因全反射X射线荧光仪中入射X射线与全反射X射线形成驻波, 导致样品被激发产生的特征X射线成周期变化, 且在驻波峰区探测器产生的核脉冲重叠严重, 有必要提高后续分析电路的重叠核脉冲幅度准确提取能力, 进一步提升全反射X射线荧光仪的性能。 由于SDD探测器产生的核脉冲下降沿依衰减系数 τ_rc呈指数规律衰减, 依据衰减补偿原则, 建立直线成形方法并推导应用于数字化核脉冲的递归公式。 实验发现: 在ADC转换器的取样周期为 T下, 当 k≥ T/τ_rc时直线成形后的核脉冲顶部才能形成一条直线; 在 k= T/τ_rc时上升沿最短, 重叠核脉冲分辨能力亦最好。 直线成形后的核脉冲上升沿及平顶分布与上升时间参数最小时的梯形成形结果相同但前者的平顶更为光滑; 由于直线成形后的核脉冲不存在下降沿, 使得其宽度更窄, 有利于分离相距更近的重叠核脉冲。 通过拟合直线成形后的核脉冲分布规律, 确定了直线成形方法可准确提取幅度值的最窄重叠核脉冲间隔; 且直线成形方法对不同幅度核脉冲的放大倍数一致, 证实直线成形方法不会改变谱仪的能量线性。 直线成形方法能可靠应用的前提是准确定位核脉冲的起始位置, 为此建立起冲击成形递归公式, 验证发现冲击成形后的核脉冲仅在其起始位置出现幅度约等于核脉冲幅度的冲击响应, 利用阈值判别就能准确定位所有核脉冲起始位置。 最后在FPGA内实现上述算法, 并进行10组比对实验, 实验结果显示直线成形方法在堆积平台与和峰压制能力、 峰背比和能量分辨率三个方面均优于梯形成形与模拟式能谱测量系统, 证实直线成形方法能极大地提升全反射X射线荧光仪的性能。

Because the interaction of incident X-ray produced from the X-ray tube and reflection X-ray from sample disk create standing wave in Total-Reflection X-ray Fluorescence (TXRF) spectrometer, the characteristic X-ray produce from sample changes periodically. So a large number of piled-up nuclear pulses will be generated at the time of wave crest. In order to improve the performance of the TXRF spectrometer, the ability to accurately extract amplitudes of piled-up pulses must be strengthened. The falling edge of every pulse from an SDD (silicon drift detector) decays with time constant τ_rc exponentially. According to the principle of attenuation compensating, a straight-line shaping method is devised, and its transfer formula is deduced for digitized pulses. When the ADC sampling period is equal to T, the method has a stronger ability to remove noise in pulse as the value of attenuation compensation coefficient k increases. The flat-top of shaping results by this method can be approximated to be straight line only when k≥ T/τ_rc. And the width of the rising edge is narrowest when k= T/τ_rc. In this case, the straight-line shaping method has the best ability to extract amplitudes of piled-up pulses. Compare the result of a trapezoidal shaper at smallest rise time with this method at k= T/τ_rc, both of their rising edge and height of flat-top are same. For the smooth degree of flat-top area, this method is better than trapezoidal shaper. This method is better than trapezoidal shaper in separating piled-up pulses. Through fitting the rising edge of pulse transformed with this method, the distinguishable narrowest interval between the start positions of two pulses is determined. Magnifications of different amplitudes of pulses are the same by experiment, which shows that this method does not change the energy linearity of the TXRF spectrometer. The start position of the pulse must be determined before the application of this method. So the transfer formula of impulse shaper is deduced. By analyzed shaping results of impulse shaper, an impulse response appears at the start position of one pulse and the value approaches to the amplitude of the pulse. A threshold method is used to locate the start position of pulse accurately. Finally, all transfer formulas are implemented in an FPGA chip. Experiment results illustrate that the straight-line shaping method has a better energy resolution, peak to background ratio and suppression ability of sum-peak than trapezoidal shaper and analog energy spectrum system. It shows that the performance of the TXRF spectrometer is greatly improved by applying the straight-line shaping method.