Laser Magnetic Field Test (March 2001)

This test was performed in collaboration with the CMS Tracker Optical Links group [72]. Two different types of semiconductor lasers were tested in the same magnetic field as the APV25S1 (see section , p. ). Similar to the APV25, the lasers were also positioned in three orientations (A, B and C) as shown in fig. .

Figure: The three different orientations (A, B and C) of the semiconductor lasers (red) with respect to the magnetic field.

In this test, the lasers were driven by a programmable current source and optically connected to the analog optical link receiver hybrid. The voltage drop at the laser diode and the output of the analog receiver (level, linearity and noise) were measured as a function of the input current within the parameter space defined by laser type, orientation and magnetic field. Additionally, the optical output power and spectrum were analyzed.

Figure: Typical laser output power vs. input current. The red bar represents the typical prototype laser driver output current range for its nominal input voltage span.

Fig.  shows the optical output power of a typical semiconductor laser depending on the input current. Above a certain threshold, the light output power linearly increases with the input current. In this plot, the input current varies between zero and $100\,\rm mA$, but only a small fraction of this range is used by the laser driver, which has a typical transconductance of $10\,\rm mS$. Thus, the prototype analog optical link input span of $800\,\rm mV$ translates to a range of $8\,\rm mA$. The bias is normally set close above the threshold, resulting in an operating range of the laser driver as shown in red.

Figure: Typical output spectrum of an edge-emitting semiconductor laser with an input current of $15\,\rm mA$.

A typical output spectrum of an edge-emitting semiconductor laser with a nominal wavelength of $1310\,\rm nm$ is shown in fig. . The lasing semiconductor is a narrow-band emitter of approximately Gaussian spectral shape enclosed in an optical resonator, consisting of two facing mirrors. One of the mirrors is semi-transparent to extract the light into the optical fiber. The optical resonator only leads to amplification when an integer multiple of half the wavelength fits in between the mirrors; other wavelengths are extinguished by destructive interference. This leads to the forked structure in the spectrum. The resonator condition for constructive interference is given by

$$N\,\frac{\lambda}{2}=D\quad,$$ (5.4)

with the wavelength $\lambda$, the resonator length $D$ and an integer multiple $N$. Knowing two neighboring spectral lines $\lambda_2 > \lambda_1$, the length of the optical resonator can be obtained by

$$D=\frac{1}{2\left( \lambda_1^{-1} - \lambda_2^{-1} \right) }\quad.$$ (5.5)

In the sample spectrum shown in fig. , the two highest adjacent peaks are found at $\lambda_1=1315.5\,\rm nm$ and $\lambda_2=1316.3\,\rm nm$. Eq.  returns a length of $1.08\,\rm mm$ for the optical resonator. The actual length is unknown, but all lasers tested for CMS have a resonator length about one millimeter.

It is obvious that the position of the spectrum peaks strongly depend on the actual length of the resonator. In fact, the method of interferometric length measurement employs this principle. When the laser is powered by a current flow, it slightly heats up compared to the zero current state. This causes the resonator to expand, resulting in a peak shift towards higher wavelengths. With an input current of $100\,\rm mA$ (corresponding to an electrical input power of approximately $176\,\rm mW$), peak shifts in the order of $1\,\rm nA$ were observed, indicating a resonator expansion of approximately $1\,\rm\mu m$.

The only effect observed in a precision scan of the magnetic field was a small shift in the laser threshold and slope. As shown in fig. , the laser threshold approximately depends on the square of the magnetic flux density, decreasing by about $4\%$ at $10\,\rm T$. The effect on the slope is even smaller. A relative change of less than $2\%$ has been observed at the maximum magnetic field of $10\,\rm T$.

Figure: Dependence of the laser threshold of the magnetic field.

Thus, the influence of the magnetic field is very small. In fact, temperature effects largely dominate the performance of the laser, such that the presence of a magnetic field is irrelevant regarding the application in CMS. Not only the wavelength is affected by temperature, but also the input-output characteristics of the laser.