单晶锗刻蚀速率的研究

A study has been made of the rate at which single-crystal germanium is etched under  various conditions. For simplicity, the principal etchants used were composed of only  H~O2, HF, and water. Data are given for the rate as a function of temperature and composition of the etchant, and as a function of the orientation and impurity content of  the crystal. An equation which assumes two reactions taking place in sequence on the  surface fits the rate vs. temperature data within the limits of experimental error. In  the range studied, the rate-controlling processes are sensitive to sample orientation.  From etching data, a value was obtained for the thickness of the disturbed surface layer  due to abrasive grinding. This was found to be in the range of 2-10 ~ and to depend on  sample orientation and abrasive particle size. Comparative data are given for two common etchants of a more complicated composition.  

INTRODUCTION  

For several years, the treatment of germanium by  etching has played an important role in the production of germanium devices and in the preparation of  samples for research purposes. In the last two years,  such treatment has become an increasingly important  tool for investigating the orientation and quality of  germanium samples, different treatments revealing  different properties of the material. But the treatments themselves have not been well understood.  In the hope of progressing toward an understanding  of these processes, a study has been made of the  etching rate of germanium crystals under various  conditions. For simplicity, most of the work has been  confined to etchants containing only HF, H202, and  water. In order to insure sample uniformity, all  samples were cut from one large single crystal of  As-doped, n-type germanium. Since the resistivity  of the crystal varied slowly along its length, resistivities between 1 and 9 ohm-cm could be obtained by selecting samples from the proper region  of the crystal. The samples were thin, usually of  the order of 30 mils thickness, so that the ratio of  face area to edge area was large. The error introduced by including the edge area as if it were of the  same orientation as the face was less than 2%.  The samples were held with polyethylene-coated  tweezers and the area of contact was small enough  to be neglected. Etching rates were obtained by  measuring the loss of weight of a sample after a  known etching time. Weights were measured to  0.2 mg and the weight differences were usually of  the order of 20 mg. Considerable attention was given  to the freshness and composition of the etchants and  to the surface preparation of the samples. In order  that the circulation of the etching solution should  not influence the results, all experiments were  performed under conditions of agitation sufficiently  violent that substantial changes in agitation made  no difference in the etching rate.  

GENERAL PRECAUTIONS  Since little was known about the effects of various  parameters on the etching rate, much effort was  devoted to insuring that only the parameter under  investigation was influencing the results. This was  done either by comparison of samples differing in  only one respect, or by direct experiment to determine what parameters could be ignored safely.  Temperature.--The temperature was maintained  constant by immersing the etching vessel in a constant temperature bath. The volume of etchant  used, about 60 cc, was large compared to the volume  of the samples. The temperature of the etchant was  measured directly to =t=0.3~ by means of a polyethylene-coated thermometer.  Resistivity.--The samples used were kept short  so that the variation in resistivity along the length  was less than 4-5 % from the mean except for the 1  ohm-cm sample. Insofar as possible, samples were  compared which were cut from the same depth in the  crystal. All samples were labeled to show from what  part of the crystal they came. Thus, the notation  4(100)3-2 means that the sample was from the  fourth resistivity region of the crystal, a (100) cut,  and the third slab from the surface. This sample was  then cut into smaller pieces, the -2 indicating the  second such piece.  

Etch composition.--Etehes were made up fresh each day in batches large enough that the composition was measured with an accuracy greater than  the assay of the reagents used. Most work was done  with the simplest etch (1 HF, 1 H202, 4H20), called  the No. 2 etch (see appendix for etch compositions).  The variation in the concentration of the reagent  H202 used (5 parts in 30) was sufficient to introduce  a rate error of about 4-4%. However, in most  experiments, the same reagent was used. Therefore,  although the absolute values may have been affected,  the accuracy of the comparisons was determined by  other factors.  

Faust (1) reported that the presence of poly-ethylene in contact with the HF-H202 etch increases the reaction rate. Both the reagent HF and  H202 in the present study were stored, prior to use,  in polyethylene bottles. To determine stability, a  stock solution of No. 2 etchant was kept in a poly-ethylene bottle for an additional six weeks; the rate  measured at the end of this period was the same  as that measured at the start. This indicates that  etching solutions may actually be made up in large  quantities and used over a considerable period of  time.  

Weighing.--Etching rates were obtained by  measuring the loss of weight of the sample during a  measured time of etching. Weights were measured  to • mg. Weight differences were usually of the  order of 20 mg.  

Agitation.--In order to make certain that the  rate of agitation used did not influence the results,  a set of measurements was made of the weight loss  of a sample etched under different conditions of  agitation. This experiment was done at 40~ a  temperature at which the rate of etching was known  to be high. The curve is shown in Fig. 1. One agitation is a movement to and fro of 1 in. in a 60-ce  beaker containing 40 cc of solution, and it is seen  that one or two agitations per second are enough to  stabilize the etching rate.