Data set for 3D Printing of Chewable Tablets Manuscript

Figure 1. Formulation and characterization of nanosuspension inks: SEM images of griseofulvin (GF), sodium starch glycolate (SSG), and milled nanosuspension, and TEM image of the nanosuspension. Data set: particle size distribution (vol.%) of the milled nanosuspension, water content (wt.%) as a function of incubation time at 35°C during preparation of various ink formulations, and mass of water evaporated as a function of time as well as % mass loss with temperature (from TGA).
Figure 2. Rheology of ink formulations: Shear viscosity with shear rate, shear stress with shear rate, shear modulus with frequency, shear modulus with shear strain, and shear modulus with time at repetitive low (0.05%) and high (300%) strain.
Figure 3. Characterization of the printability of ink formulations. B) Printed grid designs, including photographs of the scaffolds, optical images of the pores, and thresholded images of pores. D) Printed solid discs, including top-view images, cross-sectional images, and thresholded images used for contact angle calculation. F) Optical images of printed struts for the 40% ink under varying print pressures (400-600 kPa) and print speeds (3-15 mm/s) (scale bars = 500 mm). Data set: Printability index (Pr) plotted for each ink formulation, contact angle values for each ink formulation, and measured strut width for all conditions.
Figure 4. Data set including change in ink flow rate (calculated and experimentally measured) with print pressure (P), experimentally measured line width and predicted line width for P = 400, 500, and 600 kPa and nozzle offset = 400, 500, and 700 mm, and experimentally measured line width with predicted values for all conditions.
Figure 5. Data set: summary of machine learning outcomes: Model efficiency and validation for Gradient Boosting, K-Neigbors, Random Forest, and Linear Regressor models, inverse prediction of printing parameters from the target line width using the General Bossting Regressor model, actual line width measured for struts printed using a single set of predicted print parameters, actual line width measured for struts printed using multiple sets of predicted print parameters.
Figure 6. Imaging and characterization of 3D-printed tablets alongside compressed powder mixture (PM) and griseofulvin (GF): Micro-CT images of 3D-printed tablets (using 40% ink) with 6 mm diameter (100% and 50% infill) and 9 mm diameter (50% infill), along with control PM and GF samples. Data set including XRD profiles of HPC, GF, PM, and 3D-printed dose, and DSC profiles for HPC, GF, PM, and 3D-printed dose.
Figure 7. Intra-tablet homogeneity and operator effects on process quality: UV-VIS data analysis of GF content across the same layers, and as-printed weight, dried weight and GF weight (determined by UV-VIS) for 3D-printed tablets produced by operators with varying levels of expertise.
Figure 8. (A) Average compressive modulus of each sample, including 3D-printed (3DP) dose, swollen 3DP dose, compressed powder mixture (PM), GF powder, and a commercial gummy. (B) Percent dissolved GF over time for 3D-printed (3DP) tablets with 100% infill (6 mm diameter), 50% infill (6 and 9 mm diameter), as well as compressed PM and GF (Data are presented as mean ± std. for n = 3).

图1 纳米悬浮液墨水（nanosuspension inks）的制备与表征：灰黄霉素（griseofulvin, GF）、羧甲淀粉钠（sodium starch glycolate, SSG）以及研磨后纳米悬浮液的扫描电子显微镜（SEM）图像，以及纳米悬浮液的透射电子显微镜（TEM）图像。数据集包含：研磨后纳米悬浮液的粒径分布（体积百分比）、不同墨水配方制备过程中35℃下孵育时间与含水量（质量百分比）的关系、水分蒸发量随时间的变化，以及热重分析（TGA）得到的质量损失百分比与温度的关系。
图2 墨水配方的流变学特性：剪切粘度随剪切速率的变化、剪切应力随剪切速率的变化、剪切模量随频率的变化、剪切模量随剪切应变的变化，以及在重复低应变（0.05%）和高应变（300%）下剪切模量随时间的变化。
图3 墨水配方可打印性的表征。B) 印刷栅格设计，包括支架实物照片、孔隙光学图像以及用于孔隙分析的阈值化图像。D) 印刷固体圆盘，包括俯视图、横截面图以及用于接触角计算的阈值化图像。F) 40%浓度墨水在不同打印压力（400~600 kPa）和打印速度（3~15 mm/s）下印刷的支柱光学图像（比例尺=500 mm）。数据集包含：各墨水配方的可打印性指数（Printability index, Pr）、各墨水配方的接触角值，以及所有实验条件下测得的支柱宽度。
图4 数据集包含：墨水流量速率（计算值与实验测量值）随打印压力（P）的变化、在打印压力为400、500、600 kPa且喷嘴偏移为400、500、700 mm时的实验测量线宽与预测线宽，以及所有实验条件下的实验测量线宽与对应预测值。
图5 数据集：机器学习结果汇总：梯度提升（Gradient Boosting）、K近邻（K-Neighbors）、随机森林（Random Forest）以及线性回归器（Linear Regressor）模型的模型效率与验证结果；使用通用提升回归器（General Boosting Regressor）模型从目标线宽反向预测打印参数；使用单组预测打印参数印刷的支柱的实际线宽测量值；使用多组预测打印参数印刷的支柱的实际线宽测量值。
图6 3D打印片剂以及压粉混合物（PM）与灰黄霉素（GF）的成像与表征：使用40%浓度墨水印刷的直径6 mm（填充率100%和50%）、直径9 mm（填充率50%）的3D打印片剂的微计算机断层扫描（Micro-CT）图像，以及对照压粉混合物与GF样品。数据集包含：羟丙基纤维素（Hydroxypropyl Cellulose, HPC）、GF、PM以及3D打印制剂的X射线衍射（XRD）图谱，以及HPC、GF、PM以及3D打印制剂的差示扫描量热法（DSC）曲线。
图7 片内均匀性及操作者对工艺质量的影响：同一层内GF含量的紫外-可见光谱（UV-VIS）数据分析，以及由不同熟练度操作者制备的3D打印片剂的印刷后重量、干燥后重量以及通过UV-VIS测定的GF重量。
图8 (A) 各样本的平均压缩模量，包括3D打印（3DP）制剂、溶胀态3D打印制剂、压粉混合物（PM）、GF粉末以及商业软糖。(B) 填充率100%（直径6 mm）、填充率50%（直径6和9 mm）的3D打印片剂，以及压粉混合物与GF的GF累积溶出百分比随时间的变化（数据以均值±标准差表示，n=3）。