result indicated that the magnetic field decelerated the nucleation and propagation reactions in contrast to the case where the 435.8 nm light was irradiated on the same gaseous mixture.' ^ Deceleration of the reaction is in accord with the fact that the product yield of the sedimentary aerosol particles from the gaseous mixture decreased to one third under the appUcation of a magnetic field of 5 kG.

To investigate magnetic field effect on chemical compositions, a strong magnetic field of up to 5 T was applied using a helium-free superconducting magnet during light irradiation at 313 nm. Sedimentary aerosol particles deposited on a glass plate were collected and mixed with KBr powder to prepare a KBr pellet. FT-IR spectra measured with the KBr pellets are shown in Fig. 7-5-3. With increasing magnetic field up to 5 T, the 1060 and 1508 cm~ bands due to CS2 polymerization decreased their intensities and the shoulder at --1074 cm" became prominent, showing that contribution from CS2 polymerization decreased under a high magnetic field. '

As for the photochemical reaction between GLY and CS2, we propose that excited CS2 and GLY in the ground state form a complex (M"), and M" reacts with GLY and CS2 (or CS2*) molecules to induce nucleation reactions.

CS2*+GLY^M' (6)

M"+ GLY -> Nucleation + Heat release (7)

7.5 Magnetic Field Effects on Photo-induced Ultrafme Particle Formation in Gas Phase 339

0.10

0.05

0.2

^ 0.1

(b)

4000 3500 3000 2500 2000 1500 1000 500

Wave number/cm'

Fig. 7-5-3 FT-IR spectra of the sedimentary aerosol particles produced from a gaseous mixture of GLY (4.0 Torr) and CS2 (40 Torr) under light irradiation at 313 nm in the presence of a magnetic field of (a) 0, (b) 1, (c) 3 and (d) 5 T.

M"+ CS2 (or CS2 *) ^ Nucleation + Heat release (8)

A bimolecular encounter of M" also induces the nucleation reaction accompanying heat release.

M "+ M " Nucleation + Heat release (9)

It was reported previously ' ^ that fluorescence intensity from CS2* is reduced by 40% and 50%, respectively, by the application of magnetic fields of6.1kG and 10 kG.

Depopulation of the singlet excited state of CS2 by a fast intramolecular relaxation is caused by the application of a magnetic field. This result can well explain the decrease in nucleation efficiency in aerosol particle formation through Eq. (6) under a magnetic field. Considering the change in the chemical compositions of the aerosol particles, the magnetic field effect is more pronounced in Eq. (8) than in Eq. (7).

In the GLY/CS2 gaseous mixture, chemical reaction paths in aerosol particle formation change depending on which molecule is excited under light irradiation. A magnetic field influences various reaction paths


differently, and this phenomenon can be used to control the chemical compositions of ultrafme particles by applying a high magnetic field.

7.5.3 Organosilicon Compound/CS2 Mixture

Organosilicon compounds can be incorporated into aerosol particle formation processes. Trimethyl(2-propynyloxy)silane (TMPSi) is a reactive molecule which itself can produce a deposit of polytrimethylsiloxy-substituted polyhydrocarbon under light irradiation with ArF excimer laser (193 nm),' ^ and a gaseous mixture with AC produces sedimentary aerosol particles containing elemental silicon under N2 laser light irradiation (337 nm). ^

Under light irradiation with mercury lamp at 313 nm, a gaseous mixture of TMPSi and CS2 deposited sedimentary aerosol particles of yellowish brown color with a reproducible sedimentation pattern due to convection of the gaseous mixture.'^^ Under the same experimental conditions, pure CS2 produced sedimentary aerosol particles of reddish brown color with a lower product yield. Because TMPSi does not have an absorption peak at a wavelength longer than 200 nm and hence does not

3.5

0 .5-

0.0 3500 3000 2500 2000 1500 1000 500

Wave number / cm'

Fig. 7-5-4 FT-IR spectra of the sedimentary aerosol particles produced from a gaseous mixture of TMPSi (10 Torr) and CS2 (50 Torr) under light irradiation at 313 nm in the presence of a magnetic field of (a) 0, (b) 1, (c) 3 and (d) 5 T. [Reproduced from H. Morita et al., Mol Phys., 104, 1711 (2006)]


absorb light at 313 nm, slight color change of the sedimentary particles and the increase in the product yield for the gaseous mixture strongly suggest that electronically excited CS2 molecules initiated the nucleation reaction involving TMPSi molecules. This was confirmed by measuring the monitor (He-Ne laser) light intensity scattered by aerosol particles, revealing that with increasing partial pressure of TMPSi, the induction period to detect scattered light became shorter and the scattered light intensity became stronger. Both CS2 and TMPSi molecules contribute to the nucleation and propagation processes in aerosol particle formation. From analysis of X-ray photoelectron spectra (XPS) and FT-IR spectra, the chemical composition between TMPSi and CS2 of the sedimentary aerosol particles was estimated to be 1:1-1:2 in molar ratio.

Magnetic field effect was detected on chemical compositions. From the gaseous mixture of TMPSi (10 Torr) and CS2 (50 Torr), sedimentary aerosol particles were produced in a helium-free superconducting magnet under Ught irradiation at 313 nm. FT-IR spectra of the sedimentary aerosol particles (Fig. 7-5-4) showed that with increasing magnetic field up to 5 T, the band intensities at 1252, 845, and 752 cm'^ ascribed to the trimethylsilyl group in TMPSi gradually decreased.'^^ The stoichiometry of atoms from XPS analysis^^^ also supported the lower involvement of chemical species originating from TMPSi with the application of a

158 160 162 164 166 168 170 172

Binding energy / eV

Fig. 7-5-5 XPS and fitted spectra of S 2p photoelectrons of the sedimentary aerosol particles produced from a gaseous mixture of TMPSi (10 Torr) and CS: (50 Torr) under light irradiation at 313 nm (a) in the absence and (b) in the presence of a magnetic field of 3 T. [Reproduced from H. Morita et al., Mol. Phys., 104, 1711 (2006)]


magnetic field of 3 T. The XPS spectra of S 2p photoelectrons in Fig. 7-5-5 show that sedimentary aerosol particles deposited under a magnetic field of 3 T contain more sulfur atoms in -C-S- bonding at the expense of sulfur atoms in > C = S or Cu-S bonding. These results imply that the aerosol particle formation process is not dominated by a single chemical pathway but is composed of several chemical pathways competing with each other. Initially formed transient species (CS2-TMPSi) with C = S and C-S chemical bonds can react with CS2 TMPSi and (CS)2 species, and under a magnetic field it reacts more favorably with CS2 molecules, resulting in a higher abundance of C-S bonds than C = S bonds. In nucleation and propagation processes in aerosol particle formation, a chemical pathway which is favorable under the influence of a magnetic field is selected from among several chemical pathways.

7.5.4 Organometal Compound/CS2 Mixture

The gas-phase photochemical method can be applied to synthesize metal-organic composite parficles. Under UV light irradiation at 313 nm, a volatile organometal compound, Fe(C0)5, produces crystalline-like ragged deposits of ca. 20 and 4 //m in size, but a gaseous mixture of Fe(CO)5 and CS2 produces spherical sedimentary aerosol particles (with a mean diameter of 0.6 j^mY^K The morphological change in the deposits clearly

3500 3000 2500 2000 1500 1000 500

Wave number / cm '

Fig. 7-5-6 FT-IR spectra of the sedimentary aerosol particles produced from a gaseous mixture of Fe(C0)5 (1 Torr) and CS: (20 Torr) under light irradiation at 313 nm in the presence of a magnetic field of (a) 0, (b) 3 and (c) 5 T.


shows that Fe(C0)5 reacted chemically with CS2 to produce aerosol particles. As is seen from the FT-IR spectrum of the sedimentary aerosol particles (Fig. 7-5-6), Fe(CO)5 is incorporated into the sedimentary aerosol particles with chemical structures different from those deposited from pure Fe(C0)5 vapor. The chemical composition of the particles was influenced by the application of a magnetic field (Fig. 7-5-6). By increasing magnetic field up to 5 T, band intensities in the 2000 cm" region (assignable to C = O stretching vibration attached to Fe atom) increased, indicating that the chemical reaction of Fe(C0)5 was accelerated under a magnetic field.

For the metal-organic composite particles, control of the chemical structures of metal compounds is important to give some special characteristics such as magnetic properties. Application of a high magnetic field on gaseous organometal compounds during photochemical reactions is a promising way to control chemical compositions of ultrafine and nanometer-size composite particles.

7.5.5 Conclusions

A gas-phase photochemical method to produce aerosol particles from some gaseous mixtures is proposed. Gaseous mixtures involve some organosilicon compounds and organometal compounds in addition to AC and CS2, which play an important role in initiating the nucleation reaction in aerosol particle formation. In the photochemical method, nucleation and propagation processes are influenced by the application of a magnetic field. Utilization of the magnetic field effect on chemical reactions is useful to control the chemical compositions and the mean diameters of ultrafine and nanometer-size particles. The magnetic field effect will find many applications in fabricating nanometer-size devices composed of nanoparticles.

References

1. K. Ernst, J. J. Hoffman, Chem. Phys. Lett., 68, 40 (1979); K. Ernst, J. J. Hoffman, Chem. Phys. Lett., 75, 388 (1980); Y. P. Vlahoyannis, E. Patsilinakou, C. Fotakis, J. A. D. Stockdale, Radiat. Phys. Chem., 36, 523 (1990).

2. H. Morita, M. Shimizu, J. Phys. Chem., 99, 7621 (1995). 3. H. Morita, J. Photopolym. Sci. TechnoL, 12, 95 (1999). 4. H. Morita, K. Semba, Z. Bastl, J. Pola, J. Photochem. Photobiol. A: Chem., 116, 91

(1998). 5. K. Semba, H. Morita, J. Photochem. Photobiol. A.Chem., 134, 97 (2000); K. Semba,

H. Morita, J. Photochem. Photobiol. A: Chem., 146, 141 (2002); K. Semba, H. Morita, J. Photochem. Photobiol. A: Chem., 150, 7 (2002).

6. H. Morita, H. Tonooka, J. Photopolym. Sci. TechnoL, 14, 203 (2001). 7. H. Morita, R. Nozawa, Z. Bastl, submitted to Mol. Phys., 104, 1711 (2006). 8. F. E. Blacet, G. H. Fielding, J. G. Roof, J. Am. Chem. Soc, 59, 2375 (1937); H. W.

Melville, Proc. Royal Soc, A163, 511 (1937); A. Tam, G. Moe, W. Happer, Phys. Rev. Lett., 35, 1630(1975).

9. H. Morita, H. Ohmuro, Mol. Phys., 100, 1365 (2002). 10. J. T. Yardley, J. Chem. Phys., 56, 6192 (1972); J. T. Yardley, J. Am. Chem. Soc, 94,


7283(1972). 11. M. B. Robin, N. A. Kuebler, K. Kaya, G. J. Diebold, Chem. Phys. Lett.. 70, 93 (1980). 12. H. Morita, F. Matsubayashi, A. Nozue, RIKEN Rev., 44, 27 (2002); H. Morita, F.

Matsubayashi, A. Nozue, Mol. Phys., 101, 2569 (2003). 13. H. Morita, S. Kanaya, Proceedings of the 6th Meeting on New Magneto-science, Nov.

6-8, 2002, Tsukuba, Japan, p. 227 (2002). 14. H. Orita, H. Morita, S. Nagakura, Chem. Phys. Lett., 81, 29 (1981). 15. J. Pola, H. Morita, Tetrahedron Utt., 38, 7809 (1997). 16. H. Morita, R. Nozawa, Proceedings of the 6th Meeting on New Magneto-science,

Nov. 6-8, 2002, Tsukuba, Japan, p. 18 (2002). 17. H. Morita, R. Nozawa, Z. Bastl, Proceedings of the 7th Meeting on New Magneto-

science, Nov. 5-7, 2003, Tsukuba, Japan, p. 32 (2003). 18. H. Morita, H. Okamura, H. Ishikawa, Proceedings of the 7th Meeting on New

Magneto-science, Nov. 5-7, 2003, Tsukuba, Japan, p. 263 (2003).

Documents

[Springer Series in Materials Science] Magneto-Science Volume 89 || Novel Magnetic Field Effects