Temperature Drift Compensation of Bias Point Stabilization High Voltage Power Supply for Lithium Niobate Electro Optic Modulator
Lithium niobate electro optic modulators serve as critical components in optical communication systems, converting electrical signals to optical intensity modulation through the linear electro optic effect. The modulator operation requires a stable bias voltage that sets the operating point on the optical transfer function. Temperature variations cause drift in the optimal bias point through thermal effects in the lithium niobate crystal and the bias electrode structure. High voltage power supplies with temperature drift compensation maintain stable bias operation across the operating temperature range.
The Mach Zehnder interferometer structure common in lithium niobate modulators splits the input optical waveguide into two arms that are recombined at the output. Phase shifters in the arms, driven by the radio frequency signal and the bias voltage, control the relative phase between the arms. The output intensity depends on the phase difference, following a cosine squared transfer function. The bias voltage sets the quiescent operating point, typically at the quadrature point where the slope of the transfer function is maximum and the modulation is most linear.
Temperature effects on the bias point arise from several mechanisms in the lithium niobate device. The refractive index of lithium niobate changes with temperature through the thermo optic coefficient, causing thermal phase shift in the waveguide arms. Thermal expansion changes the waveguide length, contributing to the thermal phase shift. The pyroelectric effect in lithium niobate generates electric fields in response to temperature changes, which can affect the refractive index through the electro optic effect. These mechanisms combine to produce temperature dependent phase shift that drifts the bias point.
The bias electrode structure also exhibits temperature dependent characteristics that affect the bias voltage to phase shift relationship. The electrode capacitance changes with temperature due to the temperature dependence of the dielectric properties. Resistive heating in the electrode material causes temperature rise that affects the electrode characteristics. Any buffer layers or cladding materials between the electrode and the waveguide have temperature dependent properties that influence the electric field distribution in the waveguide.
Temperature drift compensation approaches include feedforward correction based on temperature measurement, feedback control based on optical output monitoring, and design techniques that minimize the inherent temperature sensitivity. Feedforward compensation measures the device temperature and applies a temperature dependent correction to the bias voltage based on a predetermined model of the temperature drift. This approach requires characterization of the temperature drift behavior and accurate temperature measurement at the relevant location.
Feedback control of the bias point monitors the optical output and adjusts the bias voltage to maintain the desired operating point. Dither techniques apply a small low frequency modulation to the bias and detect the resulting modulation in the optical output, providing an error signal that indicates deviation from the quadrature point. The feedback loop continuously adjusts the bias voltage to null the error signal, automatically compensating for temperature drift and other disturbances. This approach requires additional detection and control circuitry but provides robust compensation without requiring temperature drift characterization.
The high voltage power supply for bias application must provide stable output with the resolution and accuracy required for precise bias control. The output voltage range must cover the full range of bias voltages needed across the operating temperature range, including the compensation for temperature drift. The voltage resolution must be fine enough to set the bias point within the allowable tolerance, typically requiring millivolt scale resolution for high performance modulators. The output noise must be low enough to avoid introducing phase noise that would degrade the optical signal quality.
Temperature coefficient matching between the power supply components and the modulator can reduce the net temperature drift. If the power supply output voltage drifts with temperature in a way that partially compensates the modulator temperature drift, the overall bias point stability improves. This approach requires careful selection of power supply components and characterization of their temperature coefficients. The matching is specific to a particular modulator design and may not generalize to other designs.
Thermal design of the modulator package and the power supply housing affects the temperature dynamics and the compensation requirements. Thermal insulation slows the temperature change rate, giving the compensation system more time to respond. Thermal mass reduces the temperature excursion from transient thermal loads. Strategic placement of temperature sensors at locations with good correlation to the modulator temperature improves the feedforward compensation accuracy. The thermal design must balance the temperature stability against other requirements including size, weight, and cost.
Long term stability of the bias point requires consideration of aging effects in addition to temperature drift. The electro optic properties of lithium niobate can change over time through photorefractive damage or other degradation mechanisms. The power supply components age and drift, affecting the output voltage. The combination of temperature drift and aging drift determines the required frequency of bias recalibration or the performance of the automatic bias control system over the device lifetime.
