self heterodyne mixing method of two inter mode beat frequencies for frequency stabilization of a three mode he ne laser

Self-heterodyne mixing method of two inter-mode beatfrequencies for frequency stabilization of a three-mode He-Ne laser Jeongmin Lee and Tai Hyun Yoona Department of Physics, Korea Unive

Self-heterodyne mixing method of two inter-mode beat frequencies for frequency stabilization of a three-mode He-Ne laser

Jeongmin Lee and Tai Hyun Yoon

Citation: AIP Advances 2, 022170 (2012); doi: 10.1063/1.4733344

View online: http://dx.doi.org/10.1063/1.4733344

View Table of Contents: http://aip.scitation.org/toc/adv/2/2

Published by the American Institute of Physics

Self-heterodyne mixing method of two inter-mode beat

frequencies for frequency stabilization of a three-mode He-Ne laser

Jeongmin Lee and Tai Hyun Yoona

Department of Physics, Korea University, 145, Anam-ro, Seongbuk-gu, Seoul 136-713, Korea

(Received 27 January 2012; accepted 12 June 2012; published online 27 June 2012)

We present a robust self-heterodyne mixing method of two inter-mode beat frequen-cies suitable for frequency stabilization of a three longitudinal-mode (3-mode) He-Ne laser at 633 nm A high-contrast frequency discrimination signal with a broad locking range of 244 MHz is obtained by using a self-heterodyned double balanced mixer op-erating at the inter-mode beat frequency of 607 MHz We show that the central-mode frequency of the 3-mode He-Ne laser could be stabilized to the center of the gain profile with a frequency fluctuation less than± 5 MHz for more than 12 h Copyright

2012 Author(s) This article is distributed under a Creative Commons Attribution 3.0 Unported License [http://dx.doi.org/10.1063/1.4733344]

He-Ne laser oscillating at the vacuum wavelength of 633 nm is one of the most widely used lasers in dimensional metrology and fundamental science.1 4 For example, laser interferometers and many other dimensional measurement instruments are based on the frequency-stabilized He-Ne lasers referenced to the center frequency of the gain profile.1 7 Recently, for practical use of the unstabilized or stabilized but uncalibrated He-Ne lasers at 633 nm, the vacuum wavelength of the laser operating on the 3s2 → 2p4 transition of Ne atoms was recommended by the International Committee for Weights and Measures (CIPM) in 20078 as λU = 632.9908 nm with the relative standard uncertainty of 1.5× 10−6 The quoted uncertainty captures all sources of variability that are seen to occur in practice such as the effects of Ne isotope and gas pressure

In this paper, we present a robust self-heterodyne mixing method (SHMM) of inter-mode beat frequencies for frequency stabilization of the central-mode of a three longitudinal-mode (3-mode)

He-Ne laser to the CIPM value of c/ λU, where c is the speed of light in a vacuum The advantage

of the proposed SHMM is an ability to detect a frequency discrimination signal (FDS) with a high signal-to-noise (S/N) ratio without using a reference frequency synthesizer so that it could provide

a compact local wavelength standard traceable to the international standard.8 Two inter-mode beat

frequencies f21 and f32 in a 3-mode He-Ne laser differ slightly due to the frequency pulling effect (FPE)9 across the gain profile and a polarization anisotropy of the cavity mirrors.10 , 11 Here, f21

= f2 − f1 and f32 = f3 − f2, with f1, f2, and f3 being optical frequencies of the 3-modes in the increasing order (Fig.1(a)) We should emphasize that the FDS obtained by the proposed SHMM is not sensitive to the FPE and the polarization anisotropy effect, but sensitive to the deviations of the optical frequencies relative to the gain center In a proof-of-principle experiment, the wavelength of

a He-Ne laser at 633 nm could be stabilized to the center of the gain profile with the stability of

3× 10−9at 2.4-h averaging time One can measure the absolute frequency f

2and evaluate its uncer-tainty after stabilization by using an optical frequency comb,12for instance, in order to use the stabi-lized laser as a local wavelength standard at 633 nm We also show that the optical frequency of the He-Ne laser stabilized by the proposed method could be stabilized always within the frequency range

of 250 MHz from the CIPM value without calibration, i.e., with the relative standard uncertainty of

a Electronic mail: thyoon@korea.ac.kr

022170-2 J Lee and T H Yoon AIP Advances 2, 022170 (2012)

3-mode

D1

DBM

DL

A1

Servo P

BS

) b ( )

a

(

fcνq νq+1

νq-1

FIG 1 (a) Frequency positions of the three longitudinal modes under a Gaussian lineshape function (b) Schematic diagram

of the proposed SHMM The central-mode can be separated with a PBS for other experiments BS; beam splitter, PBS; polarizing BS, P; polarizer, D1,2; fast photo-diodes, A1,2; low-noise amplifiers, PS; phase-shifter, DL; coaxial delay-line, DBM; double balanced mixer, LF; low-pass filter, Servo; thermal-heating servo controller.

5× 10−7, since the symmetric error signal occurs only in the frequency range of 40% of the longi-tudinal mode-spacing

Figure1(b)shows a schematic diagram of the proposed SHMM proposed in this paper We split

off the output beam by using a beam splitter (BS) and two photo-currents I1(t) and I2(t) are detected independently by two fast photo-diodes D1 and D2 with a bandwidth of 1.5 GHz I1(t) is then converted into a voltage signal V1(t) by using a low-noise pre-amplifier A1 and coupled to the local oscillator port of a double balanced mixer (DBM) The total cable length in this channel is L1 On

the other hand, I2(t) is converted into a voltage signal V2(t) by using another low-noise pre-amplifier

A2 followed by a phase shifter PS and a coaxial-cable delay line DL, and finally coupled to the RF

Port of the DBM The total cable length in this channel is L2 Power levels of V1(t) and V2(t) are

0 dBm and –5 dBm, respectively, below the saturation power of +7 dBm of the DBM.16 The operating frequency of the DBM is from 1 MHz to 1 GHz (ZFM-2, Mini-Circuits) Then, the output

of the intermediate frequency (IF) port of the DBM is low-pass filtered (LF) with a cutoff frequency

of∼1 kHz and it generates an antisymmetric FDS when the central-mode frequency crosses the gain center in either direction Finally, the FDS is fed to a servo-controller that controls the temperature

of the plasma-tube by using a thin-film heater for the frequency stabilization This completes the demonstration of a novel SHMM to stabilize the central-mode frequency to the gain center without using a reference frequency synthesizer

The SHMM generates a high S/N ratio in the FDS compared to that of the well-known power-balance method (PBM),13 – 15since it works at a RF frequency of f a = (f1+ f3)/2∼ 607 MHz so that the FDS is completely decoupled from the low frequency acoustic noises It has thus a similar S/N ratio compared to that of the secondary beat-frequency stabilization method,6 , 10 , 11but without using

an external frequency synthesizer In addition, our method stabilizes the central-mode frequency to the gain center so that it has an output power of ∼2 mW, a factor of two larger than that of the PBM.6,10,11

In experiment, we used an internal-mirror 3-mode He-Ne laser with a cold-cavity free spectral

range of fFSR= c/2L0= 606.8 MHz, corresponding to the cavity length of L0= 24.7 cm Here, we shall briefly discuss analytically the principle of the SHMM shown in Fig.1(b) We write the optical frequencies of the 3-modes asν q−1,ν q, andν q+1(Fig.1(a)), where q= 780 507 is the mode number

ofν q In addition, the center frequency f cof the gain curve corresponding to the wavelengthλUmay

be written as f c ≡ c/λ U = q fFSR+ f0= 473.613 THz, where f0= 232 MHz is the offset frequency

between the q-th mode and the laser gain center.

In the actual He-Ne laser, however, the optical frequencies of the cold-cavity modes are mod-ified slightly due to the FPE across the gain center9 11 and polarization anisotropy of the cavity mirrors.6 , 10 , 11 We show below that the FDS obtained by the SHMM is not sensitive to the FPE and the polarization anisotropy of the cavity mirrors, but sensitive to the optical frequency variation

relative to f c Hereafter, we use new notations for the optical frequencies of the 3-modes, i.e., f1

= f q−1, f2= f q , and f3= f q+1 for simplicity Since the photo-current is proportional to the incident

light intensity, it can be written in terms of the optical frequencies f1, f2, and f3as

I (t)∝







3



j=1

E jexp

i (ω j t − k j z + φ j)





2

where E j is the electric field amplitude of the jth-mode, ω j = 2πf j , k j = 2πf j /c, z is the optical

path-length, and we assume that relative phases satisfy the relationφ1 = φ3 = 0 and φ2 = π.7

One should note that E jin Eq.(1)depends on the optical frequency f j , i.e., E j = E j (f j), since their intensities trace a Gaussian lineshape function as shown in Fig.1(a) Then, two voltage signals V1(t) and V2(t) at two input ports of the DBM may be written as7

V1(t)∝

2



i=1

E i E i+1cos(ω(i +1)i t − k (i+1)i L1), (2)

V2(t)∝

2



i=1

E i E i+1cos(ω(i +1)i t − k (i +1)i L2+ φ s), (3)

with L i , i = 1, 2, being the total cable length in the channel 1 and 2, ω (i+1)i = ω i+1− ω i , k (i+1)i = k i+1

− k i, andφ sbeing the phase shift by the PS Here, we ignored the terms having the inter-mode beat

frequency f31between the 1st-mode and 3rd-mode, because f31= 2fFSRis larger than the bandwidth

of the DBM We also assume that two terms in V2(t) in Eq.(3) experience the same phase shift

φ s by the PS We observed experimentally that f21= f32(see Fig.2(a)), but they were separated in average by∼ 340 kHz  f a = (f21 + f32)/2, the average inter-mode beat frequency This is due to the polarization anisotropy of the cavity mirrors as explained earlier and varies± 125 kHz from the average separation due to the FPE of the gain medium.6 , 10 , 11

Finally, the rectified voltage signal V s of two photo-voltages V1(t) and V2(t) in Eqs.(2)and(3) can be measured by using the DBM used as a linear synchronous detector.16Then, V sat the IF port

of the DBM after passing a low-pass filter with the cutoff frequency of∼1 kHz can be written as

V s = 2



i=1

 i I i ( f i )I i+1( f i+1) cos

k (i +1)i L + φ s



whereL = |L2− L1|, I j (f j ), j = 1, 2, 3, is the light intensity of the jth-mode, and  iis a constant

to be determined by the experiment and usually1= 2

Equation (4) is the main result of the SHMM of two inter-mode beat frequencies proposed

in this paper Note that V s is a function ofL and φ s, those are parameters associated with the

cable-length difference between two detection channels and the electrical phase shift between V1(t) and V2(t), respectively Note also that V s is a function of the optical frequencies only, those are

explicitly included in I j (f j), but the total phase factors 21 = k21L + φ s and 32 = k32L + φ s

remain the same for fixedL and φ s , since k21 and k32 are not changing conceivably compared to the change of optical frequency We also note that if the polarization state of the central-mode flips from the vertical to the horizontal or vice versa followed by the optical frequency change of one

fFSR, the sign of V s in Eq.(4)also flips (Fig.2(b)) In addition, the frequency locking point of the FDS can be easily controlled by adjusting eitherL or φ sor both We see that if the laser oscillates

in 2-modes, V s has only a single term in Eq.(4)with an even symmetry and it has a considerably weak frequency dependance compared to the value of a 3-mode laser (Fig.2(c)) Thus, the proposed SHMM works only for a 3-mode He-Ne laser

Figure2(a)shows a typical RF beat note spectrum at around f a = 607.0 MHz recorded in the peak-detection mode (upper trace) and the normal mode (lower trace) of a RF spectrum analyzer with a resolution bandwidth (RBW) of 20 kHz As one can see, there are two inter-mode beat

frequencies f21and f32separated by 340 kHz (see lower trace) caused by the polarization anisotropy

of the cavity mirrors.6 , 10 , 11We then examined the relation between the tuning ranges of f and f

022170-4 J Lee and T H Yoon AIP Advances 2, 022170 (2012)

606.0 606.5 607.0 607.5 608.0

-100

-80

-60

-40

-20

Frequency (MHz)

RBW = 20 kHz ( b ) )

a

(

0.0 0.2 0.4 0.6 0.8 1.0 -150

-100 -50 0 50 100 150

V s

ΔL (m)

) d ( )

c

(

-200 -100 0 100

200

D

C B

V s

Time (s)

A

-100 -50 0 50 100

-0.4

-0.2

0.0

0.2

0.4

f

2 - f

c (MHz)

3-mode

2-mode

FIG 2 (a) RF beat frequencies at around f arecorded in the peak-detection mode (upper trace) and the normal mode (lower trace) of a RF spectrum analyzer P and Q are continuous frequency tuning ranges of two inter-mode beat frequencies by the

FPE (b) V svs time when the plasma-tube temperature is decreasing Regions A and B (C and D) correspond to the values

obtained when the laser oscillates in 3-modes (2-modes) (c) Theoretical line-shapes of V sfor 3-mode (solid line) and 2-mode

(dotted line) lasers as functions of f2 (d) Average value V svs cable length differenceL.

and the optical frequency tuning range of the central-mode by measuring an optical beat frequency with an independently frequency-stabilized laser We found that the continuous optical frequency

tuning range was 244 MHz covering about 40 % of fFSRduring the 3-mode operation of the laser,

i.e., f21 and f32 existed in the regions P and Q While the optical frequency of the central-mode

varied in this period, f21 and f32 scaned about 250 kHz due to the FPE of the central-mode across the gain center During the last 60 % of period, the laser oscillates in 2-modes, resulting in only one inter-mode beat frequency either in the region P or Q depending on their polarization configuration

A typical FDS detected by the SHMM is shown in the panel (b) Note that it was recorded in the cooling phase of the plasma-tube temperature so that the time interval (or frequency interval) between two consecutive signals are increasing as time flows In the panel (b), regions A and B correspond to the regimes of 3-mode operation with an opposite polarization configuration, while regions C and D correspond to the regimes of 2-mode operation of the laser Therefore, the single

cycle of the FDS covers the frequency regions from A to D, which has 2fFSR= 2 × 607 MHz Note that the FDS either in the region A or B has an offset voltage from zero, which can be compensated

by adjusting eitherL or φ sas shown in the panel (d) of Fig.2and in the panel (b) of Fig.3 Numerical plots of the FDS (Eq.(4)) for 21 = 32 = 0 are depicted in the panel (c) for a 3-mode laser (solid line) and a 2-mode laser (dotted line) In this simulation, we used the same values

of fFSR= 607 MHz for two cases and a Gaussian lineshape function with a full-width of 1.3 GHz was assumed As one can see, the FDS corresponding to the 3-mode laser shows an antisymmetric

(odd) function suitable for the frequency stabilization to the line center where f2= f c On the other hand, that of the 2-mode laser shows an even function and almost flat frequency dependence, which

is of course not suitable for a servo control Thus, it is confirmed that the proposed SHMM works

only for 3-mode lasers The panel (d) shows an average value V s of the measured FDS, i.e., the center value of the panel (b), as a function of cable length differenceL The delay line was made

0 2 4 6 8 10 12

110

120

130

140

Time (h)

(b)

1E-11 1E-10 1E-9 1E-8

Gate time (s)

(d)

-0.2

-0.1

0.0

0.1

0.2

Power-balance (vertical)

V S

Time (s)

SHMT servo on

(a)

(c)

-15 -10 -5 0

V s

Phase shift (rad)

slope = 54.3 mV/rad

FIG 3 (a) V s before and after the stabilization detected by the SHMM (upper trace) and PBM (b) V svs phase shiftφ sfor

fixed cable length (c) f bbetween two 3-mode lasers whose frequencies are stabilized independently by using the proposed SHMM and its associated Allan deviation is shown in the panel (d).

of a 50

there is a periodicity with a periodL = 0.35 m, resulting in the phase interval of 2π for a signal oscillating at f a, i.e., 2πf a1.43L/c = 6.36 ≈ 2π At the points where V shave extreme values, the total phase differences 21and 32in Eq.(4)have the values (n + 1/2)π, n = 0, 1, 2, , so that

the amplitudes of the FDS become zero

The voltage signal V s (t) before and after the frequency stabilization is shown in the panel (a)

of Fig.3(upper trace), where an error signal obtained simultaneously by the PBM13 – 15is shown for comparison It is clearly seen that the S/N ratio of the FDS obtained by the SHMM is much higher than that of the PBM Once the frequency of the central-mode (vertically-polarized) is locked to the line center, the error signal from the PBM is locked to the maximum value corresponding to the value obtained by the central-mode only Note that the PBM does not produce any error signal for the regime of 3-mode operation As explained earlier, the frequency locking point of the FDS obtained by the SHMM can be tuned precisely by the PS as demonstrated in the panel (b), where the slope is measured to be 54.3 mV/rad We should note that the sign and the magnitude of the tuning slope in the panel (b) change depending on the value ofL, i.e., depending on the values of the total

phases 21and 32in Eq.(4)as discussed in the panel (d) of Fig.2

We constructed two identical frequency-stabilized lasers by using the proposed SHMM to evaluate the optical frequency stability The panel (c) shows the long term record of their beat

frequency at f b ∼ 126 MHz For this measurement, we intentionally adjusted V sof one laser to shift

f bfrom< 10 MHz to 126 MHz for beat frequency measurement As one can see clearly, the

long-term frequency stability of two lasers is less than± 5 MHz for 12 h, i.e., f b /(c/ λ U)= 1 × 10−8, which is two orders of magnitude smaller compared to the cited uncertainty for the wavelength

of an unstabilized He-Ne laser at 633 nm by the CIPM.8 Short-term frequency stability of f b is shown in the panel (d) From the gate time of 1 s to 2.4 h, a flicker frequency noise dominates the Allan deviation17at the level below 3× 10−9, resulting in the maximum frequency fluctuation less

022170-6 J Lee and T H Yoon AIP Advances 2, 022170 (2012)

than±1.5 MHz consistent with the long-term frequency fluctuation of ± 5 MHz measured in the panel (c)

In summary, we have proposed and demonstrated a robust SHMM of two inter-mode beat frequencies of a 3-mode He-Ne laser at 633 nm for the frequency stabilization of the central-mode

to the gain center It provides a FDS with a high S/N ratio without using a reference frequency synthesizer Long-term frequency fluctuation of a frequency-stabilized laser based on the proposed SHMM was measured to be less than±5 MHz for more than 12 h One can measure the absolute frequency and evaluate its uncertainty by using an optical frequency comb,12 for example, in order

to use the frequency-stabilized laser as a local wavelength standard at 633 nm

ACKNOWLEDGMENTS

This research was supported by the Basic Research Program (2011-0006670) of the National Research Foundation of Korea

1 T M Niebauer, S G Sasagawa, J E Faller, R Hilt, and F Klopping, “A new generation of absolute gravimeters,”

Metrologia32, 159–180 (1995).

2 J Lee, H Yoon, and T H Yoon, “High-resolution parallel multipass laser interferometer with an interference fringe spacing

of 15 nm,” Opt Commun.284, 1118–1122 (2011).

3 S Yokoyama, T Yokoyama, and T Araki, “High-speed subnanometer interferometry using an improved three-mode heterodyne interferometer,” Meas Sci Technol.16, 1841–1847 (2005).

4 M P Winters, J L Hall, and P E Toschek, “Correlated spontaneous emission in a Zeeman laser,” Phys Rev Lett.65,

3116–3119 (1990).

5 T Baer, F V Kowalski, and J L Hall, “Frequency stabilization of a 0.633-μm He-Ne longitudinal Zeeman laser,”Appl Opts.19, 3173–3177 (1980).

6 J Y Yeom and T H Yoon, “Three-longitudinal-mode He-Ne laser frequency stabilized at 633 nm by thermal phase locking

of the secondary beat frequency,” Appl Opts.44, 266–270 (2005).

7 J D Ellis, K.-N Joo, E S Buice, and J W Spronck, “Frequency stabilized three mode HeNe laser using nonlinear optical phenomena,” Opt Express18, 1373–1379 (2010).

8 CI-2007 Recommendation 2, “On the value and uncertainty of unstabilised He-Ne lasers,” CIPM, 2007.

9A E Siegman, Lasers, Chap 12, University Science Books, CA 94965 (1986).

10 H S Suh, T H Yoon, M S Chung, and O S Choi, “Frequency and power stabilization of a three longitudinal mode He-Ne laser using secondary beat freyency,” Appl Phys Lett.63, 2027–2029 (1993).

11 S Yokoyama, T Araki, and N Suzuki, “Frequency stabilization by frequency pulling for single-mode oscillation of He-Ne laser at maximum intensity,” Rev Sci Instrum.66, 2788–2795 (1995).

12 T H Yoon, J Ye, J L Hall, and J.-M Chartier, “Absolute frequency measurement of the iodine-stabilized He-Ne laser at

633 nm,” Appl Phys B72, 221–226 (2001).

13 T M Niebauer, J E Faller, H M Godwin, J L Hall, and R L Barger, “Frequency stability measurements on polarization-stabilized He-Ne lasers,” Appl Opts.27, 1285–1289 (1988).

14 T B Eom, H S Choi, and S K Lee, “Frequency stabilization of an internal mirror He-Ne laser by digital control,” Rev Sci Instrum.73, 221–224 (2002).

15 M.-S Kim and S.-W Kim, “Two-longitudinal-mode He-Ne laser for heterodyne interferometers to measure displacement,”

Appl Opts.41, 5938–5942 (2002).

16 E Rubiola, “Tutorial on the double balanced mixer,” e-print arXiv:physics/0608211, 1–52 (2006).

17E Rubiola, Phase noise and frequency stability in oscillators, (Cambridge University Press, 2009).