大爆炸是基于冷暗物质模型的标准宇宙学所预言的宇宙起点，在大爆炸发生后大约三分钟，大爆炸核合成(Big Bang Nucleosynthesis，BBN) 产生了宇宙中第一批轻核素，即2H，3He，4He 和7Li。其中2H 可直接缩写为D 或，中文称为氘。这些原初核素丰度从4He/H ≈ 0.08 到7Li/H ≈ 10−10 横跨了8 个数量级，因此对轻核素原初丰度的研究能够在核物理层面对大爆炸理论框架提出严格的限制。随着精确天文观测时代的到来，天文学家报告了越来越多的精确观测数据，这些结果对 BBN 理论计算的精度也提出了新的要求，从核物理的角度出发，我们需要为 BBN 模型的计算提供更加精确的截面数据。

本工作利用上海同步辐射光源新建成的激光伽马 (SLEGS) 装置，对大爆炸核合成中的第一个反应 H(, )D 的逆反应 D(, )H 进行了直接测量。实验共测量了 10 个能量点，束流能量范围 = 2.81–7.18 MeV，考虑束流展宽的情况下已经覆盖了 D(, )H 反应的天体物理感兴趣的能区。通过卷积迭代的方法对实验数据进行了反卷积，得到了从 = 2.4–6.9 MeV 范围内的单能截面数据。利用细致平衡原理得到了 H(, )D 反应的单能截面数据，并计算得到了其 因子() = ，利用 JAGS 软件包对 () 在质心系能量 c.m. = 10−4–5 MeV 进行了多项式拟合，并计算得到的最新的H(, )D 单能截面(即激发函数)。在考虑系统误差和统计误差的情况下，我们的数据在 = 3–5.97 MeV 的能区内是目前最精确的直接测量数据。利用此激发函数我们最终得到了 H(, )D 的反应率数据。最后，利用大爆炸核合成理论的 PRIMAT 程序研究了相应的核素丰度演化曲线以及不确定度。除此之外，我们还对氘丰度(D/H) 的不确定度进行了一系列灵敏度研究。 我们的计算结果表明，在目前的测量精度下，仅对D(, )H 反应进行测量并不能有效减小氘丰度计算值与观测值的差距。要将 BBN 的氘计算精度提升到观测水平，H(, )D，(, )3He，(, )3He，(, ) 的反应率误差都需要进一步降低，同时天文观测的重子密度 bℎ2 的误差还需要降低到 Planck 合作组 2015年报告水平的 50%。

The Big Bang, predicted by the standard cosmological model based on cold dark matter, marks the origin of the universe. Approximately three minutes after the Big Bang, Big Bang nucleosynthesis (BBN) synthesized the first batch of light elements in the universe, including 2H, 3He, 4He, and 7Li. Among them, 2H, also known as deu-terium, can be abbreviated as D or d. The abundances of these primordial nucleosyn-thesis span eight orders of magnitude, from 4He/H ≈ 0.08 to 7Li/H ≈ 10−10. Therefore, studying the primordial abundances of light elements can provide strict constraints on the framework of BBN from the nuclear physics aspect. With the advent of precise astronomical observations, astronomers have reported increasingly accurate observa-tional data, imposing new requirements on the precision of BBN theoretical calcula-tions. From the perspective of nuclear physics, more accurate reaction cross-section data are needed to provide precise BBN model calculations. This work utilized the newly constructed Shanghai Laser Electron Gamma Source (SLEGS) to directly measure the inverse reaction D(, )H of the first reaction in BBN, H(, )D. The experiment measured of ten energy points, with a beam energy range = 2.81–7.18 MeV, covering the energy region of astrophysical interest for the D(, )H reaction in the case of beam broadening. The experimental data were deconvoluted using an iterative convolution method, obtaining the single-energy cross-section within the range of = 2.4–6.9 MeV. The monoenergetic cross-section data for H(, )D were obtained according to the detailed balance principle, and the factors () = were calculated. The JAGS software package was utilized to perform polynomial fitting of () in the center-of-mass energy c.m. = 10−4–5 MeV, and thus the continuous mo-noenergetic cross-sections (i.e., excitation functions) were derived. Taking systematic and statistical errors into account, our data are the most accurate direct measurements available in the energy region of = 3–5.97 MeV. With these excitation functions, the reaction rate for H(, )D were ultimately obtained. Finally, a PRIMAT code based on BBN theory was employed to study the corresponding elemental abundance evolution curve and the associated uncertainties. In addition, the sensitivity studies on the uncer-tainty of deuterium abundance (D/H) were conducted, and the results sheds light on the future studies about the BBN. Our results indicate that, at the current level of measurement precision, measuring only the D(, )H reaction cannot effectively reduce the discrepancy between the calcu-lated value of deuterium abundance and the observed value. To improve the precision of deuterium calculation in BBN to the level of observation, the errors in reaction rates of H(, )D, (, )3He, (, )3He, and (, ) need to be further reduced. Additionally, the error in the baryon density bh2 from astronomical observations also needs to be reduced to 50% of the level reported by the Planck Collaboration in 2015.