THE CLASSICAL MICROWAVE FREQUENCY STANDARDS

Giovanni Busca, P i e r r e Thomann, Laurent-Guy B e r n i e r ,

Philippe Willemin and Hartmut Schweda
Observatoire de Neuchstel

Rue de I'Observatoire 58
CH-2000 NEUCHATEL/Switzerland

1.

Introduction we present some key problems encountered in t h e classical

In t h i s paper

microwave frequency standards which a r e still not solved today. The point of view expressed here benefits from the experience gained both in the industry and in the research lab, on the following classical microwave frequency standards: active and passive H, conventional and laser pumped Cs beam tube, small conventional and laser pumped Rubidium. The accent is put on t h e Rubidium standard, t h e other topics being covered in the following papers. Table 1 presents a simple comparison among t h e microwave standards. 2. Conventional Rubidium Standards However

Models f o r predicting signal, noise and linewidth a r e available Ill.

only crude calculations exist concerning the output frequency 121. The basis of t h i s calculation is a linearization of the equation of t h e isotopic f i l t e r induced optical displacement Av equation f o r light absorption. The light-shift Av
0P

given in REF 131 and a linearization of t h e

LS

experienced by the f i r s t layer of atoms in t h e absorption

cell can be expressed with small changes from REF 121, a s

Report Documentation Page

Form Approved OMB No. 0704-0188

Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.

1. REPORT DATE

3. DATES COVERED 2. REPORT TYPE

DEC 1990
4. TITLE AND SUBTITLE

00-00-1990 to 00-00-1990
5a. CONTRACT NUMBER 5b. GRANT NUMBER 5c. PROGRAM ELEMENT NUMBER

The Classical Microwave Frequency Standards

6. AUTHOR(S)

5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

Observatoire de Neuchatel,Rue de IObservatoire 58,CH-2000 Neuchatel//Switzerland,

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

8. PERFORMING ORGANIZATION REPORT NUMBER

10. SPONSOR/MONITORS ACRONYM(S) 11. SPONSOR/MONITORS REPORT NUMBER(S)

12. DISTRIBUTION/AVAILABILITY STATEMENT

Approved for public release; distribution unlimited

13. SUPPLEMENTARY NOTES

See also ADA239372. 22nd Annual Precise Time and Time Interval (PTTI) Applications and Planning Meeting, Vienna, VA, 4-6 Dec 1990
14. ABSTRACT

15. SUBJECT TERMS 16. SECURITY CLASSIFICATION OF:

a. REPORT b. ABSTRACT c. THIS PAGE

17. LIMITATION OF ABSTRACT

18. NUMBER OF PAGES

19a. NAME OF RESPONSIBLE PERSON

unclassified

unclassified

unclassified

Same as Report (SAR)

10

Standard Form 298 (Rev. 8-98)

Prescribed by ANSI Std Z39-18

where 9 i s t h e f i l t e r temperature

is the light intensity impinging on the f i l t e r

0 0

I [l-K (0-9 11
0

the

light

intensity

at

the

output

of

the

filter

and

entering t h e Rubidium cell,

K
1

a light shift parameter depending upon the spectral shape of t h e

t h e f i l t e r temperature f o r which the light s h i f t goes t o zero.

pumping line and

B
0

With t h e same meaning of the parameters, considering a mixture of isotopes in the lamp and, f o r t h e integrated f i l t e r approach, a mixture of isotopes in t h e cell, we have t h e general equation giving the frequency vs light intensity and f i l t e r cell temperature:

where:
v
0

i s t h e unperturbed frequency (independent f r o m t h e temperature and light intensity 1

the index - r e f e r s t o ~b~~ isotope the index + r e f e r s t o Rb and

85

isotope

9 i s t h e filter-cell temperature

rn

x m a r e respectively t h e linear and quadratic temperature dependence

given by t h e buffer g a s mixture [21.

Eq.
( 2 ) explicits

the

fact

that

the

~b~~ gives,

for

normally

used

cell

temperatures > 65"C,

a negative light shift; t h e ~b~~ gives a positive light

s h i f t which never goes t o zero.

The parabolic behaviour given in eq. ( 2 ) has a minimum f o r a temperature

3 =

emin given by

where

p. represents the r a t i o (absolute value) of the negative t o t h e positive light

1

shift,

the

index i meaning t h a t t h e r a t i o i s taken

at the input of

the

f ilter-cell. Eq. ( 3 ) shows t h a t depends upon the t o t a l light intensity I emitted by min o t h e lamp if 6 and ;y a r e different from zero. m m
3

The eq. ( 2 ) can be expressed in t e r m s of a measurable output parameter, i.e. light intensity I transmitted by t h e cell.

where

The meaning of index o i s t h a t we a r e referring t o output parameters.

Eq. (5) shows t h a t t h e frequency is a linear function of t h e t o t a l transmitted

light in agreement with experiments 121. From eq. (5) by differentiating with respect t o t h e transmitted light intensity I one obtains t h a t &/a1
= 0 f o r 9 = 8LS=O given by:

p a h a s t h e same meaning a s p. but r e f e r r e d now t o the transmitted t o t a l light 1 intensity.

For obtaining good standard performances

t h i s relation can be realised by choosing the isotopic r a t i o in the lamp and

in t h e cell and in addition t h e buffer g a s mixture.

For one commercial Rubidium this occurs a t a filter-cell temperature of 75C
121.

The model shows t h a t the main limitation of the conventional Rubidium i s t h a t temperature coefficient and light s h i f t compensation occurs only f o r a well defined light intensity and well defined cell-filter temperature.

All t h e parameters occuring in t h e previous equations a r e important f o r a Rubidium d r i f t model.

3.

Laser Pumped Rubidium

From experiments in our laboratory, and using a laser power corresponding t o t h e "saturated" light s h i f t 141 the following d a t a a r e obtained:

a) The

light

shift
-G

vs

the

laser frequency i s SX~O-''/MHZ

-10 -1/2 5x10 t

or

Rubidium

stability of 10

-11 -1/2

requires a laser stability of

b) The light-shift vs t h e laser power is:

- 3~10-~'/%.
cell has a cell temperature induced

c ) The

laser

locked

t o the

Rubidium

frequency s h i f t of 9 MHz/"C [41.

d ) From c ) and a) t h e laser

servo produce

an additional

cell temperature

coefficient of -4. S X ~ O - ~ ~ / " C . frequency range (100 Hz - 1 kHz), t h e laser while shot noise i s <

e ) In t h e

relevant Fourier

intensity noise i s > l' o/ --

f ) Aging of the laser parameters in single mode operation, locked t o t h e Rb cell, i s commonly experienced. In view of points a) t o f ) i t appears not a simple work t o reach t h e predicted performance [51 o- (TI = 3x10 Y
-14 -112

-c

However when all t h e previous problem will be solved we will certainly see t h e realization of t h e promise of the laser pumping: a very small, low power and high performance Rubidium standard.

4.

Cs standards

As a comment t o t h e Table 1 we like t o remark t h a t t h e Cs standard has the lowest intrinsic drift temperature mechanism in coefficient, to no the significant other drift
In

and

no

significant recent

compared

standards.

addition, still One

advances

microprocessor of

controlled

electronics

161 have

produced

remarkable

improvements

t h e environmental

characteristics.
5

problem t o be solved is t h e aging of t h e signal due t o the degradation of the electron multiplier. This problem exists f o r high gain (figure (1 dB) multipliers. 10 ) and low noise

For low gain and high noise figure multipliers

t h e problems seem solved [ 7 ] . However in this case much higher atomic flux i s required f o r obtaining the same short t e r m stability, i s shortened. i.e. t h e tube lifetime

The small optically pumped

stability i s < 1 0
-11
-G

Cs

is appealing. is:

The

demonstrated

short

term

-1/2

[81.

The potential

small size and light

weight.

The problems t o be solved: t h e laser frequency noise and laser aging.

Relatively t o t h e H-Maser

we believe t h a t the problem of the frequency d r i f t context, the cavity pulling has drawn

is

still

open.

In

this

lot

of

attention. Today, t o our knowledge, tuning a r e used for maser in

4 different methods of The

automatic cavity 191, t h e magnetic

the field:

SE tuning

relaxation tuning

[lo], the cavity frequency switching [I11 and the auxiliary

mode stabilization [12]. Many more have been proposed and tested [ 1 As a 2. conclusion we consider t h a t the long t e r m maser frequency d r i f t associated with t h e cavity pulling i s a well known subject due t o the high level of

precision obtainable in principle by these methods. On the contrary wall s h i f t d r i f t i s still very poorly known [131.

From our experience i t appears illusory t o derive conclusions on wall s h i f t d r i f t vs time

a ) f o r masers which a r e not in continuous operation, b) masers

which show a signal amplitude decay which is normally correlated with t h e line
Q decay, c) masers which s u f f e r from magnetic relaxation requiring neck coil

current trimming, in order t o g e t the full power operation at a low C field of

< 1 mC. Maser

amplitude, line
Q

and

magnetic

relaxation

should

be

monitored

in

addition t o t h e cavity frequency f o r correct interpretation of d r i f t data.

t h e frequency

As

a conclusion wall

shift

and the

associated

wall

relaxation

is

still

the

main problem t o be solved.

Acknowledgments We thank t h e Swiss Commission CERS, t h e European Space Agency and t h e Swiss Federal Office of Metrology f o r supporting our work on t h e atomic frequency standards. We acknowledge also
the

cooperation

of

Oscilloquartz

S.A.

for

discussing d a t a on Cs beam tubes.

TABLE 1. CLASSICAL p-WAVE STANDARDS: WHERE DO WE STAND?

Rubidium Lamp Laser

Short-term stability

Cesium (small) magnetic optical

H active

Temp. coeff.
Light-shift coeff.

10-12/K th. gain: 100

~ 3 .0-'4/~ 1 th. gain: 102-lo4

Magnetic coeff. Long-term 10-I1/month stability Life time

?
.-

>5 years

3 4 10 years

References J.C. Camparo: Proc. of t h e 4th Symposium on Frequency Standards and

Metrology, Ancona 1988, p. 62. J. Vanier and C. Audoin: The Quantum Physics

of

Atomic Frequency

Standards (A. Hilger, 1989).

G. Busca, M, TCtu, J. Vanier: Can. J. Phys.

J.T. Liu e t al.:

3,(19731,

1379.

Proceedings of t h e 4 t h European Frequency and Time

Forum, Neuchstel, 13-15 March 1990, p. 157. J.C. Camparo, R.P. Frueholz: J. Appl. Physics 59, (19861, 3313.

J. Rabian, P. Rochat: Proceedings of the 2nd European Frequency and Time

Forum, Neuchstel, 16-18 March 1988, p. 461.

R.P. Frueholz e t al.: Proceedings of t h e 18th Annual Precise Time and

Time Interval Applications and Planning Meeting, (19861, p. 11.

V. Candelier e t al.: Proceedings of the 2nd European Frequency and Time

Forum, Neuchgtel, 16-18 March 1988, p. 483.

L.J. Rueger, J. Instn. Electronics and Telecom. Engrs 27 (1981) 493.

N. Demidov, A. Uljanov: Design and Industrial Production of Frequency Standards in USSR, these Proceedings, paper no. 13.
H.E.

P e t e r s et al.:

Hydrogen Masers with

Cavity Frequency Switching

Servos, these Proceedings, paper no. 22.

G. Busca and L. Johnson: Proceedings of t h e 1st European Frequency and

Time Forum, Besanqon, '18-20 March 1987, p. 339.

E.M. Mattison

and

R.F.C.

Vessot:

Proceedings

of

the

41th

Annual

Frequency Control Symposium, 1987, p. 95.

QUESTIONS AND ANSWERS

Dave Wineland, National Institute of Standards and Technology: Concerlzing the light shift with the lasers on the rubidium standard; what about chopping the laser light to get rid of the light shift?

Mr. Thomann: Yes, of course that is a method that has already been proposed with norlrlal lamps. It could more easily done with lasers because it is vcry easy to chop lasers, at least in priizciple. I don't know of anybody that has tried tha.t, h i t it is certainly worth trying. One problem is that wc have to keep the laser frequency locked at the same time as we chop the intensity. This could bc a problem. Of course we don't want to use involvcd choppers in colnmercial rubidium standards which are traditionally the cheapest available, but therc is a choice: we could make a laboratory standard with elaborate tech~liques,but for a commercial device one should keep with very simple techniques.