We present experimental studies on the charge transfer inefficiency (CTI) of charge-coupled device (CCD) developed for the soft X-ray imaging telescope, Xtend, aboard the XRISM satellite. The CCD is equipped with a charge injection (CI) capability, in which sacrificial charge is periodically injected to fill the charge traps. By evaluating the re-emission of the trapped charge observed behind the CI rows, we find that there are at least three trap populations with different time constants. The traps with the shortest time constant, which is equivalent to a transfer time of approximately one pixel, are mainly responsible for the trailing charge of an X-ray event seen in the following pixel. A comparison of the trailing charge in two clocking modes reveals that the CTI depends not only on the transfer time but also on the area, namely the imaging or storage area. We construct a new CTI model by taking into account both transfer-time and area dependence. This model reproduces the data obtained in both clocking modes consistently. We also examine apparent flux dependence of the CTI observed without the CI technique. The higher incident X-ray flux is, the lower the CTI value becomes. It is due to a sacrificial charge effect by another X-ray photon. This effect is found to be negligible when the CI technique is used.

我们目前对为XRISM卫星上的软X射线成像望远镜Xtend开发的电荷耦合器件（CCD）的电荷转移效率（CTI）进行实验研究。CCD具备电荷注入（CI）功能，其中定期注入牺牲电荷以填充电荷陷阱。通过评估在CI行后观察到的俘获电荷的重新发射，我们发现至少存在三个具有不同时间常数的陷阱种群。具有最短时间常数的陷阱（约等于一个像素的传输时间）主要负责在后续像素中看到的X射线事件的尾随电荷。比较两种时钟模式下的尾随电荷，可以发现CTI不仅取决于传输时间，还取决于面积，即成像或存储区域。我们通过考虑转移时间和面积依赖性来构建新的CTI模型。该模型一致地再现在两种时钟模式下获得的数据。我们还检查了未使用CI技术观察到的CTI的明显通量依赖性。X射线通量越高，CTI值就越低。这是由于另一个X射线光子的牺牲电荷效应所致。当使用CI技术时，发现这种影响可以忽略不计。这是由于另一个X射线光子的牺牲电荷效应所致。当使用CI技术时，发现这种影响可以忽略不计。这是由于另一个X射线光子的牺牲电荷效应所致。当使用CI技术时，发现这种影响可以忽略不计。