Ultra Trace Determination Scheme for 26Al

by High-Resolution Resonance Ionization Mass Spectrometry

using a Pulsed Ti:Sapphire Laser

Hideki TOMITA1;2;�, Christoph MATTOLAT2, Thomas KESSLER2;3,Sebastian RAEDER2, Fabio SCHWELLNUS2, Klaus D. A. WENDT2,

Kenichi WATANABE1 and Tetsuo IGUICHI1

1Department of Quantum Engineering, Graduate School of Engineering, Nagoya University, Nagoya, 464-8603, Japan2Institute of Physics, Johannes Gutenberg-Universitat Mainz, 55099, Mainz, Germany

3Department of Physics, University of Jyvaskyla, 40014, Finland

(Received September 19, 2007 and accepted December 13, 2007)

We propose an ultra trace analysis approach for 26Al by high-resolution Resonance Ionization MassSpectrometry (RIMS) using a pulsed narrow band-width Ti:Sapphire laser. For ensuring efficient ioniza-tion and high isotopic selectivity in RIMS of Al, we developed an injection seeded pulsed Ti:Sapphirelaser with high repetition rate operation at up to 10 kHz. The laser produced an output power of 2Wand a spectral band-width of �20MHz with a repetition rate of 7 kHz. A first demonstration of its per-formance was done by detecting stable 27Al using RIMS.

KEYWORDS: resonance ionization mass spectrometry, long-lived radioisotope, trace analysis, Al-26, Ti:Sapphire laser, injection seeding

I. Introduction

Trace analysis of long-lived radioisotopes is necessary fora wide range of studies. For example, the long-lived radioi-sotope 26Al (half-life of 7� 105 years) has been used in geo-physics,1) biochemistry,2) nuclear astrophysics3) and nuclearengineering.4) In nature, 26Al is produced by cosmic ray in-duced spallation reactions on 40Ar in the earth’s atmosphereand from 28Si in minerals. The natural isotopic abundance of26Al rarely exceeds 10�14 of natural aluminum (27Al).5) Infusion reactors, 26Al is produced through the reaction27Al(n,2n)26Al induced by fast neutrons. It is suggested, thatthe product yield of 26Al can be used for diagnostics of fu-sion reactor plasma and its dosimetry.6–8) The productionrate of 26Al by an intense irradiation of neutrons at typically14MeV at ITER is high and thus, 26Al/27Al ratios in alumi-num alloy and/or ceramic might exceeded 10�9,9) i.e. up to1 ng of 26Al would be detectable in 1 g of aluminum sample.On the other hand, 26Al produced at high-energy acceleratorswas used as a tracer for research on aluminum bio-kinet-ics.10) Within aluminum absorption experiments in hu-mans,11) several tens of ng of 26Al were administered. Fromthe test persons about 1 ng of 26Al could be re-extracted fromblood and urine after bio-kinetical exchange. For 1mg of27Al added to the sample as carrier, a 26Al/27Al ratio of typ-ically 10�6 to 10�9 must be detected.

For determination of those trace amounts of 26Al in sam-

ples, detection of radiation caused by decay of 26Al is inap-propriate, because the specific decay rate of 26Al is too low.The detection limit of 26Al by coincident 511 keV annihila-tion quanta counting is 0.1 ng (0.1 Bq).12) Therefore, acceler-ator mass spectrometry (AMS) is widely used for lowest lev-el 26Al determination.13) The detection limit of 26Al in AMSis 10�17 g which is corresponding to a limiting 26Al/27Al ra-tio of 10�14 and which also permits the study of the isotopicabundance in nature. Although AMS has ultra high sensitiv-ity, high experimental effort limits wider application of 26Al.

As alternative, Resonance Ionization Mass Spectrometry(RIMS) is well-suited for determination in the above pico-gram order of long-lived radioisotopes, combining high toultra high isotopic selectivity with high efficiency. Isotopicselectivity is achieved by using a combination of multi-steplaser excitation/ionization and mass spectrometry. In orderto distinguish the different isotopes through the precise posi-tions of their individual optical resonance lines, the spectralband-width of the lasers for excitation/ionization must be aslow as possible, preferably not exceeding a few tens of MHz.Due to the availability of suitable continuous wave narrowband-width diode lasers, high-resolution RIMS based on di-ode lasers was developed for some long-lived radioisotopes,such as 41Ca, 236U and few others.14) However, high-resolu-tion RIMS based on diode lasers is not applicable for quite anumber of other elements, including Al, because suitable la-ser diodes for the necessary atomic ground state excitations,often lying in the far blue to UV regions, are not available.

Recently, a pulsed Ti:Sapphire laser (Ti:Sa) operated withrepetition rate of up to 10 kHz was developed for RIMS.15)

�Atomic Energy Society of Japan

�Corresponding author, E-mail: tomita@nagoya-u.jp

Journal of NUCLEAR SCIENCE and TECHNOLOGY, Supplement 6, p. 37–42 (September 2008)

37

ARTICLE

This powerful laser provides tunability from 670 nm to 950nm and frequency doubling, tripling or quadrupling can beeasily accomplished. In particular the high repetition rate op-eration is suitable for efficient excitation and/or ionization inRIMS. These Ti:Sa lasers are widely used as a laser sourcefor RIMS at Mainz University and elsewhere.3,16,17) Figure 1shows possible optical two- and three-step excitation andionization schemes for Al optimized for Ti:Sa laser use.Concerning the chosen first excitation step at 308 nm, T.G. Cooper et al. reported an isotope shift in the 308.3 nm2P1=2–

2D3=2 transition of ��26;27 ¼ þ616ð3ÞMHz18) betweenthe centers of gravity. Due to the non-vanishing nuclearspins of both isotopes 26Al with I ¼ 5 and 27Al withI ¼ 5=2, both spectra exhibit a hyperfine structure and theminimum distance between interfering resonance peaks of26Al and 27Al amounts to about �380MHz. Corresponding-ly, laser radiation with a narrow spectral band-width of lessthan 50MHz is required in at least one step of the excitationand ionization scheme to resolve both isotopes and signifi-cantly suppress contributions from the neighboring one inthe photo ion count rate. With about 3 to 5GHz band-width,the standard version pulsed Ti:Sa laser is by far too broad forultra trace determination of 26Al. As a solution, a narrowband-width pulsed Ti:Sa laser for high-resolution RIMSwas envisaged, which should generally enable and simplifyisotope selective trace analysis of numerous long-lived radi-oisotopes using high-resolution RIMS. One of the most com-mon techniques to achieve pulsed laser operation with re-duced spectral band-width as low as few MHz is injectionseeding.19) Low repetition rate and cw injection seeded Ti:Salasers were reported previously.20,21) In this work we presentthe development of an injection seeded Ti:Sa pulsed laserwith repetition rate of up to 10 kHz. For characterizationof its performance first RIMS investigations on stable 27Alusing the injection seeded Ti:Sa laser are discussed.

II. Principle of Injection Seeding

For injection seeding, two lasers are mandatory: One istypically a continuous-wave, narrow band-width and lowpower laser called the master laser. The other is a high powerlaser operating in free run in a broad band-width region,called the slave laser. In our case, the master laser is an ex-tended cavity diode laser and the slave laser is a Ti:Sa laser.When the master ‘‘seed’’ laser injects into a slave laser andgood spatial mode matching is ensured, one particular longi-tudinal mode within the slave laser resonator is pre-populat-ed with photons from the master laser. Therefore, injectionseeded lasing of the slave particular into this mode is highlyfavored. As a result, narrow band-width, single longitudinalmode and high power lasing is achieved. Here, the band-width of the seeded laser is dominantly governed by theFourier transform limit due to its pulse structure. Assumingthat the laser pulse has a temporal Gaussian shape, its Four-ier transform, i.e. its frequency spectrum is also Gaussian.The band-width (� fFWHM) is related to the pulse width(��FWHM) as:

� fFWHM ���FWHM ¼ 2 � ln 2=� � 0:44:

For example, a temporal pulse width of 80 ns corresponds toa Fourier transform limited band-width of �5:5MHz.

III. Experimental Set-up for the Injection SeededTi:Sapphire Laser

Figure 2 shows the experimental set-up for the injectionseeded Ti:Sa laser. A high power Nd:YAG laser (ClarkIndustries, Orc-1000) was used for pumping the Ti:Sa crystalat 532 nm with a repetition rate of 7 kHz. A Lyot filter and anetalon are used as standard frequency selective elementswithin the Ti:Sa laser resonator, while an internal Pockels

Fig. 1 Possible optical two- and three-step excitation and ionization schemes for Al optimized for Ti:Sapphire laser use

38

JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY

cell could be used to control lasing timing. A cw extendedcavity diode laser with a band-width of �5MHz wascoupled into the z-shape Ti:Sa laser resonator through theoutput coupler by using a polarizing cube, a Faraday rotatorand a half wave plate. To prevent damage of the laser diodefrom Ti:Sa laser operation two optical isolators with 60 dBsuppression each were used. The diode laser intensitytransmitted through the z-shape cavity was detected by aphoto diode.

The length of the z-shape resonator cavity was stabilizedby a dither locking technique to the wavelength of the seedlaser. For this purpose the diode laser wavelength was dith-ered across a resonance with laser current modulation. Thepeak width of the z-shape cavity longitudinal mode wasabout 30MHz. Thus, a diode laser frequency dither ampli-tude just below 30MHz was chosen.

The acute Ti:Sa laser pulse, being about 6 orders ofmagnitude stronger than the diode laser intensity, makes it

extremely difficult to keep the cavity locking electronics op-erating properly. Optical selection via filters or polarizers isexcluded, because the polarization, direction and wavelengthof the seeded Ti:Sa laser are identical to those of the masterlaser. Furthermore, it was similarly impossible, to separatethe Ti:Sa laser from the diode laser signal using only passiveelectronic filtering within the Ti:Sa pulse pauses of 140 msduration. Hence, the Ti:Sa laser pulse was actively sup-pressed during the 100 ns instant of Ti:Sa lasing from theoutput line of the photo diode by using an fast electricalswitch. These two switches were operated by TTL signalssynchronized with the instant of the Ti:Sa lasing. Figure 3shows the schematic diagram of the electrical circuitry ofswitch and amplifier module. To detect up to a few tens ofmW from the diode laser, the transimpedance gain of thecurrent-to-voltage converter was set to 106 V/A. In addition,the response time of the electrical switch module was put aslow as a few ms to rapidly remove the large charge generated

Fig. 2 Experimental set-up for the injection seeded Ti:Sapphire laser. FPI = Fabry-Perot interferometer, Ring PZT =ring piezo transducer

Fig. 3 Schematic diagram of the electrical circuitry of switch and amplifier module. OP = operational amplifier

39

SUPPLEMENT 6, SEPTEMBER 2008

by the intense Ti:Sa laser pulse in the photo diode. Afteractive filtering, the photo diode signal was fed into a lock-in amplifier with a reference signal of 60 kHz. The outputof the lock-in amplifier serves as an input error signal tothe PID circuit. The output of the PID controller was ampli-fied and used for resonator cavity adjustment via a ring piezotransducer (Piezomechanik, HPSt 150/14-10/12 VS22).This was mounted behind the high reflective end mirrorand kept the cavity length on resonance according to thefeedback signal.

IV. Characteristics of the Injection Seeded Ti:Sap-phire Laser

The wavelength of the Ti:Sa laser was tuned to the925 nm. Typical free-running (unseeded) and seededTi:Sa laser power was about 2.1W and 1.8W, respectively.Figure 4 shows the temporal profiles of the pump laser, thefree-running and the seeded Ti:Sa laser. As expected, theseeded Ti:Sa laser started lasing significantly earlier thanthe free-running Ti:Sa laser. The seeded Ti:Sa lasing timingwas controllable with the Pockels cell in a similar way as thefree-running Ti:Sa laser.22) This is a very important prereq-uisite for the synchronization of all individual laser pulsesneeded for excitation and ionization in multi-step RIMS.The intensity of the minor sub-pulse, that occurred afterthe main pulse, and which is visible as correspondingpedestal in Fig. 4 decreased with increasing delay of thelasing timing.

Figure 5 shows the spectra of the diode laser, the free-running and the seeded Ti:Sa laser, taken by a Fabry-Perotinterferometer (FPI) with a FSR of 2.0GHz. The seededTi:Sa laser in principle shows a very narrow spectrum rathersimilar to the cw diode laser, but exhibits contributions fromside-modes. These side peaks are located about �300MHzapart from the main peak and their intensity amounts toabout 20% of the one of the central peak. These peaks arecaused by spatial hole-burning of the standing wave in thez-shape cavity; they correspond to ðn� 1=2Þ � � when n � �is the seeded mode. The free running laser does not giveany useful spectrum, as its band-width exceeds the FSR ofthe etalon by far. Figure 6 shows the spectrum of the seededTi:Sa laser taken by a FPI with FSR of 300MHz. From thisdiagram the spectral band-width of the seeded Ti:Sa laser

could be estimated to be �20MHz, which is close to theFourier limit and finally given by the diode laser modulationamplitude of the dither locking.

V. RIMS on Stable 27Al with the Injection SeededTi:Sapphire Laser

For first demonstration of the laser, we demonstratedRIMS on stable 27Al using the injection seeded Ti:Sa laser.Figure 7 shows the experimental RIMS set-up used: Alumi-num nitride deposited on titanium foil in a graphite tubeserved as an aluminum atom source and provided a collimat-ed atomic beam. The sample was atomized by ohmic heatingof this graphite oven. In the measurement, a two-step UV-UV excitation and ionization scheme was used as shownin Fig. 1. One Ti:Sa laser was seeded at 924.90 nm. The fre-quency tripled Ti:Sa laser light was injected into the atomicbeam oven. Sample atoms were excited and ionized in theinteraction region. Laser ions were accelerated with highvoltage and selected with a mass separator. Laser ions weredetected in a Faraday cup. The wavelength of the seededTi:Sa laser was scanned through a variation of the diodelaser wavelength. During injection seeded Ti:Sa laser scan-ning, no etalon and Pockels cell were installed within the

Fig. 4 Temporal profiles of the pump laser, the free-running andthe seeded Ti:Sapphire laser

Fig. 5 Spectra of the diode laser, the free-running and the seededTi:Sapphire laser, taken by a FPI with a FSR of 2.0GHz

Fig. 6 Spectrum of the seeded Ti:Sapphire laser taken by a FPIwith FSR of 300MHz

40

JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY

z-shape cavity. Frequency tripled laser power of about13mW at 308.30 nm (972.39 THz) was used.

Figure 8 shows the detected 27Al ions as a function of thescanning of the seeded Ti:Sa laser. The line-width of 27Alions signal is about 7.0GHz and thus by far broader thanthe expected line-width. Obviously it is by far dominatedby longitudinal Doppler broadening within the effusingatomic beam. Doppler broadening of Al atomic vapor isestimated as follows: The Gaussian line shape function in-cluding Doppler shift is described as

Ið�Þ ¼ I0 exp �m

2kBT

�� �0

�0

� �2

c2

!;

where kB is the Boltzmann constant, T is the temperature ofatom sample, m is the atom mass, � is the frequency, �0 isthe line center frequency and c is the light velocity. Thewidth (��FWHM) of Doppler broadening is

ð��FWHMÞ ¼ 2ffiffiffiffiffiffiffiffiffiffiffi2 ln 2

p�

ffiffiffiffiffiffiffiffiffiffiffiffiffikBT�20mc2

s;

¼ 7:14� 10�7 � �0 �ffiffiffiffiffiT

M

r;

where M is the atomic mass number. When �0 ¼ 972:39THz, T ¼ 2790K (boiling point of AlN) and M ¼ 27, thewidth of Doppler broadening is estimated to be 7.1 GHz,well explaining the measured width. From this experimentwe concluded, that the laser beam must be overlapped ina perpendicular direction with the atomic beam to permitextraction of the effect of narrow band-width of the injec-tion seeded Ti:Sa laser. Corresponding adaptations of theRIMS system are presently under way, but will as a draw-back reduce ionization efficiency through the diminishedspatial overlap.

VI. Summary and Outlook

We have proposed and tested an ultra trace determinationscheme for 26Al by high-resolution resonance ionizationmass spectrometry using the frequency tripled light of apulsed Ti:Sa laser. To suppress contributions from the neigh-boring stable 27Al, laser radiation must exhibit a narrowspectral band-width of less than 50MHz for at least one stepin the excitation and ionization schemes. In addition, thelaser system must provide a high repetition rate of severalkHz for efficient resonant excitation and ionization.

Thus, we developed an injection seeding for a pulsed pow-erful Ti:Sa laser, which operates in a high repetition ratescheme. The injection seeded Ti:Sa laser produces an outputpower of �2W and a spectral band-width near �20MHzwas accomplished. Additionally, first results of RIMS on sta-ble 27Al with the injection seeded Ti:Sa laser were obtained,unfortunately not yet exhibiting the advantages of the narrowband-width due to unfortunate experimental conditions.These will be optimized soon.

The characterization of the laser radiation exhibits, thatthe spectral profile of the seeded Ti:Sa laser, even beingnarrower than 50MHz, exhibits weak side modes located�300MHz from the central mode. Thus high optical selec-tivity is not yet achievable. To remove these side modes,caused by spatial hole-burning of the standing wave in thez-shape cavity, we presently develop a ring cavity resonator

Fig. 7 Experimental RIMS set-up on stable 27Al with the injection seeded Ti:Sapphire laser

Fig. 8 Detected 27Al ions as a function of the scanning of theseeded Ti:Sapphire laser

41

SUPPLEMENT 6, SEPTEMBER 2008

for the injection seeded Ti:Sa laser.As a further step, high resolution spectroscopy and analy-

sis of 27Al with the injection seeded Ti:Sa laser is foreseenand afterwards RIMS analysis of 26Al samples of knownconcentrations will be addressed. In addition, the narrowband-width injection-seeded laser developed here shall beused for spectroscopic investigations of Rydberg and auto-ionizing states of miscellaneous other elements of interestby RIMS in the future.

Acknowledgments

This research was supported by the Ministry of Education,Culture, Sports, Science and Technology (MEXT), aGrant-in-Aid for the 21st Century COE Program, ‘‘Isotopesfor the Prosperous Future.’’ The authors are deeply gratefulto M. Hori at CERN, Ch. Geppert at Johannes Gutenberg-Universitat Mainz, and G. Ewald at GSI for the helpfuldiscussions and advices in this work. The authors thankN. Trautmann at Johannes Gutenberg-Universitat Mainzfor the chemical sample preparation for this work. Theauthors are especially grateful to H. Lenk at JohannesGutenberg-Universitat Mainz for the discussions and advicesconcerning the electronic circuitry developed within thiswork.

References

1) K. J. Kim, P. A. J. Englert, ‘‘In situ cosmogenic nuclideproduction of 10Be and 26Al in marine terraces, Fiordland,New Zealand,’’ Nucl. Instr. Meth. Phys. Res., B 223–224,639–644 (2004).

2) C. Steinhausen et al., ‘‘Medical application of 26Al,’’ Nucl.Instr. Meth. Phys. Res., B113, 479–483 (1996).

3) T. Achtzhen et al., ‘‘26Al beam production by a solid state laserion source at TRIUMF,’’ Hyperfine Interact, 174, 27–32(2007).

4) A. Wallner et al., ‘‘Precise measurement of the 27Al(n,2n)26gAlexcitation function near threshold and its relevance for fusion-plasma technology,’’ Eur. Phys. J. A, 17, 285–296 (2003).

5) R. Middleton et al., ‘‘26Al: Measurement and Applications[and Discussion],’’ Philosophical Transactions of the RoyalSociety of London. Series A, Mathematical and PhysicalSciences, 323, 121–143 (1987).

6) R. K. Smither, L. R. Greenwood, ‘‘Measurement of the

27Al(n,2n)26Al reaction cross section for fusion reactor appli-cations,’’ J. Nucl. Mater., 122–123, 1071–1077 (1984).

7) R. K. Smither, L. R. Greenwood, ‘‘New method for measuringion temperatures in hot D-T plasma,’’ Rev. Sci. Instrum., 56,1078–1080 (1985).

8) L. R. Greenwood, ‘‘New ideas in dosimetry and damagecalculations for fusion materials irradiations,’’ J. Nucl. Mater.,141–143, 654–657 (1986).

9) K. Asai, private communication, Nagoya University, (2006).10) S. N. Dmitriev, N. G. Zaitseva, ‘‘Highly-pure radioisotopes for

life science researches produced at the FLNR JINR accelera-tors,’’ Radiochim. Acta, 93, 571–573 (2005).

11) C. Steinhausen et al., ‘‘Investigation of the aluminiumbiokinetics in humans: a 26Al tracer study,’’ Food Chem. Tox-icol., 42, 363–371 (2004).

12) N. D. Priest, ‘‘The biological behaviour and bioavailability ofaluminium in man, with special reference to studies employingaluminium-26 as a tracer: review and study update,’’ J. Envi-ron. Monit., 6, 375–403 (2004).

13) O. Meirav, R. A. L. Sutton, D. Fink, ‘‘Application of AMS inAl metabolism studies,’’ Nucl. Instr. and Meth., B 52, 536–539(1990).

14) N. Trautmann, G. Passler, K. D. A. Wendt, ‘‘Ultratrace analy-sis and isotope ratio measurements of long-lived radioisotopesby resonance ionization mass spectrometry (RIMS),’’ Anal.Bioanal. Chemi., 378, 348–355 (2004).

15) C. Rauth et al., ‘‘First laser ions at an off-line mass separator ofthe ISAC facility at TRIUMF,’’ Nucl. Instr. Meth. Phys. Res.,B215, 268–277 (2004).

16) Y. Liu et al., ‘‘Laser ion source tests at the HRIBF on stableSn, Ge and Ni isotopes,’’ Nucl. Instr. Meth. Phys. Res.,B243, 442–452 (2006).

17) T. Kessler et al., ‘‘Upgrade to the IGISOL laser ion sourcetowards spectroscopy on Tc,’’ Hyperfine Interact, 171, 121–126 (2006).

18) T. G. Cooper et al., ‘‘The nuclear magnetic moment of 26Al byatomic beam laser spectroscopy,’’ J. Phys. G, 22, 99–106(1996).

19) N. P. Barnes, J. C. Barnes, ‘‘Injection seeding I: Theory,’’IEEE J. of Quant. Elec., 29, 2670–2683 (1993).

20) T. D. Raymond, A. V. Smith, ‘‘Injection-seeded titanium-doped-sapphire laser,’’ Opt. Lett., 16, 33–35 (1991).

21) E. A. Cummings, M. S. Hicken, S. D. Bergeson, ‘‘Demonstra-tion of a 1-W injection-locked continuous-wave titanium:sap-phire laser,’’ Appl. Opt., 41, 7583–7587 (2002).

22) J. Yi et al., ‘‘Temporal control of pulses from a high-repeti-tion-rate tunable Ti:Sapphire laser by active Q-switching,’’Jpn. J. Appl. Phys., 42, 5066–5070 (2003).

42

JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY